MINI GRIDS FOR HALF A BILLION PEOPLE Market Outlook and Handbook for Decision Makers MINI GRIDS FOR HALF A BILLION PEOPLE Market Outlook and Handbook for Decision Makers © 2022 | International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved. Rights and Permissions The material in this work is subject to copyright. Because the World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes if full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: +1-202-522-2625; e-mail: pubrights@worldbank.org. Furthermore, the ESMAP Program Manager would appreciate receiving a copy of the publication that uses this publication for its source sent in care of the address above, or to esmap@worldbank.org. This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:/ /creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Energy Sector Management Assistance Program. 2022. Mini Grids for Half a Billion People: Market Outlook and Handbook for Decision Makers. Washington, DC: World Bank. www.esmap.org/ mini_grids_for_half_a_billion_people. License: Creative Commons Attribution CC BY 3.0 IGO Translations—Add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—Add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author(s) of the adaptation and are not endorsed by The World Bank. Third-Party Content—The World Bank does not necessarily own each component of the content contained within the work and does not warrant that the use of any third-party owned individual component or part contained in the work will not infringe on the rights of those third parties. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. Production Credits Designer | Naylor Design, Inc. Cover Image | Image by vitranc, GettyImages. All images remain the sole property of their source and may not be used for any purpose without written permission from the source. ESMAP’s MISSION The Energy Sector Management Assistance Program reflect the views of the World Bank, its affiliated organiza- (ESMAP) is a partnership between the World Bank and tions, members of its board of executive directors for the 24 partners (https://esmap.org/donors) to help low- countries they represent, or ESMAP. The World Bank and and middle-income countries reduce poverty and boost ESMAP do not guarantee the accuracy of the data included growth through sustainable energy solutions. ESMAP’s in this publication and accept no responsibility whatsoever analytical and advisory services are fully integrated within for any consequence of their use. The boundaries, colors, the World Bank’s country financing and policy dialogue in denominations, and other information shown on any map the energy sector. in this report do not imply on the part of the World Bank Group any judgment on the legal status of any territory or Through the World Bank Group, ESMAP works to acceler- the endorsement of acceptance of such boundaries. ate the energy transition required to achieve Sustainable Development Goal 7 (SDG 7) (https://sdgs.un.org/goals/ The text of this publication may be reproduced in whole goal7) to ensure access to affordable, reliable, sustainable, or in part and in any form for educational or nonprofit and modern energy for all. It helps to shape WBG strate- uses, without special permission, provided the source is gies and programs to achieve the World Bank Group’s Cli- acknowledged. Requests for permission to reproduce por- mate Change Action Plan targets. Learn more at: https:// tions for resale or commercial purposes should be sent to esmap.org. the ESMAP Manager at esmap@worldbank.org. ESMAP encourages dissemination of its work and normally gives ESMAP reports are published to communicate the results permission promptly. The ESMAP Manager would appre- of ESMAP’s work to the development community. Some ciate receiving a copy of the publication that uses this sources cited in this report may be informal documents report for its source, sent to esmap@worldbank.org. not readily available. All images remain the sole property of their source and This work is a product of the staff of the World Bank with may not be used for any purpose without written permis- external contributions. The findings, interpretations, and sion from the source. conclusions expressed in this work do not necessarily MINI GRIDS FOR HALF A BILLION PEOPLE    iii A NOTE TO THE READER This is a big book. It is packed with actionable information • A volume of case studies on the history of mini grids in for decision-makers, and it is the World Bank’s most com- electric power systems, as well as mini grid regulations prehensive and authoritative publication on mini grids to and subsidies in Bangladesh, Cambodia, India (Uttar date. Pradesh), Kenya, Nigeria, and Tanzania. We intend this book as a reference guide to be consulted • Animations, infographics, and videos to present high- when important decisions about mini grids need to be level findings to a wide audience. made at the project, portfolio, or national program level. • Briefs in the Live Wire series that can serve as quick To that end, we have balanced cohesiveness among the reference guides for World Bank operations teams and chapters with each chapter’s ability to stand on its own as other project implementation partners. a resource. Ensuring That Regulations Evolve as Mini Grids – “ The book is structured as follows. The overview presents a Mature” (https://openknowledge.worldbank.org/ global market outlook for mini grids and introduces the 10 handle/10986/31773). building blocks that need to be in place if mini grids are to – “Investing in Mini Grids Now, Integrating with the Main be scaled up in any country. These building blocks also rep- Grid Later: A Menu of Good Policy and Regulatory resent the 10 frontiers for innovation for the sector, where, Options” (https:/ /openknowledge.worldbank.org/ with disruptive digital solutions across all 10 frontiers, the handle/10986/31772). services offered to end users can be raised to a level sub- stantially better than what would be possible with alter- • A roster of experts to provide rapid-response support natives. In the Handbook, the terms “building blocks” and for project implementation. “frontiers” are used interchangeably. Chapters 1–10 pres- The objective of this comprehensive knowledge package ent the 10 building blocks in detail and answer the question is to present road-tested options and examples from the how do we scale up mini grid deployment to connect half a leading edge of mini grid development. Decision-makers billion people by 2030? Chapter 11 is our call to action. can draw on these options and examples to scale up mini This book is part of a comprehensive knowledge package grid deployment in their own contexts. By acknowledging that the World Bank has prepared on mini grids, which con- different national approaches to mini grids and providing sists of the following elements, available online at www. context-specific considerations for implementation, this esmap.org/mini_grids_for_half_a_billion_people: suite of knowledge products offers an adaptive approach to helping countries achieve their electrification targets. • An executive summary, published separately in June 2019 (https://openknowledge.worldbank.org/handle/ 10986/31926). iv   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE OF CONTENTS Acknowledgments   xvi Abbreviations   xix MINI GRIDS BY THE NUMBERS   xxiv MAIN FINDINGS 1 The New Electricity Access Landscape 1 Building Blocks 1 through 10 and Frontiers for Sector Growth: Creating the Environment for Takeoff of Mini Grid Portfolios 4 1. Solar Mini Grid Costs, Design, and Innovation 4 2. Planning national strategies and developer portfolios with geospatial analysis and digital platforms 7 3. Transforming productive livelihoods and improving business viability 7 4. Engaging communities as valued customers 8 5. Delivering services through local and international companies and utilities 9 6. Financing solar mini grid portfolios and end user appliances 9 7. Attracting exceptional talent and scaling skills development 10 8. Supporting institutions, delivery models, and champions to create opportunities 11 9 and 10. Regulating the sector and making it easier to do business 11 A Call to Action 12 References 12 Notes 13 OVERVIEW THE NEW ELECTRICITY ACCESS LANDSCAPE AND THE GROWING SPACE FOR SOLAR MINI GRIDS 14 SDG 7: A Global Agenda Running Behind 14 Access to electricity has increased . . . 14 . . . But progress has been insufficient to meet the goal of universal access 14 More financing is needed, and it must be better targeted 15 Double down on solutions that have the potential for exponential growth curves 16 The Place for Solar Mini Grids 16 What are solar mini grids? 16 The historic role of mini grids in national electrification efforts 18 The role of solar mini grids in universal electrification 22 Where do solar mini grids fit in? 22 How to Scale Solar Mini Grid Deployment to serve Half a Billion People 25 Five drivers: Cost, quality, pace, finance, and enabling environment 25 Ten building blocks 27 Global Market Snapshot, Outlook 2030, and Call to Action 29 Projections 34 Status of the five drivers and ten building blocks 36 Call to action 42 References 43 Notes 43 MINI GRIDS FOR HALF A BILLION PEOPLE    v CHAPTER 1 REDUCING COSTS AND OPTIMIZING DESIGN AND INNOVATION FOR SOLAR MINI GRIDS 45 The Levelized Cost of Mini Grid Electricity 45 The levelized cost of energy from mini grids: Seven analytical cases 46 Modeling assumptions and scenarios 47 Modeling results 48 Effect of expected declines in capital and operating costs by 2030 51 The share of renewable energy 51 Modeling results: LCOE of optimum hybrid vs. 0 percent and 100 percent renewable energy 52 Implications for national utilities of lower mini grid LCOE 52 Implications for national utilities of improving the quality of mini grid services 52 Win-win for mini grids and national utilities 54 Current Status of Selected Mini Grid Capital and Operating Costs 54 Cost per unit of firm power output 54 Investment costs per customer 56 Cost of individual components 57 Replacement costs 66 Operating costs 67 The Outlook for Mini Grid Capital and Operating Costs 67 PV module trends 67 PV inverter trends 68 Battery trends 69 Battery inverter trends 69 Smart meter trends 69 Trends in other capital costs 69 Trends in operating costs 70 The Impact of Economies of Scale 71 Reasoning from First Principles 72 Conclusion 73 Summary of potential cost reductions 73 Government’s role in reducing mini grid costs and catalyzing innovation 74 The importance of coordinated collection of data on mini grid costs 74 References 74 Notes 75 CHAPTER 2 NATIONAL STRATEGIES AND DEVELOPER PORTFOLIOS: THE ROLES OF GEOSPATIAL ANALYSIS AND DIGITAL PLATFORMS 78 Assessing the Market Potential for Mini Grids 79 Simple exploratory spatial data analysis for Sub-Saharan Africa using GRID3 79 The Global Electrification Platform and least-cost electrification analysis for Sub-Saharan Africa 80 Maximal mini grid deployment modeled in the Global Electrification Platform 82 National Electrification Planning 85 Operational experience and the Global Electrification Platform 87 Indicative workflow for the development of a GIS-based national electrification plan 89 Analytical insights and generic observations 91 Lessons learned and challenges ahead 93 Mini Grid Portfolio Planning 94 Overview 94 The workflow phases for mini grid portfolio planning: Spatial data and analytics 95 Lessons learned and next steps 112 References 112 Notes 113 vi   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 3 PRODUCTIVE LIVELIHOODS AND BUSINESS VIABILITY 114 The Multiple Benefits of Connecting Income-Generating Machines and Appliances to Mini Grids 114 Rolling Out Programs to Promote Productive Uses and Stimulate Demand 117 Step 1. Assessing markets and demand: Geospatial analysis superimposed over mini grids, appliances, and finance for end users 118 Step 2. Surveys and workshops build on Step 1 data through community engagement 120 Step 3. Demand analysis for mini grid design and market potential for appliances and associated end-user finance 121 Step 4. Preparation of road shows involving local government, leadership communities, communities, interested appliance providers and end-user financiers, mini grid companies 124 Step 5. Road shows to load centers where mini grid developers, appliance suppliers, and end user financiers explain the value propositions to potential end users based on their current and aspirational living standards 125 Step 6. Rollout of mini grid connections, sales of appliances, and end-user finance 126 Who Organizes PUE Programs? 129 Mini grid developers 129 Government agencies and policy makers 129 Local change agents 130 Timing Productive Use Programs for Maximum Effect 130 What’s Next? 130 References 135 Notes 135 CHAPTER 4 ENGAGING COMMUNITIES AS VALUED CUSTOMERS 136 Why Is Community Engagement Important? 136 Community Engagement throughout the Mini Grid Project Cycle 139 Design and planning phase 139 Promotion and information-sharing phase 141 Financing and procurement phase 142 Implementation and construction phase 143 Operations and maintenance phase 143 Summary of community engagement over the project cycle 144 Important Gender Aspects of the Community Engagement Process 144 Removing Barriers to Scale through Innovations in Community Engagement 146 Country-level program: Smart Power India 146 ICT for community engagement at the developer level: A pilot video hub delivering mini grid stories 147 Conclusion 148 References 148 Notes 149 CHAPTER 5 DELIVERING SERVICES THROUGH LOCAL AND INTERNATIONAL ENTERPRISES AND UTILITIES 150 An Evolving Technology from First-Generation to Third-Generation Mini Grids 151 First- and second-generation mini grids 151 Second-generation mini grids 151 Third-generation mini grids 151 The Business Case for Private-Sector Participation in Third-Generation Mini Grids 152 Private-sector segmentation 155 Mini grid facilitators 165 Facilitating Collaboration Between and Among Local and International Private-Sector Entities 169 Local and International Industry Players Across the Mini Grid Industry Value Chain 170 Component manufacturing 170 Market assessment 170 MINI GRIDS FOR HALF A BILLION PEOPLE    vii Permitting and financing 172 Grid design and procurement 173 Integration and installation 174 Operations and maintenance 175 After sale 177 Profit potential of the solar mini grid value chain 177 Summary 179 References 180 Notes 181 CHAPTER 6 FINANCING SOLAR MINI GRID PORTFOLIOS AND END-USE APPLIANCES 182 What Are the Financial Needs of Mini Grid Developers? 182 What Types of Finance Are Available? 183 Mini Grid Debt and Equity Investors 183 Development Partners’ Investment in Mini Grids 187 Barriers Faced by Private Developers in Accessing Commercial Finance 190 Mini grids are perceived as risky 190 Demand uncertainty remains high 190 Consumers are unable to pay the full costs of supply and lack credit history 190 Equity is inadequate 191 The tenors of assets and liabilities are mismatched 191 Political and regulatory risks create uncertainty 191 Future macroeconomic conditions are uncertain 191 Overcoming These Barriers 192 Developer actions 192 Government and development partner actions 192 Subsidies 197 Facilitating equity 200 Mitigating risk 202 Conclusions, and General Recommendations for Governments 204 References 205 Notes 205 CHAPTER 7 ATTRACTING EXCEPTIONAL TALENT AND SCALING UP SKILLS DEVELOPMENT 206 Mini Grid Development Requires Distinctive Skills and Capacity 207 Preinvestment phase 207 Project construction 207 Project operations and maintenance 207 Identifying Skills Gaps 208 Project developers 209 Utilities 209 Banks and financial institutions 209 Engineers 209 Suppliers 209 Policy makers and regulators 210 Local communities and customers 210 Training and Skill-Building Interventions to Address Skills Gaps 210 National-level training and skill building 210 Project-level training and skills building 213 Community involvement and awareness 215 viii   MINI GRIDS FOR HALF A BILLION PEOPLE Lessons Learned from Effective Training and Skills-Building Programs 218 Database of Training Programs 219 References 220 Notes 222 CHAPTER 8 DELIVERY MODELS AND SUPPORTING INSTITUTIONS 223 Mini Grid Delivery Models 223 The build-own-operate model 223 Public-private models 223 Concession model 224 Utility models with or without private-sector involvement 224 Cooperative model 225 Comparison of various models 225 Institutional Framework 228 The desirable features of a framework to support mini grids 228 Institutions inside the energy sector 231 Institutions outside the energy sector 232 Investors’ Perspective on Institutional Frameworks 233 Role for Development Partners 233 References 234 Notes 234 CHAPTER 9 ENACTING REGULATIONS AND POLICIES THAT EMPOWER MINI GRID COMPANIES AND CUSTOMERS 235 The Importance of Workable Regulations for Scaling Up Mini Grid Development 236 Five Key Regulatory Decisions 238 Regulating entry 239 Regulating retail tariffs 242 Regulating service standards 249 Regulating technical standards 252 Regulating relationships between the main grid and mini grids 255 Combining Regulatory Elements 259 Combining regulatory elements into effective packages 259 Combining regulatory elements in a phased approach 261 Regulatory Innovations to Further Incentivize Private-Sector Investment in Mini Grids 262 Regulation by contract 263 Arbitration-style appeal mechanism to complement a light-handed regulatory approach 264 Investors’ Perspective on Mini Grid Regulations 265 Conclusion and Resources 266 Acknowledgments for the Six-Country Case Studies That Informed This Chapter 267 Bangladesh 267 Cambodia 267 Uttar Pradesh (India) 268 Kenya 268 Nigeria 268 Tanzania 268 References 268 Notes 269 MINI GRIDS FOR HALF A BILLION PEOPLE    ix CHAPTER 10 CUTTING RED TAPE FOR A DYNAMIC BUSINESS ENVIRONMENT 271 Why an Enabling Environment Matters 271 Characteristics of Doing Business as a Mini Grid 272 Long-lived sunk assets 272 Unpredictability and the need for flexibility 272 The political nature of electricity 272 Four Complementary Options to Make It Easier for Mini Grid Developers to Do Business 273 Reducing red tape through standardized, preapproved templates 275 Using technology platforms to connect developers with investors and suppliers and to conduct large-scale mini grid tenders 283 Eliminating duplication of government oversight by delegating authority to a single entity 284 Setting up e-government services to reduce overhead 286 Investors’ Perspectives on the Four Options Presented Above 287 Conclusion 288 References 288 Notes 288 CHAPTER 11 CALL TO ACTION 289 Policy Makers 289 Regulators 289 Development Partners 289 Industry Associations 289 Mini Grid Developers 290 Investors 290 Suppliers 290 Researchers 290 Topics for Future Research 290 Collecting data on installed and planned mini grids 290 Combining mini grids, solar home systems, and main grid extensions into an integrated electrification strategy at the local level 290 Business tactics and strategies for mini grid developers 291 Policies and business environment factors that affect mini grids 291 References 292 x   MINI GRIDS FOR HALF A BILLION PEOPLE BOXES Box 1.1 • The levelized cost of energy for best-in-class mini grids dropped nearly 31 percent from 2018 49 Box 1.2 • Direct current mesh grids 65 Box 2.1 • Data sources for cluster definition in the Nigeria Electrification Project   96 Box 2.2 • Finding the optimal input parameters for DBSCAN   98 Box 3.1 • How IDCOL increases productive uses of energy in solar-hybrid mini grids in Bangladesh 131 Box 3.2 • Rural, productive uses of electricity: Lessons from Ethiopia 133 Box 3.3 • Lessons from a utility-NGO partnership in Indonesia 134 Box 5..1 • Green Village Electric partners with Schneider Electric 155 Box 5.2 • Africa Minigrid Developers Association 160 Box 5.3 • Utility-led rollout of mini grids on the national scale: Case study from Ethiopia 161 Box 5.4 • Prepay is not just for mini grids—Eskom’s “Power for All” scheme 164 Box 6.1 • Women’s limited access to finance 183 Box 6.2 • Nonprice factors in renewable energy projects 194 Box 6.3 • A financial support program in Bangladesh 195 Box 6.4 • A revolving fund for consumer finance in Lao PDR 199 Box 7.1 • Policy and regulatory train-ings provided by Economic Consulting Associates 211 Box 7.2 • Lessons from community training by Trama TechnoAmbiental 216 Box 7.3 • Productive-use training from IEEE Smart Village 217 Box 7.4 • Components of IEEE Smart Village’s comprehensive training program 219 Box 8.1 • The build-own-operate model in Tanzania 224 Box 9.1 • Setting individualized tariff controls 245 Box 9.2 • From laissez-faire to comprehensive regulation: Cambodia’s successful electrification with mini grids 262 Box 9.3 • Independent appeal tribunal in Jamaica 265 Box 10.1 • The effect on investment of permits outside the electricity sector 276 Box 10.2 • World Bank experience with Odyssey Energy Solutions in Nigeria 283 FIGURES Figure MF.1 • A mini grid system (part A) and a containerized solar mini grid (part B) 5 Figure O.1 • Example of a common solar hybrid mini grid setup 17 Figure O.2 • The first, second, and third generations of mini grids 21 Figure O.3 • Matrix of market drivers and building blocks to support them 27 Figure O.4 • Number of people connected to mini grids under business-as-usual and universal access scenarios, 2020–30 34 Figure O.5 • Mini grids installed annually in each of the top 20 electricity-access-deficit countries, 2018–30 37 Figure O.6 • Average mini grid load factor, 2018–30 38 Figure O.7 • Total cumulative investment in mini grids for energy access, 2018–30 39 Figure O.8 • Average RISE score in top 20 electricity-access-deficit countries 40 Figure O.9 • LCOE of best-in-class solar hybrid mini grids 41 Figure 1.1 • Load profiles for 22 percent load factor, 22 percent load factor (sun following), 40 percent load factor, 40 percent load factor (sun following), and 80 percent load factor 47 Figure 1.2 • Economic LCOE calculations for mini grids in 7 cases based on 0 percent subsidy and load profiles described in figure 1.1 50 Figure 1.3 • Comparison of levelized cost of energy of mini grids and utilities in Africa 53 Figure 1.4 • Total economic cost of mini grids per kWfirm as a function of firm power output 55 Figure 1.5 • Mini grid economic costs per customer (left) and costs per customer for mini grids below median cost (right) 56 Figure 1.6 • Average share of component economic costs in total capital costs of mini grids 58 Figure 1.7 • Costs of solar panels (including PV inverters) for mini grids, by year, 2012–21 59 Figure 1.8 • Economic cost trends for the storage capacity ($/kWh) of lithium-ion and lead-acid batteries used in mini grids between 2012 and 2021 60 MINI GRIDS FOR HALF A BILLION PEOPLE    xi Figure 1.9 • Net present value of storage capacity for lithium-ion and lead-acid batteries, 2012–21 61 Figure 1.10 • Unit costs for inverters, energy management systems, and monitoring (blue), and balance of system (orange) 62 Figure 1.11 • Powerhouse innovations can lower costs and expedite deployment 63 Figure 1.12 • Distribution costs per customer, 2012 to 2021 64 Figure 1.13 • Distribution costs per customer as a function of customers served 65 Figure 1.14 • A 30 kWp Mandalay Yoma mini grid in Myanmar (left) and a 40 kWp Winch Energy mini grid in Uganda (right) 66 Figure 2.1 • Scatter plot of settlement population vs population density in Sub-Saharan Africa 79 Figure 2.2 • Sub-Saharan Africa’s addressable market for mini grids 81 Figure 2.3 • Sub-Saharan Africa’s addressable market for mini grids, mapped by settlement population 81 Figure 2.4 • Distribution by country of 429.5 million people served at least cost by mini grids in 58 access-deficit countries 83 Figure 2.5 • Distribution by mini grid size of 429.5 million people served at least-cost by mini grids in 58 countries with severe access deficits 85 Figure 2.6 • Geospatial least-cost rollout plans in Kenya and Rwanda 86 Figure 2.7 • Geospatial least-cost electrification plans for Myanmar and Nigeria by 2030, by technology component 87 Figure 2.8 • The GEP Explorer 88 Figure 2.9 • Typical least-cost electrification planning sequence (best practice) 89 Figure 2.10 • Distance as a function of load size: Break-even grid extension 92 Figure 2.11 • Geospatial portfolio planning sequence 95 Figure 2.12 • Methodology for the generation of population clusters in Nigeria, using the HRSL and OSM data 96 Figure 2.13 • Nigeria’s population clusters: Spatial distribution (left) and size histogram, in hectares (right) 97 Figure 2.14 • Sample outputs from the Digitize Africa building footprint data set 97 Figure 2.15 • Concept of the DBSCAN algorithm 97 Figure 2.16 • Ethiopia’s rural population settlements and mini grid deployment: Bounded DBSCAN clustering 98 Figure 2.17 • Cluster contour delineation: Convex hull (left) and alpha shapes (right) 99 Figure 2.18 • Health facilities and education facilities within 500 meters of village boundary 100 Figure 2.19 • Main grid coverage in Nigeria 100 Figure 2.20 • VIDA GridLight prediction for Ethiopia (blue) compared to Ethiopian Electric Utility data (red) 101 Figure 2.21 • Nighttime lighting in Nigeria 101 Figure 2.22 • Nigerian night lights and 20 km buffer zones 101 Figure 2.23 • Results of prioritization of clusters for mini grid electrification in Nigeria 102 Figure 2.24 • Population clusters falling in protected areas or KBAs flagged for exclusion in Nigeria 103 Figure 2.25 • Requirements for estimating load profile 104 Figure 2.26 • Indicative load profiles for various customer segments in potential mini grid locations 105 Figure 2.27 • Demand curve for a randomly selected candidate site in Ethiopia 105 Figure 2.28 • Sample load profile for a village 106 Figure 2.29 • Output of the voltage drop model 108 Figure 2.30 • Sample outputs of hyperlocal density analysis from Village Data Analytics 108 Figure 2.31 • Indicative flow of financial modeling process of mini grids 109 Figure 2.32 • VIDA interactive platform 110 Figure 2.33 • Mini grid tender preparation in Odyssey 111 Figure 3.1 • The impact of productive electricity uses on the daily load profile and levelized cost of energy 115 Figure B3.1.1 • Share of expected load achieved by selected mini grids in Bangladesh 131 Figure B3.1.2 • Effect of extensive customer awareness campaigns on uptake 132 Figure 4.1 • Typical issues hindering the mini grid development process 138 Figure 4.2 • Typical mini grid project cycle 139 Figure 4.3 • The A-B-C model 140 Figure 5.1 • Tariff-charging type by customer class 163 Figure 5.2 • Tariffs by customer class ($/kWh, charged by consumption) 163 Figure 5.3 • Tariffs by customer class ($/month, charged in flat fee) 163 xii   MINI GRIDS FOR HALF A BILLION PEOPLE Figure 5.4 • Mini grid industry value chain 171 Figure 5.5 • Customer types served by mini grids 172 Figure 5.6 • Average contribution share of each disclosed funding type 173 Figure 5.7 • Generation source of the mini grid 174 Figure 5.8 • Causes behind operators’ worst months of delivering electricity 175 Figure 5.9 • Tiers of daily availability in typical and worst months 176 Figure 5.10 • Tiers of evening availability in typical (left) and worst (right) months 176 Figure 5.11 • Daily peak hour profile 177 Figure 5.12 • Tiers of reliability in typical and worst months 178 Figure 5.13 • Profit potential for facilitator organizations across the value chain, 2019 and 2030 178 Figure 6.1 • Cumulative private-sector investment in mini grid companies 184 Figure 6.2 • Annual number of deals between investors and mini grid companies, 2010–22 185 Figure 6.3 • Top 20 investors in mini grid companies by cumulative investment 185 Figure 6.4 • Trends in debt and equity 186 Figure 7.1 • National-level training needs and relevant stakeholders 211 Figure 7.2 • Project- and portfolio-level training needs and relevant stakeholders 213 Figure B7.3.1 • Future community entrepreneurs and mini grid technicians participate in a classroom discussion at a training program offered by Igniting Africa 217 Figure 8.1 • Roles for developer and utility under different delivery models 226 Figure 8.2 • Sample ecosystem of institutions affecting mini grid developers 229 Figure 9.1 • Regulatory Indicators for Sustainable Energy (RISE) scores for mini grid framework 237 Figure 9.2 • Correlation between mini grid policies and regulations and number of mini grids planned 238 Figure 9.3 • Decision tree for regulating entry 241 Figure 9.4 • Decision tree for regulating tariffs 247 Figure 9.5 • Decision tree for regulating service standards 251 Figure 9.6 • Decision tree for regulating technical standards 254 Figure 9.7 • Decision tree for integration and exit options 258 Figure 10.1 • Nigeria’s Environmental and Social Management System for minimum-subsidy tenders for mini grid development 280 Figure 10.2 • Nigeria’s Environmental and Social Management System for performance-based grants for mini grid development 281 TABLES Table MF.1 • Market drivers and 2030 targets 3 Table MF.2 • Building blocks and identified areas for potential magnitude change 3 Table MF.3 • Sub-Saharan African mini grid markets and their progress across the 10 building blocks 4 Table MF.4 • The levelized cost of energy by load factor, 2018, 2021, and 2030 6 Table O.1 • Benchmarks and price projections, mini grid component costs, 2010–30 24 Table O.2 • SDG 7 and mini grid industry targets, 2020–30 26 Table O.3 • Installed and planned mini grid projects worldwide: A summary 30 Table O.4 • Summary of installed mini grid projects by region 31 Table O.5 • Number of installed and planned mini grids by region 32 Table O.6 • Top-10 lists for key mini grid indicators for installed mini grids 32 Table O.7 • Characteristics of installed and planned mini grids 33 Table O.8 • ESMAP mini grid outlook scenario: A regional breakdown 36 Table O.9 • Top 20 countries with energy access deficits: Doing Business and RISE scores, 2020 40 Table O.10 • The global mini grid sector and its progress across the 10 building blocks 42 Table 1.1 • Representative mini grids from seven cases: An analysis of key characteristics 46 Table B1.1.1 • Estimated and potential levelized cost of mini grid energy, 2018 and 2021 49 Table 1.2 • Optimum renewable energy share for mini grid cases considered 51 MINI GRIDS FOR HALF A BILLION PEOPLE    xiii Table 1.3 • Economic LCOE of hybrid mini grid versus diesel only and renewables only 52 Table 1.4 • Mini grid components: A summary of costs and characteristics 57 Table 1.5 • Average economic costs of key mini grid hardware components, by country 58 Table 1.6 • Performance characteristics of lead-acid and lithium-ion batteries as modeled in HOMER levelized cost of energy calculations 60 Table 1.7 • Mini grid component cost benchmarks and price projections 68 Table 1.8 • Net present value broken down by category with economies of scale 71 Table 1.9 • Change in unit costs with economies of scale, by cost category 72 Table 2.1 • Characteristics of Sub-Saharan African settlements suitable for electrification via mini grid 80 Table 2.2 • Selected electrification results for 2030 retrieved from the Global Electrification Platform, aggregated for 46 countries in Sub-Saharan Africa 82 Table 2.3 • Breakdown of electrification results from bottom-up demand scenario 82 Table 2.4 • GEP scenario codes for each country’s maximum number of new mini grid connections by 2030 83 Table 2.5 • Distribution by mini grid system size of 429.5 million people served at least-cost by mini grids in 58 countries with access deficits 85 Table 2.6 • Maximum cost-justified distance for connecting a customer as a function of the required level of service 92 Table 3.1 • The Mwenga hydro mini grid: Estimated costs and benefits 116 Table 3.2 • Six steps to roll out a PUE program 118 Table 3.3 • Example of PUE program stakeholders identified in the Democratic Republic of Congo, Ethiopia, and Nigeria 119 Table 3.4 • Power requirements, costs, and indicative payback periods of selected income-generating appliances 122 Table 3.5 • Stakeholders that could be involved in road shows and their respective roles 126 Table 4.1 • Potential for community engagement to limit mini grid risks 138 Table 4.2 • Key community engagement activities over the project cycle 144 Table 5.1 • Utility connection rates, 2004–14 153 Table 5.2 • Categories of local and international private-sector players 156 Table 5.3 • Sample mini grid developers 157 Table 5.4 • Profit potential of mini grid operators given certain tariffs and costs of service 158 Table 5.5 • Mini grid developers and large-scale IPPs: A comparison of audited financial results, 2018 159 Table 5.6 • Sample utility mini grid projects 161 Table 5.7 • Sample mini grid experience of EPC companies 166 Table 5.8 • Sample system integrators with technologies in mini grids 166 Table 5.9 • Sample original equipment manufacturers with technologies in mini grids 167 Table 5.10 • Benefits of local and international partnerships 170 Table 6.1 • Types and sources of commercial finance 184 Table 6.2 • Categories of investors in mini grid companies, 2010–22 186 Table 6.3 • World Bank mini grid investment portfolio as of June 30, 2022 188 Table 6.4 • Overcoming barriers to investments in mini grids 193 Table 6.5 • Impact of performance-based subsidies of capital expenses on the levelized cost of energy of a well-designed mini grid 198 Table 7.1 • Conducting a project- or portfolio-level capacity needs assessment 208 Table 7.2 • Selection of tools for resource planning 212 Table 7.3 • Sample tools for financial planning 212 Table 8.1 • Comparative analysis of mini grid delivery models 226 Table 8.2 • Roles of national and international institutions 230 Table 9.1 • Options for regulating entry 239 Table 9.2 • Information requirements for registration, permitting, and licensing of mini grids in five countries 240 Table 9.3 • Options for regulating tariffs 243 Table B9.1.1 • Individualized tariff features in four countries 245 Table 9.4 • Assessment of tariff options 246 Table 9.5 • Options for regulating service standards 250 xiv   MINI GRIDS FOR HALF A BILLION PEOPLE Table 9.6 • Options for regulating technical standards 253 Table 9.7 • Options for preserving value when the main grid arrives 256 Table 9.8 • Effective regulatory packages 260 Table 9.9 • Three phases of evolutionary regulation 261 Table 9.10 • Two innovations in regulation that can further incentivize private-sector investment in mini grids 263 Table 10.1 • Options for making it easier for mini grid developers to do business 274 Table 10.2 • General types of bureaucratic processes that mini grid developers navigate 276 Table 10.3 • Key provisions of power purchase agreements in Cambodia and Tanzania 277 Table 10.4 • Advantages and disadvantages of two options for mini grid oversight 285 Table 11.1 • Analytical framework to guide future research on mini grid business models 292 MINI GRIDS FOR HALF A BILLION PEOPLE    xv ACKNOWLEDGMENTS This book is the result of a collaborative effort over more (HOMER). The write-up on the first generation of mini than four years to collect and synthesize the best research, grids is based on a report on the history of mini grids data, and knowledge produced to date on the entire eco- in electric power systems prepared by Morgan Bazil- system that supports mini grids—regulations, financing, ian (Colorado School of Mines), Francesco Fuso Ner- technology, training, gender, productive uses, geospatial ini (KTH), Mark Howells (KTH), Alexandros Korkovelos planning, and more. We would like to acknowledge here (KTH), Lucille Langois (IAEA), Holger Rogner (IIASA), and those individuals and organizations that contributed to Hisham Zerriffi (University of British Columbia), which is this project. available on the companion website to this handbook: www.esmap.org/mini_grids_for_half_a_billion_people. PARTNERSHIP AND FUNDING Guidehouse (formerly Navigant Research) provided valu- This book is part of a much larger comprehensive knowl- able assistance with initial data collection on the mini edge package that was funded by World Bank/ESMAP and grid market in 2018, and we leveraged their global data- the Foreign, Commonwealth and Development Office of the base as a starting point for our data set. We are grateful United Kingdom as one of ESMAP’s key donors in particu- also to World Bank operations teams for helping us col- lar. This comprehensive knowledge package is the result of lect data on mini grids in their respective countries, and ESMAP’s collaboration with a broad set of mini grid sector to Bloomberg New Energy Finance—in particular Itamar stakeholders, including mini grid developers, regulators and Orlandi—for answering our questions and helping us other government officials, financiers, technology provid- understand their earlier outlooks for the mini grid market. ers, researchers, project implementation partners, develop- For the discussion on the Key Performance Indicators for ment partners, and internationally recognized experts. We the mini grid industry, we are grateful for contributions are particularly grateful to the following mini grid industry and guidance from the Africa Minigrid Developers Asso- leaders who not only gave us their time but also helped ciation’s leadership team. ensure our work captured ground-level realities and frontier innovations accurately: Africa Minigrid Developers Associa- CHAPTER 1: LEAD AUTHORS CHRIS GREACEN AND tion (AMDA), Engie, Havenhill, HOMER Energy, Husk Power JON EXEL Systems, INENSUS, Odyssey Energy Solutions, Power Cor- This chapter would not have been possible without the ner, PowerGen, PowerHive, Techno Hill, and Trama Tecno- detailed data provided by mini grid developers. Detailed Ambiental. data on costs and designs are commercially sensitive. The authors thank all mini grid developers who recognized PROJECT MANAGEMENT AND AUTHORSHIP the importance of research on the evolution of mini grid This book was prepared by the Global Facility on Mini Grids costs and provided the data they could. For 32 mini grids team under the overall guidance of ESMAP’s former and in diverse countries across Africa and Asia, the firm Trama current practice managers, Rohit Khanna and Gabriela TecnoAmbiental S.L. (TTA) undertook the difficult work of Elizondo Azuela, respectively. Jon Exel, Tatia Lemondzhava, requesting data and following up to iron out inconsisten- Ashish Shrestha, and Dr. James Knuckles managed the cies as they were discovered. Chris Purcell provided data project and oversaw the book’s development, from incep- for mini grids in Myanmar and Bangladesh; Sunita Chik- tion to publication. Dr. Knuckles was the book’s lead editor. katur Dubey did the same for several mini grids in Ghana. The following individuals and organizations are this book’s The Odyssey team provided extensive data for hundreds co-authors and contributors: of mini grids in Nigeria. The authors thank Andrew Pascale and Ziting Huang for their assistance and expertise in pro- OVERVIEW: LEAD AUTHORS JON EXEL, cessing data and Mr. Pascale for his work on an early draft JAMES KNUCKLES, AND TATIA LEMONDZHAVA of this chapter. Marilena Lazopoulou and Pol Arranz Piera The chapter benefitted from expert inputs from Dr. Ber- from TTA and Chris Purcell were generous and patient nard Tenenbaum, Dr. Chris Greacen, and Peter Lilienthal with our many questions. xvi   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 2: LEAD AUTHORS ASHISH ment deals featured in this chapter. Dr. Chris Greacen and SHRESTHA AND ALEXANDROS KORKOVELOS HOMER Energy conducted modeling and analysis for the The chapter benefitted from expert inputs from Philipp section of this chapter on the impacts of performance-based Blechinger (Reiner Lemoine Institut), Stewart Craine and grants on the cost of solar hybrid mini grid electricity. Yann Monty Craine (Village Infrastructure Angels), Reja Amatya Tanvez and Candice Lanoix provided a great deal of input (MIT), Xiangkun Li and Samuel Booth (NREL), Peter Lil- from the International Finance Corporation. ienthal (HOMER), Rik Wuts (Powerhive), Ibrahim Abada CHAPTER 7: LEAD AUTHORS SUNITA DUBEY AND (Engie), Emily McAteer (Odyssey Energy Solutions), Nabin CASTALIA Raj Gaihre, Philippe Raisin and Tobias Engelmeier (VIDA), and Frankie Eckersley-Carr, Imran Muhammad, Oliver Haas We are grateful for contributions from the Institute of Elec- (Integration). The writeup on least-cost electrification plan- trical and Electronics Engineers and to GIZ, HOMER, and ning was prepared with contributions from Chiara Odetta Trama TecnoAmbiental for their contributions as well. Rogate and Rhonda Lenai Jordan (World Bank). The spatial CHAPTER 8: LEAD AUTHORS SUBODH MATHUR, data analysis for Sub-Saharan Africa using GRID3 data was WITH NICO PETERSCHMIDT AND JOANIS HOLZIGEL carried out by Christopher James Arderne. OF INENSUS CHAPTER 3: LEAD AUTHORS JULIETTE BESNARD, We are grateful to the International Finance Corporation TATIA LEMONDZHAVA, JON EXEL, AND JAMES for important inputs on this chapter, including the Inves- KNUCKLES tors’ Perspective section on institutional frameworks for The chapter benefited from the support of Besnik Hyseni mini grids. (World Bank), the Communities in Kisii and Nyamira Coun- CHAPTER 9: LEAD AUTHORS CASTALIA AND ties (Kenya), CLASP, Erika Lovin (CrossBoundary), Amanda JAMES KNUCKLES DelCore (Factor[e]), GMG Facility Kenya, IDCOL, INENSUS and the World Bank RERED II team, Jon Leary, Ed Brown, This chapter draws on a research project conducted by and Simon Batchelor of Loughborough University, Pact/ ESMAP, Castalia, and Ecoligo, which included 11 field visits Smart Power Myanmar, Adriana Karpinska (Powerhive), and 70 interviews, conducted between August and Sep- SNV Netherlands Development Organisation, Barani Aung tember 2017, with key mini grid sector stakeholders in Ban- (Techno Hill), and Trama TecnoAmbiental. gladesh, Cambodia, India (Uttar Pradesh), Kenya, Nigeria, and Tanzania. Stakeholders in each country included reg- CHAPTER 4: LEAD AUTHORS TATIA LEMONDZHAVA, ulators, developers, rural electrification agencies, and con- WITH FELIX TER HEEGDE OF SNV sumers, as well as utilities and development partners, and We are grateful to Rishabh Sachdeva of Quicksand and we are very grateful for their time and insights. A full list of Elijah Siakpere of the World Bank office in Abuja, Nigeria, acknowledgments for each of the six countries is provided for additional contributions. at the end of chapter 9. We also thank the International Finance Corporation for important inputs on this chapter, CHAPTER 5: LEAD AUTHORS RICKY BUCH, JON EXEL, including the Investors’ Perspective section on mini grid JAMES KNUCKLES, AND TATIA LEMONDZHAVA regulations. Kojo Adom Quagraine led the development of the database CHAPTER 10. LEAD AUTHORS: of investment deals featured in this chapter. Bryan Koo and JAMES KNUCKLES AND CASTALIA Ziting Huang (World Bank) helped manage the nationally representative surveys of mini grid operators conducted by Castalia developed sections of the chapter as part of a the World Bank and contributed to the chapter’s analysis larger project it completed for ESMAP (see chapter 9), and writeup of survey results in Cambodia, Myanmar, and and Elijah Abiodun Siakpere contributed in major ways to Nepal. the write-up of the standardized Environmental and Social Management System in Nigeria. Dr. Chris Greacen and CHAPTER 6: LEAD AUTHORS INTERNATIONAL Anastas Mbawala provided valuable comments on ear- FINANCE CORPORATION, JAMES KNUCKLES, lier drafts of the standardized asset transfer agreement. AND SUBODH MATHUR We are grateful to the International Finance Corporation We are grateful for feedback from David Ross of Statera for important inputs on this chapter as well, including the Capital on earlier drafts of the chapter. Kojo Adom Investors’ Perspective section on factors that make it eas- Quagraine led the development of the database of invest- ier (or harder) for mini grid developers to do business. MINI GRIDS FOR HALF A BILLION PEOPLE    xvii CHAPTER 11: LEAD AUTHORS JON EXEL AND chapters throughout the book. The International Finance JAMES KNUCKLES Corporation also provided extensive comments on several chapters, which strengthened the book as a whole. To all of KEY GENDER ASPECTS THROUGHOUT THE BOOK these reviewers: thank you. Inka Schomer and Mary Dominic (World Bank) provided important contributions on the gender-related aspects STAKEHOLDER ENGAGEMENT of mini grid topics throughout the book, from access to Over the past five years, ESMAP has cohosted events and finance to productive uses to community engagement workshops on mini grids both virtually and in-person in and skills building, among others. Ethiopia, Ghana, Kenya, Myanmar, Nigeria, Spain, Tanzania, the United Kingdom, and the United States. We are par- REVIEW AND CONSULTATION ticularly grateful for the host governments as well as the This handbook underwent a formal World Bank Decision more than 2,000 participants at these events representing Review in February 2022, chaired by Demetrios Papatha- all mini grid stakeholder groups from more than 60 coun- nasiou, Global Director for the Energy & Extractives Global tries for their input, debate, and validation of the knowledge Practice. The peer reviewers were Raihan Elahi (Lead Econ- brought together in this report. The Global Facility on Mini omist, World Bank), Dana Rysankova (Lead Energy Special- Grids team would also like to express its gratitude to the ist and Global Lead for Energy Access, World Bank), Arsh ESMAP communications team, and especially to Lucie Sharma (Senior Energy Specialist, World Bank), Patrick Blyth, Nansia Constantinou, Anita Rozowska, and Janice Thaddayos Balla (Senior Energy Specialist, World Bank), Tuten for their help in synthesizing and tailoring the main and Yann Tanvez (Energy Specialist, International Finance messages of this report for a wide audience, and their over- Corporation). In addition, the following experts carefully all support in its preparation and dissemination. reviewed the book in whole or in part. For their time, exper- EDITING, GRAPHIC DESIGN, AND OTHER tise, and thoughtful comments we are exceptionally grate- CONTRIBUTORS ful. Dr. Bernard Tenenbaum, Castalia, INENSUS, the Rocky Mountain Institute, and Trama TecnoAmbiental reviewed Our editors, Joan O’Callaghan and Steven Kennedy, require the book in its entirety; their comments greatly elevated special praise. We are grateful to have been able to work the quality of the final product. Gabriela Elizondo Azuela with them. Naylor Design, Inc. typeset the manuscript and and Michael Toman (World Bank) provided provided invalu- designed every table, chart, and figure in this book. The able input on the report’s main findings that helped us con- quality and consistency they brought to this report is a tes- vey comprehensive and impactful messages throughout tament to their diligence and expertise. the book. Comments from the United Kingdom’s Foreign, The many people we interviewed or who reviewed this Commonwealth and Development Office, the United States report were gracious with their time and knowledge. They National Renewable Energy Laboratory, the Institute of were patient and accommodating with our requests under Electrical and Electronics Engineers, and the Africa Mini- tight deadlines, and we are sincerely grateful. We alone are grid Developers Association helped us transform several responsible for any errors of fact or interpretation. xviii   MINI GRIDS FOR HALF A BILLION PEOPLE ABBREVIATIONS — not available A-B-C Anchor-Business-Community AC alternating current ACP-EU Africa, Caribbean, and Pacific Group of States–European Union ADB Asian Development Bank AFD Agence Française de Développement (French Development Agency) AfDB African Development Bank AGRITEX Agricultural Technical and Extension Services Ah ampere-hour AMADER Agency for the Development of Domestic Energy and Rural Electrification AMDA Africa Minigrid Developers Association ARE Alliance for Rural Electrification ATP ability to pay BAU business as usual BBBEE Broad-Based Black Economic Empowerment Act of 2003 BERC Bangladesh Energy Regulatory Commission BLS Bureau of Labor Statistics BMZ Federal Ministry for Economic Cooperation and Development BNEF Bloomberg New Energy Finance BOO build-own-operate BRD Development Bank of Rwanda CAPEX capital expenditure CBS Central Bureau of Statistics CE community engagement CEDECAP Centre of Demonstration and Qualification in Appropriate Technologies CELAMeD community engagement, load acquisition and micro-enterprise development CIESIN Center for International Earth Science Information Network CO2 carbon dioxide cofin. cofinanced COGS cost of goods sold COP Conference of the Parties CPI Climate Policy Initiative CREDA Chhattisgarh State Renewable Energy Development Agency CSCs customer service centers DC direct current MINI GRIDS FOR HALF A BILLION PEOPLE    xix DISCO distribution company DRC Democratic Republic of Congo E4I Energy4Impact EAC Electricity Authority of Cambodia ECOWAS Economic Community of West African States ECREEE ECOWAS Centre for Renewable Energy and Energy Efficiency ECS electricity consumer society EDC Electricité du Cambodge (Cambodia Electricity) EDC Enterprise Development Cambodia EPC engineering, procurement, and construction company EEP Energy and Environment Partnership e-MPF European Microfinance Platform ERC Energy Regulatory Commission ESCO energy service company ESIA Environmental and Social Impact Assessment ESMAP Energy Sector Management Assistance Program ESMP Environmental and Social Management Plan ESMS Environmental and Social Management System EUEI PDF European Union Energy Initiative Partnership Dialogue Facility EWURA Energy and Water Utilities Regulatory Authority FCDO Foreign, Commonwealth and Development Office FCV fragility, conflict, and violence FDoE Fiji Department of Energy GEM Global Entrepreneurship Monitor GFMG Global Facility on Mini Grids GIS geographic information system GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit (German Society for International Cooperation) GMG Green Mini-Grid Help Desk GoP Government of Peru GVE Green Village Electric GW gigawatt GWh gigawatt-hours GWp gigawatts-peak ha hectare HOMER Hybrid Optimization of Multiple Energy Resources HR human resources HRSL High Resolution Settlement Layer HV high voltage ICT information and communications technology IDCOL Infrastructure Development Company Limited IDS Institute of Development Studies IEA International Energy Agency xx   MINI GRIDS FOR HALF A BILLION PEOPLE IEEE Institute of Electrical and Electronics Engineers IEG Independent Evaluation Group IFC International Finance Corporation ILO International Labour Organization IMF International Monetary Fund IPCC Intergovernmental Panel on Climate Change IPP independent power producer IQR interquartile range IRENA International Renewable Energy Agency IUCN International Union for Conservation of Nature JPS Jamaica Public Service Company KBA key biodiversity area km kilometer KPI key performance indicator KPLC Kenya Power and Lighting Company kVA kilovolts-ampere kW kilowatts kWfirm kilowatts of firm alternating current output kWh kilowatt-hour kWp kilowatts-peak L liter LCOE levelized cost of energy LCOS levelized cost of storage LED light-emitting diode LF load factor LGA local governmental area li lithium LRP Livelihood Restoration Plan LV low voltage M million MAS Multi-Agent System MDG Millennium Development Goal MFD Maximizing Finance for Development MG mini grid MGA Micro-Grid Academy mi mile MIA Microgrid Investment Accelerator MIT Massachusetts Institute of Technology MLIP Ministry of Labour, Immigration and Population MOHP Ministry of Health and Population MOI Ministry of Infrastructure MRG Minimum Revenue Guarantee MINI GRIDS FOR HALF A BILLION PEOPLE    xxi MTF Multi-Tier Framework MV medium voltage MW megawatt MWh megawatt-hour n.a. not applicable NAPTIN National Power Training Institute of Nigeria n.d. no date NEP Nigeria Electrification Project NERC Nigerian Electric Regulatory Commission NGO nongovernmental organization NIS National Institute of Statistics NPC-SPUG National Power Corporation Small Power Utility Group NREL National Renewable Energy Laboratory (US) O&M operations and maintenance ODI Overseas Development Institute OECD Organisation for Economic Co-operation and Development OEM original equipment manufacturer OPEX operational expenditure OSM OpenStreetMap OUR Office of Utilities Regulation P2PB peer-to-peer business PAYG pay-as-you-go PDR (Lao) People’s Democratic Republic PFI participating financial institution PPA power purchase agreement PPP public-private partnership PRG partial risk guarantee PU productive use PUE productive uses of energy PV photovoltaic PwC PricewaterhouseCoopers RAP Resettlement Action Plan RE renewable energy REA Rural Electrification Agency (Nigeria) REA Rural Energy Agency (Tanzania) REAs rural electrification/energy agencies REF Rural Electrification Fund (Cambodia) REF Renewable Energy Fund REM Reference Electrification Model RENAC Renewables Academy REopt Renewable Energy Integration and Optimization RES4Africa Renewable Energy Solutions for Africa xxii   MINI GRIDS FOR HALF A BILLION PEOPLE RISE Regulatory Indicators for Sustainable Energy RMI Rocky Mountain Institute RoE return on equity SAIDI System Average Interruption Duration Index SAIFI System Average Interruption Frequency Index SEforALL Sustainable Energy for All SDG Sustainable Development Goal SDG7 Sustainable Development Goal 7 SG&A selling, general, and administrative Sh shilling SI system integrator SIDA Swedish International Development Cooperation Agency SM smart meter SNV Netherlands Development Organisation SPD small power distributor SPI Smart Power India SPM Smart Power Myanmar SPP small power producer TANESCO Tanganyika Electric Supply Company TARA Society for Technology and Action for Rural Development TASF Transaction Advisory Services Facility TTA Trama TechnoAmbiental TWp terawatts-peak UCS Union of Concerned Scientists UN United Nations UNDP United Nations Development Programme UNIDO United Nations Industrial Development Organization UNOPS United Nations Office for Project Services USAID United States Agency for International Development USD United States dollar USDA United States Department of Agriculture US DOE United States Department of Energy USG United States Government VEC Village Electrification Committee VIA Village Infrastructure Angels W watt WB World Bank Wfirm watt of firm alternating current output Wdc watt of direct current Wp watts-peak WTP willingness to pay All dollar amounts are US dollars unless otherwise indicated. MINI GRIDS FOR HALF A BILLION PEOPLE    xxiii MINI GRIDS BY THE NUMBERS Where we are today 48 million people connected to 21,500 mini grids, of which half are solar PV, at an investment cost of $29 billion. 29,400 mini grids planned, 95 percent of them in Africa and South Asia, 99 percent solar PV, connecting more than 35 million people at an investment cost of $9 billion. Where we need to be to reach universal access by 2030 490 million people served at least cost by 217,000 mini grids, almost all solar-powered, requiring an investment of $127 billion. To deploy mini grids at scale, countries must act on 10 Building Blocks: (1) reducing costs and optimizing design & innova- tion for solar mini grids; (2) planning national strategies and developer portfolios with geospatial analysis and digital platforms; (3) transforming productive livelihoods and improving business viability; (4) engaging communities as valued customers; (5) delivering services through local and international companies and utilities; (6) financing solar mini grid portfolios and end user appliances; (7) attracting exceptional talent and scaling skills development; (8) supporting institutions, delivery models, and champions that create opportunities; (9) enacting regulations and policies that empower mini grid companies and customers; (10) cutting red tape for a dynamic business environment. Regional mini grid trends from ESMAP’s database of more than 50,000 mini grid projects in 138 countries Top 5 countries . . . INSTALLED PLANNED INSTALLED PLANNED (mostly first- and second- (mostly third-generation (mostly first- and sec- (mostly third-generation generation mini grids) mini grids) ond-generation mini grids) mini grids) 9,600 South Asia 19,000 South Asia 4,700 Afghanistan 18,900 India 7,200 E ast Asia and Pacific 800 East Asia and Pacific 4,000 Myanmar 2,700 Nigeria 3,100 Africa 9,000 Africa 3,200 India 1,500 Tanzania 1,200 OECD and Central Asia 400 OECD and Central Asia 1,500 Nepal 1,200 Senegal 300 Other 100 Other 1,200 China 600 Ethiopia Current financing Top 3 private-sector developers By installed $29 billion—Cumulative global investment in mini and planned mini grids grids to date 1. Tata Power Renewable Microgrids (10,000 / India) 2. Husk Power (5,000 / India & Africa) $9 billion—Cumulative global investment in Africa 3. OMC Power (5,000 / India) and South Asia in mini grids to date Top 3 utilities By installed and planned mini grids $2.6 billion—Development Partners committed, including AFD, AfDB, FCDO, the Islamic Development 1. RAO (700 / Russia) Bank, GIZ and the World Bank, among others 2. PT Perusahaan Listrik Negara (500 / Indonesia) 3. NPC-SPUG (300 / Philippines) $1.4 billion—World Bank commitment to mini grids in 31 countries through 2027 $500+ million—Private-sector investment in mini Private-sector opportunity grid developers in low-income countries since 2013 $3.3 billion annual profit potential for developers across all mini grids deployed through 2030 25 percent—Average World Bank share of total mini grid investment (government, development $5.8 billion net profit potential across all mini grid partners, and private sector) in client countries component and service suppliers in 2030 alone xxiv   MINI GRIDS FOR HALF A BILLION PEOPLE MINI GRIDS BY THE NUMBERS, continued Cost of a best-in-class Cost of unsubsidized . . . Compared with solar-hybrid mini grid electricity from a best- utilities in Africa today . . . . . . and by 2030 in-class solar hybrid $0.27/kWh average mini grid . . . across 39 utilities $3,659/kWfirm total <$2,500/kWfirm capital expense total capital expense $0.38/kWh (LCOE) 2 of 39 utilities with baseline today cost-recovery tariffs $596/kWp Solar PV $290/kWp Solar PV Module Module $0.28/kWh with income- generating machines to $297/kWh Lithium-ion $137/kWh Lithium- achieve 40 percent load factor batteries ion batteries $0.20/kWh with $265/kW battery income-generating machines inverter and expected 2030 costs Income-generating machinery 3rd generation mini . . . compared with grid service . . . typical utilities < 12 months payback period for more than 130 income-generating machines and 99 percent uptime 40–50 percent other equipment available today uptime Tier 4–5 access 84/100 customer Tier 3–4 access $3.6 billion microfinance needed for 3 million satisfaction rate 41/100 customer machines and other equipment connected satisfaction rate to third-generation mini grids in 2030 Environmental impact Typical third-generation mini grid 10–15 GW solar PV installed by 2030 $0.5– $1.0 million investment 50–110 GWh batteries mostly lithium-ion 200–800 clients connected 60 percent energy savings from energy efficient 800–4,000 people receiving electricity for the first appliances time 1.2 billion tonnes of CO2 emissions avoided 50–100 kWp solar PV installed 200–500 kWh batteries installed What is a mini grid? Mini grids are electric power generation and distribution systems that provide electricity to just a few customers in a remote settlement or bring power to hundreds of thousands of customers in a town or city. They can be fully isolated from the main grid or connected to it but able to intentionally isolate (“island”) themselves from the grid. Mini grids supply power to households, businesses, public institutions, and anchor clients, such as telecom towers and large agricultural processing facil- ities. They are designed to provide high-quality, reliable electricity. A new, “third generation” of mini grids has recently emerged. They incorporate the latest technologies, such as smart meters and remote monitoring systems; and are typically designed to interconnect with the main grid. To be considered in our analysis in the context of this report, a mini grid had to serve multiple customers. Electricity systems that service a single hospital, industrial facility, military base, university campus, mine, or other single entity, were therefore not considered mini grids. We also do not define mini grids in terms of size, although in our detailed analysis of mini grid costs and in our global database of more than 50,000 mini grid projects, the vast majority (90 percent) ranged from 10 kW to 1 MW in installed capacity. Sources and underlying analysis for the figures above are presented throughout the book. MINI GRIDS FOR HALF A BILLION PEOPLE    xxv MINI GRIDS BY THE NUMBERS, continued Key performance indicators for the mini grid industry 2018 2021 2025* Reducing cost (levelized cost of energy [$/kWh] of a best-in-class solar $0.55/kWh $0.38/kWh $0.30/kWh hybrid mini grid) Pace of deployment (mini grids built per key access-deficit country 20–75 150 450 per year) mini grids mini grids mini grids Quality of service (industry-wide standard for reliability of electricity 90–97 percent 99 percent 99 percent supply) uptime uptime uptime Access to finance for mini grids designed to boost access to energy $13 billion $16 billion $25 billion (total cumulative investment) Establish enabling environments (average RISE score for mini grids 59/100 64/100 75/100 framework in top 20 electricity access-deficit countries) Note: * projection with business-as-usual scenario. Mini grid industry progress across all 10 frontiers / building blocks 2018 2021 2025* Reducing costs and optimizing design and innovation for solar mini grids Planning national strategies and developer portfolios with geospatial analysis and digital platforms Transforming productive livelihoods and improving business viability Engaging communities as valued customers Delivering services through local and international companies and utilities Financing solar mini grid portfolios and end user appliances Attracting exceptional talent and scaling skills development Supporting institutions, delivery models, and champions that create opportunities Enacting regulations and policies that empower mini grid companies and customers Cutting red tape for a dynamic business environment Note: * projection with business-as-usual scenario. Dark green = magnitude change has been achieved; light green = irreversible progress towards magnitude change; yellow = needing attention; orange = no significant activities to date. xxvi   MINI GRIDS FOR HALF A BILLION PEOPLE MAIN FINDINGS THE NEW ELECTRICITY ACCESS with 2,500 to 10,000 residents. Finally, for nearly 3,000 settlements, each with 10,000 to 100,000 people, custom LANDSCAPE sizing of mini grids might be more suitable. To achieve Sustainable Development Goal 7 (SDG 7), Internal analysis by the World Bank team based on Multi- 930 million people will have to obtain an electricity con- Tier Framework (MTF) data suggests that users in these nection between 2022 and 2030 (IEA 2021). In 2020, load centers spend on average $5–$20 per month on the global electrification rate reached 91 percent, with alternative forms of energy such as candles, kerosene, the number of people without access dropping to around dry-cell batteries, car batteries, and petrol and diesel fuel 733 million—compared with around 1 billion people in for stand-alone gensets. The introduction of innovative 2016 and 1.2 billion in 2010 (IEA, World Bank, and others technologies in the marketplace (like solar home systems 2022). Nonetheless, the pace of electrification has slowed or mobile phones) has taught us that these new solutions in recent years. Between 2010 and 2018, an average of 130 need to be more than a little better than the current alter- million people gained access to electricity annually. From native. They need to be much better. Why else would con- 2018 to 2020, this number shrank to 109 million per year. sumers take the risk of changing their behaviors? For these While the slowdown is attributed in part to the difficulties in clusters of clients, the service provided by the solar mini reaching the remotest and most vulnerable populations, it grids should be a reliable source for their consumptive activ- was compounded by the devastating effects of the COVID- ities like lighting, charging, and radio/TV. More than that, 19 pandemic. If current policies and efforts are not ramped they need to provide for life-changing productive activities up, only 260 million people are anticipated to be electrified within the current monthly expenditure of $5–$20. From between now and 2030 (IEA 2021), and an estimated 670 the end user’s perspective, a $5–$20 monthly expenditure million people are projected to remain without access, with should cover the cost not only of reliable electricity but also 9 out of 10 of them likely to live in Sub-Saharan Africa (IEA, of transitioning to (and purchasing) electric appliances. So World Bank, and others 2022). over the lifetime of the technology, monthly payments of In Sub-Saharan Africa, nearly 291,000 population clus- about $3–$15 cover the cost of electricity, while monthly ters have profiles favoring the deployment of solar payments fall in the range of $2–$5 for appliances. These mini grids. That is, they are located more than 1km from costs pose a challenge for the mini grid industry if it is to the existing grid network and have a population density fulfil its full market potential. (>1,000 people/km2) that favors decentralized sys- Countries with a comprehensive approach involving tem deployment. More specifically, analysis conducted main grid extensions, mini grids, and solar home sys- internally by the World Bank team—based on spatial tems have achieved the fastest results in electricity distribution of digitalized settlements (GRID3, CIESIN), access (IEA, World Bank, and others 2022). Strong lead- grid network (Arderne C. et al—GridFinder) and popula- ership, supporting policies, and more private financing will tion (WorldPop) over the region—shows that more than be required if electricity access is to reach the remaining 177,000 settlements have a population of 100 to 500 unserved people—including those that depend on frail, people. These settlements could be powered by smaller overburdened urban grids and displaced people and those solar mini grids of up to 20 kilowatts (kW) each. Nearly living in hard-to-reach locations. In Sub-Saharan Africa, 96,000 settlements, each with populations of 500 to electricity services are delivered to end users by 60 utili- 2,500 people, could be powered by medium-sized solar ties, more than 80 solar mini grid companies, and almost mini grids of up to 80 kW. The larger solar mini grids, up to 90 main solar stand-alone-system companies (Balaban- 200 kW, could power more than 15,000 settlements, each yan and others 2021; GOGLA 2022). MINI GRIDS FOR HALF A BILLION PEOPLE    1 Mini grids are not a new phenomenon: nearly all cen- either because the main grid is unreliable or to boost resil- tralized electricity grid systems began as isolated mini ience in the face of climate shocks or severe weather. With grids that were connected to each other over time. more than 160,000 mini grids needed, Sub-Saharan Africa This first generation of mini grids was pivotal to the early accounts for the largest share of mini grids and investment development and industrialization of most modern econ- required to achieve universal access, at a cost of $91 billion omies, including Brazil, China, Denmark, Italy, the Nether- to connect 380 million people. These projections are based lands, Spain, Sweden, the United Kingdom, and the United on country-specific scenarios for the 58 countries with data States. Mini grid systems introduced in the late nineteenth in the Global Electrification Platform (GEP) and ESMAP and early twentieth centuries can be described as the first estimates for countries not included in the GEP. Meanwhile, generation of mini grids. Today a second generation of mini ESMAP estimates that resilience- and renewable-moti- grids is widespread in many low-income countries. These vated mini grids could serve an additional 2–3 million new systems are typically small and isolated, powered by diesel connections globally (serving 6–7 million people) per year, or hydro, and built by local communities or entrepreneurs or the equivalent of 10–15 cities or small regional utilities primarily to provide rural households with access to elec- per year deciding to strengthen their power systems by tricity, especially in areas not yet served by the main grid. developing interconnected micro/mini/metro grids. Tens of thousands of these systems were built, starting in In 2021, the global mini grid market consisted of more the 1980s and ramping up through the 1990s and early than 50,000 installed and planned mini grids in more 2000s. Many of these systems were overtaken by the than 130 countries. Although there is a clear trend toward national grids; the ones that still exist are now prime can- solar as the dominant technology, the overall pace of mini didates for hybridization with solar photovoltaic (PV) sys- grid development is not on track to achieve the 2030 mini tems to reduce the fuel cost. grid market potential. ESMAP identified 21,557 mini grids in Over the past few years, a third generation of solar mini 131 countries and territories, serving more than 48 million grids has emerged. These mini grids, mostly solar PV people. Most of these systems are first- and second-gen- hybrids, are owned and operated by private companies eration mini grids, and approximately half of installed mini that leverage transformative technologies and innovative grids are powered by solar, with hydro and fossil fuels strategies to build portfolios of mini grids instead of one-off accounting for an additional 35 percent and 10 percent, projects. The typical third-generation mini grid is grid-in- respectively. Another 29,353 mini grids are planned for terconnection ready. It also uses energy management development in 77 countries and territories, of which 99 systems, prepay smart meters, and the latest solar hybrid percent will be powered by solar. The trend toward solar has technologies. This third-generation mini grid also incor- been accelerating: more than 10 times as many solar mini porates energy-efficient appliances for productive uses of grids were built per year from 2016 to 2020 than fossil fuel electricity into its business model. These mini grids operate mini grids. Meanwhile, from 2010 to 2014, by comparison, in more favorable business environments, taking advan- about three times as many solar mini grids were built per tage of cost reductions in the latest mini grid component year than fossil fuel mini grids. This is a major acceleration technologies and regulations developed specifically for in solar and deceleration in fossil fuels. But the annual pace private-sector investment. Developers of third-generation of mini grid development worldwide—averaging between mini grids are joining industry associations to speak with 1,300 and 1,900 between 2010 and 2021—would see one voice and drive policies and regulations that favor pri- only 44,800 mini grids serving 80 million people at a total vate-sector investment. investment cost of $37 billion by 2030. This is well short of the 430 million people that could be served at least cost by The Energy Sector Management Assistance Program mini grids in order to achieve universal access. (ESMAP) analysis indicates that a combination of fall- ing costs for key components, the introduction of new Year-on-year gains needed to achieve universal access digital solutions, and early signs of favorable economies will require scaling up private-sector-led mini grid deploy- of scale, has made solar mini grids an option to connect ments from tens to hundreds to thousands of mini grids 490 million people by 2030. Achieving universal access per country per year in each of the top 20 countries with to electricity will require the construction of more than the highest electricity access deficit rates today. Exam- 217,000 mini grids by 2030 at a cumulative investment ples showing this exponential growth are with the introduc- cost of almost $127 billion. Of these totals, mini grids are tion of mobile phones, solar home systems, and electric the least-cost option for 430 million people who would gain vehicles, where the private sector, supported by public poli- access to electricity for the first time at a cost of about $105 cies, provides superior products. Does this mean that more billion. For about $22 billion, an additional 60 million peo- public-sector-led programs cannot be beneficial? Not at ple, mostly in middle- and high-income countries, could be all: these programs provide great benefits to the coun- serviced through an interconnected network of mini grids try and the end users. Yet when one must attain universal 2   MINI GRIDS FOR HALF A BILLION PEOPLE access by 2030, private-sector-led programs should be the The industry is ahead or on track to achieve most of the dominant initiative, across the board, in a country or region. key performance indicators (KPIs), but it lags in terms of number of mini grids installed per key energy access Overarching sector performance indicators and targets deficit country per year, and total cumulative invest- can help benchmark the sector. It is within this context ment. The overall cost of the delivery of electricity services of market dynamics as well as through a collaborative, by mini grids has plunged since 2018, from a levelized cost iterative process, that ESMAP and mini grid industry lead- of energy (LCOE) of a best-in-class solar hybrid mini grid ers—including the Africa Minigrid Developers Association equal to $0.55/kilowatt-hour (kWh) to $0.38/kWh in 2021, (AMDA) and development partners—jointly identified five compared with the $0.45/kWh target for 2021. The quality market drivers and associated targets that will set the sec- of mini grid electricity services is also ahead of pace, with tor on a trajectory to achieve universal electrification and AMDA members achieving uptimes of around 99 percent in its full market potential (table MF.1). 2021 compared with 90–97 percent in 2018 and the 2021 These targets are ambitious but achievable if 10 building objective of 97 percent uptime. A continuation of delivery blocks are in place at the national level. Looking through of high-quality services has improved the average load fac- the lens of innovation and the impact that, for example, dig- tor, from around 22 percent in 2018 to 30 percent in 2021, itization and technological advancement can bring to the ahead of the 2021 objective of 25 percent. Enabling envi- solar mini grid sector, our analysis identified 10 areas that ronments have also improved and are ahead of pace, with stood out where notable magnitude-level improvements the Regulatory Indicators for Sustainable Energy (RISE) can be expected to reach the abovementioned targets of score for mini grids in the top 20 access-deficit countries cost (C), pace (P), quality (Q), finance (F), and enabling rising from 59/100 in 2018 to 64/100 in 2021, on pace to environment (EE). These are identified in table MF.2. achieve 90/100 by 2030, compared with 80/100 as the TABLE MF.1 • Market drivers and 2030 targets Market Driver 2030 Target 1. Reduce the cost of solar hybrid mini grids. $0.20/kilowatt-hour (kWh).  ncrease the pace of deployment through a portfolio 2. I Building around 2,000 projects per key access-deficit country per year approach to mini grid development. by 2030. 3. Provide superior-quality service. Achieving industrywide average uptime of more than 97 percent and industrywide average load factor of 45 percent.  everage development partner funding and government 4. L Attracting approximately $127 billion of investment from development investment to “crowd in” private-sector finance. partners, governments, and the private sector, of which $105 billion for energy access mini grids.  stablish enabling mini grid business environments in 5. E Raising the average Regulatory Indicators for Sustainable Energy (RISE) key access-deficit countries. score in the top 20 electricity-access-deficit countries to 80 out of 100. TABLE MF.2 • Building blocks and identified areas for potential magnitude change Building Blocks / Identified Areas for Potential Magnitude Change Progress 2018–21 2. Planning national strategies and developer portfolios with geospatial analysis and digital platforms (C, P, F, EE) 1. Reducing costs and optimizing design and innovation for solar mini grids (C, P, Q) 9. Enacting regulations and policies that empower mini grid companies and customers (C, P, EE) 5. Delivering services through local and international companies and utilities (C, P, Q, F) 6. Financing solar mini grid portfolios and end-user appliances (P, F, EE) 8. Supporting institutions, delivery models, and champions to create opportunities (C, P, EE) 3. Transforming productive livelihoods and improving business viability (C, Q) 4. Engaging communities as valued customers (C, P, Q) 7. Attracting exceptional talent and scaling skills development (C, P, Q, F, EE) 10. Cutting red tape for a dynamic business environment (C, P, EE) Dark green = magnitude change has been achieved; light green = irreversible progress toward magnitude change; yellow = needs attention. C= cost; EE = enabling environment; F = finance; P = pace; Q = quality. MINI GRIDS FOR HALF A BILLION PEOPLE    3 2030 target. But while the industry has seen a shift from ments. The most progress has been made in geospatial the deployment of mini grids on an individual pilot basis to planning and the costing, design, and innovation of solar their deployment by service providers in portfolios of 5 to hybrid mini grids. The most work, however, is needed to sup- 10 mini grids per month in 2021, the pace across the indus- port these key countries and regions, and others, in commu- try is still behind the 200 mini grids per country per year in nity engagement, scaling up private sector participation and 2021 needed to be on track for achieving 2,000 mini grids utilities to deploy mini grids, and skills development. per country per year in 2030. In addition, while cumulative investment in mini grids for energy access rose from about $13 billion in 2018 to $16 billion in 2021, this is well behind BUILDING BLOCKS 1 THROUGH the $20 billion needed to be on pace to achieve $105 billion cumulative investment in energy access mini grids by 2030. 10 AND FRONTIERS FOR SECTOR GROWTH: Creating the Environment Over the past three years, notable progress has been evident across all building blocks. Most of the advances for Takeoff of Mini Grid Portfolios were seen in the geospatial portfolio planning and workable Building blocks 1 through 10 and the frontiers for sector regulations. Solid progress has also been evident in tech- growth are described in the sections that follow. nology and costing in the private sector and utilities, along with greater access to finance and supporting institutions. BUILDING BLOCK 1. Even though we see advances with productive uses and SOLAR MINI GRID COSTS, DESIGN, AND community involvement (and on attracting exceptional tal- INNOVATION ent and reducing red tape), these are the very areas where transitions must emphasize scale on the ground. For more Solar mini grids consist of specialized components for details on progress by building block, see each of the chap- the generation, distribution, metering, and consumption ters covering these topics. of electricity (figure MF.1). A typical third-generation mini grid comprises a solar hybrid generation system made A cohort of partners has begun to track building blocks up of solar panels, batteries, charge controllers, inverters, by country for in-country coordination and readiness and diesel backup generators. The distribution network to scale. For key access-deficit countries and regions in consists of poles and low-voltage wires; larger mini grids Sub-Saharan Africa, we see that Ethiopia, Kenya, and Nige- sometimes also have medium-voltage systems. Third-gen- ria have taken more steps toward achieving magnitude eration mini grids often use smart meters offering both changes across the 10 building blocks than the other coun- prepaid payment options for consumers and real-time, tries and regions (table MF.3). The Democratic Republic of granular information about energy consumption patterns Congo and the Sahel, despite being large potential markets and system performance. They also use remote-moni- for mini grids, need major support on almost all 10 building toring systems that allow operators to identify technical blocks to prepare the market for scaling up mini grid deploy- TABLE MF.3 • Sub-Saharan African mini grid markets and their progress across the 10 building blocks Building block DRC Ethiopia Kenya Nigeria Sahel 1. Costing, design, and innovation 2. Geospatial planning 3. Income-generating appliances and machines 4. Community engagement 5. Companies and utilities 6. Access to finance 7. Skills development 8. Institutional setup and business models 9. Regulations and policies 10. Cutting red tape Source: ESMAP analysis. Dark green = magnitude change has been achieved; light green = irreversible progress toward magnitude change; yellow = needs attention; orange = no significant activities to date. DRC = Democratic Republic of Congo. 4   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE MF.1 • A mini grid system (part A) and a containerized solar mini grid (part B) A AC load in village AC appliances Poletop hardware Smart meter Generator Distribution line AC bus Residential Smart meter AC/DC inverter Service drop Pole PV array Charge Battery block Commercial controller Smart Solar-hybrid generation system Distribution system meters Efficient productive loads AC = alternating current; DC = direct current; PV = photovoltaic. B solar battery system only, solar-biomass-based systems, AC (alternating current) or DC (direct current) systems, and high digital solution integration. The Levelized Cost of Energy of a Solar Mini Grid In 2021, the LCOE in a best-in-class, third-generation mini grid was $0.38/kWh at a 22 percent load factor, or a 31 percent reduction from 2018. This trajectory is fueled by the falling expenditures for preparation, capital, and Source: © SustainSolar. Used with permission by SustainSolar. Further operations, combined with more income-generating uses permission required for reuse. of electricity and more efficient economies of scale. The combination of expected cost reductions and higher load issues before they affect energy services and rectify prob- factors (from 22 percent to 40 percent) caused by produc- lems quickly and inexpensively, thus improving the quality tive use is expected to bring the LCOE of third-generation of customer service. Many developers of third-generation mini grids to $0.20/kWh by 2030 (table MF.4). mini grids encourage and incentivize customers to use effi- Preparation costs have been reduced by more than an cient household appliances as well as efficient machines order of magnitude to $2,300 per mini grid; however, and equipment for income-generating activities, and pro- there is limited progress in efficiency. The introduction of vide or facilitate access to financing options to help cus- geospatial and other digital technologies have decreased tomers manage upfront costs. the cost of preparation and planning by an order of magni- Smaller solar mini grids with an installed capacity of about tude. In the past, the unit cost per site was more or less the 100 kW or less are more and more standardized, ranging same, irrespective of the number of sites—about $30,000 from prefabricated components to containerized mini per site—because each one required a high level of on-site grids. Larger systems with an installed capacity of more analysis. Today, portfolios of mini grids can be prepared to than 250 kW remain designed and delivered on an individ- the point where they are ready for full feasibility assessment ual system basis. Irrespective of the installed capacity, ser- and community engagement at a cost of about $2,300 per vice providers have chosen their mini grid design around site in 2021, based on the World Bank’s recent experience one main technical and business approach, for example, in Ethiopia, Nigeria, and South Sudan. MINI GRIDS FOR HALF A BILLION PEOPLE    5 TABLE MF.4 • The levelized cost of energy by load second-generation mini grids today have a load factor of factor, 2018, 2021, and 2030 around 22 percent, indicative of low levels of income-gen- erating uses of electricity. However, third-generation mini Levelized cost of energy (US$/kWh) grids provide high-quality, reliable electricity services that Load factor (percent 2018 2021 2030 can support income-generating loads, such as agricul- 22 0.55 0.38 0.29 tural milling. If mini grids can achieve a 40 percent load 40 0.42 0.28 0.20 factor through strong daytime consumption by local busi- Source: ESMAP analysis. nesses and commercial clients, the costs of producing Note: The 2018 LCOE data are for a best-in-class 294-kWfirm solar hybrid electricity drop 25 percent compared with a load factor mini grid in Bangladesh serving more than 1,000 customers (more than of 22 percent. For an 80 percent load factor—achieved 5,000 people). LCOE data for 2021 are based on a representative mini by inclusion of a water pump with storage tank and an grid synthesized from average costs and consumption levels in three mini grids in Myanmar, Nigeria, and Ethiopia commissioned in 2020 or 2021. anchor load, such as a telecommunications tower—LCOE The 2030 LCOE is for a “best-in-class” mini grid based on projected com- reduction is 37 percent. ponent costs in 2030. A detailed description of the underlying analysis is provided in chapter 1. kWh = kilowatt-hour. Implications for national power sectors As a result of declining LCOE, increasing income-generat- ing uses of electricity, and the mainstreaming of geospa- A best-in-class solar hybrid mini grid costs about $3,700/ tial planning, solar mini grids can have transformational kWfirm,1 and the falling trend is expected to continue effects on power sectors. They are on track to provide through 2030, bringing capital expenditure (CAPEX) to power at lower cost than many utilities by 2030. At $0.40/ below $2,500/kWfirm. Components used for generating kWh, mini grid LCOE would be less than the LCOE of national and distributing electricity account for 66 percent of total utilities in 7 out of 39 countries in Africa. At $0.20/kWh, mini capital costs. The components with the largest share of grid LCOE would be less than the LCOE of national utilities in overall CAPEX were batteries (15 percent). PV modules 24 African countries (Trimble and others 2016). This would (10 percent), inverters/energy management systems (9 make mini grids the least-cost solution for grid-quality elec- percent), and distribution grids (poles, wires; 27 percent). tricity for more than 60 percent of the population in Africa in Meanwhile, component costs vary widely across countries a scenario assuming that national utilities do not dramati- and regions, mostly as a result of a combination of taxes cally change their operations—with major implications for and duties, differences in margins charged by wholesalers the allocation of both public and private investment funds. and distributors, and other costs incurred in doing busi- However, scaling up mini grids does not mean scaling back ness that vary from country to country. Downward trends the main grid. On the contrary, solar mini grids enhance the in component costs mean that the up-front investment economic viability of expanding the main grid. By designing cost of solar and solar hybrid mini grids fell from about the system from the beginning to interconnect with the $8,000–$10,000/kWfirm in 2010 to $3,900/kWfirm in main grid and by promoting income-generating uses of 2018 and less than 3,700/kWfirm in 2021. Looking ahead, electricity through effective community engagement and the expected decreases in component costs associated training, third-generation mini grids can provide early eco- with current best practices can reduce up-front investment nomic growth, so that significant load already exists by the costs to less than $2,500/kWfirm by 2030. time the main grid arrives, and customers have a greater Mini grid operating expenditure (OPEX) averages around ability to pay. New regulatory frameworks give developers $80 per customer per year. Costs are expected to decline viable options for what happens when the main grid arrives, because of technological advances over the next decade. and reductions in the cost of components enable develop- Staff costs on average account for 76 percent of operations ers to build grid-interconnection-ready systems, while still costs, but economies of scale and new remote-controlled, keeping tariffs affordable. prepay smart meters and remote-monitoring technologies Supporting solar mini grids therefore goes hand in hand have slashed labor costs per mini grid. Replacement costs with strengthening the power sector. Interconnecting have also fallen as more developers invest in lithium-ion third-generation mini grids with the main grid can increase (Li-on) batteries, which have about twice the number of the resource diversity and overall resilience and efficiency charging cycles before failure compared with conventional of the power system. However, this presents a couple of lead-acid batteries, and the costs of power electronics, such operational challenges that are better addressed in a com- as PV inverters and battery inverters, are also decreasing. prehensive strategy for developing the sector, for example, Further cost reductions per kWh are derived from for governments through their electrification strategies increasing income-generating uses of electricity, which to allow for utilities, mini grids, and off-grid companies to can decrease the LCOE by 25 percent or more. Most deliver services in the country, as well as for utilities to be 6   MINI GRIDS FOR HALF A BILLION PEOPLE able to introduce the practical technical functions to sup- turning to implementation of these plans. For example, sev- port power system operations and planning with multiple eral national electrification plans have incorporated buffer mini grids connected to the distribution grid, such as short- zones for grid extension (for example, 15 kilometers from and long-term forecasting and other procedures. the existing grid), which during implementation has created situations where the utility is delayed in certain geographical Experiences with interconnected mini grid collabora- areas and the mini grid and solar companies are not allowed tions are emerging, for example, in Nigeria and India, to sell products and services because they are prohibited to and are providing valuable lessons. These interconnected do so under the plan. It is important that during the prepa- mini grids are built to serve different market segments: ration of these plans, the different stakeholders are carefully rural and peri-urban towns and villages, large urban mar- consulted so that the underlying assumptions are based in ketplaces, commercial and industrial (C and I) installa- reality. Countries that are using advanced geospatial analy- tions, and separate urban residential communities. Early sis to develop national electrification plans include Angola, evidence seems to indicate that these interconnected mini Cambodia, the Democratic Republic of Congo, Ethiopia, The grids can create “win-win-win” economic outcomes for the Gambia, Haiti, India, Kenya, Liberia, Mozambique, Myanmar, three key parties. The arrangement can eliminate or reduce Nigeria, Pakistan, Rwanda, Sierra Leone, Somalia, Tanzania, financial losses for distribution companies (DISCOs) that Togo, Uganda, and Zambia, among others. are forced to sell electricity at non-cost-recovering retail tariffs. Interconnections also allow DISCOs to earn new Geospatial analysis is also being used as part of a port- revenues through bulk power sales to the mini grid as well folio planning approach for mini grid development. This as rental revenues from the leasing of some or all of the would complement a comprehensive national least-cost DISCO’s existing distribution system to the mini grid. For electrification planning framework or, in the absence of such the mini grid operator, a physical connection to the contig- a framework, identify portfolios of mini grid sites where grid uous DISCO offers the possibility of purchasing bulk power, extension is expected to be limited or unlikely because of whether on a firm or an “as available” basis from the inter- political considerations, insolvency of the DISCOs, and so connected DISCO or an upstream supply source. This can forth. Geospatial portfolio planning, which is already being lead to lower operating and capital costs (for example, a used by a number of established mini grid companies, util- lower LCOE) for the interconnected mini grid than if oper- ities, and governments in Sub-Saharan Africa, slashes the ates in a pure stand-alone mode. And for the mini grid’s preinvestment cost associated with preparing sites for mini customers, this should lead to lower tariffs than would be grid development compared with traditional approaches, possible than if the mini grid operated in a totally isolated which rely on the deployment of multidisciplinary teams to mode. Finally, it is well documented that mini grids, whether villages to explore the scope for mini grid electrification. interconnected or isolated, routinely achieve high levels of reliability for their customers than DISCOs do for theirs BUILDING BLOCK 3. (Tenenbaum, Greacen, and Shrestha 2022 forthcoming). Transforming productive livelihoods and improving business viability BUILDING BLOCK 2. Because of their reliability, third-generation mini grids Planning national strategies and developer can support income-generating uses of mini grid elec- portfolios with geospatial analysis and digital tricity, which creates an everyone-wins scenario for mini platforms grid developers, rural entrepreneurs, communities, and Countries are using geospatial analysis to develop national utilities over time. Increasing income-generating national electrification plans that delineate areas for uses of electricity reduces the LCOE (see table MF.4), which mini grids. Through a geospatial approach to national elec- increases the developer’s margins and therefore financial trification planning, the existing grid network is mapped and viability. Entrepreneurs and small businesses benefit from its attributes are digitalized. The supply of and demand for switching from expensive diesel generators to affordable electricity are geolocated and overlaid with supporting data, mini grid electricity. In one of the most comprehensive including demographic, social infrastructure, and economic assessments of productive-use appliances and equipment data. Spatial modeling then delivers a least-cost plan that to date, ESMAP identified more than 130 machines and identifies the optimal ranges for grid, mini grid, or off-grid appliances available today that had payback periods of less technologies. Even though these national plans now typically than 12 months. Communities benefit from the jobs cre- include all options for electrification—grid extension, mini ated and increased economic activity. The growth of rural grids, and off-grid—they still rely on chosen input assump- economies also benefits national utilities once intercon- tions that can result in a more advantaged position of one nection to the main grid is considered, because it increases solution over the other. Furthermore, the chosen param- customers’ ability to pay higher tariffs and creates a strong eters can also exclude large groups of customers when base of demand for electricity. MINI GRIDS FOR HALF A BILLION PEOPLE    7 But demand uncertainty remains a key area of risk for end user financiers, and mini grid companies. Road shows both developers and financiers. In ESMAP surveys of are the next step, where mini grid developers, appliance mini grids presented in this book—the detailed survey of suppliers, end user financiers visit load centers to explain more than 400 mini grids in Africa and Asia (see chapter the value propositions to potential end users. The final step 1), the high-level survey of installed and planned mini grids is the roll-out of mini grid connections, sales of appliances globally (see the overview), and the detailed nationally and end user finance. representative surveys of mini grid operators (see chap- A number of digital tools are emerging that also allow for ter 5)—demand per customer varied widely from one mini a more efficient and lower-cost planning and rollout of grid to another, and from one country to another. Most mini productive uses activities in conjunction with the arrival grids had demand per customer of between 5 and 35 kWh of electricity from mini grids. The above-mentioned geo- per month, but all else held equal there is a sevenfold differ- spatial tools that help to identify and prioritize mini grid ence in revenue expectations from customers consuming portfolios are now also used to share the associated mar- 5 kWh per month and those consuming 35 kWh per month. ket intelligence with the appliance providers and end user Developers use demand estimates as key inputs not only financiers. The information supports these companies to to inform the designs of their mini grids, but also to secure make an informed decision if their products have a suf- external financing. The uncertainties around future demand ficient addressable market. The tools also support the growth therefore represent a key risk area for both develop- mini grid developers, appliance providers, and end user ers and financiers. Mitigating this risk requires concerted financiers to coordinate visits to these communities, so efforts to increase the daytime use of income-generating that their collective, potential clients can learn how elec- appliances and machines, and financing mechanisms that tricity with the appropriate, affordable appliances can help de-risk some of the demand uncertainty. alter their lifestyle and business prospects for the better. Increasing the uptake of productive-use equipment Often nongovernmental organizations and social change requires access to approximately $3.6 billion in afford- organizations operate in these peri-urban, rural areas and able consumer finance and a proactive involvement of can play an important role in coordinating these efforts these financiers and appliance providers. Assuming an on the ground. average up-front cost of $1,200 and 15 appliances per mini grid for 200,000 new mini grids by 2030, approx- BUILDING BLOCK 4. imately $3.6 billion in microfinance will be needed for Engaging communities as valued customers the purchase of 3 million productive-use appliances by Community engagement strategies can help increase 2030. Although they have relatively high up-front costs, productive uses of electricity and stimulate demand for most productive-use appliances and equipment provide mini grid services. Experience from successful mini grid opportunities to generate or increase revenue. Financing developers indicates that community engagement begins the up-front purchase cost of the appliances—by the mini by raising awareness before moving to adoption, produc- grid operator via on-bill financing or by a third party, such tive operation, and word-of-mouth marketing. Community as a microfinance organization—is a good way to increase engagement requires a flexible approach; a clear under- productive uses of mini grid electricity. Both financing standing of the local socioeconomic and cultural charac- pathways have benefits and drawbacks for the mini grid teristics; and tailoring of promotional tools, materials, and operator, and both require the operator to develop new channels.2 business model capabilities. The benefits of prioritizing access to female-led house- Drawing from existing research and the World Bank’s holds and small businesses and increasing the partici- recent experience with productive uses programs across pation of women in management positions in mini grid Africa, Asia, and Latin America and the Caribbean, businesses are clear. Mini grids can greatly boost women’s ESMAP has identified six steps to roll out initiatives productivity, particularly in labor-intensive agricultural and that support the uptake of income-generating appli- food processing activities that women dominate. Women ances in towns served by mini grids. Step one is a mar- are 9–23 percent more likely to gain employment outside ket/demand assessment with geospatial analysis overlying the home following electrification (Smith 2000). Electrifica- mini grids, appliances, and end use finance. Step 2 is com- tion lowers fertility levels, through greater exposure to tele- munity engagement to confirm and improve data collected vision (Buckley 2012). Electrifying health clinics for lighting during Step 1 through survey(s) and workshops. Step 3 is and the refrigeration of medication is especially beneficial a demand analysis for mini grid design and market poten- for maternal health. Mini grid projects can create jobs for tial for appliances and associated end user finance. Step women while shaping new community decision-making and 4 is preparation of roadshows involving local government, leadership models by placing women in leadership roles. community leaders, interested appliance providers and 8   MINI GRIDS FOR HALF A BILLION PEOPLE Innovations in community engagement are emerging centers for mini grid components will be solar PV, battery that can reduce costs and improve effectiveness. One storage, and distribution infrastructure and technologies example from a few years ago was the smartphone app like smart meters. As the costs of solar PV and battery and accompanying online YouTube-like platform called Mini storage continue to fall, the fraction of energy produced by Grid Stories, developed by Quicksand Design Studio with solar PV and batteries will approach 100 percent, resulting support from ESMAP. Following simple on-screen instruc- in the profit potential for diesel dropping to nearly zero over tions, mini grid customers and staff of mini grid companies the next decade. ESMAP analysis also indicates a profit used the free smartphone app to create short videos—on potential for mini grid developers that could exceed $3.3 how a customer uses electricity in her small business, for billion on an annual basis for all third-generation mini grids example—and uploaded them to a Mini Grid Stories web- deployed between 2022 and 2030. It is important to note site, where the videos could be viewed, shared, and down- that financial support packages, including subsidies from loaded. The approach was inspired by the success of the governments and development partners, will be needed to agricultural web-based platform Digital Green, which uses unlock this profit potential, particularly over the next few videos for agricultural extension work, which was 10 times years to set the market on the trajectory of rapid scale-up. more cost efficient than traditional community engage- Public funds enabled high-income countries to achieve uni- ment services on a cost-per-adoption basis (Abate and versal electricity access; the same will be true for electricity others 2018). Another example is Smart Power India (SPI), access-deficit countries today. supported by the Rockefeller Foundation. This India-based, Even in countries in which the government leads mini Indian-led organization intermediates between key stake- grid development, the private sector is a key partner in holders, including developers, national and local govern- mini grid initiatives. Public-private partnerships are often ment entities, and community organizations (Rockefeller an effective way of distributing responsibilities to optimize Foundation 2017). SPI’s approach is called “Community government and private-sector capacities. They enable Engagement, Load Acquisition and Micro-enterprise Devel- mini grid operators that do not have substantial financial opment” (CELAMeD). With SPI support, developers have resources to enter the market. In addition, major opportu- crafted communication and marketing strategies to inform nities for partnership between local and international firms consumers about the benefits of renewable energy and exist across the mini grid industry value chain. Local enti- catalyze the growth of rural businesses (SPI 2017). ties are best positioned to focus on the aspects of the value BUILDING BLOCK 5. chain that require knowledge of local rules and regulations Delivering services through local and international or require coordination with the customer being served by companies and utilities the mini grid; international companies are best suited to perform tasks that can be replicated across geographic Connecting 490 million people by 2030 will require utili- boundaries. Recent local-international partnership agree- ties and private companies to develop and operate more ments include Caterpillar and Powerhive in Africa, ABB and than 210,000 mini grids. National utility companies in Husk Power in India, Mitsui and OMC in India, ENGIE and Kenya, Madagascar, the Philippines, Russia, and many other Mandalay Yoma Energy in Myanmar, and Schneider Elec- countries are already important developers of mini grids. tric with both EM-ONE and GVE in Nigeria. Private-sector developers—including Tata Power Renew- able Microgrids, Engie Energy Access, Havenhill, PowerGen, Industry associations can facilitate collaboration and OMC Power, Green Village Electric (GVE), and Husk Power, deal making between local and international entities. among many others—are developing large portfolios of mini AMDA comprises more than 40 developers, each operating grids. In a well-established market, private-sector-led initia- a portfolio of commercially viable mini grids in Sub-Saha- tives have a better chance of reaching exponential growth— ran Africa. AMDA helps its members present a unified voice something that is needed to reach universal access by 2030. and facilitates deals between developers and suppliers. By National utilities—including the Ethiopian Electric Utility collecting data from their members, associations can pres- (EEU), the Kenya Power and Lighting Company (KPLC), ent data-driven opportunities to investors as well as suppli- and Engie—also see an expanding role for mini grids based ers of specialized products and services. on their organizational cost-benefit analysis.3 BUILDING BLOCK 6. The mini grid industry offers major profit potential to Financing solar mini grid portfolios and end user private-sector equipment and service suppliers and appliances developers alike, but financial support packages are Private investors—both domestic and international— needed to unlock this potential. ESMAP analysis projects are financing third-generation mini grids and driving that the annual profit potential across the mini grid value innovation in financing mechanisms. Private financiers chain will be almost $5.8 billion by 2030.4 The largest profit invested more than $500 million in developers building MINI GRIDS FOR HALF A BILLION PEOPLE    9 mini grids in low-income countries between 2012 and may require an expanded support package, as women 2022, according to ESMAP’s analysis of publicly available often face additional barriers to accessing finance. data on more than 100 unique deals between developers Performance-based grants have become a mainstream and investors. Impact investors and commercial inves- subsidy mechanism, and can greatly lower the cost of tors, as well as local and national banks, have developed mini grid electricity to allow mini grid services to be equity, debt, and blended finance options to help devel- affordable to a larger group of end users. According to an opers scale up their mini grid business. Acumen, Bamboo ESMAP analysis, a 40 percent capital cost grant reduces Capital Partners, CrossBoundary Energy Access, ElectriFi, the LCOE of a best-in-class third-generation mini grid from InfraCo Africa, and Shell Foundation are just a few exam- $0.38/kWh to $0.28/kWh in a scenario with very low pro- ples of recent investors in mini grids. ductive uses of electricity. In scenarios where productive Development partners, including the World Bank, have uses increase the mini grid’s load factor to 40 percent, the increased funding for mini grids, from millions of dol- same 40 percent capital cost grant reduces the LCOE from lars in the 2000s to billions of dollars in 2018. A group $0.28/kWh to $0.22/kWh. of 15 major international donors and development part- Performance-based grants for mini grids based on a ners, including the World Bank, has collectively commit- percentage of the developer’s cost to connect new ted approximately $2.6 billion just to mini grid investment customers are often less than the implicit or explicit (that is, excluding funding for technical assistance and subsidy that the main grid receives for each new con- research). The World Bank has committed more than nection. A survey of 39 national utility companies in Africa $1.4 billion to mini grids over the next five to seven years, showed that utilities received explicit or implicit subsi- through 50 projects in 42 countries (41 projects approved dies that enabled them to sell electricity at prices that by the World Bank Board and at least 9 under preparation). were on average 41 percent—and up to 80 percent—less The investment plans of this portfolio include the deploy- than the utilities’ unsubsidized LCOE (Trimble and others ment of 3,000 mini grids by 2027, with the expectation 2016; Kojima and Trimble 2016). This would indicate that of bringing electricity to more than 11 million people. This many national utilities in Africa receive implicit subsidies investment commitment is expected to crowd in close to $1 that are more than 40 percent of the connection cost. billion of cofinancing from private-sector, government, and With national utility connection costs often exceeding development partners. $2,000 in rural areas (Trimble and others 2016; Blimpo In countries where the World Bank has an investment and Cosgrove-Davies 2019), it is therefore likely that many commitment in mini grids, the Bank’s investment rep- national utilities in Africa receive implicit cost subsidies in resents on average about 25 percent of the total invest- excess of $800 per connection. To put this in perspective, ment in mini grids in each country from governments, the a performance-based grant equivalent to 40 percent of a private sector, and development partners. On a demand typical third-generation mini grid developer’s connection basis, the World Bank will continue to provide support for costs would be about $400–$900 per connection. well-designed, new energy access projects that include Performance-based grants should be applied with cau- mini grid investments. In the broader context, the upscal- tion, however, as relying exclusively on final output ing of financing in the sector will need the involvement of makes it difficult for developers to finance their up-front the World Bank, development partners, and governments, capital costs. Therefore, it is reasonable to designate some at least at the same level of engagement over the next five intermediate results—such as purchase orders or the years, to create the leverage for exponential private-sector arrival of goods on site—as a basis for early subsidy pay- involvement. In the longer run, the percentages of public ments. Capital cost subsidies can also dilute the benefits of funds compared with overall investment should taper off increasing productive uses of electricity. Although the com- with the growth of private-sector investment. bined impact of grants and productive uses on the LCOE Different financing packages—consisting of different is typically greater than either on its own, their cumulative combinations of equity, debt, subsidy, and risk-shar- impact can increase the LCOE when OPEX costs are large ing mechanisms—are required for different types of relative to CAPEX.5 mini grid developers. In response, governments and their development partners are preparing packages of financial BUILDING BLOCK 7. support for mini grid developers that help them overcome Attracting exceptional talent and scaling skills barriers and finance the scale-up of mini grid deployments. development Larger international and local firms tend to have greater Scaling up mini grid deployments will be possible only access to equity and debt; smaller, mostly local firms usu- if human capital keeps pace with financial capital. Inno- ally do not. Female-led enterprises and project developers vative technologies and initiatives have emerged to train 10   MINI GRIDS FOR HALF A BILLION PEOPLE the stakeholders needed to support a thriving mini grid • Governments that recognize mini grids as a desirable industry. ESMAP has identified more than 50 training pro- and viable electrification option. grams for key stakeholder groups in the mini grid ecosys- • Government institutions that support mini grid develop- tem, including developers, financiers, policy makers, and ment through their actions and decisions. regulators. Many of these courses leverage new technolo- gies. For example, LED Safari’s flexible curriculum design • Flexible institutional frameworks able in principle to and remote web-based training enables developers and support various mini grid delivery models. governments to create high-quality, reputable certification • Frameworks that minimize duplication of oversight and programs. Comprehensive training programs that follow a conflicting roles. train-the-trainer approach, such as the Institute of Electri- cal and Electronics Engineers’ Smart Village’s Comprehen- BUILDING BLOCKS 9 AND 10. sive Training Program, can provide training to thousands Regulating the sector and making it easier to do of people. These programs seek to create a skilled, knowl- business edgeable ecosystem of stakeholders that can support the No single approach to regulating mini grids works best rapid scale-up of mini grids. in all settings, and regulation has costs as well as ben- Capacity needs assessments are a critical early step in efits. ESMAP has developed a series of decision trees that designing training and skills-building initiatives. They present options for how to regulate mini grids and the con- reveal gaps in key areas, including technical expertise, ditions under which each option is suitable. The decision management skills, institutional capacity, policy frame- trees are not prescriptive. They can provide guidance to works, partnerships, knowledge, and implementation help regulators and policy makers make informed deci- know-how. Needs assessments generally follow a four-step sions in five regulatory areas: market entry, tariffs, tech- process—(1) identifying key actors, (2) determining the nical specifications, service standards, and what happens capacity needs of a project or portfolio, (3) assessing exist- when the main grid arrives in the service area of a mini grid. ing capacity, and (4) identifying capacity gaps—that uses a Several countries are developing mini-grid-specific reg- mixed-methods approach using existing data or data col- ulatory frameworks that support private-sector invest- lected from key interviews with respondents and commu- ment. Across Asia and Africa, countries such as Bangla- nity members, focus group discussions, and surveys. desh, Cambodia, India, Kenya, Nigeria, Rwanda, Tanzania,6 and Zambia have developed regulatory frameworks for BUILDING BLOCK 8. mini grids that address key issues. Supporting institutions, delivery models, and champions to create opportunities The goal of a regulatory framework for mini grids should be to promote good service at the lowest cost-recov- National-level institutions are supporting the scale-up of ery tariffs. Pursuit of this goal throughout the stages of mini grids as a key element of electrification strategies. development of a country’s mini grid sector—taking into Haiti’s Ministry of Public Works has developed a special account subsidies and the broader national electrification unit, the Energy Cell, to implement a World Bank-sup- strategy—requires a regulatory framework that is predict- ported national mini grids program. Nigeria’s Rural Elec- able but flexible enough to evolve as the market does. trification Agency is implementing the largest mini grid program in Africa, targeting 850 mini grids by 2025, out of Meanwhile, innovative solutions that cut down on red an estimated potential market of 10,000 sites. Regulatory tape and make it easier for mini grid developers to do agencies in Nigeria, Rwanda, Zambia, and several other business are emerging, and include the following: countries have teams dedicated to mini grids. Ministries, • Standardized templates for key bureaucratic processes national utilities, and rural electrification agencies are col- that affect mini grids, including standardized power pur- laborating on national electrification plans, as with Kenya, chase agreements, which define the terms under which Myanmar, Nigeria, and Rwanda mentioned earlier. mini grid developers sell electricity to the main grid, ESMAP’s research identified four characteristics of an and standardized environmental and social manage- institutional framework that can support mini grids, ment systems, which identify when mini grid developers given the diversity in potential mini grid delivery mod- obtain environmental approvals. els. The most common delivery models for mini grids • Technology platforms to connect developers with inves- are build-own-operate, public-private partnerships, con- tors and suppliers and to run large-scale mini grid ten- cessions, utility models with and without private-sector ders, greatly boosting market efficiencies. involvement, and cooperative models. Strong institutional frameworks that can accommodate diverse delivery mod- • Formal delegation of mini grid industry oversight author- els are characterized by: ity to a single entity—usually the local government or a MINI GRIDS FOR HALF A BILLION PEOPLE    11 government agency that provides grants or subsidies measure the global mini grid industry’s progress against to mini grid developers (such as a rural electrification the 10 building blocks and 5 market drivers outlined above. agency)—in countries where the absence of a formal regulator increases the risk that mini grid developers face multiple layers of government oversight. REFERENCES • Introduction of e-government to reduce overhead cost Abate, G., T. Bernard, S. Makhija, and D. Spielman. 2018. “Accelerating for business registration, land and building permits, and Technical Change through Video-Mediated Agricultural Extension: environmental approvals. Evidence from Ethiopia.” Working Paper, Cornell University, Ithaca, NY. http:/ /barrett.dyson.cornell.edu/NEUDC/paper_421.pdf. Africa population count data: Linard, C., Gilbert, M., Snow, R.W., Noor, A CALL TO ACTION A.M. and Tatem, A.J., 2012, Population distribution, settlement patterns and accessibility across Africa in 2010, PLoS ONE, 7(2): Connecting half a billion people to mini grids by 2030 e31743. (WorldPop—https:/ /www.worldpop.org) is a monumental task that requires unprecedented lev- Balabanyan, Ani, Yadviga Semikolenova, Arun Singh, and Min A Lee. els of investment, innovation, and commitment from 2021. “Utility Performance and Behavior in Africa Today.” World development partners, governments, and the mini grid Bank, Washington, DC. https://openknowledge.worldbank.org/han- dle/10986/36178. industry. This book calls for action by stakeholders across the mini grid value chain. Key recommendations are for the Blimpo, M., and M. Cosgrove-Davies. 2019. Electricity Access in Sub-Sa- haran Africa: Uptake, Reliability, and Complementary Factors for Eco- following actors: nomic Impact. Africa Development Forum Series. Washington, DC: • Policy makers to leverage the latest geospatial analysis World Bank. technology to develop national electrification plans that BNEF (Bloomberg New Energy Finance). 2019. “1Q 2019 Frontier Power Market Outlook.” can guide investment in mini grids, main grid extension, and solar home systems, as well as develop initiatives Buckley, A. 2012. “Best Practice Community Engagement for Infra- structure Projects: Building Community Ties That Dig Deeper.” Pub- that promote productive uses of electricity and build lic Infrastructure Bulletin 1 (8). human capital. C. Arderne, C. Zorn, C. Nicolas, and E. E. Koks, “Predictive mapping of • Development partners to work with government coun- the global power system using open data,” Sci. Data, vol. 7, no. 1, p. terparts and the private sector to create enabling 19, Dec. 2020, doi: 10.1038/s41597-019-0347-4. (https:/ /gridfinder. org/) environments for mini grids through investments in CIESIN (Center for International Earth Science Information Network), portfolios of projects and technical assistance for devel- Columbia University, and Novel-T. 2020. “GRID3 Central African oping workable regulations and strengthening institu- Republic Settlement Extents Version 01, Alpha.” Palisades, NY: tions. Geo-Referenced Infrastructure and Demographic Data for Devel- opment (GRID3). Source of Building Footprints ‘Ecopia Vector • Regulators to adopt an evolving, light-handed approach Maps Powered by Maxar Satellite Imagery’ .” Accessed June 1, 2022. for a maturing mini grid sector, providing at each stage https://doi.org/10.7916/d8-y2ax-p859. of development clear guidance on market entry, retail GOGLA. 2022. “Our Members.” Accessed May 18, 2022. https://www. tariffs, service standards, technical standards, and gogla.org/about-us/our-members. arrival of the main grid. IEA (International Energy Agency), IRENA (International Renewable • The mini grid industry and its associations to work Energy Agency), UNSD (United Nations Statistics Division), World Bank, and WHO (World Health Organization). 2021. Tracking SDG 7: toward increasing the pace of deployment, retaining The Energy Progress Report 2021. World Bank, Washington, DC. superior-quality service delivery of third-generation mini https://www.iea.org/reports/tracking-sdg7-the-energy-progress- grids, and reducing the cost of these systems through report-2021. innovation to reach a value proposition that is affordable IEA. 2021. World Energy Outlook 2021. Paris. https://www.iea.org/ to the end users. reports/world-energy-outlook-2021 • National utilities to adopt an openness to partnerships IEA, IRENA, UNSD, World Bank, and WHO. 2022. Tracking SDG 7: The Energy Progress Report 2022. World Bank, Washington, DC. https:// with the third-generation mini grid industry on the basis www.iea.org/reports/tracking-sdg7-the-energy-progress-report- that the systems are grid-integration ready, which can 2022. provide for more financially viable grid expansion pro- Kojima, M., and C. Trimble. 2016. Making Power Affordable for Africa and grams for the utility in the long run. Viable for Its Utilities. Washington, DC: World Bank. Finally, there is a clear need for accurate, up-to-date, and Rockefeller Foundation. 2017. “Smart Power for Rural Development: Transforming Lives through Energy Access.” https:/ /www.rocke- widely available data to inform any type of initiative that fellerfoundation.org/report/smart-power-rural-development-bro- supports mini grids. To this end, we strongly recommend chure/. the development of a global tracking tool to monitor and 12   MINI GRIDS FOR HALF A BILLION PEOPLE SPI (Smart Power India). 2017. Smart Power Connect Magazine, Vol. 2 The importance of tailoring the community engagement approach 2.  (May). https://smartpowerindia.org/wp-content/uploads/2021/07/ to the local context was emphasized in an interview with Havenhill smartpowerindia_magazine_may_2017.pdf. Synergy Ltd., a Nigerian mini grid developer operating several solar Smith, J. 2000. Solar-Based Rural Electrification and Microenterprise hybrid mini grids in the Kwali and Kuje local government areas of Development in Latin America: A Gender Analysis. Golden, CO: Nigeria. National Renewable Energy Laboratory. https:/ /www.nrel.gov/docs/ RAO Energy in Russia, TANESCO in Tanzania, JIRAMA in Madagas- 3.  fy01osti/28995.pdf. car, and KPLC in Kenya are utility companies that operate dozens of Tenenbaum, B., C. Greacen, and D. Vaghela. 2018. Mini Grids and the mini grids nationwide. These mini grids are typically diesel powered Arrival of the Main Grid: Lessons from Cambodia, Sri Lanka, and (or, in the case of JIRAMA, hydro powered). They tend to be large, Indonesia. ESMAP Technical Report 013/18. Washington, DC: World typically on the order of several hundred kilowatts to a few mega- Bank. watts. Some utilities (in Niger, for example) have started to hybridize their diesel systems with solar PV panels. Tenenbaum, B., C. Greacen, and A. Shrestha. Forthcoming 2022. Undergrid Mini Grids in Nigeria and India: Interconnected and Non- Rather than provide a definitive number, this analysis is designed 4.  Interconnected. to understand the relative profit potential among different mini grid value chain stakeholders. Such an analysis can be used to determine Trimble, C., M. Kojima, I. Perez Arroyo, and F. Mohammadzadeh. 2016. the viability of establishing business lines focused on the mini grid “Financial Viability of Electricity Sectors in Sub-Saharan Africa: Qua- market. The data reflect the profit potential after all variable produc- si-Fiscal Deficits and Hidden Costs.” Policy Research Working Paper tion and manufacturing costs are taken into consideration. Detailed 7788, World Bank, Washington, DC. http:/ /documents.worldbank. assumptions and methodology are documented on the companion org/curated/en/182071470748085038/pdf/WPS7788.pdf. website to this handbook: www.esmap.org/mini_grids_for_half_a_ billion_people. On average, CAPEX accounted for about 65 percent and OPEX for 5.  NOTES about 35 percent of the fully cost-recovering tariff. While the mini grid regulations in Tanzania are some of the most 6.  Firm power output means that the peak load for which the system 1.  advanced in Africa, issues concerning implementation and enforce- was designed can be supplied by the mini grid any second of the day ment, as well as elements within the regulations themselves, have throughout the year. In solar hybrid mini grids, we approximate firm recently restricted private-sector investment in mini grids. power output as the sum of the generator capacity and 25 percent of the PV array capacity. For a more detailed description of this metric and the rationale for using it, please see chapter 1. MINI GRIDS FOR HALF A BILLION PEOPLE    13 OVERVIEW: The New Electricity Access Landscape and the Growing Space for Solar Mini Grids SDG 7: A GLOBAL AGENDA RUNNING . . . BUT PROGRESS HAS BEEN INSUFFICIENT TO MEET THE GOAL OF UNIVERSAL ACCESS BEHIND These achievements notwithstanding, progress has fallen In September 2015, the United Nations (UN) General far short of what is needed. According to the latest Track- Assembly adopted Resolution 70/1, which introduced a ing SDG 7: The Energy Progress Report, taking into account new global path for sustainable development. The 2030 population growth and recent slowdowns in access as Agenda laid out 17 ambitious Sustainable Development a result of the COVID-19 pandemic, a total of 930 million Goals (SDGs), to be achieved by 2030 (UN 2015). people will need to gain access to electricity over the next eight years if universal access is to be achieved. However, The SDGs focus on key economic and social development under current and planned policies and taking into account issues, such as education, health, and climate change. Rec- the effects of the pandemic, 670 million people are still pro- ognizing that access to basic energy services is a prereq- jected to remain without access in 2030 (IEA, World Bank, uisite for poverty alleviation, sustainable livelihoods, and and others 2022). economic growth, one of the goals (SDG 7) aims to ensure access to affordable, reliable, sustainable, and modern Meanwhile, adjusted for global population growth rates, the energy for all. Its targets include universal access to elec- annual pace of access has been steadily decreasing since tricity, clean fuels and clean cooking technologies, a dou- 2018. While the annual rate of grown in energy access was bling of the rate of improvement in energy efficiency, and a 0.8 percent between 2010 and 2018, it fell to 0.5 percent substantial increase in the share of renewables in the global in 2018–20. Furthermore, these increases have been con- energy mix. centrated in a handful of countries and very unevenly dis- tributed across regions, between rural and urban areas and ACCESS TO ELECTRICITY HAS INCREASED . . . across socioeconomic groups. Impressive advances have been made in closing the elec- tricity access gap in recent decades. Between 2010 and 2020, the share of the global population with access to electricity grew from 83 percent to 91 percent, as 1.3 billion Between 2010 and 2020, about 1.3 billion people gained access during this time period (IEA, World people gained access to electricity, but 733 Bank, and others 2022).1 million people are still currently without access. After taking into account population growth, The pace of electrification accelerated from 2000 to 2018 about 930 million people will need to gain access but has since tapered off: between the years 2000 and to electricity by 2030 to achieve SDG 7. However, if 2010, 100 million people gained access every year, ramping the current pace of electrification, current policies, up to 130 million people per year between 2010 and 2018. and current population trends continue, as many as But in the final two years of the decade, 2018–20, the num- 670 million people are predicted to remain without ber of new people gaining access dropped to 109 million access to any source of electricity by 2030. per year (IEA, World Bank, and others 2022). 14   MINI GRIDS FOR HALF A BILLION PEOPLE Countries that pursue a comprehensive approach to elec- MORE FINANCING IS NEEDED, AND IT MUST BE trification through main grid extension, mini grids, and solar BETTER TARGETED home systems achieved the fastest gains. In most of the A major cause of the present gap in electricity access is countries with the fastest gains in electrification between lack of financing. Current commitments to all electrifica- 2010 and 2020—including Bangladesh, Cambodia, Kenya, tion projects in the 20 highest-access-deficit countries— Myanmar, Nepal, Rwanda, and Tanzania—national electri- which account for 560 million people—are estimated at fication strategies leveraged a combination of main grid, $32 billion a year. This is 78 percent of the $41 billion a year mini grid, and solar home system investments. Nigeria is needed to achieve universal access by 2030, and a 27 per- another recent example of a country that has developed a cent decline from 2018, when the figure reached $43.6 bil- comprehensive national electrification strategy and imple- lion (SEforALL and CPI 2021). mentation plan. This comprehensive approach is the only way to connect the 930 million people that will need access This financing has been distributed highly unevenly, both to electricity by 2030. across the group and within different customer categories. As such, most financing targets provision of electricity ser- Meanwhile, slightly fewer than 76 percent, or 560 million vices to nonresidential customers. As of 2021, only $12.9 people living without electricity are concentrated in 20 billion—less than a third—of the funds committed in the countries with the highest absolute deficit in energy access 20 high-deficit countries—were aimed at households; the (IEA, World Bank, and others 2022).2 Closing the gaps in rest targeted commercial, industrial, and public consumers these countries is therefore essential to achieving the goal (SEforALL and CPI 2021). of universal access by 2030. Within these countries, Kenya and Uganda made the greatest gains since 2010, expanding Forecasts by the International Energy Agency (IEA) indi- access by more than 3 percentage points a year between cate that 95 percent of the additional investment in elec- 2010 and 2020 (IEA, World Bank, and others 2022). trification must target Sub-Saharan Africa if the world is to reach universal access by 2030 (IEA 2021). At present, Fragile and conflict-affected countries also require sub- however, investments in these countries are estimated at stantial support if SDG 7 is to be achieved by 2030. The 39 only approximately 15 percent of what would be required countries on the World Bank’s list of fragile and conflict-af- for them to reach full electricity access (IEA 2021). One fected countries account for well over half of the global notable example of this gap is the Democratic Republic access deficit—nearly 57 percent. The access rate in these of Congo, the country with the second-highest number of countries from 2010 to 2020 rose only by 11 percent—from people without access to electricity on the continent (72 44 to 55 percent (IEA, World Bank, and others 2022). million), which by 2021 saw only approximately $18 million The rural-urban divide in energy access is also stark. In per year committed to electricity access, compared to the 2020, global access rates were almost 97 percent in urban nearly $3 billion estimated to be needed annually to reach areas but just 83 percent in rural areas. Given that 80 per- universal electrification (SEforALL and CPI 2021). This is cent of the world’s unelectrified population reportedly lives in contrast to the progress made by India, where by 2019 in rural areas, identifying electrification solutions that meet the government declared the country to have reached a 99 rural needs is essential to reaching universal access (IEA, percent electrification rate, moving it from third to seven- World Bank, and others 2022). teenth place in the list of highest-access-deficit countries (IEA, World Bank, and others 2022). Encouragingly for the global sustainable development Populations without access to electricity agenda, electrification investments in the highest-ac- tend to be concentrated geographically. cess-deficit countries appear to be firmly shifting from Just 20 countries account for almost 76 percent fossil fuels to renewables. While grid-connected fossil of the global population without access to elec- fuels received over $21 billion in investments in 2018— tricity; fragile and conflict-affected countries col- compared to the $17 billion invested in grid-connected lectively account for well over half of the global renewables—by 2019, the numbers nearly flipped, with access deficit; and more than 80 percent of the grid-connected renewables receiving over $14 billion in world’s unelectrified population lives in rural areas. investments, compared to the under $8 billion in grid-con- As a result, identifying electrification solutions that nected fossil fuels. Much of this shift can be attributed to meet rural needs, particularly in these key electric- the firmer commitments to renewables made by the gov- ity access-deficit countries, is essential to reaching ernments in some of the key high-access-deficit coun- universal access to electricity by 2030. tries—such as Pakistan and Bangladesh—which saw them end approvals for new coal-powered projects. At MINI GRIDS FOR HALF A BILLION PEOPLE    15 Recognizing the need for increasing the impact of each dollar of international public financing and the private sec- Current financing commitments to energy tor’s growing relevance in development finance, in 2018 access in the 20 highest-access-deficit the World Bank Group adopted a new approach of Maxi- countries are estimated at $32 billion a year—just mizing Finance for Development (MFD). Otherwise known 78 percent of what is needed to achieve universal as the “cascade” approach, MFD is aimed at pursuing pri- access by 2030. Forecasts indicate that 95 percent vate-sector solutions for reaching development goals and of the additional investment has to be directed to reserving the limited public funds for key areas where the Sub-Saharan Africa, with only 15 percent channeled engagement of the private sector is not optimal or possible to the continent so far, and largely to nonresidential (World Bank and IMF 2017). customers. The guidelines for implementing the “cascade” follow a decision tree approach, designed to determine whether a new project has a sustainable private-sector solution that the same time, at 0.9 percent ($294 million) of the total limits public debt and contingent liabilities. It encourages electricity investments in the group of countries in 2019, the use of nonlending World Bank Group instruments— financing flows toward mini grid and off-grid renewable such as support for policy and regulatory reforms or solutions remain far behind what is needed to reach uni- de-risking mechanisms—to promote such private solu- versal access (SEforALL and CPI 2021). tions whenever feasible. When viewed through the lens of private and public sources DOUBLE DOWN ON SOLUTIONS THAT of funds, the financing trend appears quite uneven. By HAVE THE POTENTIAL FOR EXPONENTIAL 2018, the flow of international private financing into energy GROWTH CURVES access in the 20 highest-access-deficit countries reached nearly $11.5 billion—a major ramp-up from less than $3 Another reason for the financing gap is that electrifica- billion in 2013. However, in 2019 the number shrank to tion programs have traditionally focused on extending the under $7.5 billion. A parallel trend can be noted in domes- national grid. Doing so is often both expensive and slow in tic private financing for energy access—while it followed a remote settlements and areas with low population densi- steady growth pattern since 2013 and peaked at over $14 ties and low demand for electricity. Developing electrifica- billion in 2018; in 2019 it came down to a little over $8.5 tion models that complement grid extension is therefore billion (SEforALL and CPI 2021). Similarly, public invest- critical to achieving SDG 7. ments—both international and domestic—appear to have Mini grids and off-grid systems are two practical and com- come down slightly from a peak of over $18 billion in 2018 plementary approaches to grid extension. Recent tech- to under $15.9 billion in 2019. nological breakthroughs, the emergence of innovative Some countries are bucking the global trend by devel- business models, and enabling regulations and policies oping, in collaboration with development partners, com- have made mini grids and off-grid systems affordable, scal- prehensive support packages for all three electrification able options for expanding electricity services along expo- pathways—main grid extensions, solar home systems, nential growth curves. and mini grids. Support packages for mini grids consist of subsidies—increasingly in the form of performance-based grants —as well as debt facilitation and risk-sharing THE PLACE FOR SOLAR MINI GRIDS mechanisms, alongside private-sector debt and equity. The objective of these support packages is to increase WHAT ARE SOLAR MINI GRIDS? the affordability of mini grid electricity and incentivize pri- While there is no unanimously accepted definition of mini vate-sector investment, while ensuring that public funds grids, they are commonly described as power generation are deployed appropriately and efficiently. For example, and distribution systems built to provide electricity in areas performance-based grants are increasingly favored by that have not been reached by the main grid or whose private-sector developers and investors. The Energy Sec- costs of a grid-based connection are prohibitive. Mini grids tor Management Assistance Program (ESMAP) analysis typically supply electricity to local communities, covering presented in chapter 6 shows that these subsidies can domestic, commercial, and industrial demand. They range reduce the cost of electricity by almost 50 percent, but at in size, from systems that provide electricity to just a few the same time can dampen the effect of productive-use customers in a remote settlement to systems that bring programs and put additional pressure on developers to power to tens of thousands of customers (usually groups secure major up-front funding. 16   MINI GRIDS FOR HALF A BILLION PEOPLE of households, businesses, and public institutions) in a countries for installations of similar capacity size but serv- town or city (Tenenbaum, Greacen, and Vaghela 2018). ing somewhat different purposes and customers. In OECD Most mini grids are powered by alternating current (AC).3 countries, the term microgrid is used to refer to systems that are almost always connected to the main grid, but that Mini grids can be either fully isolated from the main grid or can operate in an “island” mode to achieve exceptionally connected to it in some capacity—to feed excess energy high levels of reliability, to supply power for applications for into it, take energy from it whenever needed, or both. Mini which a power outage would prove extremely costly or haz- grids that are connected to the main grid generally have the ardous, such as industrial processes, military or medical capacity to intentionally isolate—or “island”—themselves facilities, and data farms. Such microgrid systems are also from it. This means that they are able to disconnect and frequently individually designed as one-off bespoke proj- reconnect to it, ideally without disturbing power quality, ects with no intention for scale. By contrast, in low-income with the intention of improving power reliability, for safety reasons in the event of faults or surges on the main grid, and for the purpose of maximizing opportunities for additional revenue generation for the mini grid operator. The major- Mini grids are electric power generation and ity of mini grids in low-income countries are considered, at distribution systems that supply electricity present, to be totally isolated (for example, electronically to local communities, covering domestic, commer- disconnected) from the main grid. Figure O.1 illustrates a cial, and industrial demand. Mini grids come in all common setup for a solar hybrid mini grid. sizes, from systems that provide electricity to just In various contexts, the term mini grid is often replaced with a few customers in a remote settlement to systems or juxtaposed with the term microgrid. For instance, micro- that bring power to hundreds of thousands of cus- grids are often defined as mini grids with generation capac- tomers (usually groups of households, businesses, ity below 10 kilowatts (kW) (alternatively often referred to and public institutions) in a town or city. Mini grids as pico-grids), with mini grids described as having genera- can be fully isolated from the main grid or con- tion capacities from 1 kW up to 10 megawatts (MW) (IRENA nected to it. Those that are connected to the main 2016). The two terms are also often used differently in grid generally have the ability to intentionally iso- high-income (primarily the Organisation for Economic late, or “island,” themselves from it. Co-operation and Development [OECD]) and low-income Figure O.1 • Example of a common solar hybrid mini grid setup AC load in village AC appliances Poletop hardware Smart meter Generator Distribution line AC bus Residential Smart meter AC/DC inverter Service drop Pole PV array Charge Battery block Commercial controller Smart Solar-hybrid generation system Distribution system meters Efficient productive loads AC = alternating current; DC = direct current; PV = photovoltaic. MINI GRIDS FOR HALF A BILLION PEOPLE    17 countries, what is referred to as mini grids of similar gener- was coupled with, and amplified by, the coevolution of sup- ation capacities are generally built in areas not connected ply, demand, disruptive technology, and policy. Gradually, to the main grid and at some distance from it, and by devel- and as electricity systems became more complex, physical opers who often strive to achieve modularity, replicability, expansion and interconnection came as a natural conse- and economies of scale over time. quence, leading to today’s power systems. For the purposes of this report, the term mini grid will be Many dynamics affected these systems’ development, used to refer to all forms of mini grids, regardless of their some of which (technical advancements, innovation, entre- generation capacity or location, and regardless of whether preneurial drive, and decisions) were endogenous, while they are interconnected with the main grid, as long as they others (economic principles, legislative constraints and have the capacity to operate in an “island” mode. support, institutional structures, historical contingencies, and geography) were exogenous (Hughes 1983). Histor- Mini grids may be owned and managed by communities, ically, areas with robust socioeconomic activity were the local governments, utilities, private companies, or some earliest adopters. The first modern electric utility was the combination of the above. The delivery mechanism nec- Pearl Street Station in Manhattan, New York. Fired by coal, essary to finance, develop, operate, and maintain a mini this thermal power plant initially served electricity for lamp grid depends on characteristics like ownership structure, lighting in 1882 to about 80 customers via a direct current size, and technology (or a combination of technologies). (DC) distribution system (Hughes 1983). It was thus, by For example, diesel-only mini grids require lower up-front definition, an isolated mini grid. costs as compared with solar, solar hybrid, or hydropower mini grids, but are expected to have much higher and less From New York City and Chicago in the United States to predictable operations and maintenance (O&M) costs London and Berlin in Europe and Kimberly in South Africa, (Greacen, Nsom, and Rysankova 2015), often resulting in mini grids started to emerge and operate autonomously limitations or restrictions to the electricity supply service. in cities throughout that period. Other similar systems While government or utility-run mini grids often charge developed to provide electricity to industrial loads or to subsidized tariffs, mini grids that are owned and operated serve particular populations, such as rural US agricultural by private companies require rates of return sufficient producers. not only to cover O&M costs but also to turn a profit. This Various factors supported the early deployment of decen- means that, in the absence of a government subsidy, they tralized electricity systems in areas of high demand den- must charge cost-reflective tariffs, which are often higher sity (urban areas and industrial facilities) or low-cost than the average national electricity tariff. supply (such as hydro sites). First, DC systems and early THE HISTORIC ROLE OF MINI GRIDS IN low-voltage AC systems had physical limits that kept NATIONAL ELECTRIFICATION EFFORTS distribution local; technology in the late nineteenth and early twentieth centuries did not allow for larger systems Mini grids are not a new phenomenon: all current central- covering long distances. Second, electricity demand was ized power grid systems started with small, isolated power initially limited to a few services, such as public lighting. systems and mini grids. These systems were the initiat- Third, the capital intensity of electric power systems ing “spark” of electricity uptake some 130 years ago, and meant that cost recovery required the maximization of were pivotal to the early development and industrialization electricity output and sales. These early power systems of most modern economies, such as Spain, Sweden, the therefore sought to improve the load factor and economic United Kingdom, and the United States. Although these performance. systems were initially few and scattered, their development While government or utility-run mini grids Mini grids are not a new phenomenon: all often charge subsidized tariffs, mini grids current centralized power grid systems that are owned and operated by private companies started with small, isolated power systems and mini require rates of return sufficient to not only cover grids, which gradually interconnected. These sys- O&M costs but also turn a profit. This means that, tems were the initiating “spark” of electricity uptake in the absence of a government subsidy, they must some 130 years ago, and were pivotal to the early charge cost-reflective tariffs, which are often higher development and industrialization of most modern than the average national electricity tariff. economies. 18   MINI GRIDS FOR HALF A BILLION PEOPLE As technologies improved, demand increased, and the public systems, rural cooperatives, and large federally policy and regulatory regimes stabilized, larger generators owned power generation corporations—and supported could be built and electricity could be transmitted over through public and nonprofit entities, such as rural longer distances. These factors resulted in the emergence electrification agencies and, in the United States, the of centralized utilities (either privately or publicly owned). National Rural Electric Cooperative Association. Mini grids either became integrated with one another, • Local participation and ownership appear to be attri- forming the nucleus of a larger centralized system, or were butes of many public and cooperative efforts, par- absorbed by a larger grid system as it expanded. ticularly in small communities and rural areas. Rural The process was not always smooth. In Bolivia, for exam- communities were eager to access electricity. In most ple, lack of technical coordination meant that different mini cases, the local population was actively involved in the grids used different frequencies, making their integration in process. Community engagement and political com- a central grid challenging. In the United Kingdom, compet- mitment through financial and regulatory support ing business and institutional interests resulted in aggres- were crucial. sive competition and stranded assets. Over time, however, • Interconnecting neighboring mini grids was a way to the increasing variety of sources, loads, and control nodes cope with load variation and to increase system flexibil- created the extensive and complex grid network many ity. Increasing generating capacities (per unit) was a way countries have today. to lower costs through economies of scale. These historical systems can be described as the first gen- • Choosing between centralized grid versus mini grid elec- eration of mini grids, which faced many of the same policy, trification was a lengthy process that depended upon regulatory, and operational challenges experienced by mini technological advances, geographic factors, resource grids in developing countries in Sub-Saharan Africa and availability (for example, hydropower), sociodemo- Asia today. A retrospective overview of a number of these graphic factors (for example, demand density), and systems is available online (www.esmap.org/mini_grids_ policy. With the exception of resources and geography, for_half_a_billion_people), highlighting the origin stories of the other factors shifted and changed over time, with modern grids from isolated mini grids in Bolivia, Cambodia, accompanying changes in how electricity demand was China, India, Ireland, Spain, Sweden, the United Kingdom, met, both technically and institutionally. These factors and the United States. That brief historical review provides pushed the industry to ever-increasing interconnection, a number of insights: standardization, and centralization. • Rapid industrialization and the socioeconomic shifts Second-generation mini grids it spurred created demand for new, low-cost forms of Unlike the first generation, what has often been referred energy. to as the second generation of mini grids can be found • The electric power sector soon became a strong new in modern low-income countries (Tenenbaum, Greacen, business opportunity, attracting substantial entre- and Vaghela 2018). These systems are typically small and preneurial and investment activity. The competitive isolated, and generally built by local communities or local environment in the electric power industry promoted entrepreneurs to provide access to electricity in zones technological innovation, leading to new technical sys- with low population densities and low demand, primarily tems that were quickly adopted by utilities. in rural areas that have not yet been reached by the main • Regardless of their type or size, the earliest power sys- grid or where it would be too prohibitively expensive to tems were designed to be successful in terms of eco- extend it. Typically, such second-generation systems are nomics as well as engineering, contributing to their built to supply electricity to single villages. Tens of thou- profitability and competitiveness. • Early deployment of isolated stations and urban mini grids (and later peri-urban systems) was driven primar- Mini grid systems that came about in the ily by the growing demand for electricity. In rural areas, late nineteenth and early twentieth centu- system expansion was largely a function of an explicit ries can be described as the “first generation” of social welfare policy aimed at bridging the gap between mini grids, which faced many of the same policy, urban and rural areas. regulatory, and operational challenges as those • Private power companies would not or could not serve experienced by mini grids in developing countries all of the population and provide power at large scales. in Asia and Sub-Saharan Africa today. The unelectrified areas were filled with small municipal MINI GRIDS FOR HALF A BILLION PEOPLE    19 sands of these systems were built, starting in the 1980s to Haiti, Zambia, Rwanda, and elsewhere—to explicitly and ramping up through the 1990s and early 2000s. and preemptively provide economic options for mini grid developers when the main grid arrives. The second generation of mini grids provided important lessons about technical design, the importance of produc- Third-generation mini grids tive uses for financial viability, and economies of scale to In the past decade, a new, third generation of mini grid drive down costs. The developers of second-generation technologies and business models has emerged. These mini grids, whether public or private, were motivated by the third-generation mini grids differ from the earlier genera- overriding need to supply rural communities with a higher tions in several important ways. level of electricity service as soon as possible. Developers of such second-generation mini grids almost always relied New technologies. Technological developments have on standard existing technologies—such as diesel or mini allowed third-generation projects to use more modular hydro generation—and mini grids were built as one-off proj- technologies—especially solar photovoltaic (PV) genera- ects instead of as part of a larger portfolio. Second-gener- tion backed up with diesel, batteries, or both—and state-of- ation mini grids typically used basic meters, on-site meter the-art hydropower. In most of Africa and parts of Asia (for reading, and in-person bill collection, which was expensive example, Bangladesh, Myanmar, and India), the dominant and did not permit innovative pricing schemes that could emerging technology is solar hybrid mini grids. These new promote productive uses of electricity during the day. They systems are usually combined with sophisticated pay-as- also often charged flat monthly tariffs or postpaid fees cal- you-go (PAYG) billing, smart metering, mobile payment culated and collected at the end of each month based on options, and real-time internet-based monitoring systems, the customers’ power consumption for that month (Tenen- enabled by cellular data, allowing company engineers to baum, Greacen, and Vaghela 2018). spot problems as they start to emerge and make adjust- ments or repairs before small problems snowball into Second-generation mini grids also provided important larger ones. Some third-generation mini grid developers lessons about regulatory frameworks, particularly to use sophisticated load dispatch technologies to ensure reduce the risk of stranded assets once the main grid that priority loads always get electricity by automatically arrives (for a detailed discussion of what happened when shifting low-priority loads to times of energy surplus. the main grid arrived in Cambodia, Indonesia, and Sri Lanka, see Tenenbaum, Greacen, and Vaghela [2018]). New players. In addition to local entrepreneurs and com- When these mini grids were developed, little thought was munity organizations, new national and international pri- given to the possibility of later interconnecting with the vate companies are building or proposing to build these main grid, and many of them were simply abandoned third-generation projects. They seem to be motivated by when the main grid arrived (Tenenbaum, Greacen, and the possibility of using the modular (often proprietary) Vaghela 2018). If they did not go out of existence, these technologies that can be scaled up quickly to serve differ- mini grids often chose to become small power producers ent-sized villages and towns, providing opportunities for (particularly if using a more affordable renewable energy cost-reducing economies of scale that were not available generation source rather than diesel); or small power to second-generation developers. The very early evidence distributors, converting to buying all of their electricity suggests that this will be accomplished through joint ven- supply wholesale from the main grid and selling it to the tures with local firms. Large multinational corporations that local customers at retail prices. These options for what have previously not operated in the mini grid market—such happened when the main grid arrived in the service area as Caterpillar, Tesla, Siemens, General Electric, and ABB— of second-generation mini grids are now being codified have publicly announced their intentions to enter it. Unlike in new mini grid regulations—from Tanzania to Nigeria, second-generation local private entrepreneurs, these new third-generation companies have better access to national and international financial markets. Public-private partnerships. In Kenya, Sierra Leone, and “Second-generation” mini grids are com- elsewhere, governments have proposed public-private mon in low-income countries today. These partnerships to build and operate mini grids. This is an systems are typically small and isolated, and built alternative to pure publicly owned or pure privately owned by local communities or entrepreneurs to provide mini grid systems that have been used in second-gen- access to electricity in zones with low population eration mini grids. These new partnerships appear to be densities and low demand, primarily in rural areas motivated, in part, by the reality that it is politically easier that have not yet been reached by the main grid. to channel a subsidy through a government entity in a joint 20   MINI GRIDS FOR HALF A BILLION PEOPLE venture than to openly give the same or even a smaller sub- Access to new geospatial tools. In the last few years, low- sidy to a private company. cost geospatial planning tools have become more widely available to those planning to develop mini grids. These Not necessarily isolated. Mini grids are no longer being new tools use satellite imagery data that allow potential built only in isolated rural villages at a distance from the developers to obtain important market intelligence on the main grid. For example, in the Indian state of Uttar Pradesh, physical characteristics, likely initial customer base, and one private mini grid operator (OMC Power) has built many probable daily electricity demand profiles of individual mini grids in villages that are already served by a govern- villages. The cost of acquiring the data is rapidly coming ment-owned distribution utility, because the distribution down. Several years ago, one donor organization paid $1 utility has not been able to provide reliable service, espe- million to gather this information on 25 villages in Nigeria cially during peak evening hours (Rockefeller Foundation without the use of geospatial tools. More recently, simi- 2018). These mini grids are not currently interconnected lar information was obtained for 300 villages in Nigeria at with the main grid but have been built to be grid compati- roughly the same total cost with the application of latest ble in the future. A similar arrangement has been proposed geospatial analysis and planning applications. in the mini grid regulations recently issued by the Nigerian electricity regulator. FIGURE O.2 • The first, second, and third generations of mini grids Intake weir and setting basin Forebay tank Penstock Power house containing turbine and generator 1st generation 2nd generation 3rd generation Source: Upper left: International Magazine Co. 1925; upper right: World Bank design; bottom left: World Bank photo. MINI GRIDS FOR HALF A BILLION PEOPLE    21 reflective tariff for 39 utilities across Sub-Saharan Africa is $0.27/kilowatt-hour (kWh); 25 percent of utilities In the past decade, a new “third generation” require a cost-reflective tariff of more than $0.40/kWh, of mini grids has emerged, characterized about half require a tariff of $0.20–$0.40/kWh, and 25 by new technologies, new business models, new percent require less than $0.20/kWh. Only 2 of the 39 players, new types of partnerships, new tools, and utilities (Seychelles and Uganda) charged tariffs that tailored policy and regulatory systems. enabled them to recover their costs (Trimble and others 2016; Kojima and Trimble 2016). Mini grids are therefore often the least-cost, best solution to connect communi- ties where the cost of extending the main grid is simply Targeted regulatory systems. Until recently, developers too expensive. were “flying blind” on government policies and regulations that would apply to mini grid projects. This, too, is chang- Meanwhile, the penetration of off-grid solar—including ing. Mini grid regulatory systems have been developed by solar lanterns, pico PV systems, and solar home systems— governments in India, Kenya, Myanmar, Nigeria, Rwanda, grew rapidly over the last two decades, with more than 100 Sierra Leone, and Tanzania, among several others. These million systems sold in Africa alone. This market growth has systems reduce regulatory uncertainty for mini grid devel- been the result of increasing consumer demand for elec- opers, though there is always the remaining uncertainty tricity services in homes, as well as the pace of innovations as to whether the regulatory rules will be implemented as in telecommunications, which enabled the rise of the PAYG written. model for electricity access. Significant consumer data that emerged from the mobile money and PAYG revolution THE ROLE OF SOLAR MINI GRIDS IN provided lenders and investors with more confidence with UNIVERSAL ELECTRIFICATION regard to the credit risk of the end users, enabling them to Electrification programs have traditionally focused on raise more capital and consequently expand their services. extending the national grid, primarily through power Today, such solar home systems, depending on their size, generated from fossil fuels. Experience in electricity-ac- can typically cost $30–$200 and provide electricity ser- cess-deficit countries over the past five decades, however, vice at Tiers 1 and 2. Some larger, component-based sys- has shown that the main grid is typically unreliable. Across tems are also in use (GOGLA 2019). Sub-Saharan Africa, more than half of households con- WHERE DO SOLAR MINI GRIDS FIT IN? nected to the main grid reported receiving electricity less than half of the time (Blimpo and Cosgrove-Davies 2019). Mini grids have characteristics of both utilities and solar In most electricity-access-deficit countries, the main grid home system companies, creating both challenges and usually provides only Tier 3 or Tier 4 electricity.4 The main opportunities for their large-scale deployment. Like the reasons for this unreliability are the challenges with the main grid, mini grids have sunk cost assets, are subject national transmission and distribution networks, rather to regulatory oversight, and have the possibility of provid- than with the generation systems. Given the region’s size ing 24/7 electricity and supporting productive loads. Mini and frequently very low population densities, the vast dis- grids also have features of the solar home system industry, tances between rural economic hubs in many countries with the possibility for very rapid expansion when the value prove to be prohibitively expensive to connect to central- proposition is right for the market. ized systems. Both utilities and solar home system companies are enter- In addition, research has shown that most utilities in Africa ing the mini grid space for economic reasons, in ways that are not financially solvent. Most national utilities in Sub-Sa- mirror their respective business models, with utility mini haran Africa sell electricity at a loss, as the full cost of con- grids operating as rural distribution networks and solar necting residential customers (typically $800–$2,000 home system companies interconnecting individual stand- but often much higher for rural areas) is too expensive alone systems. This trend would lead to modest growth in for most households (Trimble and others 2016), and this the deployments of mini grids, as the two sectors develop cost is frequently subsidized by the national government. mini grids at the margins of their current target markets. If, In addition, the amounts that the rural, remote, and poor- however, the unique position of mini grids can build on the est groups of the population are able to pay for electric- strengths of both sectors—24/7 electricity from the utility ity generally do not reach the cost-recovery threshold for sector and agility and customer service from solar home national utility companies, and the tariffs charged to these system companies—mini grids will be able to bring afford- customer segments are often cross-subsidized across able access to high-quality electricity to millions of people the utilities’ large customer bases. The average fully cost- at an accelerated pace. 22   MINI GRIDS FOR HALF A BILLION PEOPLE Indeed, as a result of the declining levelized cost of energy “win-win-win” economic outcomes for the three key parties. (LCOE), increasing income-generating uses of electricity, The arrangement can eliminate or reduce financial losses and mainstreaming of geospatial planning, solar mini grids for distribution companies (DISCOs) that are forced to sell are on track to provide power at lower cost than many util- electricity at non-cost-recovering retail tariffs. Interconnec- ities by 2030. At $0.40/kWh, mini grid LCOE would be less tions also allow DISCOs to earn new revenues through bulk than the LCOE of national utilities in 7 out of 39 countries power sales to the mini grid as well as rental revenues from in Africa. At $0.20/kWh, mini grid LCOE would be less than the leasing of some or all of the DISCOs’ existing distribu- the LCOE of national utilities in 24 African countries (Trim- tion systems to the mini grid. For the mini grid operator, a ble and others 2016). This would make mini grids the least- physical connection to the contiguous DISCO offers the cost solution for grid-quality electricity for more than 60 possibility of purchasing bulk power, whether on a firm or percent of the population in Africa in a scenario assuming an “as available” basis from the interconnected DISCO or that national utilities do not dramatically change their oper- an upstream supply source. This can lead to lower operat- ations—with major implications for the allocation of both ing and capital costs (that is, lower LCOE) for the intercon- public and private investment funds. nected mini grid than if it operates in a pure stand-alone mode. And for the mini grid’s customers, this should lead to However, scaling up mini grids does not mean scaling back lower tariffs than would be possible if the mini grid operated the main grid. On the contrary, solar mini grids enhance the in a totally isolated mode. Finally, it is well documented that economic viability of expanding the main grid. By design- mini grids, whether interconnected or isolated, routinely ing the system from the beginning to interconnect with the achieve high levels of reliability for their customers than main grid and by promoting income-generating uses of DISCOs do for theirs (Tenenbaum, Greacen, and Shrestha electricity through effective community engagement and 2022 forthcoming). training, third-generation mini grids can provide early eco- nomic growth, so that significant load already exists by the Product cost: LCOE, portfolio development, CAPEX, time the main grid arrives, and customers have a greater OPEX including major replacements ability to pay. New regulatory frameworks give developers The plummeting cost of mini grid electricity is on pace to viable options for what happens when the main grid arrives, achieve $0.20/kWh by 2030. Indeed, the LCOE of “best- and reductions in the cost of components enable develop- in-class” mini grids already dropped from $0.55/kWh in ers to build grid-interconnection-ready systems while still 2018 to $0.38/kWh in 2021. Up-front investment costs per keeping tariffs affordable. customer have also fell dramatically, from around $2,000 Supporting solar mini grids therefore goes hand in hand per connection just a few years ago to $700–$800 per cus- with strengthening the power sector. Interconnecting tomer in 2021 (AMDA 2021). third-generation mini grids with the main grid can increase These cost declines are the result of decreases in the cost the resource diversity and overall resilience and efficiency of major components, and increasing economies of scale of the power system. However, this presents a couple as mini grids are built as part of ever-larger portfolios of of operational challenges that are better addressed in a projects by private-sector developers and national utilities. comprehensive strategy for developing the sector, for As table O.1 shows, prices for major components declined example, for governments through their electrification 60–90 percent between 2010 and 2020, and are projected strategies to allow for utilities, mini grid, and off-grid to decrease even further through 2030. companies to deliver services in the country, as well as for utilities to be able to introduce the practical technical In addition, economies of scale can further drive down functions to support power system operations and plan- costs for mini grids. Companies like Tata Power Renewable ning with multiple mini grids connected to the distribution Microgrids, Husk Power, Engie PowerCorner, OMC, and grid, such as short- and long-term forecasting and other others, are planning hundreds and thousands of mini grids procedures. over the next several years. At this scale, the unit costs of distribution infrastructure, batteries, solar PV modules, and First experiences with interconnected mini grid collabora- power electronics drop dramatically, as we discuss in more tions are emerging, for example, in Nigeria and India, and detail in chapter 1. are providing valuable lessons. These interconnected mini grids are built to serve different market segments: rural and With these cost declines, mini grid electricity is on pace to peri-urban towns and villages, large urban marketplaces, achieve an unsubsidized cost whereby $10 buys 50 kWh of commercial and industrial (C and I) installations, and sepa- energy each month— transformative consumption levels rate urban residential communities. Early evidence seems for hundreds of millions of people and millions of commu- to indicate that these interconnected mini grids can create nities worldwide. MINI GRIDS FOR HALF A BILLION PEOPLE    23 TABLE O.1 • Benchmarks and price projections, mini grid component costs, 2010–30 Mainstream Mainstream Best in Class industry industry 2030 LCOE Best in Class Mainstream benchmark estimate by modeling Percent Median cost 2020 LCOE industry in 2020 (% 2030 (% assumption of total in ESMAP modeling benchmark change from change from (% change Component Unit capital cost survey assumption in 2010 2010) 2020) from 2020) PV module US$/kWp 9.7 441 596 1,589 198 (–88) 114 (–42) 343 (–42) PV inverter US$/kWp * * * 320 80 (–75) 70 (–12.5) * Battery (Li-ion) US$/kWh 14.9 314 297 1,160 126 (–89) 58 (–54) 137 (–54) Battery inverter US$/kVA 8.6 † 415 303 565 113 (–63) 99 (–12.5) 265 (–12.5) Smart meters US$/ ‡ ‡ ‡ 106 40 (–62) 35 (12.5) ‡ customer Source: Bloomberg New Energy Finance Solar Spot Price Index; ESMAP analysis; Feldman and others 2021; Kairies 2017; National Renewable Energy Laboratory US Solar Photovoltaic System Cost Benchmark: Q1 2020. * PV inverter is included with PV module cost. † Battery inverter is grouped with EMS and monitoring equipment. ‡ Smart meters are included in distribution cost. Average, median, minimum, and maximum costs are all expressed in inflation-adjusted dollars. ESMAP = Energy Sector Management Assistance Program; kVA = kilowatt-ampere; kWp = kilowatt-peak; LCOE = levelized cost of energy; Li-ion = lithium-ion; PV = photovoltaic.  Addressable market and demand: Number of load mobile phones) show that these new solutions need to be centers, current expenditure, income-generating not a little, but much better than the current alternative: appliances why else would consumers take the risk to change their Meanwhile, as costs for mini grid electricity continue to fall, behavior? For these clusters of clients, the service provided the addressable market for its services remains immense by the solar mini grids should be a reliable source for their and continues to grow along with the population. Indeed, consumptive activities (lighting, charging, radio/TV) as well taking into account population growth, ESMAP’s Global as provide for life-changing productive activities within the Electrification Platform estimates that mini grids are the current expenditure of $5–$20 per month. From the end least-cost option to provide first-time access to electricity user’s perspective, a $5–$20 expenditure should not only to 430 million people. This represents around 86 million cover the cost of the reliable electricity provided but also mini grid connections and an estimated up-front invest- cover the cost of transitioning into electric appliances as ment cost of $100 billion. well as the purchase of appliances themselves. When cal- culating this over the lifetime of the technology, this would For Sub-Saharan Africa, nearly 291,000 load centers have mean that roughly $3–$15 per month covers the cost of the profile that favors the deployment of solar mini grids. electricity and about $2–$5 per month, the appliances. The analysis based on digitalized rooftops for the region This equation shows the tall order that needs to be met by shows that over 177,000 clusters consist of 100 to 500 peo- the mini grid industry to fulfil its full market potential. ple each, matching smaller solar mini grid of up to 20 kW, almost 96,000 clusters with 500 to 2,500 people matching This addressable market represents a major business medium-sized solar mini grids of up to 80 kW, more than opportunity not only for mini grid developers but also for 15,000 clusters with 2,500 to 10,000 people that would be suppliers and financiers of income-generating appliances best serviced with larger solar mini grids of up to 200 kW, and machines. Households in low- and middle-income and nearly 3,000 population centers of 10,000 to 100,000 countries receiving electricity for the first time have an people where customization rather than standard sizing of available budget for energy services of between $5 and mini grids would likely be appropriate based on GRID3 data $20 per month. If we use the 86 million connections esti- (CIESIN 2020). mate from the Global Electrification Platform, and a $10 per month expenditure in 2030, the resulting global mar- The end users in these load centers spend on average ket potential for mini grid electricity services is $10 billion an estimated $5–$20 per month on alternative forms per year by 2030. Furthermore, ESMAP’s research has of energy such as candles, kerosene, dry-cell batteries, identified more than 130 income-generating machines car-batteries, and petrol and diesel fuel for stand-alone and appliances with a payback period of less than a year, gensets. Lessons from the introduction of innovative tech- and a median up-front cost of $1,200. If we apply this nologies in the marketplace (like solar home systems or 24   MINI GRIDS FOR HALF A BILLION PEOPLE cost globally and assume 15 appliances per mini grid and properly planned and executed as part of a national elec- 200,000 new mini grids by 2030, the market potential trification strategy, it can increase the resource diversity for income-generating appliances connected to mini grid and overall resilience and efficiency of the power system. electricity is at least $3.6 billion. But this presents operational challenges such as those described earlier. This means that mini grid development— The private sector has taken note of this market opportu- as a viable strategy for helping deliver universal access to nity, and is already going after it, by treating communities electricity—also entails a greatly strengthened utility sec- as valued customers and partners rather than beneficia- tor able to accommodate interconnecting mini grids with ries. Leading developers today across Africa and Asia are the main grid. Many electricity-access-deficit countries expanding their portfolios, raising capital to deploy doz- lack clear procedures for integrating mini grids into the util- ens of new mini grids in the next two years, and hundreds ity’s system planning and operations. So national electrifi- of mini grids in the next five years, on a per-developer, cation plans will need to accommodate scenarios in which per-portfolio basis. mini grids are isolated from the main grid or connected only to other mini grids. A transformational end-user value proposition In addition to declining costs and a large, addressable mar- At the same time, mini grids developed today are challeng- ket, mini grids today are providing high-quality electricity ing the existing centralized approach to electricity service services to their customers. A recent benchmarking report delivery. The cost of mini grid electricity is expected to plum- by the Africa Minigrid Developers Association (AMDA) met over the next decade to levels that make it competitive found that modern solar hybrid mini grids in Africa had with main grid electricity in a large number of electrici- 99 percent uptimes on average, with only seven mini grids ty-access-deficit countries (more discussion on this point reporting uptimes below 95 percent. is provided in chapter 1). In addition, modern mini grids provide higher-quality service—in terms of reliability, avail- Third-generation mini grids are true engines of economic ability, and customer service—than many national utilities development, especially when taken as a package. Their in low-income countries. As mini grid developers establish declining costs are on pace to deliver 50 kWh of electric- strong reputations in their respective countries of operation, ity for $10 by 2030, while a large, addressable market is demand for their services in urban and peri-urban areas is already attracting private sector investment and superior likely to increase, incentivizing developers to target these service. They offer a powerful value proposition for end customers as well. This will put pressure on national utility users, communities, and governments. companies to evolve and improve their service offering. Strengthening the power sector: Win-win with utilities Older diesel-powered mini grids were expensive, ineffi- cient, polluting, and dangerous. Nor were they managed HOW TO SCALE SOLAR MINI GRID as businesses. The fact that they existed at all proved DEPLOYMENT TO SERVE HALF A that customers were willing to pay for electricity and suggested that demand would develop once the main BILLION PEOPLE grid arrived. Most customers used little electricity. But FIVE DRIVERS: COST, QUALITY, PACE, FINANCE, that meant the main grid would sustain ongoing financial AND ENABLING ENVIRONMENT losses when it reached areas served by these older mini grids. Modern mini grids are flipping this narrative. By Through a collaborative, iterative process, ESMAP and mini designing the system from the beginning to interconnect grid industry leaders—including AMDA5 and development with the main grid and by promoting productive uses of partners—have jointly identified five market drivers for the electricity through community engagement and training, sector to achieve its SDG 7 targets: mini grids can provide early economic growth. So by the • A more rapid deployment of mini grids through a port- time the main grid arrives, substantial load already exists folio approach; and customers are able to pay. In parallel, new regulatory frameworks give developers viable options for what hap- • Better service; pens when the main grid arrives, and lower-cost compo- • Crowding in private-sector and government finance; nents enable developers to build interconnection-ready • Creating an enabling business environment for mini grid systems while keeping tariffs affordable. grids in access-deficit countries; and To restate the important point above, supporting mini grids • Reducing the cost of solar hybrid mini grids—which the therefore goes hand in hand with supporting utilities. When other four market drivers will also support. interconnection of modern mini grids with the main grid is MINI GRIDS FOR HALF A BILLION PEOPLE    25 With support from ESMAP and the World Bank, the mini the overall value proposition of mini grids as an electrifica- grid industry can work toward clear and measurable tar- tion strategy. Mini grids can be deployed faster than main gets for these market drivers. These drivers will enable the grid extensions, often at a lower cost per connection; they sector to connect 490 million people by 2030. We summa- tend to provide better-quality electricity and customer ser- rize these targets in table O.2. vice than utility companies; they support productive uses, unlike solar home systems; and they can attract both pri- In addition to helping the mini grid sector achieve magni- vate- and public-sector finance (AMDA 2019). tude changes in scale, these five market drivers support TABLE O.2 • SDG 7 and mini grid industry targets, 2020–30 Target Objective/indicator What is measured 2018 Baseline 2020 2025 2030 1. Increase pace of mini grid development Time from purchase order to Cohort of leading private- 6–12 7 6 5 commissioning (weeks) sector developers Time from goods arriving on site to Cohort of leading private- 6–12 5 4 3 commissioning (weeks) sector developers Mini grids per key access-deficit country Portfolios from rural 20–75 150 500 2,000 per year electrification agencies, utilities, private developers, or industry associations 2. Provide superior-quality service Industrywide standard for minimum Industry associations Under Developed for Developed for Developed for technical specifications preparation solar hybrid mini solar hybrid all renewable grids and hydro mini energy mini grids grids Industrywide standard for reliability of Representative sample of 90–97 percent 97 percent 97 percent >97 percent electricity supply mini grid developers uptime uptime during uptime for 24/7 uptime for 24/7 promised electricity electricity availability times Customer satisfaction (percent) Representative sample of 82–84 85 88 90 mini grid customers Average load factor across the industry Representative sample of 22 25 35 45 (percent) mini grid developers 3. Establish enabling mini grid business environment in key access-deficit countries Average RISE score for mini grids Top 20 electricity-access- 59 60 70 80 framework in top 20 electricity-access- deficit countries deficit countries 4. Crowd in government and private-sector funding Cumulative government and development Estimated from a cohort 8 10 18 32 partner funding committed to mini grids in of leading development key access-deficit countries (US$, billions) partners Cumulative private sector debt and equity Global estimates from 5 10 27 73 invested in mini grids in key access-deficit market research countries (US$, billions) Total cumulative investment in mini grids for Sum of all funding for mini 13 20 45 105 energy access (US$, billions) grids in key energy-access- deficit countries 5. Reduce cost of solar hybrid energy Levelized cost of energy (US$/kWh) Average across a cohort 0.55 0.40 0.25 0.20 of leading mini grid developers Source: ESMAP analysis. Note: See the discussion of the underlying analysis for each target is presented in the overview. kWh = kilowatt-hour; RISE = Regulatory Indicators for Sustainable Energy. 26   MINI GRIDS FOR HALF A BILLION PEOPLE TEN BUILDING BLOCKS • Regulations and policies: Enacting regulations and pol- The World Bank’s experience over the past decade working icies that empower mini grid companies and customers. with mini grid developers, electricity regulators, investors, • Doing business: Cutting red tape for a dynamic busi- policy makers, ministries, rural electrification agencies, ness environment. experts, and donor partners has helped it identify a set of The market drivers focus mainly on PV and solar-die- 10 building blocks that need to be in place to achieve coun- sel hybrid mini grids, as these two types of mini grids are try-level scale-up in mini grid development. These 10 build- likely to be the most prevalent technologies for scaling ing blocks are as follows. up mini grid deployments in key electricity-access-defi- • Solar mini grid costs and technology: Reducing costs cit countries. A similar focus on solar PV and solar-diesel and optimizing design and innovation for solar mini hybrid mini grids is evident in the chapters on the 10 build- grids. ing blocks. The building blocks themselves, however, are conceptualized to support vibrant renewable energy mini • Geospatial planning: Planning national strategies and grid sectors at the national level, regardless of renewable developer portfolios with geospatial analysis and digital energy technology. platforms. Each building block contributes to different market drivers • Productive uses: Transforming productive livelihoods presented above. Collectively, they represent the founda- and improving business viability. tion of successful national mini grid programs. How each • Community engagement: Engaging communities as of the building blocks supports various market drivers is valued customers. shown in figure O.3. • Companies and utilities: Delivering services through As the matrix illustrates, there is a logic to the order in local and international companies and utilities. which we present these building blocks. The first six build- • Access to finance: Financing solar mini grid portfolios ing blocks, read from left to right in figure O.3, primarily and end-user appliances. support market drivers at the project and portfolio levels, while the remaining four building blocks primarily support • Skills and training: Attracting exceptional talent and the country-level market driver of establishing enabling scaling skills development. environments in key electricity-access-deficit countries. • Institutions and delivery models: Supporting institu- tions, delivery models, and champions to create oppor- The next 10 chapters of this handbook present these 10 tunities. building blocks. Each chapter presents the frontier of knowledge in its topic area and speaks directly to public- FIGURE O.3 • Matrix of market drivers and building blocks to support them Building blocks to support Solar Mini Grid Costs & Technology mini grid development at scale Institutions and Delivery Models Community Engagement Regulations and Policies Companies and Utilities Geospatial Planning Skills and Training Access to Finance Productive Uses Doing Business Market drivers of magnitude changes in scale Reducing costs Increasing the pace of deployment Providing superior quality of service Crowding in government and private-sector finance Establishing enabling environments in key countries Source: ESMAP analysis. Note: The darker the shading in the figure, the more direct the impact a building block is expected to have on a driver. MINI GRIDS FOR HALF A BILLION PEOPLE    27 and private-sector decision-makers working on mini grids Building block 4: Community engagement by trying to answer the “how” question. This building block entails engaging with local communi- ties at every stage of the mini grid development process to Building block 1: Solar mini grid technology ensure community buy-in, promote productive uses, and The objective of this building block is to benchmark and support gender equality, thus increasing the likelihood that analyze mini grid component costs and technologies to the mini grid will operate successfully over the long term. identify and promote opportunities to reduce costs and Chapter 4 presents examples and innovations from lead- improve the quality of mini grid services. Innovations pre- ing mini grid developers on their community engagement sented in the chapter include benchmarking mini grid strategies, and identifies community engagement tactics component costs based on in-depth analysis of more than at every stage of a mini grid project’s life cycle. This build- 400 operational mini grids in Africa and Asia, identifying ing block directly supports the superior quality of the ser- trends in costs and technologies, modeling the LCOE of vice market driver, and indirectly supports the pace of the solar and solar-diesel hybrid mini grids, and understand- mini grid deployment market driver by tackling an inherent ing the impacts of productive uses and subsidies on LCOE. paradox—community engagement is typically time and This building block directly supports the reduction of the resource intensive but the sector needs to scale quickly— mini grid costs market driver, and indirectly supports the by presenting innovative processes and strategies to con- pace of the mini grid deployment market driver by identi- duct community engagement at scale. fying technologies that decrease the time it takes to build a mini grid. Building block 5: Companies and utilities This building block aims to accelerate private-sector partic- Building block 2: Geospatial planning ipation in the deployment of mini grids, while strengthening The goal of deploying the latest geospatial analysis tools national-utility-led approaches to mini grid development in and techniques is to support national-scale least-cost elec- countries where this approach has a proven track record. trification planning and portfolio-scale mini grid design. Engaging the private sector also means facilitating deal Chapter 2 presents frontier knowledge on how geospatial making and collaboration between local and interna- analysis is making national electrification planning more tional industry players in the mini grid market. Chapter 5 precise and credible, thereby giving it more weight for pol- describes the industry’s value chain and companies tak- icy makers and other decision-making stakeholders, and ing part in various stages of the value chain, calculates the presents an innovative process to use geospatial analysis profit potential along the value chain, provides examples of alongside other portfolio planning and auctioning tools to effective deal making and collaboration between local and quickly and efficiently (in terms of cost per site) develop international players, and presents the results from nation- large portfolios of mini grids. Geospatial planning directly ally representative surveys of mini grid developers in three supports the pace of mini grid deployment by enabling a countries, giving readers a sense of what second-genera- portfolio approach to mini grid development, and indirectly tion mini grids look like. This building block directly sup- supports the reduction of the mini grid cost market driver ports the crowding in of the private-sector finance market by helping developers design mini grids that are appropri- driver by supporting collaborations between local and ate for their respective sites. international companies, and indirectly supports the pace of the mini grid deployment market driver by supporting Building block 3: Productive uses the upstream and downstream elements of the mini grid Given the vital importance of productive loads for the cost industry value chain. efficiency and sustainability of mini grids, ensuring that productive uses are a priority for the design and implemen- Building block 6: Access to finance tation of mini grid programs and facilitating their availabil- The objective of this building block is to develop financial ity to mini grid customers are the objectives of this building packages that complement private-sector debt and equity block. Chapter 3 answers the questions of how to increase and crowd in large private-sector investors. Chapter 6 pres- a mini grid’s load factor through productive uses of elec- ents frontier innovations in grants, equity, and debt that tricity, what impacts this can have on a mini grid’s profit- can address the different barriers that mini grid developers ability and long-term sustainability, and how to increase face when trying to finance their projects and portfolios, productive-use appliance uptake by mini grid customers. provides financing options for utility-owned mini grids and This building block directly supports the quality of the ser- public-private partnerships, and details the World Bank’s vice market driver, and indirectly supports the reduction in mini grid portfolio. In addition to directly supporting the the mini grid costs market driver because of the strong link crowding in of the private-sector and government invest- between increasing load factor and decreasing LCOE, as ment market driver, this building block indirectly supports discussed earlier. the reduction of the mini grid costs market driver by help- 28   MINI GRIDS FOR HALF A BILLION PEOPLE ing finance large portfolios of mini grids that lead to econ- presents innovative solutions to reduce red tape in two key omies of scale. areas: the long and complicated processes for environmen- tal and social approvals and for interactions with the main Building block 7: Skills and training grid, and the high costs of connecting with suppliers and The objective of this building block is to support training for investors and participating in tenders. In addition to directly key stakeholders in the mini grid sector, including develop- supporting the enabling environments market driver, this ers, financiers, technicians, regulators, and policy makers. building block indirectly supports the crowding in of the pri- Chapter 7 identifies existing training programs and training vate-sector finance market driver by making national mini program gaps for key mini grid sector stakeholders, and grid markets more attractive to private-sector investors. highlights innovative technologies and methods for large- scale training programs—an essential element of scaling up in mini grid deployment in key electricity-access-deficit GLOBAL MARKET SNAPSHOT, countries. This building block directly supports the enabling OUTLOOK 2030, AND CALL TO ACTION environments market driver by increasing human capacity across the mini grid ecosystem, and indirectly supports the Where we are, where we are headed, and where we pace of the mini grid deployment market driver by increas- need to go ing the availability of high-skilled, knowledgeable stake- The mini grid market globally is undergoing seismic trans- holders who can support a portfolio approach to mini grid formations, particularly as a source of electricity in populous development. countries with little current access. Many electricity-ac- cess-deficit countries are pursuing holistic approaches to Building block 8: Institutions and delivery models electrification, including main grid extension, mini grids, This building block aims to ensure that the agencies and solar home systems. Market analysis from Bloomberg responsible for implementing a mini grid program have New Energy Finance (BNEF 2018) predicts that in the next the required mandate and capacity and that collaboration five to seven years, decentralized renewables—both mini among agencies happens in the most effective way possible. grids and solar home systems—will bring electricity to Chapter 8 identifies the institutions both within and outside more people every year than extensions of the main grid. the energy sector with which mini grid developers interact, This section examines some key elements of the global as well as the relationships among these institutions, and mini grid sector. In particular, it presents data on installed presents country-level models that can support mini grid and planned mini grids to understand three important development at scale and the institutional arrangements questions for mini grid development that motivate the therein. This building block directly supports the establish- remainder of this report: ment of enabling environments for the private-sector mini grids market driver and, as a result, indirectly supports the • Where are we today, in terms of number of mini grids, crowding in of the private-sector finance market driver. number of connections, investment, and capacity? • Where are we headed if we keep up the current pace of Building block 9: Regulations and policies mini grid development? The goal of this building block is to ensure that mini grid • How big is the gap between where we are headed and regulations are clear, light-handed, and conducive to where we need to go to achieve the mini grid portion of private-sector participation in national-level mini grid SDG 7 (“universal access to modern energy services”) markets. Chapter 9 presents the five key decisions for reg- by 2030?6 ulators: market entry, tariffs, service standards, technical specifications, and main grid arrival. Recognizing that there Table O.3 provides a global overview of the installed and is no one-size-fits-all solution in these decision areas, the planned mini grid projects around the world.7 For the pur- chapter presents a decision tree approach as well as a way poses of our database, mini grids were defined as electricity for regulations to evolve as the sector evolves. This building systems that have both electricity generation and distribu- block directly supports the establishment of enabling envi- tion infrastructure, either isolated from the main grid or, ronments for the private-sector mini grids market driver if connected to the main grid, capable of operating as an and in doing so indirectly supports the crowding in of the electrically separate, or “islanded,” mini grid. Both alternat- private-sector finance market driver. ing current and direct current mini grids are included. To be included in our database, a mini grid had to serve multiple Building block 10: Doing business customers. Installed mini grids are those we know to have This building block focuses on reducing red tape and been built. Planned mini grids are those that developers, increasing the ease with which mini grid developers can do governments, and other organizations have said they plan business in each country where they operate. Chapter 10 to build over the next several years. MINI GRIDS FOR HALF A BILLION PEOPLE    29 TABLE O.3 • Installed and planned mini grid projects worldwide: A summary Number of Number of connections Number of Average capital Total capacity Total investment Totals calculated mini grids (millions) people (millions) cost (US$/kW) (MW) (US$, millions) Global totals: installed 21,557 10.3 47.9 3,955 7,224 28,571 Global totals: planned 29,353 8.0 35.4 3,501 2,657 9,304 Grand total 50,910 18.3 83.2 3,833 9,881 37,874 Source: ESMAP research and analysis; proprietary and in-house databases of mini grid projects from Bloomberg New Energy Finance, CLUB-ER, Guide- house, Infinergia, and Sustainable Energy for All; and, unpublished World Bank surveys. This book’s website provides a full list of sources. Note: A cascading process was followed to fill gaps in the data. When a country has many projects with usable data for a given metric (for example, connections per mini grid), that country’s median value for that metric was used to fill in gaps for the remaining projects in that country. When only few projects in a country had usable data for a given metric, or no data at all, that region’s median value for that metric was used to fill data gaps. Finally, in the rare cases when no data existed at the country or regional level for a particular metric, gaps were filled using the global median value for that metric. Mini grids smaller than 1 kW were excluded from the database, but no strict maximum capacity was used to exclude large mini grids. That said, of the more than 50,000 mini grids in our database, only 73 had an installed capacity of greater than 15 MW, and we used median values in our analyses in- stead of averages to minimize the risk that outliers skewed results. The data may be skewed toward mini grids that have a renewable energy component, as data are more abundant for renewable and hybrid mini grids. As a result, the data may underestimate the total number of mini grids globally, and may overestimate the capital costs. The data are incomplete for a number of countries where there are likely to be large numbers of mini grids, particularly countries in Eastern Europe, North Africa, and Asia. kW = kilowatt; MW = megawatt. Leveraging proprietary data from three leading market research firms—Guidehouse, BNEF, and Infinergia—as well as a large database compiled by SEforALL and BNEF, ESMAP identified 21,557 mini grids installed World Bank surveys of mini grid operators, and extensive in 131 countries and territories around the desk research, ESMAP identified 21,557 installed and world, providing electricity to 48 million people, and 29,353 planned mini grids in 138 countries and territo- an additional 29,353 mini grids being planned in 77 ries.8 countries that are expected to provide electricity to almost 35 million people. Mini grids currently provide electricity to about 48 mil- lion people worldwide. Mini grids that are currently being planned are expected to bring electricity to an additional 35 and deceleration in fossil fuels. And, almost 99 percent of million people. To put this number in perspective, the com- all planned mini grids are solar or solar hybrid. bined total number of people connected to, or expected to be connected to, a mini grid—about 83 million—is approxi- The mini grid market currently represents almost $29 mately equal to the entire population of Germany. billion of cumulative investment in capital costs, with an additional $9 billion capital cost investment expected for Installed mini grids have a combined power capacity of mini grids currently being planned. The total $38 billion more than 7 gigawatts (GW); however, the total opera- represents an average investment cost of around $2,100 tional capacity is almost certainly lower, since many mini per connection, though this is skewed higher by more grids—particularly small hydro—do not operate at their costly systems serving fewer customers in high-income full capacity.9 A further 2.7 GW of installed capacity is countries. Planned mini grids are expected to be much expected from mini grids currently being planned. less costly, with an average expected investment globally In terms of generation technology, the trend toward solar is of around $1,200 per connection, though this figure varies already clear today, and is accelerating. Approximately 51 by region. percent of installed mini grids are solar or solar hybrid, fol- Table O.4 summarizes the installed mini grid projects lowed by those powered only by hydro (35 percent), fossil regionally and table O.5 shows the breakdown of installed fuel (10 percent), and other generation technologies such and planned mini grids by region. as wind or fuel cells (5 percent). The trend is also accelerat- ing: more than 10 times as many solar mini grids were built Asia has the most mini grids installed and planned, with every year from 2016 to 2020 than fossil fuel mini grids. a combined total of 16,819 installed mini grids and 19,824 Meanwhile, from 2010 to 2014, by comparison, about three planned mini grids across South Asia and East Asia and times as many solar mini grids were built every year than Pacific. With 3,174 mini grids installed and 9,006 mini grids fossil fuel mini grids. This is a major acceleration in solar planned, Africa has many more mini grids than the com- 30   MINI GRIDS FOR HALF A BILLION PEOPLE The trend toward solar is clear already and Asia—including South Asia and East Asia is accelerating. Approximately half of all and Pacific—has a combined total of 16,819 installed mini grids to date are powered by solar, installed mini grids and 19,824 planned mini grids, but nearly 99 percent of all planned mini grids will which is 78 percent of all installed mini grids and be solar or solar hybrid. 68 percent of all planned mini grids in the ESMAP database. Meanwhile, Africa has more mini grids installed (3,174) and planned (9,006) than all other regions outside of Asia combined. bined total of all other regions outside of Asia. It is import- ant to note that the number of mini grids we identified as being planned does not equal the total market potential for mini grids. Planned mini grids have already secured or been ($6 billion), and Europe and Central Asia ($6 billion). The allocated funding. high investment figures for these regions are explained by several factors, including relatively high up-front capital Mini grids provide electricity to about 18 million people in costs (for example, Africa and Europe and Central Asia), Asia, 27 million people in Africa, and 2 million people in Latin relatively large mini grids (for example, United States and America. A further 14 million people in Asia and 20 million Canada), and a large number of mini grids (for example, people in Africa are expected to receive electricity from mini East Asia and Pacific). South Asia leads the market share grids currently being planned. In Asia and Africa, this rep- for planned mini grids, with approximately $2.8 billion, fol- resents a small but significant percentage of the region’s lowed by Africa ($2.4 billion) and the United States and total population, using World Bank population data: the Canada ($1.6 billion). combined total of installed and planned mini grids in Africa would connect less than 3 percent of the region’s current Table O.6 presents top-10 lists of countries and companies population; in Asia, installed and planned mini grids would across a set of key mini grid metrics, focusing on installed connect less than 1 percent of the region’s population. mini grids. With almost 2 GW of installed mini grid capacity, Africa Six countries in the database have more than 1,000 has the most installed capacity of any region, followed by installed mini grids: Afghanistan, Myanmar, India, Nepal, the United States and Canada (1.8 GW) and East Asia and China, and Indonesia. Afghanistan has the most mini grids Pacific (1.5 GW). South Asia leads the world in planned mini of any country in the database, with more than 4,700 grid capacity (0.87 GW), followed by Africa (0.66 GW) and installed mini grids. The top 10 countries account for 84 the United States and Canada (0.50 GW). percent of all installed mini grids. Total cumulative investment in mini grids is spread out Installed mini grids in Afghanistan, the Democratic Repub- evenly among the top four regions: Africa ($7 billion), lic of Congo, and Madagascar serve electricity to about 19 United States and Canada ($6 billion), East Asia and Pacific million people, which represents about 40 percent of all TABLE O.4 • Summary of installed mini grid projects by region Number of Number of Number of connections people Total capacity Total investment Region mini grids (millions) (millions) (MW) (US$, millions) South Asia 9,592 2 12 407 1,555 East Asia and Pacific 7,227 2 6 1,530 6,271 Africa 3,174 6 27 1,960 7,238 Europe and Central Asia 624 <1 1 1,110 6,092 United States and Canada 615 <1 1 1,783 6,447 Latin America and the Caribbean 286 <1 2 390 810 Middle East and North Africa 39 <1 <1 46 158 Source: ESMAP analysis. Note: Data remain scarce for the Europe and Central Asia, Latin America and the Caribbean, and Middle East and North Africa regions, where there are likely to be many more mini grids than this table has captured. kW = kilowatt; MW = megawatt. MINI GRIDS FOR HALF A BILLION PEOPLE    31 TABLE O.5 • Number of installed and planned mini people served by mini grids today. Collectively, the top 10 grids by region countries in terms of people served by mini grids account for about 69 percent of all people served by the mini grids Installed Planned in the database. South Asia 9,592 19,035 The United States has the highest total capacity of installed East Asia and Pacific 7,227 789 mini grids of any country for which data are available, at 1.4 Africa 3,174 9,006 GW. This is a result of the relatively large number of mini grids identified in this country, and each mini grid tends Europe and Central Asia 624 226 to have a relatively large capacity compared with mini United States and Canada 615 198 grids in other countries. With 671 MW of installed capacity Latin America and the Caribbean 286 88 for installed mini grids, Russia is second in the top-10 list. Nearly all of this capacity is from 500 diesel-powered mini Middle East and North Africa 39 11 grids in remote parts of Russia operated by a regional utility Source: ESMAP analysis. company, RAO Energy. TABLE O.6 • Top-10 lists for key mini grid indicators for installed mini grids Number of people Total Total Utility portfolios Private-sector Number of (millions, % of capacity investment (country, installed portfolios (country, minigrids population) (MW) (US$, billions) mini grids) installed mini grids) 1 Afghanistan Afghanistan United States United States RAO Energy BRAC (4,712) (7, 19%) (1,424) (4.9) (Russia, 500) (Afghanistan, 356) 2 Myanmar Congo, Dem. Rep. Russian Russian NPC-SPUG Husk Power (4,016) (7, 8%) Federation Federation (Philippines, 278) (India, 300+) (671) (3.7) 3 India Madagascar China China NIGELEC OMC (3,192) (5, 15%) (529) (1.9) (Niger, 115) (India, 280) 4 Nepal Tanzania (3, 5%) Congo, Dem. Philippines JIRAMA Tata Power (1,541) Rep. (363) (1.8) (Madagascar, 110) Renewable Microgrids (India, 163) 5 China Kenya Canada Canada Eskom MeshPower (1,236) (3, 5%) (359) (1.6) (South Africa, 100) (Rwanda, 85) 6 Indonesia Burkina Faso Philippines Congo, Dem. CREDA Optimal Power (1,190) (2, 10%) (338) Rep. (1.3) (India, 32) Solutions (India, 59) 7 Senegal India Angola Angola EEU NS RESIF (677) (2, <1%) (333) (1.2) (Ethiopia, 32) (Senegal, 53) 8 Russian Philippines Madagascar India KPLC Sud Solar Federation (1, 1%) (253) (0.9) (Kenya, 32) (Senegal, 50) (501) 9 United States Nepal Kenya Madagascar Energie du Mali Jumeme (478) (1, 5%) (239) (0.9) (Mali, 30) (Tanzania, 42) 10 Philippines Myanmar Australia Australia Alaska Village Yoma Micro Power (455) (1, 2%) (217) (0.9) Electric Coop. (Myanmar, 42) (United States, 25) Total 17,998 33 4,724 $19 1,256 1,132 (% global total) (84%) (69%) (65%) (67%) (6%) (5%) Source: ESMAP analysis. kW = kilowatts; MW = megawatts; n.a. = not applicable; NPC-SPUG = National Power Corporation Small Power Utility Group. 32   MINI GRIDS FOR HALF A BILLION PEOPLE For mini grids built as part of a developer’s portfolio, the average size of the portfolio is 33 mini grids. We defined a Six countries have more than 1,000 installed portfolio as a collection of more than two mini grids built mini grids each: Afghanistan, Myanmar, India, by the same entity. Only 258 portfolios of mini grids were Nepal, China, and Indonesia. Mini grids serve more identified. Portfolios of planned mini grids, however, were than 2 million people in seven countries: Afghanistan, an order of magnitude larger than portfolios of installed the Democratic Republic of Congo, Madagascar, Tan- mini grids. zania, Kenya, Burkina Faso, and India. Planned mini grids are expected to be larger than installed mini grids. The median installed mini grid serves 137 con- Seven countries have seen more than $1 billion of cumula- nections, while the median planned mini grid serves 386 tive investment in mini grids, led by the United States, for connections. Similarly, the median capacity of installed a combined investment of $16 billion. These seven coun- mini grids is 123 watts (W) per connection, compared to tries account for around 57 percent of the market, with the 245 W per connection for planned mini grids. To be clear, United States alone accounting for 17 percent. the capacity per connection numbers here do not reflect the capacity that every customer is guaranteed at any The utility company with the largest portfolio of installed given time. Nor are they meant to represent the tier of ser- mini grids is RAO Energy, which provides power to remote vice provided by the mini grid. Instead, they are calculated areas of the country. The second largest, National Power as total installed capacity divided by total number of con- Corporation Small Power Utility Group (NPC-SPUG), is nections. That said, the capacity per connection of installed the national utility in the Philippines, a country with more mini grids varies by two orders of magnitude across than 7,600 islands. Notably, 6 of the 10 largest utilities regions—the capacity per connection of installed mini grids by number of installed mini grids are in Africa: NIGELEC, in the United States and Canada is more than 100 times JIRAMA, Eskom, EEU, KPLC, and Energie du Mali. The larg- higher than in South Asia—likely as a result of differences est private-sector developer is BRAC in Afghanistan, but across regions in household income, and therefore ability four of the top 10 private-sector developers are in India: to pay for electricity. Tata Power Renewable Microgrids, OMC, Optimal Power Solutions, and Husk Power. One area that our data are likely to underestimate is the size of the diesel mini grid market. The primary databases Though not shown in the table, the largest portfolios of and sources used to compile the data set focus principally planned mini grids are all vertically integrated private-sec- (but not exclusively) on mini grids that contain a renew- tor developers, led by Tata Power Renewable Microgrids able energy generation source. However, thousands of (more than 9,800 planned in India), Husk Power (5,000 diesel-fired and other nonrenewable mini grids are likely in mini grids across India and Africa), OMC Power (5,000 mini operation today, for which no data are available. grids in India), Engie PowerCorner (2,000 mini grids across Africa), and Renewvia (700 across Africa). One way to estimate the number of diesel-fired mini grids Table O.7 provides a snapshot of the characteristics of the is to use global estimates for diesel generator shipments installed and planned mini grid projects around the world. using trade statistics tracked by the United Nations. In TABLE O.7 • Characteristics of installed and planned mini grids Mini grids per People per mini Connections per Capacity per Capacity per Totals calculated portfolio* grid mini grid connection (watts) mini grid (kW) Global totals: installed Median 6 1,040 137 123 20 Average 33 1,524 291 371 540 Number of observations (N) 258 portfolios 7,489 mini grids 9,601 mini grids 8,659 mini grids 19,670 mini grids Global totals: planned Median 20 780 386 245 147 Average 544 1,836 853 405 1,304 N 45 portfolios 23,827 mini grids 26,189 mini grids 24,471 mini grids 26,758 mini grids Source: ESMAP analysis. *A portfolio is defined as a collection of more than two mini grids built by the same developer. kW = kilowatt MINI GRIDS FOR HALF A BILLION PEOPLE    33 its analysis of these data, BNEF found that in the 10-year able to make updates meaningful, but is frequent enough period 2008–17, 92.49 GW of diesel generator capac- to capture trends as they happen. The database developed ity—from generators with a capacity of less than 375 for this chapter will facilitate this effort: based in Microsoft kilovolt-amperes, or 300 kW, assuming a power factor of Excel, aggregate country-level data can be shared to iden- 0.8—was shipped around the world. If we assume that 5 tify areas for improving the accuracy and completeness of percent of this capacity is used to power diesel-only mini the data. grids (4.6 GW), and that the average diesel capacity per mini grid is 150 kW, then around 30,800 diesel-only mini PROJECTIONS grids may be currently installed today. Meanwhile, our This section explains the gap between where we are database identified only about 1,400 diesel-only mini grids. headed now, in a business-as-usual (BAU) case, and where Hybridizing existing diesel-powered mini grids by adding we need to go to achieve universal access to electricity by solar PV and battery capacity represents a market oppor- 2030. ESMAP estimates that under the right conditions, tunity of $7–$18 billion. In the database, diesel-only mini mini grids have the potential to be the least-cost way to grids had a total combined capacity of 1.8 GW. As men- provide electricity to almost half a billion people by 2030 tioned above, this is likely to underestimate the global total. (figure O.4). Therefore, it can be assumed that the total installed capac- ESMAP developed four scenarios for mini grid deployment ity for diesel-only mini grids is around 1.8–4.6 GW (using between now and 2030. Each is described in turn. the BNEF analysis of diesel genset trade data). Using an estimate of $4,000/kW for the investment costs of hybrid- ESMAP Mini Grid Outlook Scenario izing diesel-fired mini grids, the total market opportunity Under this scenario, mini grids are the least-cost option for for hybridizing diesel mini grids is $7–$18 billion. 430 million people to receive electricity for the first time, A market snapshot like the one presented above should and an additional 60 million people will be serviced through be conducted every three years. This allows for enough an interconnected network with mini grids due to reliabil- preparation time and sufficient new data to become avail- ity issues on the main grid or to increase resilience in the FIGURE O.4 • Number of people connected to mini grids under business-as-usual and universal access scenarios, 2020–30 500 450 400 350 Millions of People 300 250 200 150 100 50 0 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 ESMAP Business As Usual SDG7 Tracking Stated Policies SDG7 Tracking Universal Access ESMAP Mini Grid Outlook (Mini Grids for Access) ESMAP Mini Grid Outlook (Mini Grids for Resilience) Source: ESMAP analysis. ESMAP = Energy Sector Management Assistance Program; SDG 7 = Sustainable Development Goal 7. 34   MINI GRIDS FOR HALF A BILLION PEOPLE face of climate shocks/severe weather. The resulting total together with relevant policy proposals not yet enacted. is 490 million people connected to more than 217,000 mini This scenario sees 260 million people gaining access to grids at a cumulative investment cost of approximately electricity between 2022 and 2030 (IEA 2021). We then $127 billion. These projections are based on the following conservatively estimate that 31 percent of these people will considerations. be served by mini grids—the same proportion used by the 2021 SDG 7 Tracking Report under the Sustainable Devel- • ESMAP ran country-specific scenarios for the 58 coun- opment Scenario (IEA and others, 2021, p. 160), resulting tries with data in the Global Electrification Platform and in 81 million people served by about 45,300 mini grids at a found that, under enabling circumstances, 430 million cumulative investment cost of $36 billion. people can be best served at least cost by mini grids by 2030, at a cumulative investment cost of about $105 The SDG 7 Tracking Sustainable Development billion.10 This assumes modest cost declines in key Scenario components such as batteries and solar PV and the This scenario is based on the IEA’s Sustainable Develop- main grid expanding at a rate of 2.5 percent of the pop- ment Scenario in World Energy Outlook 2021 (IEA 2021), ulation per year. which takes universal access to electricity as the point of • In addition, as more cities, islands, and utility compa- departure and thus sees 930 million people receiving elec- nies consider risks of extreme weather and invest in tricity access between 2022 and 2030, after taking into more resilient infrastructure, we expect to see more and account population growth. This scenario identifies mini more transitions toward interconnected mini grids that grids as the least-cost electrification pathway for 31 percent can isolate from the network in “island mode” if needed. of new connections (IEA and others 2021, p. 160), which This is complemented by grid-connected towns and ESMAP’s analysis and estimates indicate results in 288 mil- communities investing in grid-connected mini grids in lion people connected to approximately162,000 mini grids order to increase the fraction of renewable energy sup- at a cumulative investment cost of around $93 billion. plying their electricity. The team estimates that these Table O.8 presents a regional breakdown of the ESMAP resilience- and renewable-motivated mini grids could mini grid outlook scenario. serve about 2–3 million new connections (about 6–7 million people) per year globally, equivalent to about Upon analysis of the regional breakdown of the global 10–15 cities or small regional utilities per year deciding ESMAP Mini Grid Outlook Scenario, several important to strengthen their power systems by developing inter- points stand out. First, the largest number of mini grids, connected micro/mini/metro grids. Using costs from and associated investment, will be needed in Africa. microgrids in high-income countries, the cumulative Indeed, almost 80 percent of all access-related invest- expected investment by 2030 for these additional mini ment for mini grids between now and 2030 will need to grids is about $22 billion. go to Africa to achieve SDG 7 by 2030. This means con- necting 380 million people to 160,000 mini grids at a The ESMAP Business-As-Usual Scenario cost of about $91 billion. By contrast, the total number of The BAU scenario assumes that development in 2021– mini grids and investment required in Latin America and 30 follows the same linear growth trajectory that was the Caribbean is small relative to other regions because observed in the 2010–20 data in the ESMAP database, of this region’s current high energy access rates, with the for number of people served by mini grids, and uses exception of Haiti. Finally, mini grids for resilience and actual data from planned mini grids to estimate the total increased penetration of renewable energy represent a number of mini grids and total cumulative investment by major market opportunity, accounting for 12 percent of 2030. The 2021 baseline starting points for the scenario people connected to mini grids and 18 percent of cumu- are the totals from the database: 21,557 mini grids, 48 mil- lative investment, by 2030. lion people served by mini grids, and $29 billion of invest- ment. The results for 2030 are 80 million people served The gap between even the most optimistic BAU and univer- by almost 44,800 mini grids at a cumulative investment sal access scenarios is still vast. The gap for energy access cost of approximately $37 billion. alone (that is, counting only mini grids needed for providing first-time access to electricity, and not counting additional The SDG 7 Tracking Stated Policies Scenario mini grids built for resilience), using only ESMAP’s scenar- The basis for this scenario comes from the IEA’s World ios, is 382 million people, $76 billion, and 183,000 mini grids. Energy Outlook 2021, which developed a “Stated Policies The purpose of the remainder of this book is to identify Scenario” that accounts for policies and initiatives adopted concrete ways to bridge this gap. as of mid-2021 that have an impact on energy access, MINI GRIDS FOR HALF A BILLION PEOPLE    35 TABLE O.8 • ESMAP mini grid outlook scenario: A regional breakdown Population connected to Cumulative investment in Total number of mini mini grids (millions) mini grids (US$, billions) grids installed Region 2021 2030 2021 2030 2021 2030 Africa 27 380 7 91 3,100 160,000 South Asia 12 24 2 3 9,600 27,000 East Asia & Pacific 6 19 6 9 7,200 15,000 Latin America & Caribbean 2 6 <1 1 300 1,800 Rest of World including new mini grids for 2 60 13 22 1,400 12,300 resilience and renewable energy penetration Total 48 490 29 127 21,500 217,000 Source: ESMAP analysis. is made even more impressive given that Indonesia is an archipelago of 6,000 inhabited islands characterized by If the current pace of mini grid development jungles, mountainous terrain, and limited transport and continues, about 44,800 mini grids will be communication infrastructure. installed by 2030, serving around 80 million people. However, achieving universal access to clean and 2020 progress. In 2019–20, Tata Power Renewable Micro- reliable electricity by 2030 will require more than grids was able to achieve a similar pace of development 217,000 mini grids serving 490 million people, at a in India as the Indonesia program, constructing more cost of around $127 billion. For energy access alone, than 150 mini grids in a little more than a year. The rapid not counting new mini grids for resilience, this rep- pace of development for these projects was greatly aided resents an expected shortfall by 2030 of 382 million by the choice to use a standardized design for mini grids. people, $76 billion, and 183,000 mini grids. Standardization improved the efficiencies of carrying out tenders and enabled developers to bid on multiple sites knowing they would be installing the same type of equip- ment across all sites. Note that this does not mean that STATUS OF THE FIVE DRIVERS AND TEN every mini grid should be the same size but, instead, that BUILDING BLOCKS standardization across components facilitates modular ESMAP has begun tracking the mini grid industry’s prog- mini grid design. ress against the 5 drivers and 10 building blocks, to gauge the pace of development against the 2020 targets and Indicators embedded in mini grid development show how assess whether the sector is on track to meet the 2030 tar- long a build will take. From our conversations with AMDA, gets. More comprehensive stocktaking will be needed over we know that some are able to develop mini grids in around the next few years in order to gain a more accurate under- six weeks once initial site identification and assessment standing of the whole industry’s progress. work are completed. But the length of time from placing a purchase order to commissioning for the typical solar-die- Increasing the pace of deployment sel hybrid mini grid is usually measured in months.11 In The targets for the pace of mini grid development are general, components arrive on site at different times, while derived from what would be needed to achieve SDG 7 in delays in customs can set projects back days or weeks. each of the top 20 countries lacking access to electricity. Only a handful of the largest developers have systematized The pace of development grows from around 150 mini grids construction and installation in ways that allows quick and per country per year in 2020 to around 2,000 per country efficient deployment. The pace of deployment must speed per year by 2030 (figure O.5). up if we are to proceed from building tens of mini grids a year to building thousands by 2030. 2018 benchmark. A program in Indonesia supported by the German Society for International Cooperation (GIZ) set the One key innovation that has already been deployed to benchmark for the pace of development of mini grid port- reduce the setup time for individual mini grids is contain- folios. This Indonesia Solar Mini-Grid Programme installed erization and the associated standardization of mini grid 236 mini grids in just more than two years, from 2012 to components—the upstream integration of standardized 2014, setting the pace at approximately 100 mini grids per major mini grid components into one or two shipping con- year (Schultz, Suryani, and Puspa 2014). This achievement tainers,12 which are then delivered, unpacked, and installed 36   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE O.5 • Mini grids installed annually in each of the top 20 electricity-access-deficit countries, 2018–30 2,000 1,500 Number of Mini Grids 1,000 500 0 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 target actual target trendline actual trendline Source: ESMAP analysis. at the mini grid site. Indeed, the 2030 target for solar-die- is the average load divided by the peak load in a specified sel hybrid mini grids—working toward a five-week time- time period. The more productive uses of electricity a mini frame from purchase order to commissioning—is already grid serves during the day, the higher the mini grid’s load possible for some containerized mini grids. Both large factor, and the more economically viable the mini grid. multinational companies like General Electric and ABB, as Boosting productive uses of electricity also contributes to well as smaller, more specialized companies like Redavia, the economic development of the communities in which BoxPower, and Nayo Tropical Technology, already have off- mini grids operate and helps entrepreneurs and small busi- the-shelf containerized hybrid mini grids at the tens to hun- ness customers get the most value for their money from dreds of kilowatts scale. While not a silver bullet, this type of their connection to the mini grid. For the mini grid industry, containerization—combined with standardization of mini increasing productive uses of electricity means increasing grid components and improved efficiencies in construct- demand for mini grid electricity—a necessary component ing the distribution network—will have to transition from of growing at scale. breakthrough technology to industry norm. 2018 benchmark. For the 2018 benchmark, we use Increasing the pace of deployment for portfolios of mini HOMER’s default load factor of 22 percent. grids will also require systematized construction and project management processes. The same practices and 2020 progress. While data remain scarce on load factor for processes in use today by large construction firms that a sizeable cohort of mini grids, recent analysis of new solar manage portfolios of hundreds of small- and medium-size hybrid mini grids in Haiti offers a reference point for 2020 projects are translatable to both private-sector developers progress. These mini grids were able to achieve a load fac- and public utility companies as they seek to scale up from tor of 30 percent, due in part to the fact that they served tens to hundreds of mini grids a year by 2030. large towns with significant daytime economic activity. This is ahead of pace toward the 2030 target of 45 percent, as Providing superior service figure O.6 shows, although the 2030 target is the industry LOAD FACTOR average. As shown in chapters 1 and 3, the viability of a mini grid Achieving the load factor targets will require integrating depends on productive-use customers, those who use the imperative to promote productive uses from the outset electricity at off-peak, typically daytime, hours. This makes of every mini grid project’s development process. In addi- intuitive sense: if the mini grid is only able to sell electricity tion, it will require developers to address the appliances’ during the evening peak hours, it is earning revenue only up-front costs by partnering with local financial institutions during this limited time period. One way to determine how such as microfinance institutions or selling customers well a mini grid is performing in terms of selling electricity the appliances on credit paid back through on-bill financ- during off-peak times is a metric called load factor, which ing. To achieve the productive-use targets, developers will MINI GRIDS FOR HALF A BILLION PEOPLE    37 FIGURE O.6 • Average mini grid load factor, 2018–30 50 45 (Average Demand / Peak Demand) 40 35 Load Factor 30 25 20 15 10 5 0 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 target actual target trendline actual trendline Source: ESMAP analysis. also need to work directly with appliance suppliers to seek could agree with national governments and regulators on attractive commercial arrangements in terms of both price minimum performance requirements for mini grid batter- and after-sales service. Finally, developers must earn the ies, the mini grid market would become more attractive to trust of productive-use customers. Only if customers trust manufacturers of batteries that meet those specifications. that electricity will be supplied to them reliably and in the Third, regulatory and rural electrification agencies tasked long run will they invest in a productive-use appliance or with overseeing a national mini grid program must develop machine powered by the mini grid’s electricity. Meanwhile, their own mini grid technical specifications, which takes governments can also incentivize mini grid developers to time and resources away from other important activities. encourage productive uses of electricity in their mini grids. 2020 progress. Efforts are underway to develop an indus- In Tanzania, for example, Rule 43 of the 2018 mini grid reg- trywide standard for mini grid technical specifications. ulations in Tanzania allow developers to factor in the “asso- ESMAP has developed a set of minimum standards for ciated administrative costs” of promoting productive uses technical specifications that take a light-handed approach of electricity in their retail tariffs.13 and prioritize safety, with the goal of providing an off-the- TECHNICAL SPECIFICATIONS shelf product that countries can adopt for their mini grid programs. The ESMAP specifications are in use in Haiti and 2018 benchmark. Mini grid technical specifications typi- Rwanda. cally define the minimum power, safety, engineering, and other technical specifications for mini grid components MINIMUM QUALITY OF SERVICE STANDARDS: UPTIME and installations to which developers must adhere. Recent 2018 benchmark. Several years ago, a handful of leading examples of mini-grid-specific technical specifications can mini grid developers in AMDA’s membership were able to be found in Annex 7 of the Nigeria mini grid regulations,14 achieve around 97 percent uptime on average. This set the as part of a mini grid tender in Kenya,15 and as part of the benchmark, leading to the targets of achieving 97 percent new draft regulatory framework for mini grids in Zambia,16 uptime as the industry standard by 2025, and increasing among many other countries. Technical specifications this level of reliability through 2030.17 tend to differ in each country, if they exist at all as part of a national mini grid program, which creates at least three 2020 progress. In 2020, the average uptime for mini barriers to growth. First, developers who want to build port- grids in AMDA’s membership was already 99 percent for folios of mini grids in different countries cannot aggregate 24/7 electricity, surpassing the 2020 target of 97 percent their component orders if their portfolio spans jurisdictions uptime during promised service hours (AMDA 2021). This with different technical specifications. Second, different sets a high, but attainable, standard for new entrants to the technical specifications restrict the size of the potential market, and helps ensure that mini grids retain their good mini grid market for component manufacturers. Consider reputation as providers of reliable electricity. batteries as an example. If the global mini grid industry 38   MINI GRIDS FOR HALF A BILLION PEOPLE CUSTOMER SATISFACTION tion of development partner and government funding will Successful mini grid developers provide high-quality elec- decrease, from 60 percent or higher today to 30 percent or tricity service, delivering reliable and predictable power lower in 2030. As figure O.7 shows, total cumulative invest- while maintaining close relationships with their customers. ment in mini grids is not on track to reach the 2030 target that is necessary if SDG 7 is to be achieved. 2018 benchmark. In a study from Smart Power India (SPI 2019), mini grids scored 84 out of 100 for small business 2018 benchmark. By 2018, ESMAP data indicate that the customer satisfaction and 82 out of 100 for household cus- total cumulative investment in mini grids in Africa, South tomers. For comparison, the main grid scored 41 out of 100 Asia, East Asia and Pacific, and Latin American and the and 34 out of 100 for these two customer groups, respec- Caribbean stood at around $13 billion. We estimate that at tively (SPI 2019). least 60 percent of this funding came from governments and development partners, equal to roughly $8 billion. 2020 progress. We do not have more recent survey data to provide a quantitative update on 2020 progress for this 2020 progress. By 2020, the total cumulative investment indicator. To grow at scale, the mini grid industry will need in mini grids in these same regions was about $16 billion, to develop and sustain a reputation for high-quality cus- according to data from ESMAP’s database of mini grid tomer service, to ensure that customers feel that they are projects around the world. While we do not have good data receiving value for their money. This will require surveying on the global breakdown of public vs. private sector financ- mini grid customers on a regular basis to ascertain their ing, we estimate that still about 60 percent of this funding satisfaction with the services they receive. We know that came from governments and development partners, based many developers already invest in a variety of activities to on results-based grant programs across the World Bank’s increase customer satisfaction, including call centers for mini grid portfolio. This is behind pace toward 2020 targets, customer support and rapid response to customer com- both in terms of total cumulative investment (about $4 bil- plaints, as well as continued close engagement with the lion short of the $20 billion target for 2020), and in terms of communities they serve. the fraction of funding coming from the private sector (50 percent of funding from private sector sources by 2020). Cumulative investment in mini grids Achieving growth of two orders of magnitude in the global Establish enabling mini grid business environments mini grid industry by 2030 will require an unprecedented in key access-deficit countries level of investment from governments and their develop- Developing mini grids at scale will require major improve- ment partners. Achieving the SDG 7 targets will require ments in the regulatory environment, making it easier for total cumulative investment of around $105 billion in mini mini grids to operate as companies. A light-handed and grids for energy access by 2030, shared between the predictable business climate would address the needs of public and private sectors—but not in equal proportions. mini grids be conducive to private-sector participation in As the mini grid industry grows and matures, the propor- mini grid development. FIGURE O.7 • Total cumulative investment in mini grids for energy access, 2018–30 110 100 90 80 70 US$, b illions 60 50 40 30 20 10 0 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 target actual target trendline actual trendline Source: ESMAP analysis. MINI GRIDS FOR HALF A BILLION PEOPLE    39 The World Bank tracks these elements of an enabling envi- TABLE O.9 • Top 20 countries with energy access ronment for mini grids in its Regulatory Indicators for Sus- deficits: Doing Business and RISE scores, 2020 tainable Energy (RISE) index (World Bank 2019b), which electricity access (%) covers the regulatory environment for mini grids in more electricity access in Population without population without than 50 countries. To achieve the SDG 7 objective, countries mini grids in 2020 2020 (millions) with energy access deficits must improve their enabling Share of global RISE score for (out of 100) environments, as measured by their RISE scores. Efforts to raise these scores should focus on the 20 countries that account for almost 80 percent of the global population that does not currently have access to electricity (figure O.8). Country Nigeria 85 11 100 2018 benchmark. In 2018, the average RISE score for these Congo, Dem. Rep. 68 9 62 countries was just 59 out of 100. India 64 8 78 2020 progress. By 2020, the average RISE score for the Pakistan 61 8 60 top 20 electricity-access-deficit countries rose to 64 out Ethiopia 60 8 70 of 100, ahead of the 2020 target of 60/100. Table O.9 pro- vides the RISE scores for each of these 20 countries. Tanzania 36 5 100 Uganda 25 3 73 Reducing the cost of solar hybrid mini grids Bangladesh 24 3 80 One chief driver of growth for the mini grid market is the Mozambique 20 3 45 cost to build and operate a mini grid. The typical metric Madagascar 19 2 52 used to combine and quantify these costs is the LCOE, Niger 18 2 73 which also serves as a proxy for the average tariff at which the mini grid must sell its electricity to break even over its Myanmar 18 2 73 lifetime. For mini grid deployment to scale up rapidly, the Angola 17 2 60 LCOE of mini-grid-based electricity will need to plummet Burkina Faso 17 2 42 by 2030. Much lower LCOE would raise market demand Sudan 17 2 37 and speed deployment in low-income areas, where ability Malawi 15 2 77 to pay often limits the potential for mini grids as a solution Chad 14 2 35 for electrification, since mini grid developers often set their tariffs at or below what customers pay per month for alter- Korea, Dem. People’s Rep. 13 2 No Data natives, such as kerosene, car battery- and phone-charging Kenya 13 2 82 services, and diesel gensets. These traditional energy ser- Yemen 11 1 20 vices can be expensive: households and small businesses Total 617 78 Average 64 FIGURE O.8 • Average RISE score in top 20 electricity-access-deficit countries 100 90 80 70 60 (out of 100) RISE Score 50 40 30 20 10 0 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 target actual target trendline actual trendline Source: ESMAP analysis. 40   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE O.9 • LCOE of best-in-class solar hybrid mini grids 0.60 0.50 0.40 USD/kWh 0.30 0.20 0.10 0.00 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 target actual target trendline actual trendline Source: ESMAP analysis. routinely pay the equivalent of $1/kWh or more for lighting Mini grid industry progress across all 10 building and charging services (Tenenbaum and others 2014). blocks Table O.10 presents an overview of the global mini grid The LCOE targets that the mini grid industry can work industry’s progress across all 10 building blocks, based toward—$0.40/kWh by 2020, $0.25/kWh by 2025, and on ESMAP’s internal assessment of energy access deficit $0.20/kWh by 2030 (figure O.9)—are therefore ambi- countries. Note that each country has made its own unique tious but not impossible. They are indicative targets that progress; the table below, meanwhile, gives an overall a cohort of leading mini grid developers can realistically sense of which building blocks have seen more progress achieve, thus setting the pace for other developers in and which generally require more work. the sector. Achieving these targets would bring mini grid electricity close to universal affordability levels by 2030. Key energy-access-deficit countries have made notable In addition, they make the value proposition of mini grids progress across all 10 building blocks over the past decade. that provide reliable power 24 hours a day every day com- Arguably the most progress, particularly in just the past petitive with—if not more attractive than—other options three to four years, has been made in geospatial planning such as backup diesel generation. Achieving these LCOE and in the development of regulatory frameworks and levels would greatly limit, if not eliminate, the need for policies specific to mini grids. Advances in technology, subsidies in areas where customer ability to pay aligns combined with widespread adoption as part of technical with these targets. assistance support to client governments, have main- streamed geospatial analysis at the national level to iden- 2018 benchmark. According to detailed costing analysis tify least-cost options for electrification over the short, that ESMAP conducted in 2019, the LCOE of a best-in-class medium, and long terms. Furthermore, whereas site-spe- solar hybrid mini grid was $0.55/kWh in 2018 (ESMAP cific geospatial analysis was typically used just for feasibil- 2019). ity studies, today’s technologies and service providers are 2020 progress. As we discuss in greater detail in chapter able to identify and analyze thousands of high-potential 1, the LCOE for a best-in-class solar hybrid mini grid was mini grid sites, complete with distribution and generation $0.38/kWh in 2020. This was ahead of the 2020 target and system sizing and costing, and demand estimation that puts the industry on pace to achieve $0.20/kWh by 2030. includes productive uses and public institutions. Countries have also made strong progress on developing mini-grid- On the one hand, hitting all the targets across the five mar- specific regulatory frameworks, whether embedded into ket drivers does not guarantee the 2030 scale of deploy- licenses (for example, Rwanda) or concession contracts ment required to realize the mini grid portion of the SDG 7 (many francophone countries), developed as stand-alone objective. On the other hand, it would make reaching the regulations (for example, Nigeria), or government direc- goals much easier. tives (for example, Ethiopia). We note, of course, that there is still a long way to go toward mainstreaming regulatory MINI GRIDS FOR HALF A BILLION PEOPLE    41 TABLE O.10 • The global mini grid sector and its progress across the 10 building blocks Building block Progress Notable achievements Solar mini grid costs and technologies Large portfolios of solar hybrid mini grids in India Geospatial planning National and portfolio-level planning in Pakistan, Nigeria, Ethiopia Productive uses Productive uses embedded in mini grid planning in Ethiopia Community engagement Signed community agreements in Nigeria, Haiti, Myanmar Companies and utilities AMDA’s growing membership; utility hybridization in Niger, Ethiopia Access to finance Comprehensive financial package in the Democratic Republic of Congo Skills and training Training programs in Mali and Nigeria Institutions and delivery models Strong institutions and private-sector approach in Bangladesh Regulations and policies High RISE scores for mini grids in Nigeria, Kenya, Bangladesh Doing business e-Government initiatives in Ghana and India KEY: Source: ESMAP analysis. Significant work needed AMDA = Africa Minigrid Developers Association; RISE = Regulatory Indicators for Sustainable Energy. Some progress made Significant progress made best practices, particularly in how regulations are imple- • Development partners to work with government coun- mented and the processes underpinning them. terparts and the private sector to create enabling environments for mini grids through investments in When considering the mini grid market globally, more sup- actual portfolios of projects and technical assistance port is needed for productive uses of electricity, community for developing workable regulations and strengthening engagement, skills and training, and the overall business institutions. environment than for the other building blocks. Of these, we see the most positive momentum in productive uses, • Regulators to adopt an evolving, light-handed approach with dedicated organizations like CLASP and EnerGrow, for a maturing mini grid sector, providing at each stage as well as the importance of (and funding for) income-gen- of development clear guidance on market entry, retail erating uses of electricity embedded in a growing number tariffs, service standards, technical standards, and of national mini grid programs. However, progress toward arrival of the main grid. large-scale initiatives to engage with communities and build • The mini grid industry and its associations to work skills, as well as innovative approaches to reducing red tape, toward increasing the pace of deployment, retaining lags behind the progress in other building blocks globally. superior-quality service delivery of third-generation mini We present some ideas for how to accelerate progress in grids, and reducing the cost of these systems through each of these areas in chapters 4, 7, and 10, respectively. innovation to reach a value proposition that is affordable to the end users. CALL TO ACTION • National utilities to adopt an openness to partnerships Connecting half a billion people to mini grids by 2030 is with the third-generation mini grid industry on the basis a monumental task that requires unprecedented levels of that the systems are grid-integration ready, which can investment, innovation, and commitment from develop- provide for more financially viable grid expansion pro- ment partners, governments, and the mini grid industry. grams for the utility in the long run. This book calls for action by stakeholders across the mini grid value chain. Key recommendations are for: Lastly, there is a clear need for accurate, up-to-date, and widely available data to inform any type of initiative that • Policy makers to leverage the latest geospatial analysis supports mini grids. To this end, we strongly recommend technology to develop national electrification plans that the development of a global tracking tool to monitor and can guide investment in mini grids, main grid extension, measure the global mini grid industry’s progress against and solar home systems, as well as develop initiatives the 10 building blocks and 5 market drivers outlined above. that promote productive uses of electricity and build human capital. 42   MINI GRIDS FOR HALF A BILLION PEOPLE REFERENCES Kojima, M., and C. Trimble. 2016. Making Power Affordable for Africa and Viable for Its Utilities. Washington, DC: World Bank. https://open- AMDA (Africa Minigrid Developers Association). 2019. “Untitled Pow- knowledge.worldbank.org/handle/10986/25091. erPoint presentation to the World Bank EEX Week.” Nairobi, Kenya. Rockefeller Foundation. 2018. “Smart Power for Rural Development: AMDA. 2021. Benchmarking Africa’s Mini Grids. Nairobi, Kenya: AMDA. Accelerating Energy Access to Economically Empower and Trans- form Lives.” New York. https://www.rockefellerfoundation.org/initia- Blimpo, M., and M. Cosgrove-Davies. 2019. Electricity Access in Sub-Sa- tive/smart-power-for-rural-development/. haran Africa: Uptake, Reliability, and Complementary Factors for Eco- nomic Impact. Africa Development Forum Series. Washington, DC: Schultz, R., A. Suryani, and A. F. Puspa. 2014. “Executive Overview: Indo- World Bank. https://openknowledge.worldbank.org/bitstream/han- nesia Solar Mini-Grid Programme (PVVP/PLTS Terpusat).” Energis- dle/10986/31333/9781464813610.pdf?sequence=6&isAllowed=y. ing Development Indonesia (EnDev Indonesia), Jakarta, Indonesia. https:/ /energypedia.info/images/1/1e/Indonesia_Solar_Mini-grid_ BNEF (Bloomberg New Energy Finance). 2018. Powering the Last Bil- Programme_EnDev_Executive_Overview_2014.pdf. lion: The Outlook for Energy Access. New York: BNEF. SEforALL (Sustainable Energy for All) and CPI (Climate Policy Ini- BNEF. 2020. “BNEF Solar Spot Price Index.” Proprietary online database. tiative). 2021. Energizing Finance: Understanding the Landscape. CIESIN (Center for International Earth Science Information Network), Washington, DC: SEforALL. https:/ /www.seforall.org/publications/ Columbia University, and Novel-T. 2020. “GRID3 Central African energizing-finance-understanding-the-landscape-2021. Republic Settlement Extents Version 01, Alpha.” Palisades, NY: SPI (Smart Power India). 2019. Rural Electrification in India: Customer Geo-Referenced Infrastructure and Demographic Data for Devel- Behaviour and Demand. Gurgaon, India: Smart Power India and opment (GRID3). Source of building Footprints ‘Ecopia Vector The Rockefeller Foundation. https:/ /www.rockefellerfoundation.org/ Maps Powered by Maxar Satellite Imagery’ .” Accessed June 1, 2022. report/rural-electrification-india-customer-behaviour-demand/. https://doi.org/10.7916/d8-y2ax-p859. Tenenbaum, Bernard, Chris Greacen, and Dipti Vaghela. 2018. Mini Grids Feldman, David, Vignesh Ramasamy, Ran Fu, Ashwin Ramdas, Jal and the Arrival of the Main Grid: Lessons from Cambodia, Sri Lanka, Desai, and Robert Margolis. 2021. U.S. Solar Photovoltaic System and Indonesia. Energy Sector Management Assistance Program Cost Benchmark: Q1 2020. Technical Report NREL/TP-6A20-68925. (ESMAP) Technical Report 013/18. Washington, DC: World Bank. Golden, CO: National Renewable Energy Laboratory. https:/ /www. https://openknowledge.worldbank.org/handle/10986/29018. nrel.gov/docs/fy21osti/77324.pdf. Tenenbaum, B., C. Greacen, and A. Shrestha. Forthcoming 2022. Under- GOGLA. 2019. Global Off-Grid Solar Market Report: Semi-Annual Sales grid Mini Grids in Nigeria and India: Interconnected and Non-Inter- and Impact Data. July–December public report. Utrecht, The Nether- connected. lands: GOGLA. https:/ /www.gogla.org/global-off-grid-solar-market- report. Tenenbaum Bernard, Chris Greacen, Tilak Slyambalapitiya, and James Knuckles. 2014. From the Bottom Up: How Small Power Producers Greacen, Chris, Stephanie Nsom, and Dana Rysankova. 2015. “Scaling and Mini-Grids Can Deliver Electrification and Renewable Energy in Up Access to Electricity: Emerging Best Practices for Mini-Grid Regu- Africa. Directions in Development Series. Washington, DC: World lation.” LiveWire 2015/51, World Bank, Washington, DC. Bank. doi: 10.1596/978-1-4648-0093-1. Hughes, T. P. 1983. “Introduction.” In Networks of Power: Electrification Trimble, C., M. Kojima, I. Perez Arroyo, and F. Mohammadzadeh. 2016. in Western Society, 1880–1930. Baltimore and London: The John “Financial Viability of Electricity Sectors in Sub-Saharan Africa: Qua- Hopkins University Press. si-Fiscal Deficits and Hidden Costs.” Policy Research Working Paper IEA. 2021. World Energy Outlook 2021. Paris: IEA. 7788, World Bank, Washington, DC. http:/ /documents.worldbank. IEA (International Energy Agency), IRENA (International Renewable org/curated/en/182071470748085038/pdf/WPS7788.pdf. Energy Agency), UNSD (United Nations Statistics Division), WB UN (United Nations) 2015. “Transforming Our World: The 2030 Agenda (World Bank), and WHO (World Health Organization). 2021. Track- for Sustainable Development.” General Assembly Resolution 70/1, ing SDG 7: The Energy Progress Report 2021. Washington, DC: World A/RES/70/1 (September 25). https:/ /www.un.org/ga/search/view_ Bank. doc.asp?symbol=A/RES/70/1&Lang=E. IEA (International Energy Agency), IRENA (International Renewable World Bank. 2019b. “Regulatory Indicators for Sustainable Energy Energy Agency), UNSD (United Nations Statistics Division), World (RISE).” http://rise.worldbank.org/scores. Bank, and WHO (World Health Organization). 2022. Tracking SDG 7: World Bank and IMF (International Monetary Fund). 2017. “Maximiz- The Energy Progress Report 2022. World Bank. Washington, DC. ing Finance for Development: Leveraging the Private Sector for https://www.iea.org/reports/tracking-sdg7-the-energy-progress-re- Growth and Sustainable Development.” https:/ /www.devcommit- port-2022. tee.org/sites/dc/files/download/Documentation/DC2017-0009_ International Magazine Co. 1925. “The Location of Places Served with Maximizing_8-19.pdf. Electricity.” 119 West 40th St., New York. IRENA (International Renewable Energy Agency). 2016. Policies and Regulations for Private Sector Renewable Energy Mini-Grids. NOTES Abu Dhabi: IRENA. https://www.irena.org/publications/2016/Sep/ Policies-and-regulations-for-private-sector-renewable-energy- 1. This report defines access to electricity in accordance with the mini-grids. Multi-Tier Framework (MTF), which is elaborated further in this Kairies,Kai-Phillip.2017.“Battery StorageTechnology Improvements and section. Cost Reductions to 2030: A Deep Dive.”PowerPoint presentation at the 2. The 20 countries are Angola, Bangladesh, Burkina Faso, Chad, the International Renewable Energy Agency Workshop, March 17. https:/ / Democratic People’s Republic of Korea, the Democratic Republic of www.irena.org/-/media/Files/IRENA/Agency/Events/2017/ Congo, Ethiopia, India, Kenya, Madagascar, Malawi, Mali, Mozam- Mar/15/2017_Kairies_Battery_Cost_and_Performance_01.pdf? bique, Myanmar, Niger, Nigeria, Pakistan, Sudan, Uganda, and Tan- la=en&hash=773552B364273E0C3DB588912F234E02679CD0C2. zania (data available: https://data.worldbank.org/). MINI GRIDS FOR HALF A BILLION PEOPLE    43 3. Notable exceptions include the solar DC mini grids built, owned, a similar issue: many generation plants are not operating at full and operated by Mera Gao India in the Indian state of Uttar Pradesh, capacity, and some countries are planning to export electricity and by Devergy in Tanzania. because of poor-quality in-country distribution infrastructure. 4. The World Bank’s MTF defines electricity access by five tiers of ser- 10. This is inclusive of ESMAP estimates for countries not covered by vice provision. The tiers rank seven attributes of electricity service: the GEP. capacity, service hours, reliability, quality or voltage fluctuations, 11. Before placing a purchase order and having the goods arrive on site, affordability, legality, and safety. The MTF then assigns any given mini grid developers and their partners will have already completed household to one of the five tiers, from no meaningful access at Tier a number of time-consuming activities, from site identification 0; basic lighting and charging at Tier 1; Tier 2 households can power and assessment to feasibility studies to community agreements. a few small appliances; Tier 3 households have formal grid connec- For the purposes of tracking the sector’s progress, however, clear tions with limited service; Tier 4 access supports refrigeration; and and measurable start and end dates are required as proxies. This Tier 5 is unrestricted continuous service. requirement motivated our decision to use the purchase order and 5. More information about AMDA is available on its website: https:// goods arriving on site as start dates as initial benchmarks for the africamda.org/. pace of mini grid development. 6. Sustainable Energy for All website: https://www.seforall.org/ 12. It is important to note that shipping containers, even if modified by 7. We have less confidence in the reliability of data for planned mini cutting doors and windows into them, do not always meet the min- grids than we do for installed mini grids, and fewer data points were imum standards for powerhouse facilities, particularly where these available for planned mini grids than for installed mini grids. For this standards specify rules for preventing overheating and insulation. reason, the main text and tables related to the mini grid market 13. The 2018 mini grid regulations in Tanzania are available at http:// today focus on installed mini grids. However, we will provide tables www.ewura.go.tz/wp-content/uploads/2018/06/The-Electrici- and analysis of planned mini grids on the website associated with ty-Development-of-Small-Power-Projects-Rules-2018.pdf. this book. For installed mini grids, we made every effort to deter- See the NERC Regulation for Mini Grids, annex 7, available at 14. mine their current operating status; when we found a mini grid was https://nerc.gov.ng/index.php/library/documents/Regulations/ no longer operational, we did not include it in the database. That NERC-Mini-Grid-Regulation. said, we cannot claim that every mini grid in our database is oper- 15. The mini grid technical specifications for the tender in Kenya are ational as of January 2022 owing to the sheer number of individual available at https://tenders.go.ke/website/tender/TenderDocu- projects in the database. ment/9034. 8. It is important to note that despite the scope and depth of the 16. Technical specifications for Zambia are available on the Energy database, it is almost certainly incomplete. For example, data Regulation Board’s website as a downloadable “zip” folder from are scarce for North African, Latin American, Eastern European, http://www.erb.org.zm/content.php?viewpage=mini. and Central Asian countries. It is therefore entirely possible that The Service Standards target for 2020 focuses on increasing reli- 17. the global mini grid market is much larger than what this chapter ability during the times of day when the developer has promised describes, and what the underlying data set supports. Neverthe- to provide electricity—for example, during evening hours. The tar- less, the quality of the available data is quite high, and the result get for 2025 focuses on maintaining high reliability and increasing of the data collection is the most comprehensive database of mini availability to 24/7 electricity—which brings the standards on par grids around the world to date. with (or above) those of the main grid. 9. We learned from conversations with AMDA and energy sector experts that many of the main grids in Sub-Saharan Africa have 44   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 1 REDUCING COSTS AND OPTIMIZING DESIGN AND INNOVATION FOR SOLAR MINI GRIDS CHAPTER OVERVIEW This chapter presents the results of the deepest and most extensive survey of the costs and technology innova- tions of solar mini grids and solar-diesel hybrid mini grids in developing countries conducted by any organization to date. Detailed data were collected from 411 solar and solar-diesel hybrid mini grids in Africa and Asia. According to our analysis, a 40 percent load factor plus expected decreases in component costs would lower the levelized cost of mini grid electricity to $0.20/kilowatt-hour (kWh) by 2030. After outlining the present status of mini grids’ capital and operating costs, the chapter concludes with an outlook for these costs through 2030. The solar mini grid industry is in the early stages of scale-up. panels and a diesel generator. Some run on just solar pan- Whether solar mini grids’ potential will be fulfilled depends els. They include battery storage and deliver alternating crucially on their cost. What are the key drivers of mini grid current (AC) electricity to customers.3 (Please see https:// costs? What do the data suggest are key opportunities for www.esmap.org/mini_grids_for_half_a_billion_people lowering mini grid costs without sacrificing quality and reli- for more information about the mini grids analyzed in this ability? What is the variation in costs among projects, and chapter.) what does it suggest about best practices that should be emphasized as the rollout of mini grids scales up? To help answer these questions with real-world data, the THE LEVELIZED COST OF MINI GRID Energy Sector Management Assistance Program (ESMAP) ELECTRICITY undertook the deepest and most extensive survey of the costs and technology innovations of solar1 mini grids and This chapter focuses on the capital expenditure (CAPEX) solar-diesel hybrid mini grids in developing countries con- required to build a mini grid, including for its preparation, ducted by any organization to date. This survey of 411 mini and the operational expenditure (OPEX) required to keep grids implemented by national electrification programs it going. We can combine these costs into a single cost per supported by the World Bank probes the technology unit of energy, called the levelized cost of energy (LCOE).4 design and costs of mini grids commissioned or contracted For mini grids, LCOE pertains to the cost of electricity on between 2012 and 2021.2 The portfolio of projects encom- a per kilowatt-hour (kWh) basis delivered to mini grid cus- passes 22 countries. The survey responses include detailed tomers over the lifetime of a mini grid. LCOE considers proj- data down to the component level (solar panels, batteries, ect development costs (engineering, obtaining permits, inverters and energy management systems, distribu- management), initial costs (for example, equipment and tion networks, land, logistics and transport, and so forth), installation), the costs of operations (for example, staff and including technical specifications. fuel), and equipment replacement over the lifetime of the project. As such, it is equivalent to the minimum average The mini grids analyzed in this chapter are isolated from tariff that electricity must be sold for in order to cover proj- the main grid. Most are powered by a combination of solar ect costs, including project financing. MINI GRIDS FOR HALF A BILLION PEOPLE    45 We calculate LCOE in two different ways: financial and eco- 221 mini grids in six countries6 and is restricted to mini nomic. The financial perspective on a project is from the grids that use lithium-ion (Li-ion) batteries, whereas case point of view of the developer, and incorporates all costs 5—“Global” (355 mini grids from 19 countries7)—calculates reported by developers in constructing and operating a an average LCOE for all mini grids, including those with mini grid, including import duties and taxes. The simpli- lead-acid batteries. The distinction between battery types fied economic perspective endeavors to remove the influ- is important because our data show a major shift from the ence of taxes, duties and subsidies, and thus represents use of lead-acid batteries (comprising about 97 percent of the cost to society at large, or equivalently, the cost of mini mini grids in our database up to year 2017) to Li-ion (69 grid electrification if a country were to impose no duties or percent of mini grids in our database installed between taxes or subsidies on mini grids. A private sector operator 2018 and 2021). Because Li-ion batteries have superior competing in the marketplace and paying import duties lifetimes and performance characteristics, they lower mini and taxes must consider the financial cost. Policy makers grids’ LCOE. For more on this, including the viable role of deciding among approaches to electrification, meanwhile, lead-acid batteries in mini grids, see the discussion on bat- are most concerned by the economic cost. A separate teries later in this chapter. analysis of how subsidies affect the affordability of mini Case 6 represents the best-in-class mini grid; its LCOE is grids for end users, and the viability of their development, based on component costs and load magnitude averaged is provided in chapter 6. from three high-performing mini grids in our database, one THE LEVELIZED COST OF ENERGY FROM each from Ethiopia, Myanmar, and Nigeria. MINI GRIDS: SEVEN ANALYTICAL CASES Finally, case 7 is a best-in-class 2030 mini grid, based on We developed financial and economic LCOE estimates for case 6 but with equipment cost reductions expected in seven mini grid cases, described below (and in detail in the 2030, as the mini grid industry ramps up and other asso- website accompanying this handbook). For each of these ciated industries achieve scale that drive cost reductions cases, we collected data from multiple mini grids, in some for important components such as solar panels (driven by cases hundreds of them, to determine the average unit global solar panel deployment in solar farms) and Li-ion cost for mini grid component categories. These component batteries (driven by global expansion of electric vehicles categories included the following: and utility-scale electricity storage). Drivers of cost changes are discussed at the end of this chapter. • Solar panels • Batteries The number of mini grids in each sample used to determine • Inverters representative mini grid unit costs, the average peak load • Energy management systems (in kilowatts, kW), and average number of customers per • Backup generators mini grid are shown in table  1.1. Mini grids varied consid- • Distribution networks erably across and within countries. For example, in Nigeria • Installation the average mini grid served an average of 916 customers, • Land but only had a peak load of 69 kW—about 75 watts per cus- • Management tomer; whereas Ethiopian mini grids on average served 228 Using case-specific costs, we used HOMER® Pro (Hybrid TABLE 1.1 • Representative mini grids from seven cases: Optimization of Multiple Energy Resources) software to An analysis of key characteristics optimize a solar hybrid mini grid for each case and to calcu- late the LCOE for each optimized mini grid over a 20-year Peak Number of lifespan. load (kW) of Customers Mini grids average for average The seven cases of LCOE analysis covered three different in sample mini grid mini grid levels: country, global, and best-in-class. 1. Nigeria 150 69 916 2. Myanmar 61 96 409 For the country level (cases 1–3), we picked three coun- tries deploying mini grids at scale: Ethiopia (10 mini grids), 3. Ethiopia 10 178 228 Myanmar (61 mini grids), and Nigeria (with unit costs based 4. Global Li-ion 221 79 716 on averages of 150 mini grids).5 5. Global 355 68 587 For the global average (cases 4 and 5), we pursued the  lobal best 6. G 3 141 793 same approach, but aggregated data that met internal in class data consistency check thresholds from countries around  lobal best in 7. G 3 141 793 the world. Case 4—“Global Li-ion”—is based on data from class 2030 46   MINI GRIDS FOR HALF A BILLION PEOPLE customers but with a peak load nearly three times higher mation about the modeling and assumptions is available (178 kW), averaging 780 watts per customer. at https://www.esmap.org/mini_grids_for_half_a_billion_ people. MODELING ASSUMPTIONS AND SCENARIOS Peak hours generally occur in the evening, when house- When calculating LCOE with HOMER Pro, we included a holds use the most electricity. So for each case, we looked number of assumptions based on prior research and expe- at five scenarios to analyze the impact of daylight and non- rience, including a 20-year lifetime for the mini grid,8 a dis- peak operation of local manufacturing on productive-use count rate of 9.6 percent,9 inflation of 3 percent,10 and diesel loads. The variable adjusted in each scenario is the load prices of $1/liter.11 For all cases except Myanmar, inadequate factor—a measure of the mini grid’s utilization rate, defined data were available on non-fuel OPEX and we assumed an as average load divided by peak load over a year. The first OPEX of $1,700 per staff person with three staff ($5,100 of these scenarios is a base case (22 percent load factor) per year) required for a mini grid with up to 500 customers, representing a typical rural residential load. The third and and four staff ($6,800 per year) for a mini grid with more fifth scenarios bring the load factor up to 40 percent (the than 500 customers. Individual mini grid components medium case) or 80 percent (the high case) (figure 1.1). such as generators, batteries, and photovoltaic (PV) pan- Higher load factor scenarios (40 percent and 80 percent) els had additional variable OPEX expenses.12 More infor- represent the addition of off-peak (primarily daytime) pro- FIGURE 1.1 • Load profiles for 22 percent load factor, 22 percent load factor (sun following), 40 percent load factor, 40 percent load factor (sun following), and 80 percent load factor 100 100 90 22% of load factor 90 22% of load factor, sun following 80 80 Percent of full load Percent of full load 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Time (hour of day) Time (hour of day) 100 100 40% of load 90 40% of load factor 90 factor, sun 80 80 following Percent of full load Percent of full load 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Time (hour of day) Time (hour of day) 100 80% of 90 load factor 80 Percent of full load 70 60 50 40 30 20 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Time (hour of day) MINI GRIDS FOR HALF A BILLION PEOPLE    47 ductive-use loads such as water pumping, agricultural pro- At the global level, mini grids with Li-ion batteries had cessing, cold storage with thermal inertia, and charging of similarly low LCOE. A representative mini grid based on electric vehicles. component costs from 221 mini grids from 6 countries (but dominated by the large number of Li-ion mini grids The second and fourth scenarios illustrate the benefits of in Nigeria and Myanmar) was calculated to have an LCOE lower LCOE that arise from using electricity during sunny of $0.46 per kWh. Adding in mini grids with lead-acid bat- hours, coincident with its production by solar panels, thus teries from 16 countries brought the average LCOE up to minimizing the costly battery storage and retrieval or the $0.53 per kWh for the 22 percent load factor case. It is burning of diesel fuel in backup generators. These two noteworthy that the mini grids with lead-acid batteries “sun-following” scenarios use the same amount of daily were often more expensive, not necessarily because of load in the 22 percent (base case) and 40 percent (medium the battery type, but because they tended to be older— case) load factor scenarios but modify the timing of the built as far back as 2012—and therefore had a variety of consumption to be concentrated during sunlight hours. higher-cost components. These sun-following load profiles should be seen as aspira- tional, and are included to provide a sense of the benefits of The economic LCOE from a representative best-in-class incentivizing daytime consumption. mini grid is $0.38 per kWh, reflecting a nearly 29 percent decrease from the LCOE of current global mini grids in our In calculating the financial LCOE presented in this chap- data set and a 17 percent decrease from current global ter, we have assumed zero grants. However, a second set Li-ion mini grids. At component prices expected in 2030, of scenarios considers the impact of performance-based this drops further to $0.29/kWh, a 22.5 percent drop from capital subsidies (grants), at levels of 40 and 60 percent current best-in-class LCOE. of initial CAPEX, on seven representative mini grid cases. Because these address financing, this set of subsidy sce- The financial LCOE follows the general trends of the eco- narios is discussed in detail in chapter 6. nomic LCOE but is higher by up to 13 percent. The amount of increase depends on the country, since duties and taxes MODELING RESULTS vary from country to country. The increase also depends Our economic and financial LCOE results are for Ethiopia, on the load factor and load curve since mini grids in the Myanmar, and Nigeria as well as representatives of global highest load factor cases rely heavily on batteries (both and best-in-class cases. The LCOE and renewable energy larger battery banks and also deeper cycling and thus fraction of each varies depending on the load curve, includ- more frequent need for replacement) to provide large ing the degree to which loads occur during sunlight hours amounts electricity in the hours of little or no sunlight, and (figure 1.1). batteries are generally assessed higher duties and taxes than solar panels. The next section discusses the portion of costs that each category of expense (for example, solar panels, batteries, LCOE analysis: Adding productive use loads and management, taxes, duties, and so on) involves, how these impact load shifting costs have been trending since 2012, and the projections The mini grids in our analysis had much lower LCOE when that justify our 2030 best-in-class LCOE projections. loads shifted from evening to daytime sunlight hours (see the “sun-following” cases in figure 1.1), or when produc- LCOE analysis: Base case load profile with tive uses that increase the profile’s load factor were added 22 percent load factor (resulting in a 40 percent or 80 percent load factor). Both Our economic analysis indicates that the economic cost shifting to daylight hours and increasing the load factor are of electricity delivered to households from representative likely to be accomplished through encouraging loads such mini grids in Nigeria and Myanmar is $0.43 to $0.46 per as agricultural milling, light manufacturing, water pumping, kWh for our base-case typical village residential load profile or cold storage, for which demand is greatest during day- with a 22 percent load factor (figure 1.2). In our modeling time hours. based on winning engineering, procurement, and construc- tion contract bids, Ethiopian mini grids on average cost A simple way to understand the benefits of daytime use is somewhat less, at $0.41 per kWh. This reflects that their to consider that the marginal cost of adding new genera- commissioning was relatively recent (in 2021) in our sam- tion capacity (solar panels, PV inverters) to meet solar-co- ple, and thus they benefited most from the declining global incident consumption costs around $0.10 per kWh at mini prices of solar panels and batteries, as well as economies of grid scales. Meanwhile, the levelized cost of new capacity to scale due to being relatively large in terms of peak load and cycle electricity into and out of a battery for later use adds average customer load (figure 1.2). at least twice that per kWh (taking into account losses of 4 8   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 1.1 THE LEVELIZED COST OF ENERGY FOR BEST-IN-CLASS MINI GRIDS DROPPED NEARLY 31 PERCENT FROM 2018 Best-in-class mini grid costs have plum- TABLE B1.1.1 • Estimated and potential levelized cost of mini meted in the past few years. In 2018 grid energy, 2018 and 2021 ESMAP conducted a cost analysis of 53 LCOE ($/kWh) from best-in-class mini grid mini grids (ESMAP 2019). At that time, Percentage the best-in-class mini grid produced Load factor (%) 2018 2021 decrease (%) electricity with a levelized cost of energy of $0.55 per kilowatt-hour (kWh). By 22 $0.55 $0.38 31 2021, best-in-class costs had dropped 22 sun following — $0.30 nearly 31 percent to only $0.38 per kWh 40 $0.42 $0.28 35 in the unsubsidized 22 percent load fac- 40 sun following — $0.26 tor case, and more in cases with a higher 80 $0.35 $0.23 35 load factor. This drop in the levelized cost of energy Initial CAPEX $1,160,000 $847,000 in the past two years reflects dramatic Number of customers 1099 793 decreases in the costs of mini grid com- ponents, including solar panels (drop- Solar capacity (kWp) 228 286 ping 18 percent) and a shift in battery Battery type lead-acid OPzS Li-ion LiFePO4 type from lead-acid to lithium-ion, which Battery capacity 887 690 has similar upfront costs but superior (kWh) performance characteristics and thus Average daily load 890 758 lower life-cycle costs. While batteries (kWh) appear to be decreasing in capacity, the Firm power (kWfirm) 207 230 switch to lithium-ion batteries enables $/kWfirm $5,604 $3,659 35 deeper discharge, leading to higher Note: Levelized cost of energy data in 2018 is for the best mini grid in the ESMAP effective capacity. Cheaper storage and database at the time, representing a well-designed mini grid serving 1,100 customers PV generation, in turn, enables reduc- in Bangladesh; 2021 best-in-class data are from a representative mini grid synthesized from average costs and consumption levels in three mini grids in Myanmar, Nigeria, and tions in fuel usage: our 2021 best-in- Ethiopia commissioned in 2020 or 2021. Calculations assume that annual peak load class case used less than one-quarter of is 75% of installed battery inverter capacity and average daily load is calculated as the area under the daily load profile curve scaled to the peak load, accounting for 10 percent the diesel fuel consumption of the 2018 day-to-day and 20 percent hourly load volatility. best-in-class mini grid. These trends LCOE = levelized cost of energy; CAPEX = capital expenditure; kWh = kilowatt-hour; kWp also reflect decreased project develop- = kilowatt peak. ment and installation costs due to econ- omies of scale in deployment. electricity in the charge/discharge process). Benefits of LCOE drops by 4.5–10.6 cents per kWh (15–21 percent of solar-coincident consumption can be further maximized by igure 1.2). the total) in contemporary mini grids (f the use of dispatchable loads such as water pumping to a As the load factor is increased to 40 percent, reductions storage tank or non-time-sensitive agricultural processing in LCOE are even more substantial, shaving 8.8–16.5 US in which the activity needs to take place sometime, but can cents per kW (32–36 percent) from the base case 22 per- wait until there is an energy surplus. cent load factor scenario. If an 80 percent load factor can If the load curve remains at a 22 percent load factor but be achieved, reductions are 13.0–24.7 cents per kWh (39– demand is shifted to largely follow solar production (a 47 percent). 22 percent load factor in the sun-following case), the MINI GRIDS FOR HALF A BILLION PEOPLE    49 FIGURE 1.2 • Economic LCOE calculations for mini grids in 7 cases based on 0 percent subsidy and load profiles described in figure 1.1 Nigeria Myanmar $0.43 $0.46 $0.34 $0.37 $0.30 $0.32 $0.28 $0.29 $0.24 $0.26 22% LF 40% LF 80% LF 22% LF 40% LF 80% LF Ethiopia Global Li-ion $0.46 $0.41 $0.36 $0.33 $0.31 $0.29 $0.30 $0.27 $0.24 $0.25 22% LF 40% LF 80% LF 22% LF 40% LF 80% LF Global Global Best in Class $0.38 $0.53 $0.30 $0.42 $0.28 $0.36 $0.26 $0.34 $0.23 $0.28 22% LF 40% LF 80% LF 22% LF 40% LF 80% LF Global Best in Class 2030 $0.29 $0.25 $0.20 $0.19 $0.16 Normal daily load curve Sun-following 22% LF 40% LF 80% LF 50   MINI GRIDS FOR HALF A BILLION PEOPLE Shifting loads to daylight hours for a 22 Conservative ESMAP analysis indicates that percent load factor mini grid can decrease the combination of increased productive LCOE by up to 21 percent, while increasing the load uses and decreased component costs resulting factor of a solar-hybrid mini grid from 22 percent to from economies of scale and sector-wide technol- 40 percent decreases LCOE by up to 36 percent. ogy cost trends can bring best-in-class mini grid LCOE down to $0.20/kWh by 2030. EFFECT OF EXPECTED DECLINES IN CAPITAL AND THE SHARE OF RENEWABLE ENERGY OPERATING COSTS BY 2030 HOMER modeling calculated the optimum renewable The cost reductions seen over the past decade are expect- energy fraction for each case (table 1.2). Hybrid mini grids ed to continue until the end of the current decade for mini are largely powered by renewable energy but employ die- grid components, particularly solar panels and battery sel generators as backup during extended cloudy periods storage. In addition, ESMAP projects increased savings or for times of particularly high nighttime loads. HOMER in management, installation, and OPEX due to scaled-up modeling includes consideration of seasonal variations in deployment of clusters of mini grids. Together, these sav- sunlight as well as random day-to-day and hour-to-hour ings are expected to drive down the cost of electricity from variation in solar resources and electrical loads. Nearly mini grids. Our 2030 best-in-class case predicts mini grid all of the cases modeled have renewable energy fractions economic LCOE to reach $0.29 per kWh for a typical resi- exceeding 90 percent, meaning that over the course of the dential load curve with a 22 percent load factor. This is a 23 year less than 10 percent of the electrical energy is derived percent decrease from today’s best-in-class LCOE of $0.38 from operating the diesel generator.16 for the same load curve. When combined with a 40 percent The exception is for the global case at a very high load fac- load factor, best-in-class mini grids are expected to reach tor (80 percent), which indicates a renewable energy frac- an economic LCOE of $0.20/kWh by 2030. tion of 76 percent. The representative mini grid in the global For the best-in-class 2030 scenario, the following eco- case is derived from a large data set covering 355 mini grids, nomic cost assumptions were made: some installed as early as 2012. As such, it includes many mini grids built when solar panels were much more expen- • The costs of key mini grid components available to sive. Higher equipment costs in this global case combined developers decrease as follows:13 PV modules and PV with the high nighttime loads in the 80 percent load factor inverters (combined) cost $343 per kilowatt-peak case mean that there are more hours in the year in which (kWp), down from $596/kWp in the 2020 best-in-class the diesel generator is dispatched. This effect disappears representative mini grid;14 battery inverters cost $265 if the price of fuel is modeled at $1.50 per liter. For recent per kilovolt-ampere (kVA), down from $303/kVA; and mini grids and those in the future, high renewable energy Li-ion batteries cost $137/kWh, down from $297/kWh.15 fractions (above 90 percent) will continue to be expected. • Operation and maintenance (O&M) costs fall 50 per- cent, thanks to better bill collection through online pay- TABLE 1.2 • Optimum renewable energy share for mini as-you-go metering, and enhanced remote-monitoring grid cases considered technologies that streamline repairs and reduce staff Renewable energy share (%) costs through geographic clustering (Carlin and others 2018). 22% 40% 80% load 22% load 40% load • We conservatively assume that other CAPEX elements  Country factor sun factor sun factor and economies of scale remain constant, even though Nigeria 92 92 94 95 93 lower installation costs will be achieved by increas- Myanmar 93 95 94 95 93 ing portfolio sizes, among these, cost reductions from scalable plug-and-play building block components (the Ethiopia 93 90 94 93 91 “LEGO-fication” of mini grids), decreased management Global Li-ion 93 91 94 93 92 and engineering costs through economies of scale, and Global 87 92 91 91 76 better pricing of components through larger volumes of Best-in-class 92 90 93 93 90 purchases (Carlin and others 2018). Declines in this full range of costs are explored later in this chapter. Best-in-class 2030 94 94 95 95 94 MINI GRIDS FOR HALF A BILLION PEOPLE    51 MODELING RESULTS: LCOE OF OPTIMUM HYBRID IMPLICATIONS FOR NATIONAL UTILITIES VS. 0 PERCENT AND 100 PERCENT RENEWABLE OF IMPROVING THE QUALITY OF MINI GRID ENERGY SERVICES Powering these same loads using only a diesel generator Many of the mini grids that we analyzed in our study pro- is much more expensive. For example, using the best-in- vide 24/7 electricity and a level of service that consistently class case above, HOMER-calculated economic LCOEs are exceeds the level of service provided by the main grid. 55 percent to 126 percent higher for diesel only (table 1.3) Remote monitoring technologies and smart meters are compared to an optimized hybrid solar mini grid with bat- increasing the quality of customer service and the reliability tery storage and diesel backup. of mini grids. According to a 2022 benchmarking study by the Africa Minigrid Developers Association (AMDA), among But using a diesel generator to occasionally cover cloudy mini grid sites installed by their members in 2020, only 2 periods or periods of particularly high load lowers costs of 35 sites reported service uptime of less than 99 percent compared to the cost of a mini grid that is sized to meet (AMDA and ECA 2022). 100 percent of the load with renewable energy. The eco- nomic LCOE of an optimally sized 100 percent solar mini Across Sub-Saharan Africa, the main grid is much less reli- grid with battery storage is 24 to 39 percent higher than an able: households and small businesses typically experience optimally sized hybrid system. several hours a day of outage. In some countries—includ- ing Burundi, Ghana, Guinea, Liberia, Nigeria, and Zimba- With a renewable energy fraction of 90 percent, the deploy- bwe—more than half of households connected to the main ment of solar mini grids for half a billion people has vast grid reported receiving electricity less than half the time benefits for the environment. These systems are replacing (Blimpo and Cosgrove-Davies 2019). Disaggregated data diesel-fueled systems and/or kerosene-based appliances from the diagnostic survey reports carried out by ESMAP that on average emit 0.89 kilograms (kg) of carbon diox- in a range of countries based on the Multi-Tier Framework ide (CO2) per kWh. Assuming a rollout at scale covering the provide additional evidence of this lack of reliability, both addressable market of 217,000 systems by 2030, 1.2 billion in the Sub-Saharan region and beyond. The report from tonnes of CO2 emissions would be avoided. Rwanda indicates 97 percent of grid-connected house- IMPLICATIONS FOR NATIONAL UTILITIES OF holds experience more than four electricity disruptions a LOWER MINI GRID LCOE week (Koo and others 2018). The Ethiopia report shows that 57.6 percent of grid-connected households face 4 to 14 As a result of declining LCOE and increasing income- outages a week, and 2.8 percent face more than 14 outages generating uses of electricity, third-generation mini grids a week (Padam and others 2018). The report from Cam- can have transformational effects on power sectors. They bodia indicates that 69.3 percent of grid-connected house- are on track to provide power at costs lower than many holds face frequent, unpredictable power outages, and 9.9 utilities by 2030 (figure 1.3). At an LCOE of less than $0.30 percent of all grid-connected customers receive less than 4 per kWh, mini grids will become the least-cost solution for hours of service per day (Dave and others 2018). grid-quality electricity for more than 38 percent of African countries in a scenario in which national utilities do not dra- Utility information, while limited, corroborates this survey matically change their operations—with vast implications information. Only about a third of vertically integrated for the allocation of both public and private investment Sub-Saharan utilities reported figures for the average funds. At $0.20/kWh, electricity from mini grids is less duration and frequency of system interruptions in 2018, expensive to produce than electricity from the main grid in and only 5 of 21 distribution companies did so. Of those 24 out of 39 countries in Africa. that did, median reported duration and frequency of inter- ruptions in 2018 were 51.6 hours and 24.7, respectively. TABLE 1.3 • Economic LCOE of hybrid mini grid versus diesel only and renewables only Economic levelized cost of energy (% above hybrid)   22% load factor 22% sun 40% load factor 40% sun 80% load factor Optimized hybrid (91 to 94% 0.38 0.30 0.28 0.26 0.23 renewable energy) Diesel only (0% renewable energy) 0.85 (126%) 0.65 (115%) 0.55 (98%) 0.47 (85%) 0.35 (55%) No diesel (100% renewable energy) 0.47 (24%) 0.40 (30%) 0.37 (34%) 0.35 (38%) 0.32 (39%) Note: All equipment sizes optimized to meet load profiles based on best-in-class 2021 component pricing. 52   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 1.3 • Comparison of levelized cost of energy of mini grids and utilities in Africa Liberia 5 of 39 At $0.50/kWh: At $0.40/kWh: 7of 39 At $0.30/kWh mini grid LCOE is less than 15 of 39 Mini grid LCOE at $0.20/kWh is less than the LCOE of 24 of 39 utilities in Africa Comoros Sierra Leone São Tomé and Príncipe Cape Verde Gambia, The Rwanda Guinea Senegal Mauritania Burkina Faso Togo Mali Madagascar Seychelles Benin Gabon Kenya Botswana Nigeria Côte d'Ivoire Mauritius Burundi Central African Republic Niger Swaziland Congo, Rep. Ethiopia Tanzania Malawi Cameroon Uganda Zimbabwe Sudan Ghana Mozambique South Africa Lesotho Zambia 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 National utility LCOE ($/kWh) Source: Based on Trimble and others 2016. Note: Many customers in rural areas are charged tariffs that are much lower than the levelized costs shown above due to subsidies. kWh = kilowatt-hour; LCOE = levelized cost of energy. MINI GRIDS FOR HALF A BILLION PEOPLE    53 These are high by international standards. In order to cuted. But this presents an operational challenge, requiring receive any points under the scoring methodology used utilities to be able to introduce the practical technical func- by the World Bank’s Doing Business indicators, the maxi- tions to support power system operations and planning with mum SAIDI (System Average Interruption Duration Index) multiple mini grids connected to the distribution grid, such and SAIFI (System Average Interruption Frequency Index) as short-term and long-term forecasting and other complex is 12—equivalent to one hour-long outage each month procedures. This means that mini grid development—as a (Balabanyan and others 2021). Information on load factor viable strategy for delivering universal access to electricity— for these utilities was not available. entails a much stronger utility sector able to accommodate interconnecting mini grids with the main grid. WIN-WIN FOR MINI GRIDS AND NATIONAL UTILITIES How this can be achieved is laid out in several chapters that follow, including chapter 5, 8, and 9. However, scaling up mini grids does not mean scaling back the main grid. On the contrary, third-generation mini grids enhance the economic viability of expanding the main grid. By designing a system from the beginning to interconnect CURRENT STATUS OF SELECTED MINI with the main grid and by promoting income-generating GRID CAPITAL AND OPERATING COSTS uses of electricity through effective community engagement and training, third-generation mini grids can provide early COST PER UNIT OF FIRM POWER OUTPUT economic growth, so that significant load already exists by Of primary concern to developers and program adminis- the time the main grid arrives and customers have a greater trators alike is the total investment cost of the mini grid, by ability to pay. New regulatory frameworks give developers capacity. As a precursor to this discussion, it is important viable options for what happens when the main grid arrives, to define what we mean by mini grid capacity. How much and reductions in the cost of components enable develop- electric power (kW) can a mini grid reliably provide on an ers to build grid-interconnection-ready systems while still ongoing basis? For conventional dispatchable power plants, keeping tariffs affordable. New grid-connected mini grid this is called firm capacity, commonly understood as the business models are also providing win-win arrangements intended, sustained output of the facility at times of full load. in which utilities lease distribution assets and sell backup power to mini grid operators that take on customers that There is no consensus in the mini grid industry on how utilities have found unprofitable to serve directly, or that to define a metric comparable to firm capacity, based on require higher levels of service than the utility is able to pro- component specifications. The amount of power available vide (Tenenbaum, Greacen, and Shrestha 2022). at any given moment from a mini grid is shaped by diverse factors: the capacity of the solar array, the amount of sun- As a result, supporting third-generation mini grids goes hand light, the capacity of the inverter, the storage capacity of in hand with strengthening the utility sector. Interconnecting the battery bank, the capacity of the diesel generator, and, third-generation mini grids with the main grid as part of a as a practical matter, the availability of diesel fuel. Some of national electrification strategy can increase the resource these factors are at times limiting, and others offer alter- diversity and overall resilience and efficiency of the power native pathways to provide electricity. Some have specific system when interconnection is properly planned and exe- time durations or depend on factors such as the weather. In the absence of an industry metric, we offer an imperfect, As the cost of mini grid electricity continues but we believe useful back-of-the-envelope definition of to fall while its quality continues to rise, mini mini grid firm power. We define the firm power output of a grids will become competitive with the main grid in mini grid as the generator capacity (kW)17 plus 25 percent more and more countries. When the levelized cost of the solar array output rated peak (direct current, DC) of mini grid electricity hits $0.20/kWh, it will be less power output (kWp). In much of the world, about six hours expensive to produce than main grid electricity in of midday sunlight is the daily allotment—roughly the sun- 24 of 39 African countries. However, scaling up mini light intensity at which solar panels produce their rated grids does not mean scaling back the main grid. On output.18 Six hours is 25 percent of a full day. the contrary, third-generation mini grids enhance kWfirm = kWgen + 0.25 kWpPV the economic viability of expanding the main grid, and scaling up mini grids requires much stronger This definition assumes that the battery is sized large utilities, able to accommodate interconnection over enough to store sufficient solar electricity to redistribute time as the main grid expands. it over periods with inadequate sunlight, accounting for inefficiencies. But in some areas, especially those with a 54   MINI GRIDS FOR HALF A BILLION PEOPLE prolonged rainy season, this will not be possible. The defi- more of annual energy, solar panels or solar plus storage nition also assumes that diesel supply is not constrained. are sufficient. The low marginal cost of a diesel genset’s Thus, mini grids with large diesel generators seem to have installed capacity makes it affordable for it to carry the greater capacity than mini grids with strong solar invest- entire load. However, given high fuel costs, there is strong ment, despite the fact that operating a diesel generator for incentive to operate generator only when there is abso- anything other than backup generation is not cost effective. lutely no other choice. Within this context, the firm power The definition also ignores the effects of temperature on metric delivers valuable information on the ability of a sys- PV power output, as well as power lost through efficiencies tem to power a load for days or even weeks. Outside these in energy storage and conversion. On the other hand, the contexts, the kWfirm metric should be treated with caution. definition underestimates power available during sunlight Of 356 mini grids, there are wild cost variations per hours, which could reach as high as the sum of the genera- kilowatt of firm power (kWfirm) output (figure 1.4). The tor output, the PV array output (technically, the AC output median economic cost was $5,084 per kWfirm, while the from the PV inverter), and the battery inverter capacity. 25th and 75th percentile economic costs were $3,760 Despite these shortcomings, we find the metric useful and $6,953 per kWfirm, respectively. Most mini grids because it is easily calculated with available data, and gives below 200 kWfirm have costs around or below $5,000 an indication of the rough magnitude of a constant load per kWfirm. The economic cost of a best-in-class mini that could be powered by the mini grid for many days, if not grid in 2021 was $3,659 per kWfirm. indefinitely, if adequate diesel supply were available. The As expected, in general mini grids display economies of premise is that the provider of the electricity service can scale in generation, with smaller mini grids costing more guarantee electricity delivery upfront for any time the con- per kWfirm than larger mini grids. Most of the highest cost sumer wants it. The definition is useful for contemporary per kWfirm projects are built as individual projects and often solar mini grids that are basically designed as solar-storage lacked backup diesel generators. Many of these mini grids systems that have a diesel generator backup for less than represent early efforts to understand the marketplace and 10 percent of annual energy. For the other 90 percent or experiment with new technologies with less focus on cost FIGURE 1.4 • Total economic cost of mini grids per kWfirm as a function of firm power output 40,000 35,000 30,000 25,000 Cost per rm kW 20,000 15,000 10,000 5,000 0 0 100 200 300 400 500 600 700 800 Firm power output (kW) kW = kilowatt. MINI GRIDS FOR HALF A BILLION PEOPLE    55 reduction. Data on financing sources were not collected, INVESTMENT COSTS PER CUSTOMER but it is likely that these costlier projects had larger shares The median cost per customer for village mini grids was of grant funding. While generally only small projects (under $846, with 25th and 75th percentile costs of $468 and 50 kWfirm) had exceptionally high costs per kWfirm, many $1,413, respectively. As figure 1.5 shows, however, there are small projects also had low costs in this metric. We should notable and unsurprising outliers—many from 2020 and note that the lowest-cost mini grids from an LCOE per- 2021 mini grids reporting only a few customers—among spective are not necessarily the lowest cost in terms of firm newly commissioned mini grids. Others were relatively capacity. If a low LCOE is desired, high solar utilization is costly pilot projects, sometimes one of a kind, built with required, whereas mini grids with large diesel generators less emphasis on cost reduction. tend to score low on the $/kWfirm metric because diesel generators provide cheap capacity, albeit expensive to fuel. The total costs per customer reflect economies of scale, With this in mind, it is worth emphasizing that the $/kWfirm as mini grids serving more customers have lower costs metric is not the design parameter for optimization—but per customer on average. If only those mini grids with this indicator is useful when estimating investment. per customer costs below the median are included, every additional 100 customers lowers the per customer cost by Some high $/kWfirm projects appeared to follow a deliberate about $9. strategy of overbuilding their distribution network (in terms of both quality and scale), to easily accommodate upgrades Of the 411 mini grids in the database, a majority (217) had in generating capacity as the load grows. Other factors between 200 and 600 customers; 82 mini grids had fewer explaining the wide variation may be the amounts and ways than 200 customers, and 112 had more than 500 custom- in which project development costs are internalized into a ers. The preference for mini grids under 500 customers project or absorbed by a company and not reported as a may be because mini grids beyond this scale, both in terms mini grid development cost, the cost of doing business in of distance and cumulative consumption, require trans- the country in question, and a lack of competitive tendering. formers and medium-voltage lines to distribute power. Lack of regulatory certainty regarding grid arrival may also be a factor: in the absence of regulation, project developers may be choosing small sites farther from the main grid that Mini grid costs vary across projects and are less attractive for potential grid expansion. When a ver- countries, with a median cost per kilowatt of sion of this study was conducted in 2018 with 53 mini grids, firm power output of about $5,000. A low cost rel- the most popular size (28 mini grids) served under 200 ative to firm capacity points to the likelihood of the customers, likely reflecting relatively early preferences by mini grid having a large diesel generator relative to developers to limit risk by testing the waters with smaller its solar array, which could, in turn, raise the level- communities. ized cost of energy if it is dispatched often. FIGURE 1.5 • Mini grid economic costs per customer (left) and costs per customer for mini grids below median cost (right) 5,000 900 Cost per customer (USD) 800 Cost per customer (USD) 4,000 700 600 3,000 500 400 2,000 300 1,000 200 100 y = –0.0865x + 572.24 0 0 0 2,000 4,000 6,000 8,000 0 2,000 4,000 6,000 8,000 Number of customers Number of customers Source: ESMAP analysis. 56   MINI GRIDS FOR HALF A BILLION PEOPLE The data suggest that, on average, for every The largest cost components of the mini additional 100 customers a mini grid serves, grids in our data set were distribution (27 its per customer costs fall by about $9. Whereas in percent of total capital expenditure), batteries (15 2018 most mini grids had fewer than 200 custom- percent), import duties and taxes (12 percent), and ers, today most serve between 200 and 600 cus- installation (13 percent). tomers. cult for anyone to tell, especially early on in a project’s life COST OF INDIVIDUAL COMPONENTS cycle, how profitable the project will be and thus what the ultimate profit margin will be. For mini grids that were con- Solar mini grid components include solar panels, batteries, structed as engineering, procurement, and construction generators, inverters, other electronics, the distribution contracts, our data are derived from bid responses and in network, powerhouse, shipping and logistics, and installa- this case the profit margin is not explicitly stated, but rather tion. Mini grid total costs also include soft (but very real) is blended into line items (equipment costs, management, business and project development costs, such as site iden- installation, and so on). tification, demand assessment, design, and the process of obtaining necessary approvals. Table 1.4 provides a sum- It is important to note that the portion of reported costs that mary of the costs and characteristics of individual mini grid each component accounts for ranges widely across mini components from 351 mini grids in our database. The sec- grids (table 1.5). PV solar panels in a mini grid in Nepal, for tions below unpack the details of these components. example, were reported to cost nearly twice what they did in neighboring India. Li-ion batteries cost more than double In addition to the summary data reported in the table on a per kWh basis in Indonesia in 2017 than what they cost above, sufficient data were available from 294 mini grids to in Ethiopia for a 2021 mini grid. Distribution costs per cus- determine the average share of mini grid cost attributable tomer were more than five times higher on a per customer to each component (figure 1.6). The largest cost compo- basis in a 2016 Côte d’Ivoire project than reported in a 2021 nents were the distribution grid (26.6 percent), batteries project in India. (14.9 percent), installation (11.3 percent), PV modules (9.7 percent), and taxes and import duties (11.5 percent).19 There are likely multiple reasons for these differences. Taxes and import duties were back-calculated based on Most pronounced is that costs have come down over time tax and import duty rates provided by developers for the for major components, yet the data in the table above do countries they build and operate mini grids in, and vary not distinguish the year of project commissioning. This considerably from country to country and from compo- remarkable change in cost over time is discussed below. nent to component. The profit margin (reported at only 0.3 Another factor is that developers in different countries fold percent of economic costs in our data set) merits unpack- the profit margin into component costs in various ways. ing and further research. On the one hand, most mini grid Other reasons may include differences in interpretation projects earn revenues on electricity sales and it is diffi- by mini grid developer respondents to the surveys and TABLE 1.4 • Mini grid components: A summary of costs and characteristics Technical characteristics Costs Number of mini grids with 25th 75th 25th 75th Component available data percentile Median percentile percentile Median percentile Solar panels 351 45 kWp 76 kWp 125 kWp $388/kWp $441/kWp $599/kWp (including PV inverter) Battery Lead-acid: 133 144 kWh 288 kWh 432 kWh $154/kWh $193/kWh $224/kWh Lithium ion: 217 102 kWh 180 kWh 312 kWh $271/kWh $314/kWh $414/kWh Inverter + EMS 313 30 kW 63 kW 118 kW $325/kW $415/kW $716/kW Distribution and meters 317 N/A N/A N/A $163/kWp $250/cust $331/cust Customers 350 238 404 644 $480/cust $836/cust $1290/cust Source: ESMAP analysis. Note: percentile technical characteristics and unit costs are for each component separately. The rows should not be read together and interpreted in aggregate to represent the component capacities of a “25th percentile” or “median” mini grid. MINI GRIDS FOR HALF A BILLION PEOPLE    57 FIGURE 1.6 • Average share of component economic costs in total capital costs of mini grids 11.5% 0.3% 100% 90% 11.3% 80% 4.6% 26.6% 0.6% 70% Percent of cost 60% 50% 6.1% 8.6% 40% 14.9% 30% 20% 9.7% 10% 5.9% 0% t es y S S n nd s n s n en ic tie er io tio gi EM BO ul La st ar ut tt em du la od gi Ba M ib & al ag Lo m & tr s st er is s an PV In xe D rt M ve Ta In “Management + SG&A” (selling, general and administrative Distribution and smart meters expenses)—including engineering, planning, permits, approvals, Land licenses, and community engagement Logistics and installation Generation, including PV modules, battery, inverters, and Taxes and duties balance of system (BOS), which includes a diesel backup system Pro t margin Note: For costing purposes, PV inverters and PV controllers are grouped together. BOS = balance of system; EMS = energy management system; PV = photovoltaic. TABLE 1.5 • Average economic costs of key mini grid hardware components, by country Solar panels Lead-acid Lithium-ion Battery inverter Distribution $/kWp battery $/kWh battery $/kWh $/kW $/customer Bangladesh 622 185 — 1,242 355 Ethiopia 504 — 285 — 385 Ghana 798 143 — 1,011 520 Guinea Bissau 801 129 — 1,612 283 India 445 115 — 1,225 155 Indonesia 601 — 625 1,017 316 Ivory Coast 688 106 — 707 933 Kenya 834 142 — 928 307 Myanmar 497 231 422 467 321 Nepal 865 152 — 870 392 Nigeria 477 180 331 206 Palestine 656 158 — 1,181 303 Tanzania 585 159 614 1,431 496 Vanuatu 464 141 — 641 704 Minimum 445 106 285 467 155 Average 631 153 455 1,028 405 Maximum 865 231 625 1,612 933 Delta 94 119 119 245 504 Source: ESMAP analysis. — no data available; kVA = kilovolt-ampere; kWh = kilowatt-hour; kWp = kilowatt-peak. 58   MINI GRIDS FOR HALF A BILLION PEOPLE queries that solicited this information. In the distribution case, some projects may have needed only upgrades to existing distribution networks, whereas most others built Mini grids benefit from decreasing global new distribution networks, and there are considerable cost solar module prices, reflected in cost differences in distribution networks that are underground declines in the solar portion of mini grids of around vs. above ground, and for different sizes of poles and con- $32 per kilowatt-peak per year. ductors. The following sections explore in more detail some of the components that account for large fractions of the total Batteries investment cost. Batteries are a huge story in mini grids in the past several years. Li-ion batteries are rapidly becoming the dominant Solar panels (including racking and PV inverter) choice for new mini grids, driven by lower costs enabled Based on data from 351 projects, solar panel economic by their increasing use in consumer appliances, electrical costs for mini grids have been decreasing by about $32/ vehicles, and utility power storage. Of 211 mini grids under kWp per year on average since 2012 (figure 1.7). Based on construction or commissioned in 2020 and 2021, 145 (69 data from 278 projects built between 2019 and 2021, these percent) used Li-ion batteries while 66 (31 percent) used more recent installations had a median cost of $413 per lead-acid batteries. kWp, with a 25th percentile cost of $354 per kWp and a The common metric for battery pricing is $/kWh of battery 75th percentile cost of $599 per kWp. storage capacity. Figure 1.8 indicates economic cost trends In our data groupings, solar panel costs include not only for mini grid Li-ion (blue) and lead-acid (red) batteries. The the solar panels, but also the PV inverters.20 Both are con- graph shows lead-acid battery costs slightly increasing sistent with, yet not strictly comparable with, the mod- from our earliest project data in 2012 but holding roughly ule-level pricing that we discuss among global PV module steady at about $200 per kWh. Increasing lead-acid bat- cost trends later in this chapter. tery costs are consistent with global trends driven by the FIGURE 1.7 • Costs of solar panels (including PV inverters) for mini grids, by year, 2012–21 1,000 900 800 700 600 PV $/kWp y = –31.73x + 64568 500 R = 0.0829 400 300 200 100 0 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Source: ESMAP analysis. Note: The relative capacity (kWp) of the solar array is indicated by the relative size of the dot. kWp = kilowatt-peak. MINI GRIDS FOR HALF A BILLION PEOPLE    59 FIGURE 1.8 • Economic cost trends for the storage capacity ($/kWh) of lithium-ion and lead-acid batteries used in mini grids between 2012 and 2021 700 600 500 Cost of storage capacity ($/kWh) Lithium ion: y = –36.9x + 74962 R = 0.126 400 300 Lead Acid: y = 1.7x–3238.3 R = 0.0034 200 100 0 2011 2013 2015 2017 2019 2021 Lead $/kWh Lithium $/kWh Linear (Lead $/kWh) Linear (Lithium $/kWh) Source: ESMAP analysis. Note: The nameplate capacity (kWh) of the battery is indicated by the relative size of the dot. kWh = kilowatt-hour. increase in the commodity price of lead, especially since TABLE 1.6 • Performance characteristics of lead- mid-2015 (Trading Economics 2022). acid and lithium-ion batteries as modeled in HOMER levelized cost of energy calculations Li-ion batteries appeared first in 2016 among the mini grids we tracked, and by 2018 were in 28 new projects. In 2019 Lithium- Unit Lead-acid ion they were included in 60 new projects and this number has grown every year since. Costs have been declining substan- Cycle life kWh throughput 800 3,000 (throughput) before failure tially for Li-ion batteries and their battery management systems at a rate of nearly $37 per kWh per year. Maximum depth % 60 80 of discharge A casual glance at figure 1.8 would suggest that while Li-ion Roundtrip % 80 90 batteries are decreasing in price, they are still more costly efficiency overall than lead-acid. But this would be an incorrect inter- pretation of the data, as the nameplate kWh capacities of the HOMER LCOE calculations in the first half of this chap- lead-acid and Li-ion batteries are not comparable.21 For ter are shown in table 1.6. a given kilowatt-hour of nameplate capacity, Li-ion bat- teries can be more deeply discharged22 and thus have a The implication of these performance differences is that larger usable kilowatt-hour capacity. Moreover, Li-ion bat- a single 1 kWh lead-acid battery will, over the course of its teries have superior cycle lifetimes (the quantity of kilo- lifetime, be able to cycle 800 times to 60 percent depth of watt-hour of electricity that can be charged and discharged discharge at 80 percent efficiency, storing and releasing into the battery before failure), higher efficiencies, as well 800 x 0.6 x 0.8 = 384 kWh of electricity before it must be as decreased temperature-related degradation, which is replaced. A single 1 kWh Li-ion battery, on the other hand, problematic for lead-acid batteries in tropical countries.23 will cycle 3,000 x 0.8 x 0.9 = 2,160 kWh of electricity, over Differences in lead-acid and lithium batteries as modeled in five times more. 60   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 1.9 • Net present value of storage capacity for lithium-ion and lead-acid batteries, 2012–21 $2,500 $2,000 Lead acid NPV of storage capacity ($/kWh) y = 8.0591x – 15345 R = 0.0034 $1,500 $1,000 $500 Li-ion y = –62.74x + 127322 R = 0.126 $0 2011 2013 2015 2017 2019 2021 Corrected Lead $/kWh Corrected Li/kWh Linear (Corrected Lead $/kWh) Linear (Corrected Li/kWh) Source: ESMAP analysis. Note: The relative nameplate capacity (kWh) of the battery is indicated by the relative size of the dot. kWh = kilowatt-hour. To account for these differences, figure 1.9 compares the net present value (NPV) per kWh of energy storage capac- ity among these mini grid projects. This calculation takes Despite a higher sticker price, lithium-ion into account the factors listed in table 1.6 in a discounted batteries have replaced lead-acid batter- cashflow calculation that accommodates the battery’s ies due to their superior longevity, efficiency, and replacement schedules as predicted by the HOMER project deeper discharge capabilities. Lithium-ion battery modeling for mini grid projects serving the same load profile. costs are falling while the cost of lead-acid bat- teries is slowly increasing over time, in line with The revised figures (“Corrected Lead” and “Corrected Li” in global increases in the price of lead. Mini grids with figure 1.9) show that Li-ion batteries, despite their higher lead-acid batteries remain competitive, however, sticker price, have proven to be cost-competitive with lead- especially where strong supply chain relations can acid batteries since at least 2018. procure quality lead-acid batteries at high-volume The shift to Li-ion batteries is remarkable considering that pricing, and where discount rates are high. most subsidies for mini grids are for capital, per connection (performance-based grants), and therefore mini grid devel- opers must shoulder the higher upfront cost of Li-ion bat- Although our data set indicates that Li-ion batteries are teries at a time in the project cycle when revenue is not yet now the battery of choice in most mini grid projects, their generated. Moreover, lead-acid batteries were the incum- dominance is not complete. Our data set includes develop- bent technology and benefit from over a hundred years of ers, particularly in Nigeria and India, who are building very tried and tested operation; whereas Li-ion batteries are a competitive mini grids using lead-acid batteries at scale. new arrival accompanied by unknown technical risk as well Developers who have established lead-acid battery supply as the need to develop new supply chains. chains and low pricing through large volumes of orders will likely find it competitive to continue using lead-acid bat- MINI GRIDS FOR HALF A BILLION PEOPLE    61 teries, at least in the short term. Moreover, high discount costs are shown here on a per kilowatt-peak basis because rates reflecting high capital costs will, all things being many (solar support structure, fencing, civil works) are equal, favor lead-acid batteries since their upfront capital proportional to the size of the solar array. Somewhat low costs are lower. The country of battery manufacture is also costs in 2012 and 2014 may reflect shortcomings in data a consideration. Currently China dominates Li-ion battery collection in these categories. Likely contributors to declin- manufacture, whereas countries where lead-acid batteries ing costs are larger economies of scale through larger mini remain popular for mini grids (India and Bangladesh, for grids and clustering. example) have well-established and historically competi- Data collected in the survey circulated for the 2018 ver- tive lead-acid battery industries. sion of this analysis distinguished between conventional masonry powerhouses and powerhouses made from Battery inverters, energy management systems, and shipping containers. When a shipping container is repur- monitoring posed as a powerhouse, typically equipment arrives on site Battery inverters, energy management systems, and mon- prewired in the shipping container, which is also used to itoring compose on average 8.6 percent of project costs. transport the PV modules and racking materials to the site. Based on data from 327 mini grids, over time, these costs Measured in absolute costs, shipping containers as power- have been trending downward on a per kW basis (figure houses were, on average, the lowest cost, with an average 1.10), dropping from an average of $1,204 per kW in 2014 cost of $6,922 and a median cost of $7,235; powerhouses to $524 per kW in 2021, a decrease of nearly $100 per kW constructed on site averaged $29,700, with a median cost each year. This reflects the global decreases in the cost of of $26,253. We have anecdotal evidence, however, to the power electronics, as well as economies of scale both from contrary. Some developers have found working with local larger mini grid sizes over time as well as bulk purchases masons to be cost-efficient, particularly in areas where through expanded deployment. roads are poor, increasing shipping costs and challenges. Balance of system Shipping containers as powerhouses (figure 1.11) were also Balance of system (BOS) costs compose a catch-all cate- the lowest cost on a per kilowatt basis, accounting in the gory for the remainder of generation costs not captured in 2018 version of the study for the five mini grids with lowest the main categories of PV panels, batteries, and inverters. powerhouse cost, while conventional buildings accounted The BOS comprises the diesel generator, solar support for the most expensive five. For mini grids with shipping structures, fencing, foundations, lighting, civil works, pow- containers, the average powerhouse cost/kWfirm was $153, erhouse, and air conditioning system for the batteries, if whereas buildings constructed on site averaged $494/ installed. Based on data from 349 mini grids, average BOS kWfirm. costs are broadly trending downward (figure 1.10). BOS A recent innovation is to use weatherproof cabinets for the battery and energy management system enclosure, placed FIGURE 1.10 • Unit costs for inverters, energy manage- under a PV-paneled roof without the need for any addi- ment systems, and monitoring (blue), and balance of tional structure for a powerhouse. Even though the CAPEX system (orange) of the system is only slightly lower than its alternatives, the 1,400 major savings occur with the transportation and installa- 1,200 tion costs. These cabinets can be transported in lower-cost and more agile pickup trucks (figure 1.11, photo 6). 1,000 800 Distribution 600 For 349 mini grids with comparable data, the distribution costs of recent mini grids tend to cluster between $100 and 400 $500 per customer, with wide variation. Broadly, the distri- 200 bution costs appear to be trending downward slightly year 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Average of inverter $/kW Average of BOS $/kW Shipping containers as powerhouses were Source: ESMAP analysis. Note: Balance of system (BOS) comprises the diesel genset, solar support the lowest cost on a per kilowatt-hour basis, structure, fencing, foundation, lighting, civil works, powerhouse, and accounting for the five mini grids with the lowest cooling. powerhouse costs. kW = battery inverter rated kilowatt capacity. kWp = kilowatt peak of the solar PV array. 62   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 1.11 • Powerhouse innovations can lower costs and expedite deployment Power house: Remote Power Unit (RPU) 40-foot 1.  PV and battery inverters Inside the RPU 2.  shipping container under PV array Location: Bunjako Island, Uganda Location: Bunjako Island, Uganda Developer: Winch Energy Developer: Winch Energy  Photo credit: © Winch Energy. Used with permission by  Photo credit: © Winch Energy. Used with permission by Winch Energy. Further permission required for reuse. Winch Energy. Further permission required for reuse. Power house: 20-foot shipping container kiosk 3.  Brick power house 4.  Location: Katiko, Turkana North, Kenya Location: Kangitan Kori, Kenya. Developer: Renewvia Developer: Renewvia  Photo credit: © Jon Exel. Used with permission by Jon Exel.  Photo credit: © Jon Exel. Used with permission by Jon Exel. Further permission required for reuse. Further permission required for reuse. Micro-grid in a box (MIB) is the taller structure on the 5.  6. Power equipment in outdoor rated cabinets right. The diesel generator stands alone outside on a Location: Danchitagi, Niger state, Nigeria. platform (left). Elevating equipment protects against Developer: PowerGen flooding, increases natural cooling, and reduces risk of  Photo credit: © PowerGen. Used with permission by damage from dust, insects and animals. PowerGen. Further permission required for reuse. Location: rural India Developer: TPRMG Photo credit: © TPRMG. Used with permission by TPRMG. Further permission required for reuse. MINI GRIDS FOR HALF A BILLION PEOPLE    63 by year, with larger systems in recent years appearing to nection charge and then small monthly installments added have lower such costs per customer. to their customers’ bills for the first several months of ser- vice. While this strategy helps customers overcome what Distribution costs include poles, conductors, service drops, would otherwise be a prohibitively expensive one-time and meters, and customer wiring (or prewired “ready connection charge, implementing it adds an administrative boards” that contain a couple of light switches and one or burden—and cost—to the mini grid developer. From the two outlets). Included in this list are smart meters that can customer’s perspective, though, this pricing model is often send and receive data to and from the internet, and gener- familiar because a similar pricing model is used for most ally incorporate pay-as-you-go features by which custom- smartphones, where customers do not pay the full price of ers prepay for electricity (similar to prepaid minutes on a the phone up front but instead a portion of their monthly cell phone). Smart meters can help substantially reduce bill goes toward the cost of the phone. ongoing costs and increase revenues by lowering electric- ity theft; remove the costs of meter reading and postpay Intuition would suggest that increasing the number of cus- billing and collections; and, in some cases, provide data to tomers served would lead to decreases in costs per cus- mini grid operators on vital mini grid technical parameters tomer. Each increase of 100 customers per mini grid lowers that help operators and engineers identify and address costs by about $3 per customer, but the data suggest only problems before they become larger and more expensive. a weak correlation (figure 1.13).24 For mini grids with low consumption needs, a DC mesh can offer lower costs per Variations in cost and the technical sophistication of meter- customer (box 1.2). ing explain some of the wide variation in distribution costs per customer (figures 1.12 and 1.13). Other variation may be attributable to the fact that some mini grids provide inhouse wiring while others do not. Though not tracked in Each additional increment of 100 custom- the survey, it is nevertheless worth noting that the connec- ers correlates with declines in distribution tion fees charged to customers do not necessarily have costs per customer of about $3. But the data only a one-to-one relationship with the connection costs per weakly support this relationship. customer. Indeed, many mini grid developers choose to recoup the connection costs through a small upfront con- FIGURE 1.12 • Distribution costs per customer, 2012 to 2021 1,000 900 800 Distirbution cost per customer 700 600 500 400 y = –24.098x + 48934 300 R = 0.0379 200 100 0 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Source: ESMAP analysis. Note: The size of the installation (number of customers) is indicated by the relative size of the dot. 64   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 1.13 • Distribution costs per customer as a function of customers served 900 800 700 600 Distribution cost per customer 500 400 300 y = –0.0338x + 290.45 R² = 0.019 200 100 0 0 1000 2000 3000 4000 5000 6000 7000 Number of customers Source: ESMAP analysis. BOX 1.2 DIRECT CURRENT MESH GRIDS Although mesh grids are not included in our analysis, appliances, but they can upgrade to AC appliances they nevertheless hold promise for some communities by requesting an inverter. Larger productive use AC and states affected by fragility, conflict, and violence, appliances and higher-consumption households can with lower electricity needs. also be accommodated through networks of inter- connected customers. Approximately 90 percent of Mesh grids—or “skinny grids”—distribute DC electric- Alina’s customers are interconnected with at least one ity for lighting, electronics, and small appliances like other customer; only the remotest 10 percent of cus- fans and even efficient refrigerators or electric rick- tomers are served with isolated systems. shaws. They take the form of clusters of solar home systems made up of solar panels affixed to customers’ Alina encourages productive uses of electricity. The premises and connected in a mesh network. Special- company partners with a local appliance supplier and ized controllers allow surpluses to be shared. Examples conducts multiple community visits and workshops include Okra Solar, with installations in Cambodia, the prior to the arrival of the mesh grid arrival and as it Philippines, and Haiti; and SOLshare in Bangladesh. expands. In the rural Haitian province of Artibonite, Alina Enèji The mesh grids in Haiti typically have a capital cost has built direct current (DC) mesh grids that electrify about $800 per connection, about half of the cost for 300 households and small businesses using Okra conventional AC mini grids in Haiti. The modularity of Solar’s platform and equipment. Households start with the systems makes for quick installation and capacity systems that provide electricity for small, efficient DC upgrades as needed. a. Okra Solar, https://okrasolar.com/. b. SOLshare, https://me-solshare.com/ MINI GRIDS FOR HALF A BILLION PEOPLE    65 Land In particular, our data show how building a portfolio of mini Land comprised only a small portion (0.7 percent) of the grids can help lower costs by bundling approval processes mini grid sample’s average economic costs. Of 356 mini and exploiting economies of scale in project management, grids that provided plausible cost data, only 193 reported shipping, equipment procurement, and installation. land costs, and only 103 projects reported land costs over Project developers reported management costs (includ- $5,000. We are not sure to what extent this reflects lim- ing project development, general administration, planning, itations in data reporting. Land is often provided gratis by engineering, partnership, public relations, permits, approv- communities or local governments as part of agreements als, licenses, community engagement) for 309 sites. Logis- at project inception, although we have anecdotal evidence tics (transportation) and installation costs were reported that obtaining rights to suitable land is often a challenge. A for 327 and 297 sites, respectively. 100 kWp solar array requires about half an acre of land, or a square about 20 meters on a side. Though mini grids built separate from a portfolio tended to have fewer customers (single mini grids average 405 cus- Figure 1.14 shows two examples of solar arrays for mini tomers, while clustered projects averaged 657 customers), grids, illustrating the relatively small amount of land they had substantially higher average soft costs ($208,900, required. The array on the left is a 30 kWp mini grid devel- compared with $127,400 for portfolio projects). oped by Mandalay Yoma in Myanmar, with the powerhouse and diesel generator under the green roof at upper right. Governance requirements and developers’ internalized pre- The array on the right is a 40 kWp mini grid developed by feasibility assessments, technical standards, and account- Winch Energy in Lamwo District in northern Uganda. The ing and reporting requirements may give rise to widely solar array is built over the powerhouse to reduce land varying soft costs across similar projects. requirements, though with additional racking costs. REPLACEMENT COSTS Sales, general and administrative expenses, senior Replacement costs are for repairing worn-out or broken management, logistics, and installation equipment as the mini grid ages. These costs were not The data collected for this chapter illustrate how soft costs explicitly reported in the data that underlie this study, but (project development, logistics, and installation) might be an are nevertheless essential. For long-term sustainability, it is area where economies of scale can lower investment costs. critical to ensure that sufficient funds are available to cover replacement costs. Battery replacements in particular are problematic because it is generally necessary to replace the entire pack in order to ensure that new batteries are not Mini grids built as part of a portfolio saved electrically compromised by older batteries to which they $81,000 in soft costs on average, compared are electrically connected. In this regard, the transition with mini grids built as one-off projects. from lead-acid to Li-on batteries is important. As discussed above, although Li-on batteries have much higher upfront FIGURE 1.14 • A 30 kWp Mandalay Yoma mini grid in Myanmar (left) and a 40 kWp Winch Energy mini grid in Uganda (right) Photo credits: Left © Mandalay Yoma; used with permission; further permission required for reuse. Right: © Winch Energy; used with permission; further permission required for reuse. 66   MINI GRIDS FOR HALF A BILLION PEOPLE costs, they last much longer than lead-acid batteries, to understand whether revenues are sufficient to cover delaying the need for, and ultimately reducing, a project’s OPEX and other costs, such as debt service and equipment long-run replacement costs. replacement. While many projects are too young to report out their replacement costs, they are built into HOMER optimiza- tion and LCOE modeling. The replacement of many com- THE OUTLOOK FOR MINI GRID ponents (PV panels, inverters, monitoring equipment) is CAPITAL AND OPERATING COSTS based on years in service, diesel generators’ replacement is based on hours operated, and battery replacement is As discussed in the sections above, mini grid costs have dictated by the total throughput in kilowatt-hours based on declined substantially on an LCOE basis, including a 31 battery type. percent decline of LCOE in best-in-class mini grids since 2018. In addition, as discussed above, our data show that OPERATING COSTS the costs of key components (especially PV modules, bat- A mini grid’s operational expenses are important, espe- teries, and electronic components) reported by mini grid cially from the perspective of long-term sustainability. developers have been steadily declining. OPEX includes all costs associated with operating mini grid This section draws on research on costs in related global equipment, including fuel costs, maintenance, repairs,25 industries such as solar panels and batteries to better payment collection, and security. understand what levels component costs for mini grids OPEX was reported for 137 systems (113 in Myanmar, 4 in may reach by 2030. Industry trends suggest that compo- other Asian countries, and 20 in Africa). Reported OPEX nent costs in most key areas of generation, storage, meter- per customer varied widely from a low of $2 a year in Myan- ing, and power conversion can be expected to continue to mar to $267 a year in Kenya. decline thanks to their increasing deployment and spillover effects from technological development in much larger sis- Among the 19 mini grids (12 of them in Bangladesh) that ter industries. Table 1.7 and the sections in this chapter that reported a breakdown of staff, fuel, and other O&M costs, follow provide details of the expected cost declines for key fuel on average accounted for 30 percent of O&M, staff mini grid components. accounted for 49 percent, and other O&M accounted for 21 percent. Within this data set, there were considerable The best-in-class 2030 component price assumptions variations. In some, fuel or staff accounted for 0 percent; in used in the HOMER LCOE modeling discussed in the begin- others, 100 percent. ning of this chapter used the following approach: we started with the costs for each component in the best-in-class mini Some of the large variation in reported OPEX may reflect dif- grid from 2021, and then applied the same percentage ferences in staffing needs. Did the sale of electricity require drop to that component that is expected industry-wide. For staff, or was it accomplished automatically through a cell- example, PV costs in the best-in-class, representative mini phone-based prepayment system? Does the site require grid were $596 per kWp in 2021. Global industry PV, with security guards? How is the O&M of the mini grid plant a 2020 benchmarked cost of $198 per kWp, is expected accomplished? Are some staff responsibilities conducted to drop another 42 percent to $114 per kWp by 2030. For on an unpaid basis? Did the mini grid initially not work prop- the 2030 mini grid cost estimate, the same 42 reduction erly and therefore require more intense support until the is applied to the 2021 best-in-class price, yielding a 2030 system was operating robustly? How was OPEX allocated estimate of $343 per kWp (including PV inverters). This on a component basis? The data set does not provide suffi- is still several times higher than the industry benchmark ciently detailed information to answer these questions. price, reflecting the realistic cost multipliers that translate Developers face choices between CAPEX-intensive and an industry spot market price into the cost at a remote mini low OPEX installations (for example, a contemporary solar grid site far from a factory. hybrid system) versus those involving low CAPEX and high PV MODULE TRENDS OPEX (for example, diesel-fueled mini grids). With the availability of subsidies to help cover CAPEX for renewable Mini grids benefit from decreasing solar module prices, energy mini grids, effort is often made to set affordable tar- driven mostly by large grid-connected installations. PV iffs that cover OPEX and replacement costs. prices have fallen faster and lower than nearly any forecast. As of April 2021, global spot prices averaged $198 per kWp Further research to revisit the OPEX costs of the ana- for poly-crystalline modules (Energy Trend 2021).26 By the lyzed mini grids would be useful to understand how OPEX time PV modules arrive on the project site and are included changes over time, and how staff, fuel, and other OPEX with PV inverters they cost considerably more than global components evolve. Further research is also necessary MINI GRIDS FOR HALF A BILLION PEOPLE    67 TABLE 1.7 • Mini grid component cost benchmarks and price projections Component Unit Share Median Best-in-class Mainstream Mainstream Mainstream Best-in-class of total cost in 2021 LCOE industry industry industry 2030 LCOE capital ESMAP modeling benchmark benchmark estimate by modeling cost survey assumption in 2010 in 2020 (% 2030 (% assumption (%) change from change from (% change 2010) 2020) from 2020) PV module $/kWp 9.7 $441 $596 $1,589 $198 (–88) $114 (–42) $343 (–42) PV inverter $/kWp * * * $320 $80 (–75) $70 (–12.5) * Battery (Li-ion) $/kWh 14.9 $314 $297 $1,160 $126 (–89) $58 (–54) $137 (–54) Battery inverter $/kVA 8.6 $415 $303 $565 $113 (–63) $99 (–12.5 $265 (–12.5) Smart meters $/ †‡ ‡ ‡ $106 $40 (–62) $35 (12.5) ‡ customer Sources: ESMAP analysis; Bloomberg New Energy Finance Solar Spot Price Index; National Renewable Energy Laboratory U.S. Solar Photovoltaic System Cost Benchmark: Q1 2020; Feldman and others 2021; Kairies 2017. * PV inverter is included with PV module cost. † Battery inverter is grouped with EMS and monitoring equipment. ‡ Smart meters are included in distribution cost. Average, median, minimum, and maximum costs are all expressed in inflation-adjusted dollars. kVA = kilowatt-ampere; kWh = kilowatt-hour; kWp = kilowatt-peak; Li-ion = lithium-ion; PV = photovoltaic. spot prices. The median cost of PV (with PV inverters) for all From a materials perspective, new cell technologies like mini grids in our database was $441 per kWp. Just counting perovskite cells promise to radically reduce the amount of those mini grids commissioned between 2019 and 2021 highly pure silicon material required in solar cells, as well gives us a median cost of $413 per kWp, with 25th and 75th as improving efficiencies, paving the way to lower produc- percentile costs of $354 and $599 per kWp, respectively. tion costs (US DOE n.d.). Currently about 41 percent of the world’s supply of high grade polysilicon for solar panels Module prices have been roughly following Wright’s Law,27 comes from Xinjiang, China, a region where human rights falling 18–22 percent for every doubling of installed capac- groups and numerous governments have reported ongoing ity (Yu 2018). With growth rates averaging about 40 per- forced labor violations, specifically against the Uyghur eth- cent a year through 2017,28 production doubled about every nic minority (Jenkins 2022). 1.8 years. In recent years, growth slowed to about 24–27 percent, reflecting a doubling every 2.8 years. PV INVERTER TRENDS At the end of 2021, a cumulative total of 843 GWp of PV had PV inverters used in mini grids are similar (or in many been deployed, with 133 GWp commissioned in 2021 alone cases identical) to those used in residential and commer- (IRENA 2022). Bloomberg New Energy Finance (BNEF) cial grid-connected installations, which are projected to projects that solar PV prices will drop to $114 per kWp by increase by about 200 GW in China alone between 2021 2030 (BNEF 2020a), with a cumulative 2.4 terrawatts-peak and 2026 (IEA 2021). PV inverters used in mini grids are (TWp) of PV installed by that year. This reflects a compound also smaller cousins to the grid-tie PV inverters used in average 13 percent annual growth rate for solar PV, a con- utility-scale PV installations that account for the lion’s siderable decrease from contemporary growth levels. share of the 204 to 252 GW of PV expected in 2022 (BNEF 2022). These large volumes, together with cost declines There are wide variations in estimates of total PV that will associated with rapid expansion of other power electronic be added by 2030. As of 2020, new builds of solar PV farms markets, such as motor drives for electric vehicles, will con- are competitive with the marginal cost of existing conven- tinue to drive down costs for PV inverters and controllers. tional generation such as coal, nuclear, and combined cycle natural gas (Lazard 2020). With decreasing PV costs and A study by the National Renewable Energy Laboratory increasing electrification of transportation, heating, and (Feldman and others 2021) identifies a benchmark price industry, some scientists are envisioning that PV’s current of $80 per kWAC for three-phase string inverters (including annual percentage growth will be maintained for the next the cost of monitoring equipment, in 2019 US dollars) in decade, hitting 10 TWp of PV by 2030. To accommodate this the first quarter (Q1) of 2020 for commercial scale PV (100 level of PV would require considerable utility-level storage kW to 2 MW). Inverters are less than a third what they cost a and expanded ability to dispatch load (Haegel and others decade ago: in Q1 of 2010, three-phase inverters cost about 2019). If Wright’s Law continues to hold, expansion to 10 $270 per kWAC (also in 2019 US dollars). TWp is consistent with a price drop to below $90 per kWp. 68   MINI GRIDS FOR HALF A BILLION PEOPLE Assuming a conservative slowdown in the rate of decrease BATTERY INVERTER TRENDS in costs between 2021 and 2030, ESMAP estimates that PV Decreasing battery inverter costs are consistent with inverter costs could reach $70 per kilowatt-peak by 2030. broader trends in power electronics, driven by synergies with PV inverters and electric vehicle motor drives. While BATTERY TRENDS broader industry data were not available for battery invert- Costs for Li-ion batteries have declined dramatically since ers, using PV inverter costs as a proxy, ESMAP estimates 2010 and are expected to continue to decrease substan- that battery inverter costs will reach $99/kVA by 2030, tially. BNEF reported in December 2020 that Li-ion (pack assuming the same 12.5 percent decline by 2030 expected level) battery benchmark costs were at $126 per kWh on for PV inverters. a volume-weighted average basis, down from $1,100 per kWh in 2010. Even with rising commodity prices in the wake SMART METER TRENDS of COVID, BNEF predicts an average cost below $100/kWh The global market for smart meters has seen rapid growth for batteries by 2024 (BNEF 2021). It predicts an average in recent years, driven by strong policy support in China cost of $58/kWh in 2030 (BNEF 2020b), reflecting a cost and Europe. European utilities are projected to install 182 reduction of 54 percent from 2021 benchmark prices. million smart meters between 2016 and 2020, totaling Many Li-ion batteries used in high-end electric cars use nearly $38 billion in investment. Likewise, in Japan, 55 mil- cobalt in their cathodes to increase their energy density to lion meters costing $16.6 billion were installed between provide greater ranger or power output. Lower-end electric 2016 and 2020. Globally, the smart meter industry is a $20 vehicles (such as those increasingly sold in Chinese mar- billion a year market, expected to reach $30 billion by 2026 kets) and stationary power applications such as mini grids (Smart Energy International 2018; Global Industry Analysts or utility storage use lithium iron phosphate (LFP) batter- 2022). Smart meters used in mini grids are in some cases ies, which are less expensive, and less exposed to the risk identical to those deployed in large numbers by utilities. In of supply shortages for cobalt. Bloomberg found that LFP other cases, they are built specifically for the mini grid mar- cells were almost 30 percent cheaper than batteries with ket, with functionalities such as load dispatching, which are cobalt (BNEF 2021). Cobalt is a rare metal and the Dem- included to optimize mini grid load factor by reducing peak ocratic Republic of Congo accounted for 70 percent of demand but still benefit from technology improvements global production in 2019, and substantial concerns have and component cost reductions driven by the larger smart been raised concerning child labor and other human rights meter industry. abuses in cobalt mines in that country (Sanderson 2019). Competition among prepayment metering manufactur- The industry benchmark cost for lead-acid batteries is ers and increasing scale will allow development costs to $143 to $147/kWh (Kairies 2017; Wagman 2020). Lead-acid be spread over a larger product base. BNEF tracks smart technology is largely developed, but the industry makes meter installations and investments globally and found that improvements every year. For example, carbon added to the global average cost per smart meter in 2010 was about the negative electrode will reduce sulfation and increase $106 per unit; 2021 benchmark costs per smart meter in charge rates. ESMAP was unable to find cost projections low-income countries were around $40. ESMAP expects for lead-acid batteries by 2030, but efforts to increase their costs to continue to decline at a rate equal to inverters, life cycles beyond 800 and into the thousands would have reaching unit costs of $35 in 2030. the effect of reducing lead-acid battery storage costs even if TRENDS IN OTHER CAPITAL COSTS nominal costs remain the same. Even so, lead-acid batteries are unlikely to reclaim the mantle of battery of choice from One cost-saving trend in low-voltage distribution is the Li-ion based on expected cost decreases in Li-ion batteries. increase in local factories that build hollow reinforced con- crete poles.29 These poles are manufactured through a Other battery chemistries promise long-term energy stor- process in which concrete is poured into a mold together age that may allow solar mini grids to remove diesel gen- with reinforcing steel. The mold is spun, while centrifugal erators entirely and yet maintain high reliability. Recent force compacts the concrete into a smooth hollow cylin- innovations in iron air batteries have led to price targets der (Taizhou Amity Care 2013). Spun poles have higher for this technology at less than $20/kWh for 100 hours strength per unit weight than solid poles and require much of storage. Because of their lower-current, higher internal less concrete, lowering transportation costs by 25 percent impedance, and longer-duration chemistry, iron air batter- or more compared with solid concrete poles. Developers ies would not replace but could complement Li-ion batter- from Myanmar report that depending on local availability ies, allowing mini grids to weather a week or more of cloudy of sand and gravel, on-site construction of concrete poles weather (Plautz 2021). may also lower costs (Zaw Min 2019). MINI GRIDS FOR HALF A BILLION PEOPLE    69 For powerhouses, though the data on shipping container • The availability of big data that provide geotagged powerhouses appear promising, more research is needed points of interest that can be used to prepare a detailed to understand whether the construction cost savings of demand assessment of prospective load centers shipping containers outweigh the thermal management • Affordable high-resolution satellite imagery issues that arise from their use in hot, sunny environ- ments, and the other engineering issues (for example, • Easy-to-use but sophisticated software that can be stackable batteries) required to repurpose these contain- used to design hybrid generation systems together ers. Another option, still in its infancy, is building mini grids with the design of the distribution network that need no powerhouse, in which components are shel- • Data-driven web-based platforms that compile large tered under the solar array. The Rockefeller Foundation– amounts of geotagged market intelligence that can be supported Smart Power India program has partnered configured in different ways to be useful for mini grid with the Institute for Transformative Technologies in an developers, financiers, and government agencies approach that combines a 10 kW PV array with all neces- The introduction of geospatial and other digital technolo- sary electronics into modular units that can be scaled up, gies has decreased the cost of preparation and planning depending on the situation. In India, Tata Power Renew- by an order of magnitude (see chapter 2 for more details able Microgrids has targeted the installation of 10,000 on geospatial planning). In the past, the unit cost per microgrids using standardized equipment packages built site was more or less the same, irrespective of the num- around a mass-produced “micro-grid in a box” (figure 1.11). ber of sites—about $30,000 per site—because each site Increased factory integration of components in a “utility required a high level of on-site analysis. Today, portfolios in a box” model will lower on-site assembly requirements. of mini grids can be prepared to the point where they Through these economies of scale, the Rocky Mountain are ready for full feasibility assessment and community Institute expects these other CAPEX components to drop engagement at a cost of about $2,300 per site, based on by 15 percent (Carlin and others 2018). the World Bank’s recent experience in Nigeria. Meanwhile, the next generation of diesel generator incor- The socioeconomic surveys and energy audits (looking at porates power electronics in ways that allow engines to demand and willingness/ability to pay) make up 58 per- operate at variable speed as needed, increasing energy cent of per-site costs, which are largely linear since human efficiency (AP News 2018). Variable-speed generators, resources are the primary drivers. The time required together with a dump load and short-term battery storage for a household survey will not change with the scale of buffer, can accommodate up to 100 percent renewable the exercise, although streamlined travel logistics might energy penetration (Innovus 2015). One approach uses a produce savings. Nevertheless, technology can expedite fast-acting clutch that can disengage the motor from the these labor-intensive tasks—for example, through the use alternator when the renewables can fully support the load. of drones to map out a village and sequence household The alternator remains spinning, providing reactive power visits by enumerators. Tablet-based software can swiftly and voltage and frequency regulation (Danvest Energy and accurately capture survey data. Partnerships with 2019). cell-phone-based electronic payment companies can Local production of relatively low-tech items like PV racks obtain market data from targeted rural customers on has the advantage of low labor costs, low shipping costs, appliance purchases or other spending patterns. and local economic development. The mass produc- tion of these items in large factories can, however, take TRENDS IN OPERATING COSTS advantage of economies of scale. As mini grids scale up The introduction of remote-controlled, prepay smart and competition intensifies, mini grid developers in each meters has slashed labor costs. Reaching delayed or non- country can be expected to find context-specific solutions paying customers can now be done remotely. Consump- that optimize the costs of these components. tion patterns can also be tracked and analyzed remotely. In addition, preparation and planning costs have declined. Smart meters and cell-phone carrier-based, real-time data In the past, multidisciplinary teams prepared electrifi- collection enable detailed monitoring of system parame- cation plans, scoped sites, and conducted prefeasibility ters. When parameters exceed programmable thresholds, studies, at considerable cost. Today, most of this work,30 alarms alert technicians of problems that are much easier to address before they grow and cascade into expensive all the way up to feasibility-level analysis—including com- equipment failures and prolonged downtime. In addition, piling bills of quantity and bid documents or purchase smart meters enable developers to easily collect and ana- orders—can be done from behind a desk, thanks to the lyze their performance data, which can be aggregated and following factors: anonymized to share with development partners, indus- 70   MINI GRIDS FOR HALF A BILLION PEOPLE try associations, investors, and other stakeholders. Data grids themselves benefit from economies of scale due to uploaded to the cloud can be analyzed by machine learn- increasing portfolio size and from industry scaling at the ing algorithms, and allow early identification of problem- country level. As mini grid developers scale their portfolios atic patterns. Some companies are planning to respond from 10 to 100 and then to 1,000 or 10,000 mini grids, fixed to customer inquiries with artificial intelligence systems. costs like administration and management are spread over more units of production; sometimes a company can nego- Replacement costs have also fallen. Projects installed with tiate lower per-unit costs enabled by bulk purchases. lead-acid batteries that last three to six years can, if the bat- tery inverters are compatible, be replaced with Li-ion bat- To explore this effect, we analyzed mini grid cost data to teries with a life of ten or more years. Developers building discern changes across categories arising from portfolio mini grids would be wise to choose battery inverters (and and in-country market sizes. Categories included hardware battery chargers in the case of DC-coupled systems) com- (PV modules, batteries, inverters, and so on), manage- patible with Li-ion batteries. Replacement costs for elec- ment, logistics, and installation. We also made estimates of tronics, such as PV inverters and battery inverters, are also the net present value of the ongoing costs of O&M, major falling as they are manufactured at larger and larger scales. equipment replacements, and engaging with customers. To align portfolio-scale projections with declines in equipment Other costs are incurred in dealing with bureaucratic spot prices (see above), we assumed a representative port- processes, such as obtaining licenses, approvals, and folio size of 100 mini grids and doubled this at each time permits. These costs depend on a country’s enabling interval to 200, 400, 800, and 1,600 mini grids per portfo- environment. Several governments have incorporated lio. We also had the portfolio grow by orders of magnitude to mini grids as part of their energy policy, giving them and stress-test the boundaries of the different cost categories. the industry a place in the energy sector. Some countries have adopted mini grid regulations that allow for a light- The results show that with economies of scale, significant handed approach. In some countries, e-government has shifts are taking place in the three cost categories (table streamlined the process for obtaining location and build- 1.8). Overall, the portfolio development and management ing permits. Even though these costs are important, they cost category remains small, with less than 5 percent of are not expected to change much over the next decade the cost over the lifetime of the portfolio. This indicates in countries with high energy deficits, not unless enabling that additional cost reductions will have limited impact environments are introduced in these countries. on the overall LCOE of the portfolio. What is not incorpo- rated in the calculation is the cost of delay in processing for permits, licenses, approvals, and other red tape. More THE IMPACT OF ECONOMIES OF surprising, perhaps, is the minimal difference between the extended CAPEX and OPEX. On average the CAPEX con- SCALE tributes a little more than half the cost of the LCOE, while In addition to the benefits of decreasing spot market prices the OPEX is close to 45 percent, suggesting that the LCOE from the deployment of PV panels and batteries in large is sensitive to the makeup and design of the cost structure global industries like solar farms and electric vehicles, mini of O&M and major repairs over the lifetime of a project. In TABLE 1.8 • Net present value broken down by category with economies of scale Procurement, construction, Operation, maintenance, Portfolio size (number Portfolio development installation, and customer major replacements, and of mini grids) and management (%) engagement (%) customer engagement (%) 100 4.8 53.0 42.2 200 4.0 53.3 42.7 400 4.0 53.1 43.0 800 3.7 52.9 43.4 1,600 3.5 52.6 44.0 100 4.8 53.0 42.2 1,000 3.6 52.5 43.9 10,000 3.4 56.8 39.8 100,000 2.9 48.3 48.8 Source: ESMAP calculations and analysis using costing data described in this chapter. MINI GRIDS FOR HALF A BILLION PEOPLE    7 1 part due to difficulties in obtaining OPEX data, this topic REASONING FROM FIRST has not received the same level of attention in this hand- book and deserves more scrutiny in future work. PRINCIPLES Closer scrutiny of unit costs reveal important changes To further break down the complex setup of a solar mini (table 1.9). When doubling the size of the portfolio stepwise, grid, we tried to reason from first principles and analyze the from 100 to 1,600 mini grids per portfolio, unit costs plunge system’s basic elements. This analysis is a first attempt to across categories. Also, the LCOE falls from $0.36/kWh determine the cost asymptote for the hardware of a solar with a load factor of 22 percent for a portfolio of 100 mini mini grid. It is also an invitation to interested experts, stu- grids to $0.21/kWh for a portfolio with 1,600 mini grids. A dents, and professionals to elaborate further. We took the load factor of 40 percent produces a similar trend. typical system (see box 1.1) that consists of a 285 kWp solar system, a 690 kWh Li-ion (LiFePO4) battery, and a The analysis suggests that all component costs of mini 285 KVA back-up generator set. grids will see declines, but as imported equipment costs (PV modules, batteries, electronics) tumble downward As the generator set is expected to phase out over time through spot markets, the remaining components will due to economic forces, and optimization of this system assume more of the share of overall costs. The NPV of has been ongoing for more than a century, we have used ongoing major replacements such as batteries benefits a specific cost for the full system of $100/kW. For the solar from the future size of portfolios: they will be larger and unit and battery systems we looked into the composition of the costs lower. For example, the batteries installed at year 7 basic elements and found on the commodity market the to replace a failing pack will be part of a scaled-up battery estimated cost for each material. purchase to build 5,000 mini grids. A typical solar mini grid system needs an estimated 20 tons When moving from 100 to 100,000 mini grids in a portfolio, of glass, 16 tons of steel, 13 tons of concrete, 5 tons of alu- the marginal gain diminishes in terms of percentages. The minum, 2 tons of silicon, 2 tons of copper, 2 tons of plastic; largest gain is made from 100 to 1,000 and from 1,000 to the Li-ion batteries require an estimated 650 kg of alu- 10,000 systems per portfolio. Growing beyond this scale minum parts, 450 kg of graphite, 400 kg of copper parts, might call for closer scrutiny; perhaps multiple, smaller 250 kg of iron, and about 50 kg of lithium. Adding value to portfolios (several of 10,000 mini grids) might be optimum. these raw materials resulted in a total cost of $157k for a Additional research will need to be conducted to obtain solar-battery-genset power plant. This is a 53 percent cost more specific insights for the industry. reduction from what is reported in box 1.1 ($333,000 for the generation system). TABLE 1.9 • Change in unit costs with economies of scale, by cost category Portfolio development Procurement, construction, and management installation, and customer Total NPV per LCOE with 22 LCOE with 40 Portfolio size per mini grid (US$, engagement per mini grid mini grid (US$, percent load percent load (# of mini grids) thousands) (US$, thousands) thousands) factor ($/kWh) factor ($/kWh) 100 23 251 473 0.36 0.20 200 17 220 412 0.31 0.17 400 14 192 363 0.27 0.15 800 18 168 319 0.24 0.13 1,600 10 148 281 0.21 0.12 100 23 251 473 0.36 0.20 1,000 11 163 310 0.23 0.13 10,000 6 106 187 0.14 0.08 100,000 4 69 143 0.11 0.06 Source: ESMAP calculations and analysis using costing data described in this chapter. NPV = net present value; LCOE = levelized cost of energy; kWh = kilowatt-hour. 7 2   MINI GRIDS FOR HALF A BILLION PEOPLE When maintaining the rest of the upfront cost (manage- and commercial clients, can reduce their LCOE by up to ment, distribution system, land and logistics, installation 30 percent. When combined with the expected declines cost, taxes, and duties), the cost reduction is 21 percent in CAPEX and OPEX, the cost of electricity from a best- of the total upfront cost for the power plant. Carrying this in-class third-generation system will be $0.20 per kWh forward into the LCOE calculation, assuming a 75 percent by 2030. This is for mini grids with productive applica- CAPEX, 20 percent OPEX and 5 percent for preparation tions that enable a 40 percent load factor. costs, the power plant cost savings lowers the LCOE by • Expected decreases in component costs can reduce 15 percent, from $0.38/kWh to $0.32/kWh. If we use upfront investment costs to less than $2,500/kWfirm the breakdown as found in a database of 440 projects by 2030. In improving the design of a race car, a designer (CAPEX, 64 percent; OPEX, 31 percent; and 5 percent for might find it impossible to shave 1 kg off in a single loca- preparation costs), the reduction of the LCOE is less, and tion but could identify 20 places in the car where she when we use the breakdown as calculated in the “econ- could reduce 50 grams. Cost reductions in mini grids omies of scale” analysis (CAPEX, 52 percent; OPEX, 44 work the same way, with cost reductions in many dif- percent; and 4 percent for preparation costs), the impact ferent components adding up to a substantial overall of the power system’s cost reduction on the LCOE falls to cost reduction. If the prices that mini grid developers 11 percent. pay for the PV array, Li-ion batteries, and inverters and As also mentioned under the “economies of scale” analy- associated electronics decline by the same proportion sis, the overall reduction in power plant costs is essential as mainstream industry benchmarks between 2020 for an overall competitive product in the marketplace. and 2030, the upfront capital cost per kWfirm of a solar Equally important, and a topic that has not received the hybrid mini grid would fall by almost 25 percent. same level of attention in this handbook, is the innovation • Economies of scale will reduce the LCOE of mini grids necessary to also reduce the OPEX, including the cost of even further. As developers build portfolios of mini grids major replacements. instead of one-off projects, they benefit from increased economies of scale—primarily as a result of bulk pur- chases of components and increased efficiencies CONCLUSION through standardized processes and increased know- how. Analysis of the data collected in ESMAP’s survey Best-in-class mini grid costs have plummeted in the past of mini grids in Africa and Asia indicates that economies few years. In 2018 ESMAP conducted a cost analysis of 53 of scale can greatly reduce capital costs. As we describe mini grids published in the executive summary of Mini Grids in this chapter, for every additional 100 customers a for Half a Billion (ESMAP 2019). At that time, the “best-in- mini grid serves, its cost per customer falls on average class” mini grid produced electricity with a (financial) LCOE by about $9. Cost reductions from economies of scale of $0.55 per kWh. In the three years since this analysis, complement the downward effect on costs from greater best-in-class costs have dropped nearly 31 percent to only recourse to productive uses of electricity. $0.38 per kWh, thanks to decreases in the cost of solar • Using geospatial and other digital tools to develop panels, batteries, inverters, and efficiencies through econ- portfolios of mini grids will also reduce costs. Geo- omies of scale. spatial analysis allows developers to assess mini grid SUMMARY OF POTENTIAL COST REDUCTIONS sites at a fraction of the cost of traditional site assess- ment activities—from around $30,000 per site with- The trends in CAPEX and OPEX highlighted above will lead out using geospatial analysis, to approximately $2,300 to major cost reductions in four areas for third-generation per site using geospatial analysis. A number of estab- mini grids through 2030: lished mini grid developers in Sub-Saharan Africa use • Increasing income-generating uses of electricity can geospatial and other analytical software to plan their decrease the LCOE by 25 percent or more and, when portfolios remotely. They prioritize sites for mini grid combined with the expected cost declines described development and use technology-enabled processes below, will bring the economic cost of mini grid elec- to estimate demand, allowing them to optimize sys- tricity to almost $0.20/kWh by 2030. The baseline load tem design across their portfolios. Where government factor for mini grids of 22 percent reflects low levels of or donor entities are conducting the portfolio-level income-generating uses of electricity. Mini grids that analysis, the data can be analyzed and disseminated can increase their load factors to 40 percent through to developers on a web-based platform like Odyssey significant daytime consumption by local businesses Energy Solutions. MINI GRIDS FOR HALF A BILLION PEOPLE    7 3 GOVERNMENT’S ROLE IN REDUCING MINI GRID costing data for use in understanding detailed mini grid COSTS AND CATALYZING INNOVATION costs as they evolve in different markets. Governments can help keep the path open for mini grid More data are also needed on the standards to which component technology innovation and cost decreases by mini grids are built. For example, are poles and wires built designing and implementing regulatory frameworks and to standards that a utility would use? Or are cheaper, mini grid programs that provide light-handed regulation untreated wooden poles used to save costs? What is and exempting mini grid components from import taxes the expected life cycle of the battery? Mini grids built to (see chapter 9 for a detailed discussion of mini grid regu- different standards will naturally report different costs, lations). It is important to design standards that leave open reflecting these different standards. Without improved opportunities for innovation and not to assume (and thus knowledge of the underlying standards for each mini lock in) a particular technology or configuration. grid, variations that currently appear to be noise in data could more meaningfully and accurately reflect the real- Rural electrification agencies can harness these cost sav- ities on the ground and help identify areas where action ings by designing programs that provide opportunities for is warranted to reduce mini grid costs, improve quality, or capable developers to develop multiple nearby sites as part both. This chapter and the underlying database should be of a larger, comprehensive program. Doing so allows for viewed as living documents, which will benefit from better, economies of scale in project identification (especially har- and more, data over time. nessing geospatial information), engineering and design, site assessment and community negotiations, equipment procurement and installation, O&M, and tariff collection. REFERENCES Ensuring a competitive marketplace for mini grids will be important to promoting innovation and continued cost AMDA (Africa Minigrid Developers Association) and ECA (Economic Consulting Associates). 2022. Benchmarking Africa’s Minigrids. declines. The data presented in this chapter show sizable Nairobi, Kenya: African Minigrids Developers Association. https:// cost variations, implying in part the ability of mini grid devel- africamda.org/. opers to procure equipment at internationally competitive AP News. 2018. “Global Diesel Generator Market Size, Share & prices.31 In cases where costs are on the high end of the Trends Analysis Report, 2013–2022.” ResearchAndMarkets.Com, mini grids we analyzed, the systems were clearly overbuilt, October 19, 2018. https://www.businesswire.com/news/home/201 designed to meet a load that may not materialize for years. 81019005370/en/Global-Diesel-Generator-Market-Size-Share- Trends-Analysis-Report-2013-2022---ResearchAndMarkets.com Some subsidy programs, particularly those that subsidize a portion of renewable energy generation investments, Balabanyan, Ani,Yadviga Semikolenova, Arun Singh, and Min A Lee. 2021. “Utility Performance and Behavior in Africa Today.”World Bank, Wash- incentivize oversizing mini grids. Costs reported at the low ington, DC. https://openknowledge.worldbank.org/handle/10986/ end in this study indicate the best possible practice at the 36178. frontiers in a competitive market, keeping in mind the need Blimpo, M., and Malcolm Cosgrove-Davies. 2019. Electricity Access in to specify minimum customer-service levels and not stint Sub-Saharan Africa: Uptake, Reliability, and Complementary Factors on quality. As mini grids are deployed in larger quantities for Economic Impact. Africa Development Forum Series. Washing- and markets become more competitive, costs will trend ton, DC: World Bank. downward toward, and beyond, the best-in-class cost and BNEF (Bloomberg New Energy Finance). 2020a. “Deep Dive into Utili- performance benchmarks revealed in this study. ty-Scale PV System Cost.” BNEF. 2020b. “Battery Pack Prices Cited below $100/KWh for the THE IMPORTANCE OF COORDINATED First Time in 2020, While Market Average Sits at $137/KWh.” Bloom- COLLECTION OF DATA ON MINI GRID COSTS bergNEF (blog), December 16, 2020. https:/ /about.bnef.com/blog/ battery-pack-prices-cited-below-100-kwh-for-the-first-time-in- Data collection on mini grids is at an early stage. Better and 2020-while-market-average-sits-at-137-kwh/. more uniform data will produce more useful results and BNEF. 2021. “Battery Pack Prices Fall to an Average of $132/KWh, observations. More effort should be spent on standardizing but Rising Commodity Prices Start to Bite.” BloombergNEF (blog), data collection and integrating data collection into report- November 30, 2021. ing requirements into mini grid programs. One branch of BNEF. 2022.“Solar – 10 Predictions for 2022.” January 26, 2022. https:// about.bnef.com/blog/solar-10-predictions-for-2022/. this effort could take the form of a plug-in into a standard- ized mini grid bidding and accounting software package Carlin, Kelly, Josh Agenbroad, Stephen Doig, and Kendall Ernst. 2018. Minigrids in the Money: Six Ways to Reduce Minigrid Costs by 60% that provides developers front-end geospatial information for Rural Electrification. Boulder, CO: Rocky Mountain Institute. on prospective villages and markets, optimizes mini grid https://www.rmi.org/insight/minigrids-money/rmi-seeds-mini- system design, helps link developers with equipment sup- grid-report/. pliers and financiers, helps keep track of key milestones in Danvest Energy. 2019. “Danvest Energy A/S—Wind Diesel Systems, project development, and provides suitably anonymized Solar Diesel Systems.” https://www.danvest.com/how-does-it-work 74   MINI GRIDS FOR HALF A BILLION PEOPLE Dave, Rutu, Sandra Keller, Bryan Bonsuk Koo, Gina Fleurantin, Elisa Por- Framework. Washington, DC: World Bank. http://hdl.handle.net/ tale, and Dana Rysankova. 2018. Cambodia—Beyond Connections: 10986/30102. Energy Access Diagnostic Report Based on the Multi-Tier Framework. Plautz, Jason. 2021. “Form Energy’s $20/KWh, 100-Hour Iron-Air Bat- Washington, DC: World Bank. http:/ /hdl.handle.net/10986/29512 tery Could Be a ‘Substantial Breakthrough.’” Utility Dive, July 26, Energy Trend. 2021. “Solar Price | EnergyTrend.” April 6, 2021. https:// 2021. https://www.utilitydive.com/news/form-energys-20kwh-100- www.energytrend.com/solar-price.html. hour-iron-air-battery-could-be-a-substantial-br/603877/. Energy Trend. 2022. “Solar Price | EnergyTrend.” April 19, 2022. https:// Sanderson, Henry. 2019. “Congo, Child Labour and Your Electric Car.” www.energytrend.com/solar-price.html. Financial Times,July 7, 2019. https://www.ft.com/content/c6909812- Feldman, David, Vignesh Ramasamy, Ran Fu, Ashwin Ramdas, Jal 9ce4-11e9-9c06-a4640c9feebb. Desai, and Robert Margolis. 2021. U.S. Solar Photovoltaic System Smart Energy International. 2018. “Global Trends in Smart Meter- Cost Benchmark: Q1 2020. Technical Report NREL/TP-6A20-68925. ing.” Smart Energy International, December 31, 2018. https://www. Golden, CO: National Renewable Energy Laboratory. https:/ /www. smart-energy.com/magazine-article/global-trends-in-smart-me- nrel.gov/docs/fy21osti/77324.pdf. tering/ Global Industry Analysts. 2022. Global Smart Meters Industry. February Solarpower Europe. 2018. “Global Solar Market Grows over 29% in 2022. 2017 with Even More to Come in 2018.” PV Magazine, March 14, Haegel, Nancy M., Harry Atwater, Teresa Barnes, Christian Breyer, 2018. https://www.pv-magazine.com/press-releases/global-solar- Anthony Burrell, Yet-Ming Chiang, Stefaan De Wolf, et al. 2019. “Ter- market-grows-over-29-in-2017-with-even-more-to-come-in-2018/. awatt-Scale Photovoltaics: Transform Global Energy.” Science 364 Stevens, Pippa. 2021. “More than Half of 2022’s Solar Projects Threat- (6443): 836–38. https:/ /doi.org/10.1126/science.aaw1845.IEA ened by Spiking Costs, New Report Finds.” CNBC, October 26, 2021. (International Energy Agency). 2021. Renewables 2021. Paris: IEA. https://www.cnbc.com/2021/10/26/more-than-half-of-2022s-so- https://iea.blob.core.windows.net/assets/5ae32253-7409-4f9a- lar-projects-threatened-by-spiking-costs-new-report-finds.html. a91d-1493ffb9777a/Renewables2021-Analysisandforecastto2026. Taizhou Amity Care International Co., Ltd. 2013. “Prestressed Concrete pdf. Spun Pile Pole Centrifugal Spinning Machine.” https://www.youtube. Innovus. 2015. “Innovus Power Microgrid Platforms: Delivering the com/watch?v=fY4hirhg3c4. Highest Penetration with the Lowest Levelized Cost of Energy.” Tenenbaum, Bernard, Chris Greacen, and Ashish Shrestha. 2022. Inter- http://www.innovus-power.com/wp-content/uploads/2015/08/ connected and Non-Interconnected Mini Grids in Undergrid Areas of innovus-power_hybrid_v082015.pdf. Nigeria and India. ESMAP Technical Report. Washington, DC: World IRENA (International Renewable Energy Agency). 2022. “World Energy Bank. Transitions Outlook 2022: 1.5°C Pathway.” Abu Dhabi: IRENA. Trading Economics. 2022. “Lead - 2022 Data - 1993-2021 Historical - Jäger-Waldau, A. 2017. PV Status Report 2017. EUR 28817 EN. Luxem- 2023 Forecast - Price - Quote - Chart.” bourg: Publications Office of the European Union. https://publica- Trimble, Chris, Masami Kojima, Ines Perez Arroyo, Farah Mohammad- tions.jrc.ec.europa.eu/repository/handle/JRC108105. zadeh. 2016. Financial Viability of Electricity Sectors in Sub-Saha- Jenkins, Lisa Martine. 2022. “Chinese Solar Panels Seized at US Bor- ran Africa : Quasi-Fiscal Deficits and Hidden Costs. Policy Research der over Possible Human Rights Abuses.” Protocol, August 9, 2022. Working Paper;No. 7788. World Bank, Washington, DC. https:/ / https://www.protocol.com/bulletins/xinjiang-solar-panels-uy- openknowledge.worldbank.org/handle/10986/24869 US DOE (US ghur-enforcement. Department of Energy). N.d. “Perovskite Solar Cells.” Energy.Gov. Kairies, Kai-Phillip. 2017. “Battery Storage Technology Improvements Accessed May 1, 2021. https:/ /www.energy.gov/eere/solar/per- and Cost Reductions to 2030: A Deep Dive.” PowerPoint presen- ovskite-solar-cells. tation at the International Renewable Energy Agency Workshop, Wagman, David. 2020. “US Energy Storage Strategy Includes Tech Cost March 17. https:/ /www.irena.org/-/media/Files/IRENA/Agency/ Estimates.” PV Magazine International, December 23, 2020. https:// Events/2017/Mar/15/2017_Kairies_Battery_Cost_and_Perfor- www.pv-magazine.com/2020/12/23/us-energy-storage-strate- mance_01.pdf?la=en&hash=773552B364273E0C3DB588912F- gy-includes-tech-cost-estimates/. 234E02679CD0C2. Yu, Hyun Jin Julie. 2018. “A Prospective Economic Assessment of Koo, Bryan Bonsuk, Dana Rysankova, Elisa Portale, Niki Angelou, Residential PV Self-Consumption with Batteries and Its Systemic Sandra Keller, and Gouthami Padam. 2018. Rwanda—Beyond Con- Effects: The French Case in 2030.” Energy Policy 113 (February): nections: Energy Access Diagnostic Report Based on the Multi-Tier 673–87. https://doi.org/10.1016/j.enpol.2017.11.005. Framework. Washington, DC: World Bank. https:/ /openknowledge. Zaw Min. 2019. Interview. Kyaw Soe Win (KSW) Hydropower Company. worldbank.org/handle/10986/30101. March 18. Lazard. 2018. “Lazard’s Levelized Cost of Storage Analysis—Version 4.0.” https:/ /www.lazard.com/perspective/levelized-cost-of-ener- gy-and-levelized-cost-of-storage-2018/. NOTES Lazard. 2020.“Lazard’s Levelized Cost of Energy Analysis—Version 14.0.” https://www.lazard.com/perspective/levelized-cost-of-energy-and- 1. This chapter uses the terms solar and photovoltaic interchangeably levelized-cost-of-storage-2020/. to mean generation of electricity from sunlight. Lazard. 2021. “Lazard’s Levelized Cost of Energy Analysis—Version 2. For some 2021 mini grids, contracted costs were used rather than 15.0.” https://www.lazard.com/perspective/levelized-cost-of-ener- post-commissioning costs. gy-levelized-cost-of-storage-and-levelized-cost-of-hydrogen/. 3. Not included for detailed analysis in this chapter, but nonethe- Padam, Gouthami, Dana Rysankova, Elisa Portale, Bryan Bonsuk Koo, less promising (especially for communities with needs for smaller Sandra Keller, and Gina Fleurantin. 2018. Ethiopia—Beyond Con- amounts of electricity) are lower-cost direct current (DC) “mesh nections: Energy Access Diagnostic Report Based on the Multi-Tier grids” or “skinny grids” that distribute DC electricity for lighting, MINI GRIDS FOR HALF A BILLION PEOPLE    75 electronics, and small appliances like fans and even efficient refrig- average PV module costs. For our 2030 calculations we used aver- erators or electric rickshaws. See box 1.2. age cost of PV modules in these Li-ion battery mini grids and then 4. In the world of grid-connected power plants, LCOE is used to com- applied industry-projected cost declines discussed in the “PV Mod- pare on an apples-to-apples basis the cost of energy delivered to ule Trends” section of this chapter. the grid network from generating assets that have different capital 15. Because of Li-ion batteries’ ability to discharge energy more deeply, costs, fuel costs, and lifetimes. LCOE is typically expressed in cur- they can have a nameplate capacity that is 25 percent smaller than rency per kilowatt-hour. if lead-acid batteries were used. 5. In these three countries we restricted our analysis to mini grids 16. Somewhat counterintuitively, the optimum renewable energy frac- with lithium-ion batteries because they are the most common type tion decreases slightly in some of the cases as the load curve shifts (accounting for 76 percent of battery types in mini grids from our from normal to sun-following. For example, for cases 3 and 4 (Ethi- data set in Nigeria, 50 percent in Myanmar, and 100 percent in Ethi- opia and global Li-ion, respectively), the renewable energy fraction opia), and also have lower LCOE, on average, than mini grids with falls from a 22 percent load factor to a 22 percent sun-following lead-acid batteries. scenario. Why would the renewable energy fraction decrease when 6. Bolivia, Ethiopia, Indonesia, Myanmar, Nigeria, and Tanzania. shifting to a more solar-coincident load? The answer lies in the component sizing of the optimal mini grid in each case. Moving 7. Bangladesh, Bolivia, Chad, Ethiopia, Ghana, Guinea Bissau, India, more solar-coincident loads allows the system to rely less on bat- Indonesia, Ivory Coast, Kenya, Liberia, Myanmar, Nepal, Nigeria, tery storage, with the consequence that the system becomes a bit Palestine, Sierra Leone, Tanzania, Vanuatu, and Vietnam. more reliant on backup diesel during occasional rainy periods. 8. In addition to 20-year project economic life, we have also modeled 17. Generators are often rated in apparent power (kVA). The genera- the impact of 15- and 25-year lifetime assumptions on LCOE. Add- tor’s real power output (kW) is the apparent power multiplied by ing five years decreases LCOE by about 2.2 US cents per kWh in the power factor, typically assumed to be 0.8 for design purposes. the 22 percent load factor / 0 percent subsidy case, with a smaller reduction in other cases. Subtracting five years increases LCOE by 18. Peak power output of a solar panel is the power output at a solar about 2.7 US cents per kWh for the same case, with lesser impact irradiance of 1,000 watts per square meter, 1.5 air mass, and a tem- in other cases. perature of 25º C. 9. The weighted average cost of capital of a capital structure compris- 19. In 2019 an early version of ESMAP’s Mini Grids for Half a Billion ing 40 percent equity at 12 percent return and 60 percent debt at 8 included a similar waterfall graph based on data from 36 mini grids percent interest is 9.6 percent. Assumptions consistent with Lazard commissioned between 2012 and 2018. A comparison of the 2019 (2021). graph with this 2022 version reveals that PV costs (including for PV inverters) plummeted, from 16 percent to about 10 percent of 10. The combination of a 9.6 percent nominal discount rate and 3 per- total project costs, reflecting lower PV costs in recent years. Bat- cent inflation yields a real discount rate of 6.41 percent. tery and battery inverter + EMS costs remained the same. Distribu- 11. The World Bank tracks pump prices for diesel from Sub-Saharan tion and meters as a portion of total project costs increased from Africa at https:/ /data.worldbank.org/indicator/EP.PMP.DESL.CD? 21.0 percent to 26.6 percent as other costs fell, as did installation, end=2016&locations=ZG&start=2010. Corrected for inflation, the which increased to 11.3 percent from 8.0 percent. On the other most recent (2016) pump price is $1.08 per liter. The fuel cost sen- hand, project development costs dropped from 9.0 to 5.9 percent, sitivity analysis investigated $0.75 and $1.50 per liter of diesel fuel which likely reflects benefits of clustering and perhaps also a trend in addition to the base case of $1.00 per liter. Because the mini grids in more recent projects to fold project development costs into have high penetrations of renewable energy, the cost of diesel had a reported equipment costs. relatively small effect on LCOE. In HOMER sensitivity runs with die- 20. PV inverters convert the direct current (DC) electricity produced sel fuel costs of $0.75 and $1.50 per liter, the variation in LCOE was by the solar array into alternating current (AC) power on the mini less than +6 percent of the $1.00 per liter base case. The project grid’s network. lifetime analysis considered project economic lifetimes of 15 and 25 years in addition to the base case of 20 years. The +5 year project 21. Battery nameplate capacity is typically indicated in ampere-hours lifetime assumptions affected LCOE by less than +7 percent, with (Ah). Battery nameplate kWh is calculated as the Ah multiplied by the strongest impacts in the 22 percent load factor case. the battery’s nominal voltage. 12. Diesel generators’ nonfuel OPEX is estimated at $0.03 per kWh; 22. Lead-acid batteries are typically not discharged more than 50–60 solar PV OPEX, at $10 per kW a year; and battery OPEX, $10 per kW percent, whereas lithium-ion batteries can be discharged to 80 per- a year. These variable O&M assumptions are held constant across cent depth of discharge. all HOMER modeling runs. 23. For an apples-to-apples comparison across different battery chem- 13. The prices that developers pay for individual components are typ- istries, it is useful to compare the levelized cost of storage (LCOS). ically higher than the wholesale price direct from the factory. Even LCOS is analogous to the levelized cost of energy (LCOE) but uses as third-generation mini grid developers build larger portfolios, the discounted cost of purchasing and operating the battery over all but the very biggest developers will still pay higher prices than the course of its lifetime (in lieu of the cost of generating and distrib- those available at the factory door. As a result, we conservatively uting electricity), divided by the discounted discharged electricity. assumed that by 2030, the typical third-generation mini grid devel- It is the levelized cost associated with storing and withdrawing one oper would be able to purchase components at their 2020 factory kWh of electricity. The data set does not provide sufficient data for spot prices. an LCOS calculation. Lazard (2018) finds that for US applications in 2018 at the scale of 40 kWh of storage capacity (“residential scale” 14. In our data set, the average cost of PV modules in mini grids built in their analysis, equivalent in storage capacity to the smallest mini with Li-ion batteries (mostly built in 2019 to 2021) was $534 per grids considered in this chapter), the LCOS for lithium-ion batteries kWp, reflecting the fact that our best-in-class mini grid, while best- was $0.476–$0.735/kWh. Lead-acid batteries have a comparable in-class overall and for other equipment costs, had higher than range of LCOS values at this scale, of $0.512–$0.707/kWh. For proj- 76   MINI GRIDS FOR HALF A BILLION PEOPLE ects at the scale of 2 megawatt-hours of energy storage (Lazard’s (Stevens 2021). As of April 2022, the global average spot price “Commercial & Industrial scale,” about twice as much storage as for PV was $230 per kWp for 330–335 W multi-crystalline mod- the largest village mini grids studied in this chapter), the LCOS for ules. Industry experts expects these bottlenecks to be transitory lithium-ion batteries was $0.315–$0.366/kWh, several cents lower (Energy Trend 2022). than the $0.382–$0.399/kWh LCOS for lead-acid batteries. The 27. Wright’s Law posits that every cumulative doubling in the cumula- analysis assumed a 20:80 percent debt to equity ratio, with debt at tive amount of a product produced leads to a consistent percent- 8 percent and the cost of equity at 12 percent (Lazard 2018). age cost decline. 24. Parameters not captured in the distribution network cost data are 28. PV deployment has been growing at an average of 40 percent in the standards to which the low-voltage distribution network is built. recent years (50 percent in 2016, 29 percent in 2017) (Solarpower Projects built to a high standard or a grid-ready standard will have Europe 2018) and a compound average growth rate of more than much higher costs per customer and per kilometer than those built 40 percent over the past 15 years (Jäger-Waldau 2017). to lower standards (for example, using untreated wooden poles, 29. See http://haiyuindustry.sell.everychina.com/p-101752826-concrete- low pole heights, and undersized conductors and hardware), as will spun-electric-pole-production-machine.html for photos and a distribution grids that are deliberately oversized to accommodate more detailed description of this process. future growth. Other factors influencing the wide variations in cost per kilometer and per customer that are not captured in the survey 30. Field visits at the preparation stage are needed to engage with likely include whether poles for distribution were constructed of local communities to discuss agreements, such as the terms of land pur- materials and not costed as part of the project, accounting practices chases or leases, and to verify the geospatial analysis data—and related to in-kind labor and materials supplied by local communities, they can be handled by a much leaner team. whether service is single phase or three phase, and whether public 31. A related issue more germane to the data is that some companies lighting costs were bundled into this category by a developer. appear to report project development and business development 25. Replacement of large assets such as batteries is not included in costs explicitly, others blend them into equipment costs in the these O&M costs. These costs are included, however, in LCOE cal- form of markups, and still others internalize these costs and do not culations earlier in this chapter. report them at all. The background study carried out for this chap- ter did not include data gathering on subsidy amount or in-kind 26. With high commodity prices for polysilicon, aluminum, and other accounting (for local materials and community contribution) or raw materials due to post-COVID supply bottlenecks prices were address competition in markets. pushed higher for PV modules in 2021 and the first part of 2022 MINI GRIDS FOR HALF A BILLION PEOPLE    7 7 CHAPTER 2 NATIONAL STRATEGIES AND DEVELOPER PORTFOLIOS: THE ROLES OF GEOSPATIAL ANALYSIS AND DIGITAL PLATFORMS CHAPTER OVERVIEW This chapter discusses geospatial analysis and other digital tools that can support electrification planning at both the national and portfolio levels. Drawing on real-world examples from Nigeria and Ethiopia, and leading mini grid developers, the chapter lays out how to use cutting-edge technologies like geospatial software and online platforms to develop large portfolios of mini grids. It also introduces some of the leading technology providers for such planning tools. Thanks to new geospatial analysis technologies, a port- household income, poverty, commercial activities, willing- folio approach to mini grid development is becoming ness to pay). Spatial modeling delivers a least-cost plan by mainstream in the industry and in national electrification identifying beforehand the technology best suited to local planning. This is occurring at the national level for least-cost circumstances—technically feasible and economically via- electrification planning and among mini grid companies ble. At the same time, geospatial plans can also identify themselves. Geographic information system (GIS) software communities requiring decentralized solutions (mini grids) and geospatial data are becoming key tools for planning as they wait for the grid. electrification at the national level and performing rapid site Geospatial plans are essential in siting mini grids and sig- assessments. Mainstream digital tools are expediting tech- naling the likelihood of grid arrival, information that cur- nological advances and cost reductions, including: tails asset stranding. The identification of communities • Satellite imagery and spatial products for which mini grids offer the optimal technology solution • Big data and cloud-based computing requires at least the following: • More sophisticated algorithms and analytical solutions • Electricity demand estimates, including for productive (for example, heuristics and machine learning) uses; • Global positioning system devices and the proliferation • The location of existing infrastructure and modeling of of web-based and mobile technologies grid rollout; and • Higher-quality open-source software • Estimation of local renewable generation potential. A geospatial approach ensures that national electrification Using geospatial analysis in planning mini grid portfolios is mapped cost-effectively on the existing grid network and could cut the time spent on deployment. its attributes digitalized. Demand and supply of electricity can be geolocated by overlaying demographic data (such Geospatial planning cannot replace field-based feasibility as population density and growth patterns) on social infra- studies, but it can determine mini grid potential. It does structure (for example, schools, health centers, admin- this by evaluating current and anticipated service needs istrative offices) and the economic landscape (such as (including productive use) and the time frame for grid 78   MINI GRIDS FOR HALF A BILLION PEOPLE arrival. This exercise prepares engineers and policy makers understand how mini grids could support a speedy roll- for planning electricity services and allocating public fund- out of electrification; how many people or households can ing, ensuring that public interventions (where and why) are mini grids serve with high-quality and sustainable elec- done with equity foremost in mind. System optimization, tricity over the long term? Investors and financiers, on the network design tools, and online platforms analyze data, other hand, are interested in the addressable market and develop project proposals, select developers, solicit financ- the economically viable potential of mini grids in Sub-Sa- ing, and monitor and verify implementation. haran Africa. This section briefly describes how new spa- tial data and analysis can help address these questions, This chapter assesses the market potential for mini grid providing qualitative and, to the extent possible, quanti- sites selected by geospatial data. It then looks at tools and tative data. analyses that support least-cost electrification and plan- ning exercises. We then assess how geospatial and other The single most critical data set required for this analysis digital tools are being used to save both time and costs is the settlement distribution—that is, the location of set- in mini grid project development, from site prospecting tlements or buildings over the area of interest. Over the and analyzing demand to right-sizing solutions, packag- past few years, several data sets have been developed in ing projects, and taking them to market. Examples from this regard, based on high-resolution satellite imagery and the “frontier” of these planning exercises, particularly the processing techniques (for example, machine learning). We World Bank’s Nigeria Electrification Project and its Ethio- describe and use some of them below. pia project, Access to Distributed Electricity and Lighting in Ethiopia, illustrate their practical application. SIMPLE EXPLORATORY SPATIAL DATA ANALYSIS FOR SUB-SAHARAN AFRICA USING GRID3 In the first example, we explore the settlement distribution ASSESSING THE MARKET layer for Sub-Saharan Africa provided by GRID3 (CIESIN POTENTIAL FOR MINI GRIDS 2020). The layer constitutes a comprehensive set of set- tlement polygons classified into built-up areas, small set- With programs ramping up worldwide as countries seek tlement areas, and hamlets; 326,000 settlements were to meet their Sustainable Development Goal (SDG) 7 tar- found to be more than 1 kilometer (km) from the existing gets, what roles do mini grid systems play in that process? grid1 (figure 2.1), including a preponderance of settlements Different stakeholders come at this question from various in the 100–1,000 population range. Extracting additional standpoints. Governments and policy makers want to information about the settlements to facilitate site selec- FIGURE 2.1 • Scatter plot of settlement population vs population density in Sub-Saharan Africa All 326,000 settlements more than 1 km from the grid 10k 8k Settlement density (people/km2) 6k 4k 2k 0 0 100 1k 10k 100k Source: ESMAP analysis of GRID3 data. Settlement population Note: Point size reflects the area of the settlement. MINI GRIDS FOR HALF A BILLION PEOPLE    7 9 tion, the following criteria were thus set to define a settle- 10,000 to 100,000 people, likely requiring mini grids at the ment suitable for mini grid electrification: 500 kW to 1 megawatt (MW) scale. • Number of people: more than 100 and less than 100,000 Settlement population distribution is presented by country in figure 2.2, with the estimated addressable market for • Distance from the existing grid: more than 1 km mini grids presented in absolute numbers atop the bars for • Population density: more than 1,000 people/km2 each country. The height of the bars represents the share of Settlements with populations of more than 100,000, or this segment as a percentage of the total population of the any settlements located less than 1 km from the grid, are country. This addressable market for mini grids is further considered either already electrified or candidates for grid disaggregated into the same settlement sizes as described electrification and thus excluded from this analysis. On a in table 2.1. similar note, any settlement with less than 100 people is Using the same selection criteria and settlement sizes, fig- considered a better candidate for solar home systems ure 2.3 maps and visualizes their population distribution. (SHSs) than mini grids. The density assumption was set as such to satisfy mini grid design criteria; that is, to avoid THE GLOBAL ELECTRIFICATION PLATFORM AND the selection of settlements that have sparse populations. LEAST-COST ELECTRIFICATION ANALYSIS FOR These criteria reflect the prevailing view of what constitutes SUB-SAHARAN AFRICA a good candidate site for a mini grid; they were based on The Global Electrification Platform (GEP) is an open-ac- past projects and experience on the ground. The results cess, interactive, online platform that models and visu- presented below are bound to these selection criteria and alizes pathways toward universal access, split into an should be interpreted with caution accordingly. intermediate strategy for 2025 and full electrification by As table 2.1 indicates, about 23.7 percent of Sub-Saharan 2030, for countries marked by severe access deficits. The Africa’s population (that is, 276.9 million of 1.17 billion peo- current version, GEP V.2.0, officially launched in April 2022, ple) lives in settlements that in theory could be markets for explores 96 unique scenarios. The set of results was mod- mini grids. Assuming an average size of 5 people/house- eled with a modified version of the Open Source Spatial hold, about 55.4 million households on the subcontinent, Electrification Tool (OnSSET). This is a flexible and mod- or about 291,000 clusters, could be served by mini grids. ular GIS-based energy modeling tool developed to support electrification planning and decision making by estimating, As an initial assessment of the potential for standardiza- analyzing, and visualizing the most cost-effective electrifi- tion in the rollout of mini grids in Sub-Saharan Africa, we cation strategy. In doing so, it takes into account spatially also explored whether we could find convergence around explicit characteristics related to energy, such as popula- certain sizes of mini grids to serve the clusters described tion density and distribution, proximity to transmission and above. The growing pipeline of mini grid projects under road network, night-time lights, and local renewable energy development under the World Bank’s Nigeria Electrification potential, among others. Project provided some data on the sizing of private-sec- tor-led mini grids vis-à-vis the customer base or settlement The GEP considers current and projected values of key size, indicating an average firm power allocation of roughly parameters such as population growth, demand level, 100 watts (W) per connection. We used this as a bench- technology costs, and other policy/planning limitations mark, while acknowledging variations from project to proj- in generating electrification scenarios. Results indicate ect, including in the ratios of commercial and productive the least-cost electrification technology per settlement end users to residential customers, and associated power (or cluster)2 across millions of clusters in 46 countries requirements. We anticipate that mini grids of 20 kilowatts of Sub-Saharan Africa. The addressable market for mini (kW), 80 kW, and 200 kW may be best suited to serve grids falls within a range that depends on the input param- the settlement sizes listed in table 2.1. Custom solutions eters and assumptions, with the key parameter being the will continue to be the preferred option for settlements of level of demand for unelectrified households. Usually, the Table 2.1 • Characteristics of Sub-Saharan African settlements suitable for electrification via mini grid Size (average pop.) 100–500 500–2,500 2,500–10,000 10,000–100,000 Total: Settlements 177,087 95,702 15,188 2,948 290,925 Population total 46,886,543 97,073,397 67,627,653 65,303,591 276,891,184 Share of total population (%) 4.01 8.31 5.79 5.59 23.71 Optimum mini grid sizing (kW) 20 80 200 Custom 80   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.2 • Sub-Saharan Africa’s addressable market for mini grids 70% 5M 0. 60% Percent of total population 4M 5. M 4M M .3 50% .0 4. 23 24 11 M .5 26 M M 4M 40% 1.3 .8 M 9M 16 .4 2. .7M M 12 .1M 6. .5 M 42 7M 13 11 0M .2 .1M M 28 6. .6 0M 4. M 13 21 M 16 30% .5 2M .4 4. 16 M 6M 11 0M 4. M 1.2 M .0 4. 3. .0 2M 31 8M 33 0. M M 20% 8. .8 M .5 22 M .9 18 .5 27 M 36 .0 M 5M 9M 60 .5 M 5M M 0. 56 3. .4 .2 10% 4. 10 14 .7M 42 M 4M .2 18 2. 55 4M M M 1.9 1.1 2. 0 ep de lic ali C to d'I r rit ia ba ia Ug we So l Gu ire Za da ts ia Ke a h S ea Ga an Gu rra inea Su -Bis e n( u ia Si Gu ia So he) Rw alia er a Be n a Na nya Le ibia Es tho i Ni nin ria Sa a Co d Se go Bu gal Er di Dj rea Lib ti oz Gab ia biq n ue Pr ana ipe ina go M Ma o ag wi S A ar la tin ua te ige a on rn fric Ta ar n Ca and s da sa oo am o a u DR M nzan an Zi hiop Bo mb b er W out ngo n c ad la ub M Fa ut in wa ud n R Ver an Ch rk To n ria vo ge ibo b ne m ínc ru wa as it h so T m m nd Gh ine Le N A m Et au es h fri bo m l A Ca e te Eq Co Bu ca M éa om ra oT nt Sã Ce 100.0–500 500–2,500 2,500–10,000 10,000–100,000 Source: ESMAP analysis of GRID3 data. Note: Total country population shown in values atop each bar. Results indicate the percentage of people located in settlements (clusters) that fulfil the selection criteria for mini grid candidacy (that is, between 100 and 100,000 people, located more than 1 km from the main grid, with a distribution of more than 1,000 people/km2). Clusters derived from GRID3 (CIESIN 2020). FIGURE 2.3 • Sub-Saharan Africa’s addressable greater the targeted demand, the greater the share of market for mini grids, mapped by settlement mini grid potential. Table 2.2 presents how three different population levels of demand can dictate the share of mini grids in the least-cost mix. When demand is low, mini grids are the least-cost option for about 66.0 million people (or 13.2 million connections). Intermediate demand pushes mini grid potential to about 87.2 million people (17.5 million connections), while with high-demand scenarios, the potential is estimated at about 131 million people (26.2 million connections). Mini grids can also serve as pre-electrification solutions, which is to say, they could be least-cost options for settle- ments expecting the arrival of the main grid. Political, eco- nomic, and other considerations will ultimately determine if and when the grid reaches these communities. On the one hand, the newer mini grids mostly meet code and could connect to the main grid once it arrives. On the other hand, falling costs and decentralized renewable technologies tell us that not every community may need to connect to the Settlement population main grid. 100–500 2,500–10,000 Including pre-electrification, in the low demand scenario, 500–2,500 10,000–100,000 mini grids are cumulatively the least-cost option for 66 mil- lion people (or 13.2 million connections)—the same as the Source: ESMAP analysis of GRID3 data. count in 2030. In the bottom-up and high-demand scenar- Note: Results indicate the number of people located in settlements ios, we see huge increases in the population served at least- (clusters) that fulfil the selection criteria for mini grid candidacy (that is, cost by mini grids once we account for mini grids that are between 100 and 100,000 people, located more than 1 km from the main grid, with a distribution of more than 1,000 people/km2). Clusters derived eventually connected to the grid: 105 million people (or 21 from GRID3 (CIESIN 2020). million connections) and 325 million people (or 65 million MINI GRIDS FOR HALF A BILLION PEOPLE    81 TABLE 2.2 • Selected electrification results for 2030 retrieved from the Global Electrification Platform, aggregated for 46 countries in Sub-Saharan Africa Mini grid potential Mini grid potential Demand target a (including pre-electrification role) People Connections People Connections Low demand 66,004,359 13,200,872 66,004,359 13,200,872 Bottom-up demand 87,249,707 17,449,941 104,716,450 20,943,290 High demand 131,052,705 26,210,541 325,025,815 65,005,163 Source: ESMAP analysis of Global Electrification Platform results. Note: GEP V.2.0 released in April 2022. Low demand reflects targets equivalent to Tier 3-4 for urban households and Tier 1 for rural households. High demand indicates Tier 4-5 for urban a.  households while Tier 2-3 for rural. Tier values differ per country depending on the current electrification status and/or goals. The bottom-up value reflects an intermediate level of demand that is based on the combination of socio-economic indicators that vary spatially (poverty rate and GDP). TABLE 2.3 • Breakdown of electrification results from bottom-up demand scenario <10 kW 10–100 kW 0.1–1 MW 1–10 MW 10–100 MW >100 MW Total Settlements 36,914 136,465 19,313 1,056 94 3 193,845 Population 1,781,648 39,190,271 35,259,640 17,100,927 9,281,040 2,102,924 104,716,450 Households 356,330 7,838,054 7,051,928 3,420,185 1,856,208 420,585 20,943,290 Percentage of total new 0.19 4.16 3.75 1.82 0.99 0.22 11.13 connections in Sub-Saharan Africa Source: ESMAP analysis of OnSSET data and Global Electrification Platform results. connections) respectively. Note that all values are inclusive generation for grid electricity and the cost of solar pho- of population growth and reflect aggregated data for all tovoltaic (PV) systems, affect the least-cost option and modeling years between 2020 and 2030. together account for the 96 unique scenarios modeled in the GEP. The scenario most favorable for the deployment Looking more closely at the bottom-up demand scenario, of mini grids as the least-cost solution thus varies from we see that least-cost mini grids serve close to 200,000 country to country. settlements, numbers that correspond to the aforemen- tioned 105 million people (or 21 million households). Table The finding from this exercise is that 430 million people 2.3 displays the distribution of this population by the size can receive access at least cost via mini grids. This includes of the mini grids projected to serve them. One can see that 380 million people in Sub-Saharan Africa living in the 58 most settlements (and their mini grids) fall in the 10 to access-deficit countries covered by the GEP, which rep- 100 kW range, while mini grids in the 10 kW to 1 MW range resent nearly 40 percent of all new connections achieved serve about 7.9 percent of all newly electrified population in in these countries. See figure 2.4 for a breakdown of this Sub-Saharan Africa. population by region and country, and table 2.4 for the GEP scenario codes for readers interested in exploring country- Based on the estimated new capacity in settlements where specific scenarios. mini grids are the least-cost option for all or part of the expected population between 2020 and 2030 For those interested in electricity access or the mini grid industry, it will come as no surprise that Sub-Saharan MAXIMAL MINI GRID DEPLOYMENT MODELED IN Africa is by far the most important market for mini grid THE GLOBAL ELECTRIFICATION PLATFORM electrification. While three African countries—Ethiopia, the Rather than selecting a particular demand scenario (low, Democratic Republic of Congo, and Nigeria—stand out for bottom-up, or high) and modeling results for all coun- their massive mini grid potential (49.4 million, 47.9 million, tries for that scenario, we use the GEP to review which of and 42.9 million people, respectively), one can also see the 96 modeled scenarios for each country deploys the from figure 2.4 that many others have huge populations most mini grids. While high-demand scenarios tend to that could be served by mini grids. Elsewhere, the electri- favor mini grids as the least-cost solution, they also tend fication potential is vast: 20 million people in Pakistan, 15 to label grid densification or extension as the least-cost million in Myanmar, and, in Haiti, almost 5 million people option. Besides, other parameters, like those referring to could gain least-cost electricity by mini grid. 82   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.4 • Distribution by country of 429.5 million people served at least cost by mini grids in 58 access- deficit countries Regional population totals • SSA: 380.4 million • SAR: 24 million • EAP: 19.5 million • LAC: 5.6 million SSA SAR Niger; 15.8 Chad; 15.0 Uganda; 14.3 Sudan; 16.6 Pakistan; 20.3 Mali; 8.1 Angola; 7.8 Kenya; 7.4 Bangladesh; 3.6 Ethiopia; 49.4 EAP Somalia; 14.0 Zimbabwe; 4.7 Senegal; 4.6 Burundi; 4.6 Guinea; 6.7 Sierra DRC; 42.9 Madagascar; 16.6 Leone; Rwanda; 2.9 2.7 Zambia; 4.2 South Sudan; 12.7 Myanmar; 15.0 Malawi; 5.8 Other Papua New Cameroon; EAP; Guinea; 3.3 1.1 3.7 LAC Côte Nigeria; 47.9 Tanzania; 20.7 Mozambique; 15.8 Burkina Faso; 10.4 d’Ivoire; 5.6 CAR; 3.3 Other SSA: 16.4 Haiti; 4.9 Other LAC; 0.7 n Sub-Saharan Africa (SSA) South Asia (SAR) East Asia & Paci c (EAP) Latin America & Caribbean (LAC) Source: ESMAP analysis of Global Electrification Platform results. Note: Under the scenario most favorable for mini grids. Data from the Global Electrification Platform. SSA: Sub-Saharan Africa; SAR: South Asia; EAP: East Asia & Pacific; LAC: Latin America & Caribbean. Other SSA: Ghana (2.5); Eritrea (2.3); Benin (2.2); Liberia (1.9); Congo (1.5); Mauritania (1.4); Togo (1.1); Guinea-Bissau (1.0); Gambia (0.9); Lesotho (0.5); Equatorial Guinea (0.3); Comoros (0.2); South Africa (0.2); Eswatini (0.2); Namibia (0.1); Djibouti (0.1); Gabon (0). Other LAC: Nicaragua (0.4); Honduras (0.3). Other EAP: Cambodia (0.6); Solomon Islands (0.4); Vanuatu (0.1). TABLE 2.4 • GEP scenario codes for each country’s maximum number of new mini grid connections by 2030 People newly connected to Region Country Name GEP Scenario mini grids (millions) SSA Ethiopia et-2-2_0_1_0_0_0 49.4 SSA Nigeria ng-2-2_1_1_0_1_0 47.9 SSA Democratic Republic of Congo cd-2-2_1_1_0_1_1 42.9 SSA United Republic of Tanzania tz-2-2_1_1_0_0_0 20.7 SSA Sudan sd-2-2_1_1_0_1_1 16.6 SSA Madagascar mg-2-2_1_1_0_1_0 16.6 SSA Mozambique mz-2-2_0_1_0_0_0 15.8 SSA Nigeria ne-2-2_1_1_0_0_0 15.8 SSA Chad td-2-2_0_1_0_0_0 15.0 SSA Uganda ug-2-2_1_1_0_1_1 14.3 SSA Somalia so-2-2_0_0 14.0 SSA South Sudan ss-2-2_0_1_0_0_0 12.7 SSA Burkina Faso bf-2-2_1_1_0_1_0 10.4 SSA Mali ml-2-2_1_1_0_1_0 8.2 SSA Angola ao-2-2_1_1_0_1_0 7.8 SSA Kenya ke-2-2_1_1_0_1_1 7.4 SSA Guinea gn-2-2_1_1_0_1_1 6.7 SSA Malawi mw-2-2_1_1_0_0_0 5.8 continued MINI GRIDS FOR HALF A BILLION PEOPLE    83 TABLE 2.4, continued People newly connected to Region Country Name GEP Scenario mini grids (millions) SSA Côte d’Ivoire ci-2-2_1_1_0_1_0 5.6 SSA Zimbabwe zw-2-2_1_1_0_1_1 4.7 SSA Senegal sn-2-2_1_1_0_1_0 4.6 SSA Burundi bi-2-2_1_1_0_0_0 4.6 SSA Zambia zm-2-2_1_1_0_1_1 4.2 SSA Cameroon cm-2-2_1_1_0_1_0 3.7 SSA Central African Republic cf-2-2_0_1_0_0_0 3.3 SSA Rwanda rw-2-2_0_1_0_1_0 2.9 SSA Sierra Leone sl-2-2_1_1_0_1_0 2.7 SSA Ghana gh-2-2_1_1_0_1_0 2.6 SSA Eritrea er-2-0_0_1_0_0_0 2.3 SSA Benin bj-2-2_1_1_0_1_0 2.2 SSA Liberia lr-2-2_1_1_0_0_0 1.9 SSA Congo cg-2-2_1_1_0_1_0 1.5 SSA Mauritania mr-2-0_1_1_0_1_0 1.4 SSA Togo tg-2-2_1_1_0_0_0 1.1 SSA Guinea Bissau gw-2-2_1_1_0_1_0 1.0 SSA The Gambia gm-2-2_1_1_0_1_1 0.9 SSA Lesotho ls-2-2_1_1_0_1_0 0.5 SSA Equatorial Guinea gq-2-2_1_1_0_1_0 0.3 SSA Comoros km-2-2_0_0_0_0_0 0.2 SSA South Africa za-2-2_0_0_0_1_1 0.2 SSA Eswatini sz-2-2_0_1_0_1_0 0.2 SSA Namibia na-2-0_1_1_0_1_0 0.1 SSA Djibouti dj-2-2_0_1_0_0_0 0.1 SSA Gabon ga-2-0_1_1_0_1_1 0.0 SSA Botswana bw-2-2_1_1_0_0_0 0.0 SSA Sao Tome and Principe st-2-0_0_0_0_0_0 0.0 SAR Pakistan pk-2-2_1_1_0_1_0 20.3 SAR Bangladesh bd-2-0_0_1_0_1_0 3.6 EAP Myanmar mm-2-2_1_1_0_1_1 15.0 EAP Papua New Guinea pg-2-0_1_1_0_1_1 3.3 EAP Cambodia kh-2-2_1_1_0_1_1 0.6 EAP Solomon Islands sb-2-0_0_0 0.4 EAP Vanuatu vu-2-0_0_0 0.1 EAP Timor-Leste tl-2-2_0_0_0_0_0 0.0 EAP Federated States of Micronesia fm-2-0_0_0 0.0 LAC Haiti ht-2-2_0_1_0_0_0 4.9 LAC Nicaragua ni-2-2_1_1_0_0_0 0.4 LAC Honduras hn-2-0_0_1_0_0_0 0.3 Source: ESMAP analysis of Global Electrification Platform results. GEP Scenario code definition: cc-: two letter country code 2-: default value indicating GEP V.2.0 1st value: [0: “Bottom up” , 1: “Top-down low”, 2: “Top-down high”] for “Electricity demand target” 2nd value: [0: “Social and productive uses demand included” , 1: “Residential demand only”] for “Productive uses inclusion” 3rd value: [0: “Estimated” , 1: “High”] for “Grid generation cost” 4th value: [0: “Estimated” , 1: “High”, 2: “Low”] for “PV cost” 5th value: [0: “No connections cap” , 1: “Capped connections in 2025”] for “Intermediate investment & Grid connection Cap” 6th value: [0: “Least-cost nationwide” , 1: “Only grid within 2 km”] for “Rollout Plan” 84   MINI GRIDS FOR HALF A BILLION PEOPLE What size mini grids could deliver electricity to these 430 potential for 8.6 percent and 11.7 percent of new mini grid million people? Figure 2.5 displays the population distri- connections to be served by systems in the 500 kW to 1 bution by applicable mini grid size. For example, systems MW and 1 MW+ range, respectively, which corresponds to of less than 20 kW are expected to serve 10.5 percent of more than 87 million people and aligns with discussions new mini grid connections. Mini grids roughly correspond- about so-called metro grids in some markets. ing to 20 kW, 80 kW, and 200 kW systems can serve about The GEP also estimates the investment required at about 52 percent of all new mini grid connections, which gener- $100 billion, or 66 gigawatts (GW) of installed capacity, ally accords with most mini grids built to date. The GEP almost all of it solar hybrid, with over 90% of both this analysis also suggests there is a great deal of scope for installed capacity and investment needed in Africa. The mini grids in the 200 to 500 kW range, which could serve GEP, however, defines3 the installed capacity of solar 17.5 percent of all new mini grid connections. There is also hybrid systems as the PV capacity plus diesel generator capacity, which results in a higher measure of installed FIGURE 2.5 • Distribution by mini grid size of 429.5 capacity than if measuring firm power as defined in chap- million people served at least-cost by mini grids in ter 1. Table 2.5 presents the number of settlements, elec- 58 countries with severe access deficits tricity connections (or households), population, installed 11.7% capacity in MWs, and the investment requirement to real- 429.5 million people newly connected to mini grids ize the delivery of electricity to 430 million people by mini 8.6% grid system size. 17.5% NATIONAL ELECTRIFICATION PLANNING 22.9% Geospatial plans represent a data-driven approach to planning for the efficient and effective deployment of lim- ited resources, particularly aimed at supporting countries 28.8% with low rates of electrification. Spatial modeling delivers a least-cost plan that identifies the optimal grid or off- grid technology tailored to local circumstances (including local cost parameters) and appropriate in its technical feasibility and economic viability. It also integrates social and economic planning objectives, like equity, which may 10.5% target universal service delivery or priority access for schools and clinics. The (local) costing associated with the deployment of different technology solutions (for ≤20 kW (20, 80] (80, 200] (200, 500] (500, 1000] >1000 kW example, grid, mini grid, or SHS) is triangulated and com- kW kW kW kW pared across various dimensions. The most important of Source: ESMAP analysis of Global Electrification Platform results. these are population (or institutional) density, distance, Note: Under the scenario most favorable for mini grids. Data from the and isolation from the main grid, in addition to current and Global Electrification Platform. ] means up to and including; ( means greater than but not equal to. forecasted demand. TABLE 2.5 • Distribution by mini grid system size of 429.5 million people served at least-cost by mini grids in 58 countries with access deficits <20 kW 20–80 kW 80–200 kW 200–500 kW 500–1,000 kW >1,000 kW Total Settlements 622,061 421,358 104,279 33,202 7,322 3,634 1,191,856 Connections 8,986,188 24,767,050 19,694,967 15,035,819 7,384,990 10,024,393 85,893,408 Population 44,930,942 123,835,249 98,474,836 75,179,094 36,924,951 50,121,965 429,467,039 Capacity (MW) 5,617.70 16,817.20 12,646.50 9,878.00 4,972.60 16,063.10 65,995 Investment (US$, millions) 14,513.80 28,311.50 20,408.40 15,261.70 7,296.80 14,412.80 100,205 Source: ESMAP analysis of Global Electrification Platform results. Note: Under the scenario most favorable for mini grids. Data from the Global Electrification Platform. MINI GRIDS FOR HALF A BILLION PEOPLE    85 FIGURE 2.6 • Geospatial least-cost rollout plans in Kenya and Rwanda a. Kenya b. Rwanda Sources: Kenya: World Bank 2007; Rwanda: World Bank 2009. GIS = geographic information system; km = kilometers; kW = kilowatt; mi = miles. Geospatial analysis surfaces the most efficient technol- therefore help rural electrification agencies and mini grid ogy solution by using not only location (where do pro- developers define the addressable market for mini grids spective beneficiaries reside?) but also over time. Hence both as interim and permanent electrification solutions in the focus on off- and mini grid programs. Even if the grid the country. may, in some cases, be the least-cost solution, decentral- Examples of geospatial planning exercises for least-cost ized solutions have a role in providing nationwide access electrification can be drawn from Kenya and Rwanda (fig- in the short term, making up for what could be halting ure 2.6) and Myanmar and Nigeria (figure 2.7). Their expe- progress in network extension, and providing backup riences led to the development of the GEP, also presented solutions. National least-cost electrification planning can below in more detail. 86   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.7 • Geospatial least-cost electrification plans for Myanmar and Nigeria by 2030, by technology component Myanmar Nigeria (with a focus on 4 states) Source: Castalia 2014; World Bank 2015. HV = high voltage; MV = medium voltage. OPERATIONAL EXPERIENCE AND THE GLOBAL on-grid/off-grid distinction and indicating least-cost solu- ELECTRIFICATION PLATFORM tions between mini grids and SHSs. Figure 2.7 shows the least-cost access solutions for Myanmar and four states in Kenya and Rwanda were early adopters in the use of GIS Nigeria by 2030, broken down by technology component. tools for electrification planning. Electrification programs in Such analyses provide first-order estimates of potential both countries were informed in 2009 by investment pro- sites for mini grid projects or SHS programs. spectuses relying on the results of geospatial analysis. These early experiences with least-cost electrification planning Building on this momentum, the World Bank has engaged focused on the least-cost rollout for grid extension without in national least-cost electrification planning with its giving explicit insights about the size and space for off- Global Electrification Platform (GEP), a multiphase proj- grid solutions. In particular, transitional off-grid solutions ect.4 It will improve, standardize, and simplify the use of (whereby no distinction was made between mini grids geospatial tools in least-cost electrification planning. To and stand-alone solar) were assumed to be inversely pro- achieve this, it is designed and developed at two levels, portional in terms of required space and time to progress briefly described below. in grid expansion, while the long-term targets for off-grid electrification were presumed to lie in areas not expected The GEP Explorer to be connected to the grid even in the long term. The GEP Explorer (figure 2.8) is an open access, interac- tive, online platform that provides overviews of electrifica- Since then, several countries have undertaken geospatial tion investment scenarios for all countries with less than least-cost planning, with accurately sized components of 90 percent electrification, which now includes 58 countries electrification programs, which helps countries update worldwide. The GEP Explorer allows the user, in two steps, their existing plans or develop new ones. Initially, the loca- to navigate nearly 100 electrification scenarios to meet tion and sizing of decentralized electrification were based access goals in those countries. The first step is an outlook on short-term grid extension. For example, a five-year roll- for an intermediate investment strategy (up to 2025). The out plan for grid densification and extension (prospectuses second explores full electrification by 2030. The number, typically have a five-year overview) indicates the space for type, and parameters of investment scenarios, along with transitional off-grid solutions, whereas a long-term plan for their inherent assumptions, are presented in the form of six the rollout of connections indicates the space for long-term levers designed to reflect different socio-techno-economic off-grid solutions. assumptions about the country context. Gradually, least-cost geospatial plans have achieved further All scenarios indicate the least-cost option, investment, sophistication in geospatial planning by going beyond an and capacity required to achieve full electrification at MINI GRIDS FOR HALF A BILLION PEOPLE    87 FIGURE 2.8 • The GEP Explorer Source: https://electrifynow.energydata.info/. both settlement and national levels. Results include three users to discover the values for the levers outside the pre- types of technologies, namely grid extension, mini grids, scribed values in the GEP Explorer and to investigate the and stand-alone systems. The user can also apply filters many variables not exposed as levers. to narrow results as well as toggle on different base lay- Other code modules. A number of modules can be used ers (for example, distribution network, location of health to further customize GEP elements. Examples include facilities, MapBox satellite) that can help better assess the the backend code here, the code for estimating custom modeling results. demand for settlements here, the code for population The GEP Explorer targets high-level decision makers in cluster generation here, the code for GIS data extraction addition to policy and investment analysts that can use to those clusters here, and the code for a high-level result its output to assess geo-infographic electrification invest- analysis here. The list is expected to expand as the project ment for an area of interest. It does provide some flexibility evolves. through scenario selection, but all scenarios are pre-run Training/teaching hub (access here). Online videos, with no option to customize on the fly. presentations, short lectures, and training material (for example, exercises) support capacity-building activities The GEP Toolbox around the GEP. The GEP team has run four capacity- The backbone of the GEP initiative, the GEP Toolbox, offers building events and has established an annual training in a range of tools and material that support reproducibility, Trieste, Italy, every June. Recently, the training material replicability of the GEP Explorer, as well as capacity build- has been bundled into a self-paced, online, open course ing, dissemination, and inter-organizational collaboration.5 offered by Open University6 (access here). The course A few components are described below. seeks to introduce trainees to geospatial electrification The GEP-OnSSET code. The GEP Explorer displays results modeling and planning by providing lectures on theoret- developed in conjunction with the Royal Institute of Tech- ical concepts and practical exposure through hands-on nology (KTH), building on a special version of the Open exercises. Source Spatial Electrification Tool (OnSSET). This is called Finally, it is worth mentioning that the GEP is part of a GEP-OnSSET and is available on GitHub along with online continuous data and model discovery process. To ensure documentation that supports its installation, setup, and that new data and models can easily be integrated into the use. GEP ecosystem, guidelines for its form and description, The GEP Generator. This open access, user-friendly Jupy- as well as handling protocols, have been developed more ter notebook allows a user to reproduce and customize the here. Based on these, expect annual updates of the GEP electrification models behind GEP Explorer. The notebook to reflect advances in algorithms and models, better data requires little to no programming experience to operate; input, and more scenarios defined by increasingly relevant it hides coding complexities and presents only key input and available levers. decision parameters to the user. The GEP Generator allows 88   MINI GRIDS FOR HALF A BILLION PEOPLE INDICATIVE WORKFLOW FOR THE DEVELOPMENT geospatial analysis. These data may come from national OF A GIS-BASED NATIONAL ELECTRIFICATION agencies, such as the Census Bureau, public statistics, PLAN the survey department, and other departments/minis- tries; international agencies such as the World Bank, IEA, As described above, a spatial least-cost electrification UN, FAO, IRENA, EU JRC, WRI, and so forth; open access scheme could support planning undertaken by various databases such as ENERGYDATA.info, OpenStreetMap, stakeholders; it could help form policy and design around HOTOSM and so forth; and in some cases proprietary nationwide pathways for electrification. Such modeling sources (for example, satellite imagery, Maxar’s building activities—and the plans they might inform—should be footprint). based on rigorous models and analytics as well as good governance principles. The literature indicates some over- The type of data required depends on the modeling frame- arching principles to guide such initiatives—for example, work, but usually it covers infrastructure (for example, U4RIA (DeCarolis and others 2017; Howells and others the power network, roads, settlements, public facilities), 2021). natural resources (solar irradiation, wind speed, hydro resources, land cover, protected land) and socioeconomic ESMAP has adopted those principles and, based on its activity (night lights, population, travel time, gross domes- operational experience, converted them into a more practi- tic product, electricity demand, affordability). Note that cal, five-step workflow presented in figure 2.9. The activities non-GIS data are also collected at this stage in order to are often linear; however, in reality, the process depends on support model calibration. These may include population country-specific conditions, including feedback loops and growth, urbanization, and electrification rates; household reiterations. For example, in some cases, capacity building size; electrification targets; and so on. Finally, planners may take priority over analytical work. should collect the technical and costing parameters for the technologies used—namely, medium-voltage (MV) lines, Diagnostic and preliminary analysis mini grids, and SHSs—as well as cost curves/projections, The first step in the workflow includes a thorough investiga- and discount rates. tion of data availability and know-how over the area of inter- est. Any existing and/or past applications of electrification Once collected, cleared, and compiled, all data and infor- planning techniques should be reviewed. The public or pri- mation should be reported to and shared with stakehold- vate stakeholders involved in the project should be listed; ers. Wherever technically feasible with regard to security their capacity in the use of GIS-based analytics to support and privacy, data should be shared on an open-source data electrification planning should likewise be assessed. The repository such as ENERGYDATA.INFO. Any missing data, status of the assessment should be documented in great- like the locations of productive activities (potential and est possible detail (for example, data types, quality, meta- existing), energy expenditure, and ability to pay may be col- data, level of knowledge, etc.) as this will determine the lected by site visits, geolocated surveys, or top-down sec- level of effort required in the following steps. Therefore, the tor-based analysis working through regional government diagnostic and preliminary analysis should delineate any and private organizations and commercial associations. analytical gaps and guide the project structure. This is usu- ally presented in a short yet concise inception report that Development of least-cost electrification plan guides the project thereafter. Here, the information collected in the previous step(s) informs the analytical work. To bound the least-cost plan- Data collection, mapping, and database preparation ning exercise, planners must identify the constraining The next step involves the collection, review, and com- parameters. Then they need to explain how each parame- pilation of the best readily available data required for the ter is defined or measured. Such parameters may include FIGURE 2.9 • Typical least-cost electrification planning sequence (best practice) Diagnostic Data collection, Development Reporting, Capacity and mapping, and of least-cost data transfer, building and preliminary database electri cation and knowledge analysis preparation plan dissemination transfer Source: ESMAP analysis. MINI GRIDS FOR HALF A BILLION PEOPLE    89 definitions for (1) starts, plateaus, and endpoints; (2) elec- renewable energy. The output is a prospectus with details tricity demand targets and projections; (3) costs and ser- on the upgrades necessary to achieve on-grid targets and vice standards for networks and individual systems; (4) the associated financing requirements. availability of renewable energy resources; (5) ability of Analysis of mini grid and off-grid solutions. Alongside consumers to afford upfront investments (such as connec- this probe of on-grid solutions, an economic analysis of the tion charges) and recurring expenditures (such as monthly potential for mini grid and stand-alone systems (namely, tariffs); and (6) criteria for temporal and spatial prioritiza- SHS, diesel gensets, and so forth) is recommended as part tion. These parameters are just indicative and depend on of least-cost electrification planning. Aimed at securing the scope of the analysis. The combination of those param- sector-wide support, the study might look at representa- eters creates scenarios that can be used to assess the sen- tive samples of high-potential off-grid sites, using data and sitivity of results to different input values. Typical sensitivity comparisons from existing sites. This analysis can help analyses examine the impact of various electrification tar- articulate the most important considerations for both pri- gets (following the Multi-Tier Framework) or demand lev- vate and government stakeholders in pursuing the off-grid els (based on demand sensitivity); different commodity sector, such as the potential profitability of different busi- prices; economic forecasts; and other variables (such as ness models and technologies, the tariffs and subsidies grid supply cost, technology costs, and service standards). required to achieve profitability, and promising sites for The analytical work provides a basis for the systematic roll- public-private partnerships. Prospectuses could be devel- out of a least-cost national electrification program for both oped using the results of this analysis. urban and rural areas and aims to either maximize cover- age for a given investment level or minimize investment for Reporting results, including technical model, data the targeted coverage. The objective function depends on transfer, and dissemination the model used or the scope of the analysis. Key outputs The national least-cost electrification planning exercise include the following: usually produces the following output: • A technology mix (grid connections, mini grids, and • A well-structured and -informed database that includes stand-alone systems) that fulfils the objective function input and output data (both GIS and non-GIS) and is subject to parameters and/or constraints. • An electrification model built and customized for the • System components’ characteristics (for example, size, country (or area of interest) capacity, investment, service quality, and other opera- • Documentation related to the project; this might tional features) required to implement the least-cost include an inception report, an intermediate report or technology mix. a final report that describes the methodological frame- The results of the least-cost model can be overlayed with work, key assumptions, results of the analysis, lessons some of the input data (or other information) and provide learned, and recommendations. It may also include any a greater level of analysis as per need. The following para- user guide for data processing or model running. graphs show how this can further support on- and/or off- The output must comply with global best practices and grid rollout plans in particular. ensure the project’s long-term sustainability. The U4RIA Detailed analysis of on-grid solutions. Building on the framework is highly recommended. Its output and pro- least-cost electrification, planners could articulate the need cesses can be retrieved, repeated, and rebuilt. They can to expand generation capacity (or electricity trading with be audited and are interoperable. All stakeholders should neighboring countries) and upgrades to grid infrastructure be consulted before recommending any institutional and needed to support the stated targets. Doing so might require organizational arrangements. This will help to ensure that access to data, such as installed generation capacity, exist- the GIS database is maintained and regularly updated ing transmission network, geotagged on-grid demand and and that the GIS electrification planning exercise can be demand projection, potential generation capacity, reserve replicated. Determining the appropriate institutional and constraints and operational constraints, and interconnec- organizational arrangements involves identifying the orga- tions with neighboring countries. The sequencing of new nization responsible for hosting the national power sector connections (and related costing) and extensions, along GIS database and the arrangements by which stakehold- with other needed changes to the supply system (namely, ers will update their database. Furthermore, which organi- network reinforcement, increased generation, and trans- zation will house the electrification planning models? Who mission), can be elaborated at this stage. A power-flow anal- will be responsible for replicating the geospatial electrifica- ysis might also reveal needed grid infrastructure upgrades tion planning exercise in the future? These decisions will to support targets and potential integration of variable need to be made. 90   MINI GRIDS FOR HALF A BILLION PEOPLE Capacity building and knowledge transfer do not require distribution networks. But its recurring costs Finally, all output, as described above, should be trans- are relatively steep because of battery storage needs over ferred both to the government (or its designated counter- the long term. Mini grids typically offer an intermediate parts) and to the institution funding the geospatial work. option to serve demand levels that are too high for SHS but Those who are analyzing the least-cost electrification plan- not great enough (or too remote) to justify connection to ning should be asked to: the main grid. (See chapter 1 for more on mini grid costs.) • Train professional staff throughout the assignment; Unlike stand-alone solar systems, mini grids and grid extension both require the installation of an electrical dis- • Familiarize them with the capabilities of the models; tribution system throughout the village in addition to a • Teach them about the methodology and analysis frame- minimum density of customers to justify this installation. work for updating the geospatial high-level analysis in Table 2.6 shows the maximum distance justifying the cus- the future; and tomer-connection cost as a function of the level of ser- • Explain the key variables, such as technology costs, for vice that the customer requires. For example, for a mini future sensitivity analysis. grid or main grid distribution system to be cost-effective, a group of customers requiring Tier 1 service would have The consultant should list any licenses needed to ensure to be densely co-located (within approximately 3.3 meters the functionality of the GIS planning platform and provide of one another). By way of contrast, a group of customers estimated costs for acquiring them and also instructional requiring Tier 5 service can be about 1.7 kilometers (km) materials for ongoing capacity building and knowledge distant from the other group members for a distribution transfer efforts. system to make economic sense. ANALYTICAL INSIGHTS AND GENERIC In practice, communities that require only Tier 1 service can OBSERVATIONS almost never justify a distribution system, and communi- Although geospatial electrification plans are country and ties requiring only Tier 2 service will rarely justify a distribu- context specific, some insights with general application tion system unless one or more customers require Tier 4 or can be gleaned from experience. They are presented below. 5 service. Distribution systems, whether powered by mini grids or the main grid, are generally justified for areas that Estimated (or targeted) electricity demand of beneficia- require Tier 3 and higher levels of service. ries shapes the cost-effectiveness of various technologies. Varying demand also affects the type of system recom- Figure 2.10 presents the indicative results from a simula- mended by electrification modeling tools: household/cus- tion run using the Hybrid Optimization Model for Multiple tomers with strong demand typically favor grid extension Energy Resources (HOMER) planning tool. It indicates that if the load centers are close to the grid, and mini grids if large loads close to an existing grid are more cost-effec- they are farther away, whereas low demand favors off-grid/ tively served by a grid extension. Small loads far from an SHSs. existing grid are more cost-effectively served by a mini grid. For this exercise, the same level of service was assumed The different balances of initial and recurring technology from both approaches, and the same cost for the distribu- costs affect how economies of scale are leveraged. To illus- tion system and for operation and maintenance. trate: grid electrification has relatively high initial costs but lower recurring costs. By way of contrast, SHS has lower Both electricity demand and customer density thresholds initial costs, at least for small, remote communities, as they presented above refer to residential loads. The addition of Least-cost electrification planning at the Mini grids are rarely justified from an eco- national level using geospatial analysis tools nomic standpoint in areas with demand typically follows a five-step process: (1) diagnostic for electricity that correspond to Tiers 1 and 2. In and preliminary analysis; (2) data collection, map- contrast, distribution systems, whether powered ping, and preparation of the database; (3) develop- by mini grids or grid extensions, generally make ment of least-cost electrification planning, including sense from an economic standpoint for Tier 3 and detailed analyses of main grid, mini grid, and off- higher levels of electricity demand, all other things grid solutions; (4) reporting results; and (5) capac- being equal. ity building and knowledge transfer. MINI GRIDS FOR HALF A BILLION PEOPLE    91 TABLE 2.6 • Maximum cost-justified distance for connecting a customer as a function of the required level of service Service tier 1 2 3 4 5 PV size (Wp) 10 100 1,000 3,000 10,000 Energy requirement (kWh/month) 1 10 100 300 1,000 Generation and storage technology Solar Solar home PV, battery, PV, battery, PV, battery, inverter, Lantern system inverter inverter, backup backup generator generator Capital cost $50 $300 $3,000 $9,000 $25,000 Maximum distance between customers to 3.3 20 200 600 1,667 cost-justify a distribution system (meters) Source: HOMER Energy. kWh = kilowatt-hour; PV = photovoltaic; Wp = watts peak. FIGURE 2.10 • Distance as a function of load size: therefore crucial in efforts to boost the rate of load growth Break-even grid extension (IEG 2015). Chapter 3 discusses this topic in further detail. 250 The quality of services that beneficiaries receive is another Breakeven grid extension distance (km) useful parameter when comparing the technologies in the Mini grid is cheaper least-cost plans. 200 Stand-alone solar systems tend to have smaller capacity 150 (10–200 watts [W]) and provide on average 1–20 kilo- watt-hours (kWh) per month. Within the Multi-Tier Frame- work for measuring household electricity access, this is 100 Tier 1 or 2 access. Stand-alone systems generally don’t use an inverter or a backup generator. This means that they 50 Grid extension is cheaper provide direct current (DC) power whose availability on any given day may be determined by the weather. The supply of 0 electricity may be sufficient for households that need only 0 1,000 2,000 3,000 4,000 5,000 light-emitting diode (LED) lighting and cell phone charging Load size (kWh per day) or possibly some other small DC appliance, such as a radio. Source: HOMER Energy. By way of contrast, the national grid extension can supply km = kilometer; kWh = kilowatt-hour. 24-hour power. In practice, however, many of these grids, particularly those in low-income countries, cannot meet productive activities could change those dynamics. That is, this level of reliability, and their customers suffer frequent communities where households have low levels of power outages and load shedding. Third-generation mini grids, on demand but are close to productive loads might also be the other hand, can provide high-quality electricity service. good candidates for mini grids or grid extension. Therefore, Members of the Africa Minigrid Developers Association electrification technologies should be compared not only (AMDA) report an average of 97 percent system uptime. according to the number of households they serve but also Both the main grid and mini grids can also supply suffi- according to the productive uses and community services cient alternating current (AC) power for productive use (for they enable. Productive uses have only recently been incor- example, grain milling, water pumping, sewing, woodwork- porated into geospatial electrification modeling efforts; ing), businesses (for example, telecommunications towers; thus, their impact on national least-cost electrification local, small and mid-size enterprises), and public services plans is not yet directly quantifiable. like schools and hospitals. It should also be noted that estimating the required level Note that the cost of electrification can vary widely depend- of service of unelectrified beneficiaries can be difficult, as ing on local subsidies, but the true unsubsidized cost of demand for energy tends to grow once energy becomes power is the appropriate metric for comparing options. As available. But the rate of growth varies and depends on mentioned before, the reliability of grids in many low-in- whether and how productive uses are promoted. Outreach come countries can vary significantly. In most places, out- efforts that demonstrate productive appliances and com- ages are a common occurrence, and mini grids are often mercial opportunities enabled by reliable electricity are deployed in areas already connected to an unreliable main 92   MINI GRIDS FOR HALF A BILLION PEOPLE Because of the vast differences in the qual- Early experiences with least-cost electrifica- ity and reliability of the energy service pro- tion planning have demonstrated the impor- vided by mini grids, solar home systems, and the tance of engaging with the private sector during the main grid, it is not appropriate to compare them diagnostic and preliminary analysis phase of a proj- according to their respective costs for the provision ect. This ensures that realistic assumptions about of electricity to the same groups of customers. costs and demand growth over time, among other assumptions, are built into the least-cost model’s calculations. grid to ensure reliable electricity service. The mini grids operated by OMC Power in Uttar Pradesh offer one such by technology, location, and sizing of the different compo- example. They serve villages where a government-owned nents of electrification programs. In addition, a plethora distribution utility is already present, but with low service of spatial data is increasingly available and continues to reliability, particularly during peak evening hours. improve in quality, coverage, and availability. Amid wild variations in the quality and reliability of the Nevertheless, there is always room for improvement. For energy service provided by mini grids, SHSs, and the example, better MV line mapping and improved demand main grid, comparisons based solely on their respective estimation are essential in order to improve the sophistica- costs for energy provision are inappropriate. Any com- tion of planning and tailor services to beneficiaries’ needs. parison should—to the extent possible—internalize costs associated with the reliability of supply (for example, From a planning perspective, knowledge of existing elec- value of lost load). tricity infrastructure is fundamental to ensuring that the results of geospatial modeling tools reflect conditions on LESSONS LEARNED AND CHALLENGES AHEAD the ground. Knowing the reach of electricity infrastructure Drawing on new developments in geospatial analytics, is critical if developers are to, first, identify who already has many countries are updating their geospatial least-cost a connection and, second, to cost the investment neces- plans, taking stock of the results achieved so far with grid sary for access provision. This knowledge is based on the extension, and analyzing the off-grid space to inform their location of the beneficiaries and their distance from exist- electrification programs. These updated programs can ing infrastructure. In the absence of this information, plan- provide more guidance on the design of implementation ning tools may overestimate the number of beneficiaries. frameworks and modalities for scaling up off-grid solutions, Mapping of MV lines is not yet common in most developing as they will be more specific about location and sizing of countries. Analysts may infer the extent of the MV network off-grid and mini grid potential in the country; the location from other parameters, but this approach would result in and sizing of long-term off-grid beneficiaries; and the pre- more errors in determination of electrification status than liminary location and sizing of mini grid solutions (to be if reliable maps of MV networks are available. followed by feasibility studies on the ground), based on Demand forecasting is perhaps the single most critical population density and loads (including their forecasting), modeling parameter for electrification planning, from geo- and local renewable generation resources. spatial least-cost plans to power sector planning, although Previous national least-cost geospatial planning exercises the willingness and ability of customers to pay for electricity have taught us the need to engage the private sector during are also critical from the developer/investor point of view. the diagnostic and preliminary analysis phase of a project. Improved demand estimates are also crucial to support As early least-cost planning activities have demonstrated, existing economic centers (and maximize the economic projections of the costs of mini grid electricity, main grid returns of electricity access) through adequate access to expansion plans, and demand growth in areas not expected electricity services and to forecast locations for productive to be connected to the main grid in the near to medium uses (and future economic growth potential) that may be terms have underestimated—by far—the actual potential prioritized by electrification programs. for mini grids. Engaging with the private sector early in a Finally, the models themselves need to improve constantly national least-cost electrification plan can enable the inte- in order stay current with new data and policy/planning gration of more realistic assumptions about mini grids, needs. Better methods could help internalize the costs of SHSs, and the main grid. reliability (for example, the value of lost load) and other Most modeling frameworks available at the moment have policy mandates (like energy access equity or equality tar- evolved to provide an explicit analysis of electricity access iff and subsidy schemes). They could also better accom- MINI GRIDS FOR HALF A BILLION PEOPLE    93 Looking ahead, the next critical advances Geospatial analysis provides a broader pic- in geospatial planning are improvements to ture of communities’ locations and char- network mapping and demand estimation, which acteristics a portfolio can consider, a picture that will further increase the accuracy of national least- enables mini grid developers to exploit economies cost electrification plans. of scale and prepare quicker, more cost-effective rollout plans and plans for service and maintenance. modate new configurations like hybrid or biomass-based and reduces the time and resources spent on prospect- systems and climate aspects (for example, resilience to ing for such communities, and it can mitigate the risk of climate change or disasters). demand uncertainty by incorporating a larger number of customers in a single investment (the portfolio) as com- pared with a single mini grid. MINI GRID PORTFOLIO PLANNING Without geospatial tools, developers must often rely on anecdotal suggestions from local governments to identify OVERVIEW promising communities to visit and investigate further for Geospatial analysis can also be used as part of a portfolio consideration. While such human intelligence is still useful planning approach for mini grid development, to comple- and can complement or be used to validate the recom- ment a comprehensive national least-cost electrification mendations from geospatial portfolio planning, a broader planning framework and, in the absence of such a frame- picture of the locations and characteristics of communities work, where grid extension is expected to be limited or that can be considered at a portfolio level will enable mini unlikely because of political considerations, insolvency of grid developers to exploit economies of scale and prepare the distribution companies, and so forth. If national least- quicker, more cost-effective rollout plans and plans for ser- cost electrification planning exercises have carved out vice and maintenance. At a more micro level, geospatial areas that mini grids can serve as the least-cost solution, tools can be used for mini grid generation sizing and distri- mini grid developers and electrification agencies may bution network planning. wish to focus their time and resources on investigating the Technological advances and cost reductions in satellite potential for developing mini grids to serve communities in imagery and in machine learning, increased sophistication these areas. of algorithms and analytical approaches, and the prolifer- Developers would do well to remember that political and ation of web-based technologies have made available a other considerations may affect the likelihood of grid host of new digital tools to improve the efficiency of mini extension regardless of the underlying economics. The grid development. This section examines how some of grid may be extended to areas where it might make more these tools can expedite the process of identifying poten- sense to pursue decentralized solutions and, inversely, dis- tial sites for mini grids; collecting, estimating, and analyz- tribution companies may not be in a position to extend the ing customer data; optimizing mini grid system designs; grid to areas even when it may be the least-cost solution. and finding and selecting developers and investors, using Nevertheless, national least-cost electrification plans can innovations from the frontier as examples. These exam- serve as a guide and a starting point when prospecting for ples from the frontier include private developers, who are suitable sites for mini grid deployment. using geospatial and other digital technologies to improve preparation of portfolios of mini grid projects, as well as Geospatial portfolio planning, which is already being public-sector programs that are taking advantage of such used by a number of established mini grid companies in disruptive technologies to facilitate implementation of Sub-Saharan Africa, greatly reduces the pre-investment mini grid projects. cost associated with preparing sites for mini grid devel- opment compared with traditional approaches, which rely Figure 2.11 presents an indicating sequence of activities heavily on the deployment of full multidisciplinary teams (workflow) involved in geospatial portfolio planning for mini to villages to explore the scope for mini grid electrification. grids. The rest of this chapter describes each of these steps Geospatial portfolio planning does not eliminate the need and looks at how disruptive technologies are helping gov- to conduct feasibility studies or engage with beneficiary ernments and private developers prepare and implement communities, but it does provide guidance on where com- mini grid portfolios. The Access to Distributed Electricity munities suitable for mini grid electrification are located and Lighting in Ethiopia (ADELE) project and the Nigeria 94   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.11 • Geospatial portfolio planning sequence Site identi cation, Analysis, sensitivity, Characterization System design Financial attribution, and visualization, and of load pro le and optimization modeling prioritization dissemination Source: ESMAP analysis. Electrification Project (NEP), a flagship initiative of the fed- The cluster data set improves when it is merged with OSM eral government of Nigeria, supported by the World Bank, land-use and mapped-buildings data. are two especially apt exemplars of geospatial portfolio This process identified nearly 200,000 clusters across planning. Lessons from studies of these two projects are the country—188,014 clusters, to be precise—occupying used interchangeably, hereafter, to illustrate the implemen- from 1.1 hectares to 48,500 hectares. Figure 2.13 visualizes tation of the suggested workflow in the real world. the distribution of these clusters across Nigeria, and rep- THE WORKFLOW PHASES FOR MINI GRID resents their distribution by size (in hectares). PORTFOLIO PLANNING: SPATIAL DATA AND More recently, alternative methods have been generating ANALYTICS population clusters with vector data (polygons or cen- troids) that show the distribution of buildings (or rooftops). PART 1. Site identification One example is the building data set developed by NRECA The initial phase of the workflow involves the collection and as part of the USAID Distribution Systems Strengthening processing of geospatial (and other) data and information Project. This visualizes digitized housing structures in Ethi- for the identification of sites with potential for mini grid opia based on satellite imagery (NRECA 2019). In addition, development. The process is split into three main activities: the Digitize Africa building footprint data set has been pro- namely, the generation of population clusters and the attri- duced for Sub-Saharan Africa by Ecopia.AI and Maxar with bution of those clusters and their prioritization based on a funding from the Bill and Melinda Gates Foundation (Eco- set of criteria. This activity produces a list of possible mini pia AI and Maxar Technologies 2021) (figure 2.14). The first grid sites. data set identified about 13.7 million buildings across Ethi- opia; the latter identified about 32.8 million rooftops in the CONVERTING RAW GIS DATA AND SATELLITE IMAGERY INTO country with a high estimated accuracy (>95 percent valid POPULATION CLUSTERS precision and recall). A key input for electrification planning and projects is to understand where people live. The location of settlements The new methods grouped rooftops into clusters using and their boundaries is therefore the baseline for any geo- density-based clustering (DBSCAN) and three parameters. spatial planning exercise. Rooftops in proximity are called “core,” whereas dispersed single rooftops are called “noise” and omitted from the In Nigeria, the administrative areas (from higher to lower cluster. This identifies more populous areas (clusters) with levels) are federal states, local governmental areas (LGAs), enough density to justify mini grid development. and wards. Population statistics and administrative bound- aries are well known at the LGA level, but not at the ward The DBSCAN algorithm is based on two parameters. The level. But having the exact population figure and boundar- first is the maximum distance between rooftops (or, eps), ies for wards is not sufficient for designating suitable sites and the second is either the density threshold (or, minPts) for mini grid electrification. That exercise would require or the minimum number of neighboring rooftops each having the exact locations of the buildings and settlements. building needs to have within the maximum distance in a So a cluster identification algorithm was developed for the potential cluster. The process is iterative; the clustering NEP to automatically identify the location and boundaries algorithm needs to be run for different parameter combi- of population settlements.7 nations in order to identify those yielding the most repre- sentative results. Based on the data sets described in box 2.1, it is possible to identify clusters with great accuracy following the pro- Note that there is a modified version of the DBSCAN cess illustrated in figure 2.12. First, the HRSL buildup raster algorithm (bounded DBSCAN suggested by Village Data is vectorized, buffered, and dissolved to define boundaries. Analytics, VIDA) that includes an optional parameter MINI GRIDS FOR HALF A BILLION PEOPLE    95 BOX 2.1 DATA SOURCES FOR CLUSTER DEFINITION IN THE NIGERIA ELECTRIFICATION PROJECT A population cluster is an area that could be supplied ments are located. Combining those estimates with by a single distribution network. All households that OSM building features presents a clearer picture of are “relatively” close to one another form one cluster. the settlement boundaries and building counts within A cluster, in principle, could be as small as a hamlet, or them. Note that some of these data sets are incom- as large as a city. Key data sets were used for cluster plete—for example, in Nigeria, Niger state is mapped identification: more precisely than its neighboring states. So addi- tional sources of data should complement the meth- • OpenStreetMap (OSM) data (contains vector layers odology, when and if they are available. with buildings, residential land use, roads, water- ways, and so forth) a For example, the WorldPop’s peanutButter web appli- cation, a more recent development, allows the cus- • A high-resolution settlement layer (HRSL) b tom generation of gridded population estimates at • Administrative boundaries c 100-meter spatial resolution. e The application builds A HRSL population raster estimates population distri- on the high-resolution building footprint data set and bution accurately. d Spatial processing of the HRSL can complements both population and building counts in thus help roughly delineate where population settle- the candidate sites. In the case of Nigeria, OSM data sets were retrieved from the open access Geofabric Server, available at https:/ a.  /download. geofabrik.de/africa/nigeria.html. Vector data list coordinates that define points, lines, or polygons. The HRSL layer was retrieved from the Facebook Connectivity Lab and Center for International Earth Science Information b.  Network (CIESIN), Columbia University, available at https://www.ciesin.columbia.edu/data/hrsl/. In the case of Nigeria, administrative boundaries have been retrieved from the Database of Global Administrative Areas c.  (GADM), available at https://gadm.org/download_country_v3.html. d. Raster data consist of pixels (or cells) where each pixel has an associated value. e. The WorldPop’s peanutButter web application is available as a beta version at https://apps.worldpop.org/peanutButter/ FIGURE 2.12 • Methodology for the generation of population clusters in Nigeria, using the HRSL and OSM data Input data Processing Result Built-up raster data Vectorizing, buffering, Settlement clusters Pixels with presence of and dissolving in order Outline of settlement structures built-up structures are to indicate precise result from merging the three input depicted in light colors settlement boundaries data types OSM land use Filtering, buffering and Land use types are colored dissolve individually (residential in Extracting residential land use red, industrial in blue) and reducing polygon count OSM buildings Clipping and clustering Open source mapped Finding additional buildings polygons . . . not covered in other data sets Source: Integration and Reiner Lemoine Institut 2016. Note: OSM = OpenStreetMap. 96   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.13 • Nigeria’s population clusters: Spatial distribution (left) and size histogram, in hectares (right) Cluster size distribution 25 20 Cluster count (thousands) 15 10 5 0 Source: Integration and Reiner Lemoine Institut 2016. 1 10 100 1,000 10,000 ha = hectare. Cluster size [ha] FIGURE 2.14 • Sample outputs from the Digitize Africa building footprint data set  axar’s building footprint (blue polygons) and a. M  axar’s building centroids and HRSL’s population b. M NRECA’s building centroid (red dots) distribution (raster data set) Source: VIDA 2021. FIGURE 2.15 • Concept of the DBSCAN algorithm Border p and q are Noise p q density-reachable from o Core o Therefore p and q are density-connected Eps = 1 Eps = 1 minPts = 4 minPts = 4 Border Source: Hahsler, Piekenbrock, and Doran 2019. MINI GRIDS FOR HALF A BILLION PEOPLE    97 (minClusterSize) related to the minimum number of Once the set of parameters are selected and the rooftops grouped into clusters, then the boundaries are drawn to buildings a cluster requires in order to be considered. This parameter filters out settlements below a certain buildingidentify the exact contours of the potential village. This count and can be used in cases where this is needed (as is usually done by using an alpha shape generating algo- shown in figure 2.16). rithm, which estimates a village’s size, area, and density (by comparison, a simple convex hull-based village boundary always overestimates the village area and underestimates FIGURE 2.16 • Ethiopia’s rural population settlements and density). Figure 2.17 offers a comparison of convex hull and mini grid deployment: Bounded DBSCAN clustering alpha shapes. Number of o -grid settlements (with N>100) found across Ethiopia ATTRIBUTING POPULATION CLUSTERS 6,000 Once the clusters have been identified, attributes are Maximum distance between 50 730 393 192 72 23 8 added. These are used to rank the most suitable locations 5,000 for mini grid deployment. Common attributes in this rank- rooftops (meters) 2,231 1,496 1,114 883 713 578 ing exercise are shown below: 100 4,000 • Cluster name (if available, or its ID) 3,000 150 4,541 3,216 2,369 1,836 1,496 1,265 • Cluster size 2,000 • Administrative division(s) (municipality, district, region, country) 200 6,464 5,327 4,108 3222 2615 2,206 1,000 • Building count 5 10 15 20 25 30 • Population Minimum number of neighbors • Power situation and nighttime light intensity (electrified Source: VIDA 2021. or not, existing mini grids or SHS) Note: The y axis represents the eps parameter, maximum distance • Distance to infrastructure (grid network, roads, substa- between rooftops (in meters); the x axis, the minPts, or minimum number tion/transformer, etc.) of neighbors. The minimum minClusterSize is set to 100. Highlighted box indicates the reference pair of input parameters as derived from experience (eps: 150, minPts: 20, minClusterSize: 100). BOX 2.2 FINDING THE OPTIMAL INPUT PARAMETERS FOR DBSCAN The modeling exercise in Ethiopia has made it hard ematically good cluster will not necessarily capture a to identify a single set of input parameters that gen- physical feature such as a village. erate highly accurate results for the whole country. Yet another approach might validate results against What works well for one area might be suboptimal in real settlements via statistical validation. This would another owing to the differences in geography, topol- require a training data set and machine-learning- ogy, population densities, or even building structure based and -supervised validation methods. The latter and distribution. may contain boundaries and characteristics (such But a purely mathematical method might help identify as population and building counts) of actual villages. robust clusters. One can, for example, use a density The comparison could calibrate the input parameters based cluster validation (DBCV) approach (or similar, under a framework that reflects the situation on the for example, SC or AMI) and compare parameter val- ground more accurately. ues (Moulavi and others 2014). DBCV yields a valida- Finally, it should be noted that it is imperative to engage tion index ranging between -1 and +1, with higher values local counterparts in calibration and validation. indicating better clustering performance. But a math- The implementation code of the density-based cluster validation (DBCV) methodology is available at https:/ a.  /github.com/christopherjen- ness/DBCV; another implementation is available at HDBSCAN’s original GitHub repository at https://github.com/scikit-learn-contrib/hdb- scan/blob/master/hdbscan/validity.py. 98   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.17 • Cluster contour delineation: Ministry of Power and the Nigerian Energy Support Pro- Convex hull (left) and alpha shapes (right) gramme have compiled and collected data from the field Convex hull Alpha shape on the extent of the medium voltage grid as well as other electricity infrastructure and made this information avail- able on a web-based mapping platform. Finally, some attributes will need to be calculated based on available data. For example, the area of each cluster can be estimated simply using spatial analysis. Or, as another example, determining the reach and extent of the main grid can help calculate its distance from a cluster. This kind of information is usually available in developed economies, Source: VIDA 2021. but in developing countries may be hard to find; one may need to estimate or infer it from other data, like nighttime • Number of public institutions (for example, health facil- lighting satellite imagery. See Figure 2.18 for an example. ities, education institutions, other services, and admin- Where possible, knowledge of the grid distribution and istrative offices) grid-connection status of population clusters can be • Commercial buildings, stores, and other anchor loads improved by better collaboration with distribution com- (mobile tower) panies or by site visits. In case data on the existing grid • Agriculture (crops, harvested area, production, yield, network is not available for the country or area of interest, and so forth) it can be simulated using other available GIS data sets as proxy (for example, night lights, road networks). In fact, • Post-harvest activity (milling, drying, cooling, storage, gridfinder.org12 has developed a methodology that predicts and so forth) the routing of the transmission network using the above • Resource availability (solar, wind, hydro, biomass etc.) data and a minimum spanning tree approach. Results are • Other socioeconomic characteristics (poverty rate, available globally and the model is open source (Arderne income level, household profile, and assets) and others 2020). A growing pool of databases has recently been providing VIDA has further refined the code in the Gridfinder repos- access to this information. Some examples worth men- itory. As a result, the predictions have improved, making tioning: the code deployable across the globe for any given time • The open data platform Energydata.info8 is a good place period. A sample VIDA GridLight is shown in figure 2.20. to start finding geospatial data for electricity and energy The GridLight algorithm has been extensively tested and access planning. utilized in several countries. In Nigeria the algorithm was tested against 300 on-ground survey data points with an • OpenStreetMap (OSM) and Humanitarian OpenStreet- accuracy level of 82 percent. In Ethiopia, we identified an Map (HOTOSM)9 are open-source projects with millions overlap of 80 percent with the latest utility data. of geospatial features pertinent to population distribu- tion, commercial and public buildings, infrastructure, Apart from a proxy for the existing grid network, nighttime and resources. lighting13 can be used, provisionally, to classify electrified population clusters in areas with high night light emis- • FAO’s GAEZ Data Portal10 and the International Food Pol- sions. These regularly visible light emissions are usually icy Research Institute’s MapSPAM & Harvest Choice11 based on the number of streetlamps or other permanent series provide a suite of data sets related to agricul- lights in villages or communities. Highly visible nighttime tural activity and productivity (for example, land, water, light activity suggests electricity sources. This logical soil, terrain and agro-climatic resources, protected assumption (of available electricity) should then be vali- areas, actual and potential production/yields as well as dated. In Nigeria, for example, data on education facilities selected socioeconomic and demographic data). from the Nigeria Millennium Development Goals Infor- On top of publicly available resources, one could also con- mation System, which includes information on whether sult with local counterparts who might share additional, a facility is grid connected, have been used to verify the proprietary information that might be useful later in the electrification status of the population cluster where the ranking process. Often, geospatial data do exist in differ- educational facility is located. ent governmental agencies, but their permission may be needed in order to use these datasets. Figure 2.19 illus- trates a great example from Nigeria, where the Federal MINI GRIDS FOR HALF A BILLION PEOPLE    99 FIGURE 2.18 • Health facilities and education facilities within 500 meters of village boundary Source: VIDA 2021. FIGURE 2.19 • Main grid coverage in Nigeria Source: Screenshot from https://nigeriase4all.gov.ng/map. 100   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.20 • VIDA GridLight prediction for Ethiopia FIGURE 2.21 • Nighttime lighting in Nigeria (blue) compared to Ethiopian Electric Utility data (red) Source: Integration and Reiner Lemoine Institut 2016. Note: Nighttime lighting shown in white. Electrified consumer clusters appear in purple. Source: VIDA 2021. Figure 2.21 presents an example of the application of night FIGURE 2.22 • Nigerian night lights and 20 km buffer lights’ data, validated using data from other sources, to pre- zones dict the electrification status of population clusters. Electri- fied clusters can then be interconnected with an automatic grid-extension algorithm, which takes several factors into account—including topography, roads, and water bodies— to derive the most realistic grid connections. The informa- tion on population or population density and electrification status and distance from the grid may be supplemented with additional socioeconomic data to allow for more cri- teria when prioritizing the population clusters for mini grid electrification. LONG AND SHORT LISTS OF POTENTIAL MINI GRID SITES After the identified clusters are attributed, they can be prioritized according to certain criteria to generate long lists and short lists stating the project’s requirements and needs. The scope of this step is to eliminate clusters that don’t meet certain criteria. Elimination saves compu- Source: Integration and Reiner Lemoine Institut 2016. tational time and effort in the workflow steps that follow. Note: Purple areas show 20 km buffer zones. Night lights are shown in yellow with population clusters in black. Below are some examples. Once the locations of electrified clusters are known, a buf- fer zone can be applied around the clusters assumed to be depends on the local context. In an analysis by the Nige- electrified via grid connection. This represents a method- rian Energy Support Programme, for example, a 20 km ological shortcut to a least-cost electrification modeling radius around electrified clusters was taken to be suitable approach: clusters within the buffer zone are considered for electrification via grid connection in the base scenario. likely to be subject to grid extension and therefore unlikely Meanwhile, a 10 km buffer zone was thought appropri- to attract investment from private mini grid developers. ate in a low-grid electrification scenario (Integration and Meanwhile, clusters outside the buffer zone are unlikely to Reiner Lemoine Institut 2016). Figure 2.22 illustrates the be served by the grid within the time horizon considered 20 km buffer zones around electrified areas (based on and may thus be good candidates for mini grid electrifi- night lights), where grid extension would likely be the pre- cation. The decision on the size of a suitable buffer zone ferred option. MINI GRIDS FOR HALF A BILLION PEOPLE    101 Another approach would be to take the known grid cov- Environmental issues might also impose additional con- erage map, such as that in figure 2.22, and apply buffer straints in the prioritization process. For example, poten- zones directly around the network. This was done for the tial complications to implementing mini grid projects in NEP, and given the poor financial health of the distribution environmentally sensitive areas, and the time and cost of companies in Nigeria, a more aggressive scenario for mini obtaining necessary environmental permits and approvals, grids—where grid extensions are stalled—was consid- can be avoided by determining as early as possible whether ered, and the buffer zone was reduced to 5 km. Clusters a population cluster being considered lies in such an area, outside these buffer zones were considered suitable for so that an informed decision can be made on whether to decentralized electrification, but further screening based pursue a mini grid project at that site. on population was conducted. The economic viability of Key biodiversity areas (KBAs) are designated by the KBA mini grids was deemed unlikely for population clusters of Partnership,14 a consortium of wildlife conservation groups less than 1,000, which would be more effectively served including the World Wildlife Fund, while the International via SHS. Union for Conservation of Nature (IUCN) maintains a com- After sites within a specified distance from the grid (5 km prehensive database of legally protected areas around the for Nigeria) and with a population below a certain thresh- world. Protected areas are classified into different catego- old (1,000 people) are excluded from consideration, the ries, such as IUCN I-VI, Ramsar, and World Heritage sites, remaining clusters can be ranked based on scored criteria. which are all legally protected areas, while KBAs are not Recall the additional data collected on the clusters (pres- necessarily legally protected (though some protected ence and location of schools, clinics, and telecommuni- areas coincide with KBAs). cations towers). These can now be used to prioritize the The boundaries of population clusters can be compared clusters for mini grid electrification. with those of legally protected areas such as forest reserves The prioritization categories for the NEP were population, and national parks, as well as KBAs, using the Integrated density, distance to grid, and presence of telecommuni- Biodiversity Assessment Tool.15 This online geospatial tool cations towers, schools, and health facilities. Normalized hosts and maintains the three key global biodiversity data values for the prioritization categories were created and sets: the IUCN Red List of Threatened Species, the World summed up with weighting factors, and the clusters were Database on Protected Areas, and the World Database of then evaluated according to the prioritization criteria and Key Biodiversity Areas. ranked by state. Figure 2.23 exhibits the top 100 clusters The NEP decided to avoid the additional scrutiny and com- projected to be suitable for mini grid electrification for each pliance costs that come with implementing mini grid proj- federal state. FIGURE 2.23 • Results of prioritization of clusters for mini grid electrification in Nigeria Source: Integration and Reiner Lemoine Institut 2016. 102   MINI GRIDS FOR HALF A BILLION PEOPLE ects in protected areas and KBAs (if they are permitted at sible, to verify whether the community is served by the grid, all); therefore, any population clusters that intersected with and deploying agents (for example, on motorbikes where them were excluded from further consideration. Figure this is a swift and safe option) to quickly check on the elec- 2.24 shows an instance of population clusters within pro- trification status of all the population clusters shortlisted tected areas or KBAs being flagged for exclusion. for potential mini grid projects. FIELD VERIFICATION Part 2. Load profile characterization While the site identification and screening methodology The first part of the analysis identified, attributed, and vali- described in this section uses available geospatial data to dated—to the extent possible—the candidate sites for mini propose communities where mini grid electrification may grid deployment. The following step focuses on estimating be suitable, in the absence of reliable data on the reach the load profile for each candidate site. Estimating the load of the main grid, it does so partly by making assumptions profile is usually subject to requirements, as listed below about the electrification status of the population clusters. and illustrated in figure 2.25: Before deploying multidisciplinary survey teams to collect • Identify the composition of potential customers data on these communities, wasting resources on false positives (communities thought to be off grid that are actu- • Estimate the daily consumption of each customer seg- ally on grid) may be avoided with validation exercises. Such ment exercises include calling someone in the community, if pos- • Estimate the load profile in the cluster • Incorporate seasonality of demand over the year • Estimate growth rate (forecasting) The site-screening phase of a geospatial plan for portfolio development includes There are two paths. The first is to utilize all collected data identifying, characterizing, and ranking popula- and information, measure them against past experience, tion clusters, as well as undertaking environmental and estimate the load (or simulate it with machine learn- screening and field verification. Throughout this ing). The second is to survey all the candidate sites—if phase, the criteria used for all these tasks need to they’ve been shortlisted—or, failing that, conduct a ground be carefully considered. survey of sample sites on the load in those locations. The selection depends on the scope of the project as well as the available resources. FIGURE 2.24 • Population clusters falling in protected areas or KBAs flagged for exclusion in Nigeria Source: Screenshot from https:/ /worldbank.africageoportal.com/. IUCN = International Union for Conservation of Nature. MINI GRIDS FOR HALF A BILLION PEOPLE    103 FIGURE 2.25 • Requirements for estimating load profile 1. Composition of custmers (% split and quantity) Additional demand 2. Daily consumption A daily kWh value for each of the customer segments 3. Load pro le Residential Commercial Public Productive Irrigation Averaged daily load pro le 12 10 Low 8 Load kW 6 Medium 4 2 Agricultural 0 High equipment 2 4 6 8 10 12 14 16 18 20 22 24 Hour of day For each of these activities 4. Seasonality 5. Growth rate the secondary datasets will • Years to plan the location for be assessed and used • Presumed growth rate for those years where applicable Source: Integration 2021. kWh = kilowatt-hour. ROUGH ASSESSMENT BASED ON COLLECTED GIS DATA • Water pump (if existing) demand estimated at 5,000 (EASILY SCALABLE) kWh/day The rough assessment is based on general assumptions. • Telecommunications tower (if existing) demand esti- Some of them have been used (for example, in the case of mated at 84 kWh/day Ethiopia) to provide a quick, high-level estimate of the total demand in the candidate sites. These assumptions are MORE ACCURATE ASSESSMENT FOCUSED ON A FEW PILOT listed below: SITES (SURVEY DATA COLLECTION, MANUAL INTERVENTION) In the second approach—and for the shortlisted sites—it • Percentage of households connected to the mini grid is recommended that survey teams be deployed to col- (58 percent) lect data. Sending these teams only to those communities • Residential vs commercial customers likely to host a viable mini grid project will save time and – residential = (small rooftops and 50 percent medi- resources, and sending teams after the building mapping is um-sized rooftops) x 0.58 (connection rate) completed will enable them to conduct their surveys more – commercial = (large rooftops and 50 percent medi- efficiently. um-sized houses) x 0.58 (connection rate) Under the NEP in Nigeria, survey teams were deployed to • Residential customer demand, estimated at ~0.22– collect data on the community, households, public institu- 0.32 kWh/day tions, and commercial and productive users of energy in the prioritized unelectrified communities. Each survey cov- • Commercial customer demand, estimated at ~1.1 kWh/ ered community, commercial, household, and environmen- day tal-social data. The survey teams geotagged and surveyed • Public institution demand (if existing) estimated at: nonresidential structures in each community to identify – 2.97 kWh/day for primary schools the key loads (large, daytime, productive and commercial) – 11.23 kWh/day for health clinics that might be critical to mini grid viability. They recorded the count and wattage of large appliances, light bulbs, and • Flour mill (if existing) demand estimated at about 43.77 fans for each geotagged building. They also geotagged a kWh/day subset of households, categorizing each as large, medium, or small. 104   MINI GRIDS FOR HALF A BILLION PEOPLE Watts Power (kW) Watts Watts 10 5 0 0 0 15 10 50 70 25 35 20 30 40 60 80 90 20 30 200 400 600 800 0 10 20 30 1,000 0: 00 – 0:00–1:00 0:00–1:00 0:00–1:00 1:0 1:0 0– 0 1:00–2:00 1:00–2:00 1:00–2:00 2: 2:0 2:00–3:00 2:00–3:00 2:00–3:00 Source: VIA 2021. 00 0 – 3:00–4:00 3:00–4:00 3:00–4:00 Source: VIDA 2021. 3: 3:0 00 0 4:00–5:00 4:00–5:00 4:00–5:00 –4 5:00–6:00 5:00–6:00 5:00–6:00 Commercial 4: :0 00 0 6:00–7:00 6:00–7:00 6:00–7:00 –5 7:00–8:00 7:00–8:00 7:00–8:00 Shops load pro le 5: :0 00 0 8:00–9:00 8:00–9:00 8:00–9:00 –6 6: :0 9:00–10:00 9:00–10:00 9:00–10:00 Health clinics load pro le 00 0 10:00–11:00 10:00–11:00 10:00–11:00 –7 7: : 11:00–12:00 11:00–12:00 11:00–12:00 00 00 Households –8 12:00–13:00 12:00–13:00 12:00–13:00 Average residential load pro le :0 Time of day Time of day 8: 00 0 13:00–14:00 13:00–14:00 13:00–14:00 Time of day 9: –9 14:00–15:00 14:00–15:00 14:00–15:00 00 :00 – 15:00–16:00 15:00–16:00 15:00–16:00 10 10: :0 0 16:00–17:00 16:00–17:00 16:00–17:00 0– 0 17:00–18:00 17:00–18:00 17:00–18:00 11 11 :0 :0 18:00–19:00 18:00–19:00 18:00–19:00 0– 0 12 12 19:00–20:00 19:00–20:00 19:00–20:00 :0 :0 0– 0 20:00–21:00 20:00–21:00 20:00–21:00 13 13 21:00–22:00 21:00–22:00 21:00–22:00 :0 :0 Time 0– 0 22:00–23:00 22:00–23:00 22:00–23:00 14 14: :0 0 23:00–0:00 23:00–0:00 23:00–0:00 0– 0 15 15 :0 :0 0– 0 16 16: Watts :0 0 Watts 0– 0 Watts 17 17 :0 :0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 0 50 0 100 200 300 400 500 150 100 250 200 0– 0 18 18 :0 :0 0 0 0:00–1:00 0:00–1:00 0:00–1:00 19 –19 :0 :0 1:00–2:00 1:00–2:00 1:00–2:00 0– 0 2:00–3:00 2:00–3:00 2:00–3:00 20 20: 3:00–4:00 3:00–4:00 3:00–4:00 :0 0 0– 0 21 4:00–5:00 4:00–5:00 4:00–5:00 21 :0 :0 5:00–6:00 5:00–6:00 5:00–6:00 0– 0 6:00–7:00 6:00–7:00 6:00–7:00 FIGURE 2.27 • Demand curve for a randomly selected candidate site in Ethiopia 22 22 :0 :0 0 0 7:00–8:00 7:00–8:00 7:00–8:00 23 –23 8:00–9:00 8:00–9:00 8:00–9:00 Flour mill load pro le :0 :0 0– 0 9:00–10:00 9:00–10:00 9:00–10:00 24 :0 10:00–11:00 10:00–11:00 10:00–11:00 0 11:00–12:00 11:00–12:00 11:00–12:00 12:00–13:00 12:00–13:00 12:00–13:00 Drinking water pump load pro le 13:00–14:00 13:00–14:00 13:00–14:00 Time of day Time of day 14:00–15:00 14:00–15:00 14:00–15:00 15:00–16:00 15:00–16:00 15:00–16:00 Elementary school (Yr 1-8) load pro le 16:00–17:00 16:00–17:00 16:00–17:00 17:00–18:00 17:00–18:00 17:00–18:00 FIGURE 2.26 • Indicative load profiles for various customer segments in potential mini grid locations 18:00–19:00 18:00–19:00 18:00–19:00 MINI GRIDS FOR HALF A BILLION PEOPLE   19:00–20:00 19:00–20:00 19:00–20:00 20:00–21:00 20:00–21:00 20:00–21:00 21:00–22:00 21:00–22:00 21:00–22:00 22:00–23:00 22:00–23:00 22:00–23:00  105 23:00–0:00 23:00–0:00 23:00–0:00 FIGURE 2.28 • Sample load profile for a village 120 100 80 Power (kW) 60 40 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour of the day Source: ESMAP analysis. kW = kilowatt. The surveys capture crucial information on electricity con- sumption, including: During the survey collection and load model- • Community willingness and ability to pay ing phase, survey teams should be sent only • Current expenditure on electricity generation using pet- to those sites that have been prioritized as having rol or diesel generators great potential for a viable mini grid, because of the costs of conducting on-site surveys. Survey teams • Expenditures on candles, kerosene, and dry cell batteries typically collect data that can inform the estimation • Other data pertinent to demand estimates of potential demand from households, commercial We recommend that, with smartphones and tablets so and productive loads, and public and community ubiquitous, personal interviews be conducted with com- institutions such as churches and hospitals. puter assistance. This practice ensures that data are col- lected with optimal efficiency and quality, minimizing the time and effort spent on data cleaning. user, recording the wattage and count of lightbulbs and Survey data, along with data from the literature, should fans at each location. The data were cleaned to adjust for inform the development of load profiles for different cus- outliers and to remedy survey errors. Equipment-use pro- tomer segments. The load profile is a view of estimated files emerged from applying utilization factors to each hour electrical demand at a given hour or day. This information of the day (varying by “high” and “low” months), and each will be vital to designing the mini grid. Figure 2.28 presents piece of large equipment was assigned a unique use profile a sample load profile. by state and community size (small communities and large communities use commercial/productive equipment dif- Load profiles for the NEP were developed by aggregating the ferently), and customer-type load profiles were generated expected demand from households, commercial and pro- by combining the equipment profiles with the survey data. ductive loads, and public infrastructure, as described below. It was assumed that all existing commercial lights would be Residential sector loads. Residential load profiles were replaced by 18W LED bulbs and residential lights by 6 W. All developed with Rural Electrification Agency (REA) classi- geotagged commercial and productive loads were included fications for small, medium, and large households. These in each site’s load profile. were also based on assumptions about the number of Public sector loads. The locations of public institutions appliances per household type and the use profiles for such as schools, health facilities, and religious centers each appliance (total watts and utilization factors for each in each community were determined from a geospatial hour of the day, considering seasonal variations). database of such points of interest16 and the REA geotag Commercial and other productive loads. In the geo- survey. Generic load profiles were written up for each type tag survey, surveyors assigned commercial and produc- of public institution, based on consultations with a devel- tive loads to 13 different business categories. Then they oper experienced with sizing and installing photovoltaic recorded the count and wattage of the high-load equip- systems for public institutions,17 and applied to the public ment associated with each commercial and productive end loads at each site. 106   MINI GRIDS FOR HALF A BILLION PEOPLE Part 3. System design and optimization tive function to design the generation as well as network assets for rural electrification. REM can help project Once the power load is estimated, then follows the system developers and investors with initial technical design and design and/or optimization. There are multiple options cost estimations, including both capital and operational for system optimization. Based on the load profiles devel- expenditures. The modeling tool seeks to aid developers oped for each community, modeling tools, each running on in making viable decisions regarding mini grid design by unique algorithms, calculate optimal solutions—least-cost providing the analytics needed to conduct technical and mini grid system designs that meet predefined parame- financial evaluations. REM offers a single package capable ters—for each community. of computing generation investment, operational perfor- For example, for mini grids to be included in the tender mance, and detailed design of the network starting from for the Nigerian NEP, a renewable fraction threshold of 60 the building level. percent was prescribed (that is, the mini grid design would Renewable Energy Integration and Optimization (REopt). need to produce at least 60 percent of its annual energy NREL has developed a tool called REopt™, which it uses output from renewable sources). The NEP also has min- to provide decision support, analyzing and optimizing imum technical requirements specific to several distinct mini grid designs for different systems.20 REopt is a criti- system architectures. Depending on the system archi- cal tool for understanding the technoeconomic trade-offs tecture selected for a particular community, the mini grid in the mini grid sector, which can lead to more sustainable would need to meet the minimum technical requirements business models and promote universal energy access. for that particular system architecture. These technical Formulated as a mixed-integer linear program, it solves a requirements cover both quality and sizing of components deterministic optimization problem to establish the opti- as well as quality of service, such as the number of hours mal selection, sizing, and dispatch strategy of technologies the mini grid can be offline for scheduled or unscheduled chosen from a candidate pool, such that electrical, thermal, maintenance. These design constraints are incorporated and or water loads are met at every step in the minimum into the optimization exercise. life-cycle cost. REopt is a time series model that looks at OPTIMIZATION MODELS a full-year energy balance to determine multiyear cash flows by applying appropriate discount and cost-escalation For the NEP in Nigeria, three optimization models (HOMER, rates. As opposed to algorithmic dispatch strategies, REopt REM, and REopt) were used to propose optimal mini grid finds the global optimum by anticipating load and resource designs for each community, given the load profile of that changes over the full analysis period. In the mini grid con- community and some design constraints, such as those text, this allows REopt to dispatch batteries to maximize described above. Each model takes a different approach renewable energy utilization and minimize generator run to mini grid optimization. All three models provide gener- time, maximizing economic efficiency. ation system sizing, but REM also generates a customized distribution design for each mini grid. Each of these three OTHER APPROACHES optimization models is briefly described below. Additional Other approaches, such as those described below, can also details are available at the companion website to this book: be implemented in the manner just presented. www.esmap.org/mini_grids_for_half_a_billion_people. Village Infrastructure Angels (VIA) employs another meth- Hybrid Optimization Model for Multiple Energy Resources odology to propose least-cost distribution network designs (HOMER). Originally developed at the U.S. National Renew- for mini grids, which excludes households that might be able Energy Laboratory (NREL), and enhanced and dis- better candidates for SHSs. VIA has designed electrical tributed by HOMER® Energy, HOMER software nests three power distribution networks for mini grids in the Philip- powerful tools—simulation, optimization, and sensitivity pines and Haiti (Craine, S) using a minimum spanning tree analysis—in one software product, so that engineering algorithm, which solves for the shortest network of lines to and economics work side by side.18 HOMER is the industry connect a given set of points. VIA conducts load-flow analy- standard for optimizing mini grid design in all sectors, from ses, assuming an average load per building and a maximum village power and island utilities to grid-connected cam- 10 percent voltage drop between the power source and the puses and military bases. end user, to determine what size wire (conductor) is nec- The Reference Electrification Model (REM). Developed essary to distribute power to households, the distribution by the MIT-Comillas Universal Energy Access Labora- losses involved, and the cost. A Critical Distance analysis, tory, REM is a computational modeling tool designed to based on the prevailing cost of SHSs, suggests the length help plan detailed medium- and low-voltage distribution of the mini grid or grid extension line that is viable before networks, with an implementation focus on developing a SHS makes much more sense. This guides which lines countries.19 REM uses cost minimization as the objec- of the distribution network should be kept or discarded in MINI GRIDS FOR HALF A BILLION PEOPLE    107 favor of SHSs. For the mini grids that remain, VIA lists wire special consideration because they require higher peak sizes, poles, and any transformers (if required) and pub- power than a standard user. The power asset systems will lishes them on a GIS platform. This gives aspiring develop- then be sized according to HOMER demand analysis. The ers, perhaps considering a project at the site, a useful bill scope of the study employs a standardized simulation to of materials. scale up the modeling to more sites. Integration combines manual designs and HOMER opti- Village Data Analytics (VIDA) conducts a hyperlocal den- mization. Based upon the analysis of building footprints sity analysis to identify village cores, outskirts, and outly- and satellite imagery, a grid design is created remotely, and ing areas. Within the core area, VIDA’s algorithm identifies potential locations for the power asset suggested. The grid high-value mini grid customers. using building category, design can be made to assume that all or a portion of the the location of buildings with respect to roads, and the den- buildings are connected within a certain distance from the sity of the built-up area (figure 2.30a). Special importance central village cluster. From this data, a voltage drop model is given to anchor/institutional loads. Once demand is iden- is generated to approximate the cable sizes required (fig- tified for the different customer segments, an algorithm ure 2.29). Specific load centers or anchor loads receive then generates a distribution layout that connects the high-value mini grid customers. First, a minimum spanning tree is generated to get an upper limit for the connection FIGURE 2.29 • Output of the voltage drop model density. Next, the trunk lines are generated (figure 2.30b). These are either single or three-phase and typically follow the main roads through the village. The poles are located There are multiple options for optimizing mini grid system design at each site. Based on the load profiles developed for each community, modeling tools, each running on unique algorithms, calculate the least-cost mini grid system designs that meet predefined community parameters. The most common system optimization tools are those from HOMER, MIT’s REM, and NREL’s REopt. Source: Integration 2021. FIGURE 2.30 • Sample outputs of hyperlocal density analysis from Village Data Analytics  igh-value mini grid customers a. H  ini grid distribution layout with trunk line, poles, and b. M (colored buildings) dropdown lines connecting high-value mini grid customers Source: VIDA 2021. 108   MINI GRIDS FOR HALF A BILLION PEOPLE at equal distances and then dropdown lines are generated would provide a first indication of sites where affordable so they connect every high-value mini grid customer to the tariffs can be expected. nearest poles. The bill of materials for distribution is then • Viability-gap calculations. Here, reverse logic can be estimated, which leads to estimations of distribution cost, applied whereby an agreed tariff is used as input to including the length of the predicted trunk line, the number determine the viability gap of the projects and grant of poles, and the length of the dropdown lines. The algo- share needed in CAPEX to reach the required financial rithm uses average cost per meter of wiring (of different returns. This differentiates sites according to grant lev- sizes) and the average cost of poles to generate this infor- els needed for sustainable mini grid operation. mation on a village-by-village basis. Part 5. Result visualization and dissemination via an Part 4. Sensitivity analysis and financial modeling online platform More scenarios are generated during the final part of the workflow, and these assess how sensitive the optimal solu- Geospatially planned portfolios of mini grids contain a tion is to input parameters and financial assessments. They wealth of information for developers. When produced as might also assess sensitivity to demand levels, quality of ser- part of a mini grid program in partnership with a govern- vice, level of renewable generation, to name a few. The sensi- ment or a developer, these portfolios should be shared with tivity analysis goes hand in hand with the financial modeling, mini grid developers, presenting relevant information in an which can be developed for site-specific characteristics accurate and transparent way. allowing for outputs of tariff, capital expenditure (CAPEX), One example is the VIDA software, a tool for site identifi- viability-gap analysis, and other relevant outputs in view cation, selection, prioritization, visualization, and collabo- of project needs (figure 2.31). Besides the usual financial ration (figure 2.32). The software visualizes the mini grid model parameters, some context-specific aspects include: portfolio analysis on an interactive platform that offers both • Scenarios with and without PUE. The definition of project a high-level overview of the modeling exercise (national and CAPEX could omit investments in productive activities regional levels) but also granular descriptions of all candi- and run the model under both scenarios. Such an anal- date sites. VIDA provides access to mini grid viability indi- ysis could, for example, identify viable sites (1) without cators and distribution layout characteristics. Users can depending on productive loads, (2) only if a minimum download results, upload data, share information, and col- basic productive load is ensured, (3) only if extensive laborate on the software platform. investment is undertaken to promote large-scale pro- Another example of such a dissemination tool is Odyssey ductive loads. Energy Solutions, which provides a web-based data plat- • Tariff-based calculations. In this mode the financial form that facilitates deployment of mini grids in emerging model can use predetermined CAPEX structure (equity, markets for developers, financiers, vendors, and govern- debt, and grant) and operating expenditure as inputs ments and donors. It developed a customized version of its to drive the tariff needed to meet the required finan- online platform for Nigeria’s REA to manage data analysis cial returns. Furthermore, time-of-use tariffs and tar- and the bidding process for the sites included in the mini iffs based on customer type can also be modeled. This grid tender of the NEP. FIGURE 2.31 • Indicative flow of financial modeling process of mini grids • Finacial paramaters (discount • Excel based financial model • Income statement rate, interest rate) • Cashflow statement • Income statement • Financing structure • Balance sheet • Cashflow statement • Generation system assets • Balance sheet • Key performance indicators • Distribution system assets – LCOE • Key performance indicators • Consumer assets ? – Profit margin • Depreciations – NPV • Energy sales per category – IRR • Tariff structure – Payback Source: Integration 2021. MINI GRIDS FOR HALF A BILLION PEOPLE    109 FIGURE 2.32 • VIDA interactive platform Source: VIDA 2021. The user interface organizes data into modules that con- tain the technical and financial dimensions of each mini Geospatially planned mini grid portfolios grid site: location, load forecast, generation system and offer a wealth of information for developers. distribution design, costing, tariffs, and financial model. When produced as part of a government or devel- The online platform has detailed, site-specific informa- opment partner mini grid program, this information tion on the customers and loads as well as suggested should present the relevant information to devel- system sizing. opers in an accurate and transparent fashion. The Once officially registered, program bidders gain access Village Data Analytics platform and Odyssey both to information posted at the tender sites, where they offer excellent ways to visualize and disseminate can register to participate in the tender. Registered bid- the modeled results. ders may access key program information within the 110   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 2.33 • Mini grid tender preparation in Odyssey Source: Screenshot of Nigeria Electrification Project portal on https://www.odysseyenergysolutions.com/. NEP = Nigeria Electrification Project. platform, including all tender documents, deadlines, and instructions for submitting a bid. They may also view a list Geospatial analysis provides a broader pic- of all sites and run analytics within the software to view ture of the locations and characteristics of aggregate statistics about the sites (grouped by state, communities that can be considered at a for example). Bidders can copy all site data into their own portfolio level, which enables mini grid developers accounts in the platform, allowing them to use the soft- to exploit economies of scale and prepare quicker, ware to assemble a technical and financial proposal for more cost-effective rollout plans and plans for ser- each lot using Odyssey’s software tool suite and struc- vice and maintenance. Governments can also use tured workflow (figure 2.33). geospatial and other digital tools to catalyze deploy- Once the technical analysis for each site is complete, sites ment of mini grids led by the private sector to sup- are bundled into portfolios for each of the tender lots. Bid- ply electricity to off-grid communities. For example, ders must provide the business plan, financing approach, Nigeria’s Rural Electrification Agency is holding and required documents uploaded to the data room. minimum-subsidy tenders for portfolios of promis- Incomplete proposals cannot be submitted; a final check- ing mini grid sites that it has identified and for which list ensures all sections are complete and all required files it has collected market intelligence. uploaded to the data room. Only final portfolios for the tender lot are submitted to the evaluation committee. The VIDA and Odyssey platforms are compatible and can form a strong toolkit for governments, REA, development sey could navigate to VIDA to view and access granular financial institutions, or private organizations. VIDA’s soft- geospatial and on-ground data of the villages being ten- ware can host village-level data, geospatial, and others in dered. Similarly, a user in VIDA software could push the an interactive user interface that can then be accessed by village-level data to the Odyssey platform to tender and Odyssey for tendering and deployment. A user in Odys- deploy mini grids. MINI GRIDS FOR HALF A BILLION PEOPLE    111 LESSONS LEARNED AND NEXT STEPS REFERENCES The introduction of geospatial and other digital tech- Arderne, C., C. Zorn, C. Nicolas, and E. E. Koks. 2020. “Predictive nologies has lowered preparation and planning costs by Mapping of the Global Power System Using Open Data.” Scientific an order of magnitude: from about $30,000 per site— Data  7: Article  19. https://doi.org/10.1038/s41597-019-0347-4. because each site required high-level on-site analysis—to https://www.nature.com/articles/s41597-019-0347-4 around $2,300 per site, based on the World Bank’s recent Castalia. 2014. Myanmar National Electrification Program (NEP) Road- experience in Nigeria. Furthermore, by 2024, high-resolu- map and Investment Prospectus. Washington, DC: Castalia. https:/ / www.seforall.org/sites/default/files/Myanmar_IP_EN_Released. tion satellite imagery is expected to fall by nearly 60 per- pdf. cent from 2014 levels (Selding 2015). This too will drive CIESEN (Center for International Earth Science Information Net- down the cost of ever more accurate geospatially planned work), Columbia University and Novel-T. 2020. “GRID3 Central portfolios. Meanwhile, taking a portfolio approach to mini African Republic Settlement Extents Version 01, Alpha. Palisades, grid development, instead of building mini grids as one- NY: Geo-Referenced Infrastructure and Demographic Data for off projects, can slash upfront capital costs by around Development (GRID3). Source of building Footprints ‘Ecopia Vector Maps Powered by Maxar Satellite Imagery’.” Accessed June 1, 2022. $100/kW, according to analysis of the ESMAP’s data- https://doi.org/10.7916/d8-y2ax-p859. base of installed and planned mini grids presented in the Craine, S. 2019. Haiti Geospatial Exercise for the Global Facility on Mini overview to this handbook. In addition, geospatial analy- Grids of the Energy Sector Management Assistance Program: Incep- sis can help identify potential productive-use customers, tion Report. Washington, DC: World Bank. Unpublished. thereby shaping developers’ community engagement DeCarolis, Joseph, Hannah Daly, Paul Dodds, Ilkka Keppo, Francis Li, strategies to promote income-generating uses of mini Will McDowall, Steve Pye, Neil Strachan, Evelina Trutnevyte, Will grid electricity. Usher, Matthew Winning, Sonia Yeh, and Marianne Zeyringer. 2017. “Formalizing Best Practice for Energy System Optimization Model- As developers achieve economies of scale by developing ling.” Applied Energy 194 (May): 184–98. https://www.sciencedirect. economically viable mini grid portfolios that support pro- com/science/article/pii/S0306261917302192. ductive uses of electricity, and as mini grid component Ecopia AI and Maxar Technologies. 2021. “Digitize Africa.” https:// costs plummet over the next decade, as discussed in chap- blog.maxar.com/earth-intelligence/2018/gis-ready-building-foot- ter 1, the cost of mini grid electricity is on pace to reach print-shapefiles-for-accelerated-analysis. Internal resource. $0.20/kWh by 2030. As mini grid electricity approaches Hahsler, Michael, Matthew Piekenbrock, and Derek Doran. 2019. “dbscan: Fast Density-Based Clustering with R.” Journal of Statistical this cost threshold, mini grids become the least-cost option Software 91 (1): 1–30. DOI: 10.18637/jss.v091.i01. for more and more people. This means national least-cost HOMER Energy. N.d. “Hybrid Optimization Model for Multiple Energy electrification plans will need to weigh expected cost Resources (HOMER).” https://www.homerenergy.com/. declines in mini grid electricity as they anticipate main grid Howells, Mark, Jairo Quiros-Tortos, Robbie Morrison, Holder Rogner, expansions, more mini grids, and SHSs. Taco Niet, Luca Petrafulo, Will Usher, William Blyth, Guico Godinez, Luis F. Victor, Jam Angulo, Franziska Bock, Eunice Ramos, Fran- To catalyze deployment of private-sector-led mini grids to cesco Gardumi, Ludwig Hulk, Patrick Van-Hove, Estathios Peteves, supply off-grid communities, Nigeria’s REA, the implement- Felipe de Leon, Andrea Meza, Thomas Alfstad, Constantinos Talio- ing agency for this project, is holding minimum subsidy tis, George Partasides, Nicolina Lindblad, Benjamin Stewart, Ash- tenders21 for portfolios of promising mini grid sites it has ish Shrestha, Dana Rysankova, Adrien Vogt-Schilb, Chris Bataille, identified. The World Bank and the REA have developed an Henri Waisman, Asami Miketa, Pablo Carvajal, Daniel Russo, Morgan Bazilian, Andrii Gritsevskyi, Mario Tot, and Adrian Tompkins. 2021. innovative protocol for mini grid site identification, screen- “Energy System Analytics and Good Governance—U4RIA Goals of ing, and analysis using geospatial tools, including a geospa- Energy Modelling for Policy Support.” Unpublished preprint version. tial portfolio planning methodology to assess and select https://www.researchsquare.com/article/rs-311311/v1. the communities to be included in the minimum subsidy IEG (Independent Evaluation Group). 2015. World Bank Group Sup- tenders. The protocol enables governments, development port to Electricity Access, FY 2000–14. Washington, DC: World Bank. partners, or other public institutions to prepare portfolios https://ieg.worldbankgroup.org/sites/default/files/Data/reports/ Electricity_Access.pdf. of mini grid projects and “crowd in” private sector cofinanc- ing. We hope this may offer useful guidance to those seek- Integration. 2021. “Pilot Location Modelling Results”, (Internal report). ing to develop mini grid projects at scale elsewhere. For Integration and Reiner Lemoine Institut. 2016. “Preliminary Modelling of Off-grid PV Capacities for the Whole of Nigeria.” Nigerian Energy Sup- example, governments could replicate these steps in other port Program and the Federal Ministry of Power, (Internal report). countries interested in competitive tenders to kickstart or Khavari, B., A. Korkovelos, A. Sahlberg, M. Howells, and F. F. Nerini. scale up the market for mini grids. 2021. “Population Cluster Data to Assess the Urban-Rural Split and Electrification in Sub-Saharan Africa.” Scientific Data 8: Article 117. https://www.nature.com/articles/s41597-021-00897-9. 112   MINI GRIDS FOR HALF A BILLION PEOPLE Moulavi, Davoud, Pablo A. Jaskowiak, Ricardo Campello, Arthur Zimek, 5. This material is organized and shared through the GEP’s GitHub and Joerg Sander. 2014. “Density-Based Clustering Validation.” workspace. https://epubs.siam.org/doi/pdf/10.1137/1.9781611973440.96. 6. The development of open access training material has been under- NRECA International. 2019. “National Electrification Investment Plan. taken by the Climate Compatible Growth (CCG) program (https:/ / USAID Ethiopia—Distribution Systems Strengthening Project” , climatecompatiblegrowth.com/). (Internal report). 7. In the case of Nigeria, this task was carried out by Integration and Odyssey Energy Solutions. N.d. “Finance, Build and Operate Distrib- the Reiner Lemoine Institut. A similar approach has been employed uted Infrastructure at Scale.” https://www.odysseyenergysolutions. at the Global Electrification Platform (GEP) following a methodol- com/. ogy suggested by Khavari and others (2021) (https:/ /www.nature. Selding, Peter. 2015. “Established Imagery Providers Face Changing com/articles/s41597-021-00897-9). Competitive Landscape.” SpaceNews, September 24. https:/ /space- 8. Energydata.info is an open data platform providing access to data news.com/established-imagery-providers-face-changing-competi- sets and data analytics that are relevant to the energy sector, avail- tive-landscape/. able at https://energydata.info/. VIA (Village Infrastructure Angels). 2021. Geospatial Mapping of Least 9. HOTOSM is an international team dedicated to humanitarian action Cost Mini Grid Potential in Ethiopia. Project Completion Report for and community development through open mapping; the database the Energy Sector Management Assistance Program (ESMAP)/ is open and available at https://www.hotosm.org/tools-and-data. World Bank. Internal report. 10. The Food and Agriculture Organization of the United Nations (FAO) VIDA (Village Data Analytics). 2021. Geospatial Mapping of Least Cost Agro-Ecological Zones (AEZ) database is available at http://www. Mini Grid Potential in Ethiopia. Project Completion Report for the fao.org/nr/gaez/en/#. Energy Sector Management Assistance Program (ESMAP)/World 11. The International Food Policy Research Institute’s HarvestChoice Bank, (Internal report). products are available at https://dataverse.harvard.edu/dataverse/ World Bank. 2007. “Geospatial Plan for Kenya (National Electrification harvestchoice. Coverage Planning). Investment Costing Estimation Model.” Unpub- 12. The Gridfinder.org visualization can be found at https:/ /gridfinder. lished paper, World Bank, Washington, DC. org/; the modeled results for all countries are available at Zenodo World Bank. 2009. Project Appraisal Document on a Proposed Credit (2020). https://zenodo.org/record/3628142#.YIgO85BKhPY. in the Amount of SDR 45.1 Million (US$70 Million Equivalent) to the In the case of Nigeria, the night lights’ data set was retrieved from the 13. Republic of Rwanda for a Rwanda Electricity Access Scale-Up and NOAA Earth Observation Group (https:/ /ngdc.noaa.gov/eog/). The Sector-Wide Approach (SWAP) Development Project. Washington, World Bank’s Light Every Night initiative provides open access to DC: World Bank. all nightly imagery and data from the Visible Infrared Imaging Radi- World Bank. 2015. Achieving Universal Access in the Kaduna Electric ometer Suite Day-Night Band (VIIRS DNB) from 2012 to 2020 and Service Area. Energy Sector Management Assistance Program. the Defense Meteorological Satellite Program Operational Linescan Washington, DC: World Bank. https:/ /documents1.worldbank.org/ System (DMSP-OLS) from 1992 to 2013. You may find more infor- curated/en/782491487138851242/pdf/112802-Report-Kadu- mation at https:/ /registry.opendata.aws/wb-light-every-night/. na-Electric.pdf. More information about the source is available at http:/ 14. /www.key- Zenodo. 2020. “Data from: Predictive Mapping of the Global Power Sys- biodiversityareas.org/assets/8f1535aed3316ae2b720364019f- tem Using Open Data.” Zenodo, January 16, 2020. https:/ /zenodo. 8cb1c. org/record/3628142#.YIgO85BKhPY. 15. Tool available at https://ibat-alliance.org/. 16. The NGO ehealth Nigeria gave the authors access to this informa- tion from its database. NOTES Em-One designed and installed solar solutions in public institutions 17. such as schools and health centers in Lagos, Kaduna state, and 1. The electricity distribution grid has not been mapped or the data northeastern Nigeria. are not available in many countries. Gridfinder, an open-source tool 18. HOMER® Energy’s software suite consists of two desktop prod- for predicting the location of electricity network lines, using night- ucts—HOMER Pro and HOMER Grid—and application program- time lights satellite imagery and OpenStreetMap data, was used ming interfaces for building web-based tools, such as HOMER to determine the location of the grid for countries where the data QuickStart and QuickGrid. weren’t available. See https:/ /gridfinder.org/ and https:/ /github. com/carderne/gridfinder. 19. More information about MIT’s REM model and application is avail- able at http://universalaccess.mit.edu/#/main. 2. The GEP processes HRSL population data to transform them into population clusters (similar to GRID3) based on proximity and den- 20. More information about NREL’s REopt model and application is sity. You may read more about this approach at https://www.nature. available at https://reopt.nrel.gov/. com/articles/s41597-021-00897-9 21. REA has grouped these potential mini grid sites into lots by state 3. Firm power output of a mini grid is defined in chapter 1 as the gen- and will invite private developers to build, own, and operate these erator capacity (kW) plus 25 percent of the solar array output rated portfolios of mini grid projects. Through a competitive process, peak (DC) power output (kWp). REA will award grants per connections to the private developers selected to implement these projects. Their bids will be evaluated 4. Consortium members include KTH Royal Institute of Technology, based on the quality of their technical proposal and on their subsidy Development Seed & Derilinx, World Resources Institute (WRI), and requirement. Cambridge University. MINI GRIDS FOR HALF A BILLION PEOPLE    113 CHAPTER 3 PRODUCTIVE LIVELIHOODS AND BUSINESS VIABILITY CHAPTER OVERVIEW This chapter highlights why productive uses of electricity can be a game changer for both mini grid developers and socioeconomic development. It presents an everyone-wins scenario for developers, local entrepreneurs, commu- nities, and national utilities. Using real-world examples, the chapter outlines a six-step approach to implementing productive-use interventions and discusses who can organize such interventions. THE MULTIPLE BENEFITS OF dred times more than the energy consumed by a nearby mini grid customer. Grain mills have power ratings in the CONNECTING INCOME-GENERATING 1,000–10,000 watt (W) range, about ten times more than MACHINES AND APPLIANCES TO commercial appliances like refrigerators and a hundred MINI GRIDS times more than household appliances (CrossBoundary 2020). Figure 3.1 shows the load profiles for mini grids Increasing productive uses1 of mini grid electricity cre- operating with 22 percent and 40 percent load factors. ates an “everyone-wins” scenario for mini grid developers, Greater demand for electricity generates additional reve- rural entrepreneurs, communities, and national utilities nues for mini grid operators while improving the utilization over time. It reduces the levelized cost of energy, which capacity of their systems, which reduces the unit cost of increases the mini grid developer’s margins and therefore electricity (per kilowatt-hour) and ensures efficient use of financial viability. Entrepreneurs and small businesses ben- the mini grid’s assets. Although increased demand requires efit from switching from expensive diesel generators to more capital investment, it can optimize the use of sys- affordable mini grid electricity. Communities benefit from tems, especially during daytime, when the residential load the new jobs that mini grids create and the increased eco- is small and systems are underused. Daytime use is critical nomic activity. The growth of rural economies also benefits for solar-based mini grids, which produce electricity at min- national utilities once interconnection to the main grid is imal marginal cost during the day. considered, because it increases customers’ demand for Furthermore, an Energy Sector Management Assistance high-quality electricity and their ability to pay for it. Program (ESMAP) analysis of 1,028 mini grids in Cambo- Boosting productive uses benefits mini grids and their dia, Myanmar, and Nepal indicates that every additional 1 operational efficiency and financial viability (see also percent of nonhousehold customers (for example, micro- chapter 1). When income-generating machines and appli- entrepreneurs and small businesses) served by a mini grid ances boost demand for mini grid electricity, a mini grid’s adds 20 percent to the mini grid’s total average monthly load factor gets a corresponding boost too. Meanwhile, a consumption in terms of kilowatt-hours sold. So, if a mini higher load factor (jumping from 22 percent to 40 percent) grid serves 1,000 connections, going from 10 nonhouse- cuts the cost of electricity by 27 percent (see chapter 1). hold customers (1 percent of total customers) to 50 non- On a site analyzed by CrossBoundary Lab, a grain mill oper- household customers (5 percent of total customers) would ator consumes 300 kilowatt-hours (kWh)/month, a hun- increase the mini grid’s average monthly consumption 114   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 3.1 • The impact of productive electricity uses on the daily load profile and levelized cost of energy 100 LCOE = $0.23/kWh 80 Percentage of peak load LCOE = $0.28/kWh 60 40 20 LCOE = $0.38/kWh 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 22% load factor 40% load factor 80% load factor Source: ESMAP analysis. kWh = kilowatt-hour; LCOE = levelized cost of energy; PUE = productive use of electricity. by 80 percent (5 percent minus 1 percent, times 20). This In addition, more than 130 income-generating appliances additionality seems to be fairly consistent until 15 percent have a payback period of less than 12 months, according of a mini grid’s customers are nonhouseholds, after which to an analysis that ESMAP conducted in Ethiopia. The point, the effect begins to dissipate. It is worth noting here up-front investment costs, power consumption, pay- that most of the mini grids (about 60 percent) surveyed had back periods, and revenue potential for some of these a customer base that consisted of 1–5 percent nonhouse- income-generating machines and other appliances are hold customers, indicating substantial room to increase the presented in table 3.4. Up-front investment costs typ- share of nonhousehold customers in the mini grids’ cus- ically ranged from $500 to $1,500, with an average of tomer base. about $1,200. These appliances and machines generate between $50 and $500 of revenue per month after the The combination of savings and reliability brought by conclusion of the payback period, with an average of the mini grid makes business sense for local entrepre- $300 per month across the range of appliances identified. neurs. Reliable electricity will reduce costs for businesses as power outages and unreliable supply detract from reve- Finally, mini grid electricity enables entrepreneurs to earn nues. In Sub-Saharan Africa, outages inflict sales losses of additional profits by extending the shelf life of goods, mak- 30 percent on businesses. In some of the region’s largest ing productive processes more efficient, increasing output, economies—Angola, Ghana, and Nigeria—more than 25 and improving access to information and markets (IEG percent of businesses lose more than 10 percent in sales 2008). In Tanzania’s Ludewa District, access to the 300 because of power outages, with individual firms reporting kilowatt (kW) hydro mini grid operated by Lumama cut losing more than 70 percent. The firms with the greatest milling costs in half (USAID 2018). challenges average more than 200 hours a month without Stimulating demand for electricity from productive activities power, while even the companies getting highly reliable ser- can, in particular, assist women-run enterprises to boost vice still report more than 10 hours a month without elec- their earnings through the use of lighting, electrical equip- tricity (Ramachandran, Shah, and Moss 2018). ment for cottage industries, baking, ceramics, and so on. MINI GRIDS FOR HALF A BILLION PEOPLE    115 TABLE 3.1 • The Mwenga hydro mini grid: Estimated costs and benefits Productive uses of electricity can assist women in notable ways through their higher Present value of cost or benefit earnings achieved through better lighting and appli- Type of benefit (US$, millions) ances for cottage industries, baking, ceramics, and so on. Development subsidies received by project –7.1 Household cost savings a 6.4 Tea company savings from reduced 1.4 diesel backupb Households and communities also benefit from produc- Jobs created by electrifying villagesc 8.6 tive uses of energy, which bring socioeconomic gains in Economic net present value 9.3 addition to better opportunities to sell goods and services. Source: Banerjee and others 2017. If these opportunities existed prior to the mini grid, they a. Monthly savings of $14 for 5,600 households. were constrained by the need to rely on expensive diesel b. Diesel backup requirement of 10 percent of total power consumption. generation. Food and farm-related goods and services— Assuming that 65 percent of businesses create 1.5 jobs each, and that c.  for example, cooling, drying, processing, and so on—are each job created is valued at the average expected annual salary of two affected areas. Communications and connectivity are $1,500 a year. another (internet points), along with mechanical power Note: The estimated peak load averages about 700 kilowatts (kW) with a summer peak of 90 kW and a winter peak of 400 kW, and annual power (woodworking and metalworking machines), as well as consumption of 2,880 megawatt-hours. lighting and entertainment. These goods and services lead to newly available and improved outputs, including: backup, and job creation from new electrified businesses— • Household and community well-being; are estimated at about $9 million (table 3.1) (Banerjee and • Longer life of, and added value to, agricultural products; others 2017). • Higher productivity; Increasing the productive use of energy (PUE) also has important benefits for women in the communities served • Better-quality manufactured goods, particularly in car- by mini grids. The physical and time burdens of some pro- pentry, upholstery and tailoring, and metalworking; ductive activities mainly run by women can be alleviated by • Reduced costs; and ensuring that power goes to shared community facilities • Service availability after dark. such as mills. Women’s labor is dominated by the drudg- ery of preparing grain for household consumption (called These outputs in turn lead to more jobs, higher incomes, “agro-processing” in the literature), particularly in Sub- time saved, and improved well-being (GIZ and others 2013). Saharan Africa. For a family to eat over four to five days, a The developer Mlinda has installed almost 100 mini grids woman (and her daughters) will spend up to 13 hours to in India. Load analysis and grid design have enabled pound enough maize. Time-use estimates obtained for Mlinda to power single- and three-phase electrical devices Nigeria show that two to three hours are spent each day targeting productive and residential end users. The three- just to prepare grains for pounding—that is, threshing and phase loads go to income generation. To support pro- milling. Eighty-two woman-hours are spent processing one ductive uses within communities and identify business drum of oil palm fruits. It takes two hours to grate a basin opportunities, Mlinda set up a team to assist local busi- of cassava; a grating machine can process a basin in one ness development. An impact assessment of the first 24 minute (Kes and Swaminathan 2006). operating mini grids showed that microenterprise reve- But it takes more than the introduction of electricity to nues rose 28 percent; 115 new local jobs have been cre- boost enterprise generally and women-led businesses in ated (Mlinda 2021). particular. Outreach and capacity building are also needed. An economic impact assessment offers more evidence. A SolarAid study of the solar lighting market in East Africa in Tanzania’s Mwenga hydro mini grid, with a capacity of 4 2012–15 found that 38 percent of households interviewed megawatts, was commissioned in 2012 and operated by reinvested their energy savings into agricultural production Rift Valley Energy. Mufindi Tea Estates and Coffee Ltd. is the or to seed other small enterprises (ODI and others 2016). main client and anchor load, requiring electricity mainly to The World Bank’s work in Mali reveals some of the chal- process tea leaves and power large motors, fans, and sieves. lenges in applying a gender lens to foster productive uses Over a 20-year project life cycle, economic benefits—from of rural electrification. household energy cost savings, reduced reliance on diesel 116   MINI GRIDS FOR HALF A BILLION PEOPLE The benefits of increased productive uses of Increasing the productive uses of a mini energy are especially profound for women, grid’s electricity presents an opportunity for who spend a disproportionate amount of their time everyone to win. Mini grid developers can grow their on farm labor (agro-processing chores), particularly revenues and lower their costs. Local businesses and in Sub-Saharan Africa. Transitioning from manual entrepreneurs can transition from expensive on-site labor to machine-assisted processing can boost diesel generation to less costly mini grid electricity, or productivity and save many hours of drudgery each develop new businesses that use electricity services week. But evidence suggests that productivity and to generate revenue. Local communities benefit from time savings do not automatically follow from the the creation of new jobs and greater economic activ- introduction of electric machines and appliances. ity. National utilities benefit from the growth of rural Investments in outreach and capacity building are economies and demand for electricity once intercon- also needed. nection to the main grid is considered. Increased PUE in mini grids can also enhance the eco- nomic viability of expanding the main grid. In the absence The most effective interventions to foster of PUE, most mini grid customers in low-income areas use productive uses of energy acknowledge that little electricity. So the main grid would sustain ongoing men and women occupy different spheres in the financial losses when it reached the mini grid’s service area. productive economy. Men and women also benefit Mini grids that stimulate demand for electricity through from electricity in different ways. Equitable interven- income-generating appliances and machines flip this tions are designed to overcome the gender-based narrative. They can provide economic growth in rural and barriers around productive use. They need to deliver peri-urban areas when they are designed to connect with communitywide welfare improvements and grow the main grid and when PUE has been promoted through the productive customer base—helping electricity community engagement and training. By the time the grid suppliers become financially sustainable. arrives, a substantial load already exists and customers are better able to pay. ROLLING OUT PROGRAMS TO PROMOTE PRODUCTIVE USES returns, and profitability; and the setting of tariffs condu- cive to productive-use appliances. Finally, financial guar- AND STIMULATE DEMAND antees and up-front financing can mitigate default risks Electricity demand does not rise automatically with the (expanded from RMI [2018]). arrival of a mini grid. The barriers to demand are numer- In addition, interventions should acknowledge that men ous, among them limited markets, information, lack of and women occupy different spheres in the productive skills, up-front costs, inefficient appliances, and scant economy. Yet measures to increase productive uses of access to financing. But efforts to promote PUE will pay off. energy tend to be gender blind and assume that men In Indonesia, for example, local nongovernmental organi- and women benefit from electricity in the same way. For zations (NGOs) promoted productive use at the outset of instance, women are less likely to be employed than men, a rural electrification program, quadrupling annual elec- more likely to run informal enterprises from their homes, tricity use from the main grid (World Bank 2000). and are overrepresented in low-productivity businesses, The adoption of electricity-powered productive equipment while men tend to engage in mechanized, electricity-inten- depends on demand, competition, and other sources of sive sectors such as construction, welding, manufacturing, power—like diesel gensets and manual labor. What are and repair. So, acknowledging the gender-based barriers the up-front costs of equipment? The “Diffusion of Innova- to productive use must be part of equitable interventions tions” theory (Rogers 2003) highlights the importance of that deliver communitywide welfare improvements and peer-to-peer conversations and peer networks to stimu- expand the productive customer base. Ultimately these late productive use. These include end-user education and will improve the electricity suppliers’ financial sustainabil- vocational training; manufacturers’ road shows for poten- ity. Community-based institutions such as self-help, sav- tial entrepreneurs; more end-user engagement to raise ings, and farmers associations can become platforms for awareness about what has worked, expected investment productive-use discussions. MINI GRIDS FOR HALF A BILLION PEOPLE    117 TABLE 3.2 • Six steps to roll out a PUE program Description Outputs Step 1 Market/demand assessment with geospatial analysis Online data platform overlying mini grids, appliances, and end-use finance List of key stakeholders List of high-impact opportunity areas Step 2 Community engagement confirming and improving List of communities validated as areas of high-impact opportunity, data from Step 1 through survey(s) and workshops combined with community-specific market data List of appliances that are relevant for these communities based on local contexts List of potential PUE customers in these communities List of local providers of microfinance in or near these communities List of local suppliers of appliances that serve these communities List of community leaders and district-level government officials who are supportive of the PUE program Step 3 Demand analysis for mini grid design and market Detailed characteristics of an initial set of community-relevant potential for appliances and associated end-user appliances finance List of prioritized appliances Step 4 Preparation of road shows involving local government, Road show logistics finalized: who, what, where, when, and how community leaders, interested appliance providers Information and marketing campaign launched ahead of the road and end-user financiers, mini grid companies shows Step 5 Road shows to load centers explaining the value Road shows propositions to potential end users by mini grid devel- Customer sign-up for mini grid connections, appliances, and end- opers, appliance suppliers, and end-user financiers user finance based on current and aspiration lifestyles of the end users; document sign-ups by end users for mini grid connections, appliances, and end-user finance Step 6 Rollout of mini grid connections, sales of appliances, Customers connected and end-user finance Appliances sold and connected Financing secured PUE = productive use of energy. This next section presents six steps to roll out initiatives • Building footprints, supporting the uptake of income-generating appliances in • Road and electricity networks, towns served by mini grids. • Nightlight imagery, STEP 1 • Crop cover and yields, Assessing markets and demand: Geospatial • Socioeconomic and demographic data (like gender analysis superimposed over mini grids, appliances, composition and household income), and finance for end users • Cell-phone coverage, A good rollout can accelerate uptake of productive uses • Topography, and of energy, especially when developers understand the sector’s market potential and engage the stakeholders. In • Precipitation. 2022, leveraging geospatial analysis is the state-of-the-art Then, to identify market opportunities, the machine- way to assess the productive-use market and its demand learning algorithms highlight possible high-impact collab- potential. This data informs collaborations among mini grid orations for mini grid developers, appliance suppliers, and developers, appliance suppliers, and end-user financiers. local finance providers. Exemplifying these high-impact opportunities are town clusters that lack electricity but Machine-learning algorithms now analyze geospatial data are located in highly productive agricultural regions with sets and produce maps that highlight areas suitable for PUE good cell-phone coverage and good roads to major cities programs. This technology-enabled, data-driven approach or trade hubs. But high-impact areas are everywhere: in to market assessment first collects a range of geotagged arid regions, coastal settlements, agricultural towns, and in data, including: rural and peri-urban locales. 118   MINI GRIDS FOR HALF A BILLION PEOPLE Once uploaded to an online platform, this geospatial data tially enabled can become a platform for partnerships and (and their maps) can be presented to governments and collaborations likely to increase sales of PUE appliances their PUE development partners, who can then share the and machines. information with mini grid developers, local microfinance The outputs of Step 1 are: (1) an online platform that pres- providers, and appliance suppliers. Stakeholders can then ents the data in a user-friendly interface; (2) a long list of overlay the opportunity areas with their own operations— key stakeholders, including mini grid developers, local pro- for developers, their mini grid sites (existing and planned); viders of microfinance, and appliances suppliers; and (3) a for microfinance institutions (MFIs), their branch network; long list of high-impact opportunity areas. for appliance suppliers, their distribution networks and hubs. In addition, developers, suppliers, and MFIs can be Working with TFE Energy’s Village Data Analytics, ESMAP matched when mutually beneficial opportunities are iden- is deploying this approach in the Democratic Republic of tified. For example, a high-impact opportunity might fall Congo, Ethiopia, and Nigeria. Table 3.3 presents some of within the expansion plans of a mini grid developer, close the PUE program stakeholders—appliance suppliers and to a local MFI branch and an appliance supplier. In this local finance providers. These lists will grow alongside the way, a market assessment that is data driven and geospa- ESMAP project. TABLE 3.3 • Example of PUE program stakeholders identified in the Democratic Republic of Congo, Ethiopia, and Nigeria Stakeholders Congo, Dem. Rep. Ethiopia Nigeria Appliance Agrimont Group “Tsehay Roschli Industrial Adebash Manufacturing Hanigha Nigeria suppliers Association of Cocoa and and Agricultural Engineering Company Ibraham Onsachi Coffee Exporters of the (Selam Children Village)” Alanco & Son Steel Fabricator Kenny Construction Company DRD (ASSECAF)” Afesol Technology Plc Alaral Tech Engineering Design Kola Adekunku AVSI Foundation Amio Engineering Plc & Fabrication Koolboks CONAPAC Beza Industry Plc Alayan Metals Fabrication Nig Koolmill Fouani Bumzal Amadis Technical Company Lawod Metal Nig Lushebere Dairy and DYD Trading Plc Apexskill Works Magi Rches Limited Cheese Factory Arcadem Ethio-Mercantile MCAN Makita DRC Basicon Engineering Company General Mercantile Plc Muharib Machine Strategos Plantations Bennie Agro ltd. Kaleb Service Farmer’s Muhat Nigeria The Breeders’ Society of House PLC Besuga Global Investment Bandundu (SEBO) N.C. Gilbert Ind Dev Co Kalmeks Engineering Bifem Technologies Nigeria The Society of Livestock Nanyang Goodway Machinery & Manufacturing Limited Farms of Bas-Congo (GEL Equipment Co., Ltd Lacomelza Plc Blessed Silver Brothers Bas-Congo) Niji Lukas Nig Marast Bomik Adeyeera Engineering Nova Technologies Roda Business Group Plc Camco Nui-Lukas Ten Tools Chinige Technology Services Oladimeji Success Tsion Industrial Engineering LTD Olaleye Eliseri Coldbox PAF Metal Fabrication and Youth Coldhubs Development Confidence Technical Work Peak Products Enterprise Pentawork Technical Work Dangote Process Concepts & Technologies Deban Faith Agro Allied Ventures S. Adiss Engineering Works DEE Technica Sakilan Engineering Company Doing Segun Towoju E. K. Fabricating Engineering Sominie Nigeria Limited Eamak Technical Services Starron Ecotutu—Interview Sunday Omowaye Emeka & Sons Construction Talitha Fabrication Company Company Teekay Tronics ESE Engineering Service Tropical Development Engineering Ltd Fatoroy Steel Industry UNIC & Sons Gensaes Enterprises Weilai Machinery Zheng Zhou Sida continued MINI GRIDS FOR HALF A BILLION PEOPLE    119 TABLE 3.3 • continued Stakeholders Congo, Dem. Rep. Ethiopia Nigeria Finance Advans Bank Addis Credit & Saving Justice Development and Peace Commission (JDPC) microfinance / providers Altech Institution (ADCSI) Integrated Development Programme Business MFI Agar Microfinance Addosser Microfinance Bank Credit YA MPA Amhara Credit and Savings Afex Institution (ACSI) Babban Gona Finca Benshangul MFI Baobab Microfinance Bank Hekima/Goma Buusaa Gonofaa Barnawa Microfinance Bank IFOD Dedebit Credit and Savings Chase Microfinance Bank Kitumaini Institution (DECSI) Light in Business SA/ Corebank / CORESTEP MICROFINANCE BANK Dire Microfinance Butembo Development Exchange Center (DEC) Microfinance Metemamen Microfinance  Mam Tombuama GRASSROOTS DEVELOPMENT MICROFINANCE BANK Omo Microfinance MFI APE Ibile Microfinance Bank Oromiya Credit and Savings Micropop Life Above Poverty Organization (LAPO) microfinance Share Company (Ocssco) Paderu Nirsal Microfinance Bank Poverty Eradication and Paidek SA Community Empowerment Richway Microfinance Bank Procfin (PEACE) SMICO S.A./Goma Sidama Tout Pour le Genre Dans Somali Microfinance Le Developpement (TGD) Specialized Financial and Trust Investment Promotional Institution Development TID SA / (SFPI) Butembo Vision Fund Microfinance Tujenge Pamoja/Goma (VFMFI) Tujenge S.A. Wasasa Microfinance Vision Fund Yoasi Source: ESMAP and TFE-VIDA analysis. STEP 2 assessments of productive demand. The systematic Surveys and workshops build on Step 1 data approach maps all potential activities in a community that through community engagement could benefit from access to electricity. Every stage of the existing and potential value chains—from inputs and pro- Once the high-impact opportunity areas are identified, plan- cessing to outputs and end uses—is screened to capture ners need to “ground truth” by engaging with communities. actors, market dynamics, cycles, and seasonality. This This engagement entails multiple rounds of site visits to approach assesses the role electricity already plays and identify the context-specific potential for PUE. Each succes- could play in the prioritized sectors. sive round would home-in on top-priority locations, appli- ances, and stakeholders. In practice, this means sending The pragmatic approach, by way of contrast, does not look teams of surveyors and community engagement special- at the spectrum of productive uses. Instead, it takes advan- ists into the high-impact areas identified in Step 1. Equipped tage of existing opportunities to ease the interface between with smartphones and tablets, the team will collect data and mini grid developers and productive sectors. The process take photos and videos, obtaining a deeper understanding of identifying PUE is speeding up, so moving on to imple- of local socioeconomic dynamics, critical for effective inter- mentation takes less time. Although less comprehensive, ventions that promote productive uses and build a support- the pragmatic approach is faster and more affordable and ive ecosystem. If the mini grid developer leads this effort, builds on extant cross-sector ties. it could gain support from the rural electrification agency, After gathering information on appliance usage (current NGOs, and development partners in facilitated interaction and potential), teams could tell productive users they may with productive sectors and local businesses. sign up for mini grid electricity at a later site visit. Further- Two complementary methods—systematic and prag- more, survey teams could also approach local suppliers of matic—exist to identify opportunities for productive uses microfinance and appliances to gauge their interest in a at the community level (de Gouvello and Durix 2008). PUE program. This might require visits to MFI branches and Field investigation is, in both cases, a requirement to refine appliance or hardware shops in nearby towns to interview 120   MINI GRIDS FOR HALF A BILLION PEOPLE Identifying context-specific potential for At the community level, demand assess- productive uses should begin as early ment should also look into the gender gaps as possible in the mini grid development project in entrepreneurial activity. What is constraining the cycle. Two complementary methods—systematic ability of women-owned enterprises to thrive? and pragmatic—exist to identify opportunities for productive uses. The systematic approach maps activities that might benefit from access to electric- ity. The pragmatic approach appraises a sector (or STEP3 location) first, and then identifies economic activi- Demand analysis for mini grid design and market ties that are improvable through mini grid access. potential for appliances and associated end-user finance During the early stages of the PUE program, a team’s direct engagement with communities is important in identifying staff. Finally, the community engagement teams will need high-priority appliances and machines. Once these appli- to meet with community leaders and district-level govern- ances have been identified, their economic and technical ment officials to introduce them to, and garner their sup- attributes should inform mini grid design, so it may more port for, the PUE program. precisely ascertain the market potential for appliance sup- pliers and local providers of microfinance. Demand assessment should also assess the gender gaps that women and men face as they start up enterprises in The range of energy-efficient appliances on the market an effort to identify the chief constraints on women-owned today is continuing to expand. Table 3.4 presents indicative enterprises. For example, in the agriculture sector, men information on power requirements, costs, and payback often mediate women’s access to resources and commu- periods for a sample set of widely distributed income-gen- nity participation, and these men tend to be fathers or hus- erating appliances and machines. Of the more than 160 bands. So the agricultural contributions that women make appliances that ESMAP identified as being available on the often go unrecognized. market today, more than 130 have a payback period of less than 12 months. The typical up-front investment ranged Gender gaps are apparent everywhere, of course. But in from $50 to $1,500, with an average of about $1,200. At Nigeria, the gender productivity gap for agriculture stands the conclusion of the payback period, the appliances gen- at 18.6 percent. Another notable gap is found in Dominica, erate between $50 and $500 in monthly revenues, with an a Caribbean island state, where women work mostly in average of $300. the informal market in microenterprises and subsistence farming, which constrains their ability to advance (Mukasa Once the technical and economic information about pro- and Salami 2016; ILO 2014). Gender gaps are also driven ductive-use appliances is better understood, potential cus- by the cultivation of smaller land parcels, discriminatory tomers will need to decide if acquiring the appliance is right land laws, constrained access to financial products and for them. CrossBoundary’s Mini Grid Innovation Lab has services, snubbed by extension or business development shared some of the questions that customers can ask to services, and restricted to older technologies. help them make this decision. These include: Step 2 outputs are lists of: • What level of electricity service—primarily in terms of capacity and reliability—is required to produce the final • Communities validated as areas of high-impact oppor- product desired, and is the mini grid capable of provid- tunity, combined with community-specific market data; ing this level of service? • Appliances relevant in terms of the communities’ local • How will converting from diesel or using an energy-ef- context; ficient appliance affect the technical aspects of the • Potential PUE customers in these communities; machine? • Local providers of microfinance in or near these com- • What tariff will make converting to electric power from munities; diesel cost-effective? • Local suppliers of appliances that serve these commu- • What time of consumption (across the day or night) will nities; and yield the greatest profits? • Community leaders and district-level government offi- • Would it be more profitable for me to be located closer cials who are supportive of the PUE program. to the generation source? MINI GRIDS FOR HALF A BILLION PEOPLE    121 TABLE 3.4 • Power requirements, costs, and indicative payback periods of selected income-generating appliances Average monthly Power required revenue after (kW unless speci- Cost from Payback period payback period Sector Activities and appliances fied otherwise) supplier (US$) (months) (US$) Primary Crop dryer 6–1 1,000–5,000 12–18 236 industries Egg incubator 80–16 watts (W) 50–100 1–3 58 (agriculture, Hammer mills of various types for the 5–10 700–1,200 6–12 129 fishing) respective grains Electric sawmill 1.5–3.0 500–800 12–24 44 Threshing machine 7–9 500–1,000 3–6 208 Grinder for pulses and beans 5.2 1,500–4,000 6–12 396 Garlic/ginger paste machine 2–7 1,500–4,000 6–12 396 Tomato ketchup/paste–making machine 5 800–2,500 3–6 483 Water irrigation pump 3.7–22.4 200–1,000 3–6 183 Oil expeller 6 (600 W for 800–2,000 for 3–6 400 ordinary oil press) ordinary oil press Sterilizer (for dairy processing) 3–6 600–2,000 1–3 1,100 Packager 250 W–3 kW 500–1,000 6–12 104 Peanut roaster 1–10 1,000–5,000 6–12 458 Electric crusher for peanut paster 3 1,000–3,000 3–6 583 Automatic measurer/bagger 1 3,000–5,000 12–18 292 Light Electronic welding machines 3–7.5 200–300 6–12 33 manufacturing Angular grinder 2 100–200 6–12 21 Circular saw for wood 1–2 150–200 6–12 23 Electric smoothing plane 200–300 W 50–100 3–6 21 Jigsaw 400 W 100 3–6 25 Electric drilling machine 400 W 20–50 3–6 10 Ice maker 4–6 500–1,500 6–12 146 Socks knitter 4 2,000–5,000 6–12 500 Biscuit maker 2 3,500–6,500 6–12 688 C-brick maker 9 2,000–5,000 3–6 1,000 PET bottle maker 24 3,000–10,000 3–6 1,917 Popcorn maker 1.5–2.1 50 1–3 33 Commercial and Phone charger 5–10 W 10 1–3 7 retail activities Refrigerator for cold drinks in cafeterias 100 W 150–300 1–3 175 or medicines in dispensaries Computer 15–100 W 250–800 3–6 154 Printer/scanner for stationary 0.5–2 150–250 3–6 54 Appliances for hairdressers’ shops 1.5–2.5 (hair 40–60 (hair 1–3 19 (hair clipper, hair dryer) dryer) dryer) 10–20 W (hair 15–30 (hair clipper) clipper) Sewing machine 200 W 30–100 3–6 17 Television for local cinemas and bars 50–200 W 100–200 1–3 46 (including decoder) Hi-fi stereo system 20–500 W 20–200 A hi-fi system n.a. in itself does not provide a payback period Electric cookstove 3.5 50–100 Does not have a n.a. payback period due to high consumption Deep fryer 2 100–500 6–12 46 Sources: de Gouvello and Durix 2008; Alibaba 2022; ESMAP and INENSUS analysis. Note: Battery chargers come in a range of sizes. The one presented here would charge from 50 percent to 100 percent in about three hours. The ice maker can produce 1,000 kilograms of ice per day. kW = kilowatts; W = watts; n.a. = not applicable. 122   MINI GRIDS FOR HALF A BILLION PEOPLE is important to enable developers to effectively promote their uptake among customers, design their tariffs appro- Of the 160 income-generating machines and priately, provide financing options to end users, and design appliances on the market today, more than their mini grids. 130 have a payback period of less than 12 months. The typical up-front investment cost ranged from The different productive-use appliances come with a range $50 to $1,500, with an average of about $1,200. The of load profiles, and developers can design their mini grids monthly revenues generated by these appliances to prevent under- or oversizing the system to strengthen its after the payback period typically ranged from $50 economic viability. Undersizing could restrict revenues for to $500, with an average of $300 across the identi- the mini grid operator and push consumers toward alterna- fied appliances. tive sources of energy. Excess capacity raises investment costs above revenues, lengthening the payback period, hiking operational costs, and reducing overall efficiency. Because prediction of demand tends to be unreliable, over- For mini grid developers, adding productive users to their sizing is common for mini grids in Africa. customer base adds complexity to the project design, because they have to determine whether and how to The key elements of system design include: connect these loads, which differ in terms of time of use, • Peak power, magnitude of power and energy demand, and seasonality. As such, mini grid developers will also need answers to a • Reactive power, series of questions on how to include them in their busi- • Single- or three-phase distribution networks, ness models as early as possible in the project cycle. Cross- • Capacity utilization, and Boundary’s work with developers has helped identify some of these questions, including: • Incorporation of backup generators versus batteries. Three-phase distribution systems are able to power large • How much additional generating capacity is needed to productive-use loads across a wider range of machinery, support the load demand? especially equipment above 10 kW or 10 kilovolts-ampere, • What inverter size and distribution system will allow or rural clusters of smaller appliances in a productive facility multiple productive-use machines to operate simulta- at an isolated site. Productive users usually require a min- neously? imum service at Tier 3, or peak available capacity of more • From how far away can the grid support a large produc- than 200 W and eight-plus hours of energy supply, including tive load running on three-phase power? at least two hours in the evening (Bhatia and Angelou 2015). • How low can the tariff be while still proving sustainable? Mini grid developers generally have two ways to manage electricity demand from PUE appliances and machines: • What time of consumption will allow least-cost genera- tariff incentives and contractual obligations. tion? Tariff incentives can stimulate demand and greatly boost • How should the tariff structure be adjusted to account the use of income-generating appliances and machines. for seasonality? In regulatory environments where developers are free to Adding electric appliances to the systems is often not tech- charge cost-recovery tariffs (see chapter 9 for a detailed nically straightforward and requires demand-side manage- discussion on mini grid tariffs), the guiding principle of set- ment skills. Income-generating appliances typically require ting an appropriate tariff should be to keep it easily man- more power and energy to operate than consumer appli- ageable for the operator and easily understandable for end ances (fans or televisions). Some types of income-gener- users. The tariff structure should also allow a reasonable ating appliances are already in use in the community but return on investment while remaining attractive and afford- are powered by diesel engines and may not be the most able for productive end users. It should take into account efficient options to maximize limited load profiles. Others the ability and willingness to pay of productive clients as will be introduced into the community for the first time. well as the availability of alternative power sources or back- As a result, mini grid developers often lack information on ups, the predictability of supply, and the possibility of com- the power and energy requirements and load profiles of bining different sources of electricity. appliances and machines, which could lead to inaccurate demand forecasts and wrong-sizing the mini grid’s gener- Tariff structures with a range of applications exist, including ation and storage capacity. Accurate information on the tariffs that are: technical characteristics of income-generating appliances • Capacity-based, MINI GRIDS FOR HALF A BILLION PEOPLE    123 • Inverted block, • Per device, Demand and load management are essen- • Time-of-use, and tial for productive users, who depend on Tier 3+ service. Two strategies are available to man- • Seasonal. age demand: tariff incentives and contractual obli- When tariffs are tailored to end users, rates can be adjusted gations. Tariffs can incentivize productive uses. One by load location and size and by type of connection and common approach is to charge low daytime energy business. The mini grid operator can also decide to intro- tariffs but, during peak morning and evening hours, duce lower, curtailable load tariffs for customers who agree to charge higher energy tariffs. Other approaches to be curtailed. include allowing productive users to postpay for their electricity consumption, while waiving stand- In Tanzania, Jumeme and Energy 4 Impact designed a tariff ing charges for productive users with seasonal busi- schedule aligned with end users’ needs after collecting data ness activities. Operators can also encourage or on businesses’ energy consumption. They determined that, require a shift of some productive consumption to to be able to compete with diesel gensets, the mini grid tar- daytime or other nonpeak periods, through service iff to power a maize mill should be about $0.32/kWh. This agreement contracts. rate was much lower than what mini grid developers usu- ally charged in Sub-Saharan Africa ($0.50–$1.20/kWh). Based on an incentivized use pattern, daytime tariffs were set at about a quarter of evening tariffs. Ultimately, one of the key outputs of Step 3 is to iden- tify a set of high-priority appliances that appeal to both Payment terms should also be adjusted to reflect produc- mini grid developers and customers. For developers, tive users’ constraints. Some users may be reluctant to such appliances tend to be energy hungry at predictable pay a connection charge (or a flat standing charge) when daytime hours (or that run consistently over 24 hours). their activities are highly seasonal, with no electricity needs For customers, high-priority appliances have the short- during certain periods of the year. Prepayment can also be est payback periods while generating income over time. problematic for productive users with limited cash flows, Step 3 outputs also document community-relevant appli- especially early in their development. As a result, two ways ances—knowledge that informs the go-to market strate- to accommodate PUE customers are, (1) waiving stand- gies of mini grid developers, local microfinance providers, ing charges and (2) allowing productive-use customers to and appliance suppliers. postpay for their electricity. Operators can also use contracts to shift some productive STEP 4 consumption to daytime or other nonpeak periods. The Preparation of road shows involving local Tanzanian villages of Lupande, Mawengi, and Madunda are government, leadership communities, connected to a 300 kW hydropower mini grid. Corn milling communities, interested appliance providers and and welding are permitted only during business hours (9 end-user financiers, mini grid companies a.m. to 6 p.m.) so households have enough electricity at Steps 1–3 focus on gathering data and information that will night (USAID 2018). inform Steps 4–6, which are focused on implementing road shows and other community engagement activities and Contracts can make demand-side management more ultimately deploying income-generating appliances and efficient. They mitigate demand risk and anticipate the machines in partnership with mini grid developers, local effects both on capital expenditures and the levelized cost finance providers, and appliance suppliers. of electricity. In doing so, contracts define the tariff struc- ture, which determines what productive users should be Step 4 consists of preparing the logistics for road shows connected when, and at what tariff. Beyond collaborations and other community engagement events. Ever since elec- between mini grid operators and productive users—and tricity systems were first installed outside major cities—for in order to better secure the balance between demand example, in the 1920s and 1930s in the United States— and supply—developers can secure a guaranteed level of electricity providers, governments, and other community demand through contractual agreements with customers, development organizations have organized road shows, which increase load predictability and therefore revenue competitions, and other events that introduce communi- streams. Signing up the load before the final system design ties to electrical appliances and machines (Duke University improves the ability to right-size the system and gives the 2022). Demonstrations were key. How do electrical appli- end user the option of securing the amount and quality of ances and machines work? Productive users of electricity power it desires. share their experiences with prospective users, and explain 124   MINI GRIDS FOR HALF A BILLION PEOPLE the availability and terms of appliance financing, describing Information and marketing campaigns should preview the mini grid developer’s plans (timing, tariffs, and so on). the road shows, which are integral to Step 4. Notices on Finally, potential customers sign up for mini grid electricity, local radio, posters, leaflets, and other materials are vital appliance ownership, and consumer financing. Road shows, for community presentations. Centralized resource cen- supported by information and marketing campaigns, have ters or NGOs can handle this part, and they can gather proved their effectiveness for more than a century, increas- information on productive equipment and techniques, ing uptake of income-generating machines and appliances. advising on business capacity and productive processes. Analyzing these rural road shows in the United States Campaigns should disseminate high-quality information between 1938 and 1945, Duke University found that they materials, messages, and methods that focus on market increased consumption by 64 kWh per customer per year surveys and targeting. (Plutshak, Free, and Fetter 2020). The outputs of Step 4 are finalized logistics (who, what, But these events require careful planning and coordination where, when, and how) for road-show marketing and infor- across a diverse mix of stakeholders. In fact, six core stake- mation campaigns that target the high-impact opportu- holder groups that will need to come together are: nity communities identified in Steps 1 and 2. These steps demonstrate the first set of machines and appliances iden- • Current and potential users of income-generating tified and analyzed in Steps 2 and 3. appliances, • Mini grid developers, STEP 5 Road shows to load centers where mini grid • Local providers of consumer finance, developers, appliance suppliers, and end user • Appliance suppliers, financiers explain the value propositions to • Community leaders, and potential end users based on their current and aspirational living standards • District government leaders. With the logistics in place and the marketing and infor- All of these stakeholders should have been identified at mation campaigns implemented, the road shows can be Step 2—the task during Step 4 is to coordinate collec- deployed to the high-impact opportunity communities. tive action to plan the road shows and other community Road shows should be fun, informative, safe, and memo- engagement events. Additional stakeholders may also rable events. Indeed, the road shows implemented by the need to participate depending on the context, such as agri- Rural Electrification Administration of the United States culture agencies and local training institutes. were affectionately referred to by participants as the “Elec- The high-impact opportunity communities identified in tric Circus” (Duke University 2022). Step 1 and validated in Step 2 are the geographic targets for District-level government officials often serve as strong road shows. Events should take place in areas accessible voices of support to open the event. A successful road show to several different high-impact opportunity communities. will offer: The geospatial data collected in Step 1 should guide road- show activities, taking into account access, community • Highly interactive and participatory environments that clusters, and other relevant information. engage the mini grid developer, local authorities, local community organizations, financing agencies, vendors For road shows, good rules of thumb (de Gouvello and of electrical equipment, and potential customers; Durix 2008) are: • Events or demonstrations that target women. Because • Integrate initiatives from other sectors like agriculture to women are less likely to be employed than men, and avoid duplication; more likely than men to run informal businesses, they • Anticipate the effects of PUE on people, which is to say, are less familiar with mechanized work. They need to be aware of how increased use changes household and see the benefits of the productive uses of appliances. social behaviors; In addition to the road shows themselves, key outputs of • Favor local initiatives and equipment; Step 5 are concluding agreements with potential customers for the suite of products and services so they can acquire • Tailor activities to targeted sectors and regions and and use income-generating machines and appliances. Step adapt them to the literacy levels and habits of the tar- 5 should therefore conclude with signing customers up for geted audiences; mini grid electricity; for an appliance; and for consumer • Design all activities to include women, and develop finance (if necessary and desired by the customer). activities that specifically target women. MINI GRIDS FOR HALF A BILLION PEOPLE    125 TABLE 3.5 • Stakeholders that could be involved in road shows and their respective roles Stakeholders Roles Local distributors of targeted equipment, local Demonstrate equipment cottage industries Provide technical adaptation to local uses Train users Develop retail network to facilitate local purchase of appliances Energy service providers and their Build consumer awareness and support; help providers adhere to standards associations Educate and train customers Adapt energy infrastructure to local growth of energy demand Microfinance and credit unions Provide customized financing to potential users Conduct risk analysis Build sector awareness Educate customers about loans Agriculture and other institutes, enterprises, Provide sector knowledge, field presence, networking and outreach capacities, and promotion centers demonstration capacity, and vocational training Build user confidence Train users Training and vocational centers Develop vocational curriculum and train teachers Source: De Gouvello and Durix 2008. Clustering productive users in an area close to the source of generation is one way to increase the connectivity of Road shows, combined with information and businesses. There are different ways to encourage clus- marketing campaigns, are effective ways to tering, including multifunction platforms, solar kiosks, increase the uptake of income-generating appli- productive-use centers have been experimented. Beyond ances and machines. They require careful planning the technical and financial advantages to the developer and collective action by six core stakeholder groups: (higher-quality service, limited distribution investment), current and potential users of income-generating clustering facilitates day-to-day interactions, strengthens appliances, mini grid developers, local providers of innovation and business development, and increases tech- consumer finance, appliance suppliers, commu- nology and knowledge sharing within the productive com- nity leaders, and district government leaders. Road munity. The role of public and central authorities, assisted shows should be preceded by information and mar- by multilateral partners and donors, is key in clustering keting campaigns, should be highly interactive and small business activities by enabling land tenure, facilitat- participatory, and should actively target women, ing permitting, and providing infrastructure (roads). (See who are less likely to be employed than men, more chapter 8 for a discussion of the different institutions that likely to run informal businesses than men, and less are relevant for mini grids.) likely to engage in mechanized work than men. In addition, to unlock the entrepreneurial potential of communities, local authorities, NGOs, or the developers themselves need to provide gender-specific business STEP 6 development services and mentoring activities for targeted Rollout of mini grid connections, sales of local entrepreneurs. A study that covered Tanzania, Ghana, appliances, and end-user finance and Myanmar, highlighted differences found between Upon completion of a road show, the mini grid developers, enterprises run by men and women and the need for gen- appliance suppliers, and local finance organizations can der-specific support (IDS and GIZ 2019). Men own more begin to deploy their products and services to the custom- businesses and spend more on electricity than women, ers they signed up at Step 5. It is rarely easy, however, for who spend more on cooking fuels. Women operate in less microentrepreneurs and local small businesses to make electricity-intensive sectors that are mainly devoted to food the leap from sign-up to acquisition. Two strategies are preparation, hospitality, tailoring, hairdressing, and retail, emerging to directly support local entrepreneurial uptake while men are familiar with mechanized and electricity- of income-generating machines and appliances as they intensive work. This gap between women and men and arrive in the community: clustering, and advisory support. their respective patterns of productive energy use is eas- 126   MINI GRIDS FOR HALF A BILLION PEOPLE Two strategies to improve small business Most productive-use appliances and equip- uptake of mini grid electricity services are ment have relatively high up-front costs clustering productive users in an area close to the compared with the disposable-income levels of source of generation and providing gender-specific prospective entrepreneurs and small businesses, advisory support to entrepreneurs and small busi- but they provide clear opportunities to generate ness managers. or increase revenue. Financing the up-front costs of the appliance—whether from a mini grid opera- tor or a third party—will be necessary to increase productive uses of mini grid electricity. ESMAP estimates that approximately $3.6 billion in micro- Women often operate in jobs related to food finance for 3 million income-generating appliances preparation, hospitality, tailoring, hairdress- is needed under the scenario in which Sustainable ing, and retail, which tend to be less electricity inten- Development Goal 7 is achieved, in part through the sive. Meanwhile, men take jobs that tend to be more development of 200,000 new mini grids. mechanized and electricity intensive. They also enjoy better starting conditions in terms of capital, resources, and skills. By way of contrast, women 2018). Appliance financing schemes are therefore needed are constrained by household responsibilities and for end users to overcome up-front costs of appliances restrictive social norms. Nevertheless, access to through either the mini grid developer’s “own-managed” electricity can generate income for both men and financing facility or mechanisms managed by third par- women who own and run enterprises, highlighting ties, such as financing agencies, MFIs, or equipment sell- the need for gender-specific advisory support to ers (NREL and E4I 2018). boost the PUE. Financing by third parties Involving third parties as asset-financing companies lets ily explained: men start their working lives with more capi- developers remain focused on their primary activity—lever- tal, resources, and skills, while women take on demanding aging technology, competency, and capacity (Factor[e] household-care responsibilities, while social norms restrict Ventures 2020). Partnerships with pay-as-you-go com- their participation in certain occupations, such as fishing. panies for appliance financing services have been tested. Yet access to electricity led to increased profits for both Asset-financing companies such as Rent-to-Own in Zambia men and women, thus highlighting the need for targeted, and EnerGrow in Uganda are emerging with a specific focus gender-specific advisory support to increase overall PUE. on financing appliances for on- and off-grid customers. Consumer financing will need to arrive alongside mini Third parties, however, should tackle the issue of financial grid electricity and income-generating appliances and inclusion with end users, many of whom operate outside machines. The market potential is substantial. If mini grids the formal financial system and often lack collateral. They are to reach their full potential—serving half a billion people often face high costs and short repayment periods for loans by 2030 is a core solution for achieving universal access from the commercial banking sector in remote areas, espe- to electricity—then $3.6 billion in microfinance is needed cially when they want to develop new business activities. for 3 million income-generating appliances, assuming an Women are particularly burdened by the lack of collateral average of 15 productive-use appliances per mini grid for and are less likely to have a bank account.1 Many coun- 200,000 new mini grids at an average cost of $1,200 per tries continue to have laws that restrict women’s access appliance (see table 3.4). to inheritance and land titling, which hinders their ability to access assets that can be used as collateral when securing Expensive electrical appliances and high connection fees a loan (World Bank 2020). can impose a financial burden on entrepreneurs and small businesses, slowing or preventing the decision to invest Third parties that finance productive uses include MFIs, in productive-use appliances and equipment, especially community savings groups, rural electrification agencies, when no formal financing institution exists. A study from and ad hoc structures locally established by development the Rocky Mountain Institute showed that when end users partners, NGOs, and other local entities. As deposit-taking were offered 12-month loan terms to finance produc- institutions targeting low-income people and microenter- tive-use appliances, consumption almost doubled. After prises in developing countries, MFIs are well established 11 months, mini grid revenues grew by 18 percent (RMI and offer a variety of financing products. They also ben- MINI GRIDS FOR HALF A BILLION PEOPLE    127 Offering financing directly to productive users enables the developer to better monitor its strategy toward productive When trying to secure finance to purchase clients, including whom to connect and what appliances productive-use appliances and equipment, they choose. Controlling the type of appliances connected women are particularly burdened by the lack of col- to the mini grid and favoring high-quality equipment may lateral and are less likely to have a bank account. facilitate system management and result in better opera- Many countries continue to have laws that restrict tional efficiencies. As an example, a lease-to-own model (or women’s access to inheritance and land titling, on-bill financing) could enable end users to pay 30 percent which hinders their ability to access assets that can of the purchase price up front, paying the rest to the mini be used as collateral when securing a loan. grid operator month by month for a certain period through the electricity bill (NREL and E4I 2018). This mechanism helps end users, who pay less interest to the operator than efit from existing customer bases and loan distribution they would to a commercial lender. networks. Not yet seen are equity investments by impact In Tanzania, Jumeme, a private operator, has run a 90 kW investors in productive-use technologies. solar mini grid in Bwisya on Lake Victoria since 2016. Its Development partners can support access to credit for presence has allowed for the automation and expansion of equipment by increasing the awareness of financing agen- existing businesses (grain milling, carpentry, and bicycle cies or equipment sellers regarding nontraditional proxies repair) and supported the emergence of new businesses for creditworthiness and new approaches to assessing con- (egg incubation, ice block production, and metal welding). sumer risk (Cheney 2016). Lenders can use new sources of The company runs a shop selling appliances in the larg- data—such as records of timeliness of phone bill payments, est village. It has helped Jumeme avoid technical issues social network data, and mobile phone use—to determine and control consumption by ensuring the use of effective the ability and willingness to repay of unbanked users (Baer, and adapted equipment. Small and medium enterprises Tony, and Schiff 2013). Another option is to work through acquire appliances on credit (usually for about six months) local governance structures. provided by the mini grid operator. About 20 businesses have received appliances through in-house financing Financing by the mini grid developers (USAID 2018). In some cases, mini grid developers offer financing to their This model may impose a financial burden on the mini grid customers to support the acquisition of income-generating developers’ balance sheets, however, and divert the devel- appliances and machines. One approach is to offer custom- oper from its core business. It implies managing software ers a fee-for-service model. In one example of this model, systems for managing appliance loans and “pay-as-you-go” the mini grid company provides a productive hub and offers lockout systems in parallel with maintaining a billing sys- the use of income-generating appliances against the pay- tem (Factor[e] Ventures 2020). Mini grid developers do not ment of a fee. necessarily have the skills or expertise to handle in-house Another example comes from a deployment of 600 financing. solar mills in Indonesia, Vanuatu, and Papua New Guinea Once operational, mini grid developers should monitor funded by the United States Agency for International users for one to two years after their connections. Moni- Development and implemented by Village Infrastructure toring will enrich their knowledge of the demand dynamics Angels and Sumba Sustainable Solutions. Customers from productive-use appliances and help them address could trade goods and services (that is, share the produc- customer issues as—or even before—they arise. By mon- tion increase through sellable products made possible itoring demand, Vulcan Impact Investing (which owns by time savings) for the use of income-generating appli- about 10 mini grids in rural Kenya) found that the average ances. This approach has demonstrated positive results, revenue per user generated from the 10 percent of its cli- guiding end users through time savings gained from the ents that are small businesses was five times greater than mills and longer productive hours gained through night- the revenue generated from the other 90 percent. Even time lighting to make tradeable products. The approach though most customers consumed less than 250 watt- proved particularly efficient and robust during the COVID- hours a day, they were still critical to mitigating the risk of 19 lockdown, when cash in the communities was tight and losing larger clients (Blodgett and others 2017). village agents who disbursed noncash payments fared better than those dealing only in cash (Village Infrastruc- ture Angels 2019). 128   MINI GRIDS FOR HALF A BILLION PEOPLE other entities to take a coleading role in organizing the PUE program alongside the mini grid developers– notably, gov- Monitoring productive users for one to two ernments and local change agents. The following sections years after their connection to the mini grid describe how these two stakeholder groups can organize, can enrich a developer’s knowledge of the demand or help organize, PUE programs. dynamics from productive-use appliances and enable the developer to quickly address customer GOVERNMENT AGENCIES AND POLICY MAKERS issues as—or even before—they arise. Government agencies such as rural electrification author- ities or agricultural extension programs hold a strategic position in fostering productive use demand because of their national-level influence. They can serve as a coordinat- WHO ORGANIZES PUE PROGRAMS? ing agency to supervise and facilitate collaboration among stakeholders involved in promoting PUE (such as NGOs, MINI GRID DEVELOPERS developers, local communities, and equipment suppliers). As the providers of electricity to the end users of Operational support from rural electrification agencies income-generating machines and appliances, mini grid can include designing and implementing comprehensive developers are one of the main entities that can lead a PUE approaches that enhance the PUE in agricultural, indus- program. In particular, they can take a lead role in all six trial, and service sectors, for example, by enhancing the steps outlined above. In Step 1, they can inform the long knowledge and skills of small and microbusinesses on list of high-impact opportunity communities by directing how to use their newfound electrical and motive power analysis toward communities that they are planning to for profitable enterprise. Additional enabling interventions serve. Similarly, they can use the data collected in Step 1 could include, for example, strengthening the technical and to inform their expansion plans. In Step 2, mini grid devel- financial management capacity of women’s enterprises, opers will be engaging with communities already, as part expanding access to markets, creating linkages and access of their community outreach activities, so they can play a to financial products and services, enhancing extension lead role in identifying prospective PUE customers, local or business development services, and possibly address- community leaders, local finance providers, appliance sup- ing discriminatory land laws. As an interface between mini pliers, and even district-level government officials. Mini grid grid developers and other stakeholders, rural electrification developers also have the technical knowledge to assess the agencies can pursue these activities through the signature technical and economic impacts of connecting different of individual memoranda of understanding with developers appliances and machines to their networks, and thus are in to formalize their collaboration and precisely define roles a good position to lead on Step 3. This can also help ensure and responsibilities. A deep understanding of local socio- that the PUE program targets appliances that are attractive economic dynamics is critical, however, to ensure these to the mini grid developer. For Steps 4 and 5, the developer interventions succeed. can also take a leading position in organizing and imple- Partnering with other stakeholders enables rural elec- menting the road shows, although logistical and marketing trification agencies to improve such understanding and support from local, regional, or national governments can build their capacity on the topic of productive uses. Part- facilitate large-scale rollout of road shows. Lastly, the mini nerships could include engagement with in-country sec- grid developer is one of the three main actors in Step 6, and toral associations, microbusiness support entities, aid can coordinate the on-the-ground activities of appliance agencies and donors, governments, NGOs, private-sector suppliers and local finance entities after a road show. firms, and researchers. For example, a rural agency could Even if the mini grid developer takes the lead in organiz- contract directly with competitively selected local NGOs ing the PUE program, they will still need the support of the to assess the market and identify and promote activities other stakeholders—indeed, it requires collective action in close collaboration with developers. These contracts toward a common goal. Orchestrating a PUE program, par- can be structured in two phases. The first phase includes ticularly on a large scale, is a resource-intensive activity, conducting surveys to identify and assess productive requiring not just money but also capabilities, networks, potential and building the interface with other sectors and legwork. As a result, the overarching recommendation or programs. During the second phase, the NGO designs is that mini grid developers can be the driving force behind and launches marketing and promotional campaigns to a PUE program, but the program itself will need resources build capacity and awareness among entrepreneurs. The (time, money, and people) from a variety of stakeholder contract with the NGO would define targets, objectives, groups. And, in some cases, it might also make sense for and implementation frameworks to measure the NGO’s MINI GRIDS FOR HALF A BILLION PEOPLE    129 Government agencies like rural electrifi- Coordinated, concerted efforts to promote cation agencies and agricultural extension productive uses of electricity have been programs hold a strategic position with regard to quite successful, in one case (Bangladesh) increas- the productive use of energy, able to design and ing customer uptake by almost 500 percent. implement comprehensive, multistakeholder pro- ductive use in agricultural, industrial, and service sectors. These government-supported initiatives development programs, microfinance organizations, appli- can also impart knowledge and skills to micro and ance companies, energy service companies, municipali- small business, including, for example, how to use ties, regional/district officials, and local associations are all newfound electrical and motive power for profitable important stakeholders. Coordinating all of them requires enterprise, strengthen the management capacity of institutional support, which could be achieved through a women’s enterprises, improve access to markets, platform that facilitates dialogue. create linkages and access to financial products and services, and enhance services. TIMING PRODUCTIVE USE performance against indicators such as the amount of PROGRAMS FOR MAXIMUM EFFECT investment made in productive equipment, the amount of When should the programs launch? The answer is short electricity supplied to productive users, and the number and simple: Steps 1–5 should begin as soon as planning for of additional businesses connected. The contract should mini grid deployments starts. Step 6 should coincide with provide enough room for NGOs to tailor their strategies the arrival of the mini grid. and approaches to address local constraints and the Steps 1 through 6 can then be repeated throughout imple- specificities of each community. mentation not only to consider new data and learn from Two examples of concerted, coordinated, government-sup- previous iterations but also to maintain momentum and ported programs are offered in box 3.1, on the Infrastructure customer anticipation for expanded electricity access and Development Company Limited’s program in Bangladesh, its benefits. and box 3.2, on Ethiopia’s efforts to introduce productive Indeed, the activities presented in this chapter should be uses into its mini grid–based rural electrification program. sustained throughout the lifetime of a mini grid system, LOCAL CHANGE AGENTS with new geospatial analyses to be conducted every year or every other year, and Steps 2 through 6 undertaken at Local change agents such as village councils and commit- least once per year. tees, NGOs, and civil society organizations can also support productive uses and facilitate interactions among local businesses, mini grid operators, and equipment suppliers. WHAT’S NEXT? NGOs and civil society organizations have demonstrated Boosting productive use nearly ensures the chances of a that they can be successful partners for tasks that require mini grid’s success by offering an everyone-wins scenario substantial and continued support. Their diverse skills for developers, entrepreneurs, households and communi- (technical, social, and financial) combined with their local ties, and national utilities. More systematic and concerted presence allow them to work with entrepreneurs and coor- efforts are required, however, to promote, finance, and ulti- dinate with stakeholders in the field. Beyond identification mately boost the uptake of PUE. At the same time, devel- of productive uses and promotional activities, NGOs could opers and potential productive-use customers need to do business development, advising small enterprises on understand the economic and technical characteristics of business challenges. Field-based teams give NGOs the income-generating machines and appliances to be able to ability to analyze market opportunities and assist entre- make investment decisions accordingly. Successful exam- preneurs in preparing their business models and apply for ples of PUE programs, such as those described in boxes credit, while coordinating with mini grid developers on how 3.1, 3.2, and 3.3 below, show what is already possible, but a to provide adequate connections. dramatic scale-up of national mini grid markets will require MFIs, small business development centers, chambers of greater efforts from government agencies across sectors, commerce, and small business accelerators could also be developers, MFIs, NGOs, and other stakeholders along the mobilized. Implementation units of agriculture and rural income-generating appliance value chains. 130   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 3.1 HOW IDCOL INCREASES PRODUCTIVE USES OF ENERGY IN SOLAR-HYBRID MINI GRIDS IN BANGLADESH The Infrastructure Development Company Limited Mini grids have clear, long-term development impact (IDCOL) is providing concessional project financing potential, but uptake of productive uses varies widely to enable mini grid developers to deliver improved and—despite careful and detailed consumer surveys energy services via solar and solar-hybrid mini grids in and expected load analysis—customer uptake was remote parts of Bangladesh. More than 20 mini grids lower than predicted (figure B3.1.1). After three years, are in operation, with a total capacity of almost 5 mega- the financials under the IDCOL package show that only watts-peak, and many more mini grids are planned for two out of seven mini grids reached their expected development. Sites are located primarily on islands not level of demand. Eleven were only recently commis- reachable by grid extension because of rivers often sev- sioned, while the remaining mini grids for which data eral kilometers wide in the monsoon season. are available, struggle to reach the expected level of demand. Mini grid sites have their own microeconomies, which are usually a mix of seasonal activities (such as fisher- Uptake lags were particularly striking among larger ies, agriculture, milling, husking, and oil pressing) and mini grids (>200 kilowatts-peak) at which productive wood production and sawmills. They also host a range energy use was expected to account for 40–65 per- of supporting businesses: carpentry and metal work- cent of demand. Daytime productive energy users ing shops; diesel engine repair shops; and many small were not connecting as planned, and in some cases, retailers of services, products, and foods in the local larger nighttime customers were saturating plant bazaars, which are open in the evenings. Almost all capacity more quickly than expected. Higher invest- productive energy use is provided by dedicated diesel ment in addition to low demand and underutilization engines with belt drives. Most households and shops of the plant exposed these mini grids to negative cash have had access to stand-alone home systems for sev- flows and risks. eral years, under IDCOL’s home system program. FIGURE B3.1.1 • Share of expected load achieved by selected mini grids in Bangladesh 3-month moving average of percent of expected load achieved 180% 160% 140% 120% 100% 80% 60% 40% 20% 0% 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 Months after commissioning Site A Site B Site C Site D Site E Site F Site G Source: IDCOL analysis. Note: Figures show three-month moving averages. Site names are not mentioned for confidentiality purposes. MINI GRIDS FOR HALF A BILLION PEOPLE    131 BOX 3.1, continued FIGURE B3.1.2 • Effect of extensive customer awareness campaigns on uptake 120 100 Number of customers 80 60 40 20 0 Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Site 8 Site 9 Site 10 Site 11 Site 12 Site 13 Site 14 Mini grid sites Monthly asquisition before customer training (3 months average) Customer asquisition one month after training Source: IDCOL. Note: For purposes of confidentiality, sites are not named. To increase the uptake of productive uses of electricity, is under cultivation and has three cropping seasons. IDCOL launched intensive, three-day customer aware- Guidance was provided on irrigation services for farm- ness campaigns starting in October 2017, conducted ers at lower cost than their diesel-powered pumps. by international experts and trainers from major equip- IDCOL continues to broaden its training program to ment manufacturers. These campaigns combined cus- sensitize developers and operators on leveraging the tomer training with public shows featuring folk singing use of electricity to maximize socioeconomic benefit. and street theater, increasing productive-use custom- Training content is under development for the gen- ers and total customer acquisition by nearly 500 per- der dimension in mini grid interventions, highlighting cent (figure B3.1.2). the benefits of social facilities, streetlighting, schools IDCOL also arranged for training in management skills and clinics, rickshaw charging stations, and so forth. and development to strengthen understanding among This information will enable entrepreneurs to connect mini grid developers about business opportunities. It community benefits with their own sustainability as also required all funded mini grids to install remote, businesses. In addition, IDCOL plans to incentivize real-time monitoring capabilities for instantaneous daytime loads via time-of-use packages and financ- troubleshooting and tracking of historical trends. ing conversion packages ($120–$400, depending on Meanwhile, IDCOL provides assistance in mini grid– industry and load). Most of the larger sites need softer powered irrigation, which can greatly increase utiliza- loan finance terms (longer grace periods and terms) tion rates. In-house agriculturists identified irrigation to be viable for the sponsor. A sponsor will need to play potential for the grid area, since almost all arable land the lead microfinancing role. Source: Analysis by IDCOL and the World Bank RERED II team. 132   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 3.2 RURAL, PRODUCTIVE USES OF ELECTRICITY: LESSONS FROM ETHIOPIA In 2019, the government of Ethiopia issued the updated ductive potential, to plan targeted mini grid develop- National Electrification Programme (NEP 2.0), which ments. This comprehensive approach has emerged formalized its ambitious goal of universal electrifica- from partnering with the World Bank, the Ethiopian tion by 2030. The NEP 2.0 detailed a large-scale rollout Agricultural Transformation Agency, the Rockefeller of mini grids both through the national electrical util- Foundation, AfDB SEFA, and others, putting together ity (the Ethiopian Electric Utility, or EEU) and through cutting-edge geospatial analyses and overlaying many engaging the local and international private sectors. layers of data and maps. Since then, the government, primarily through the To foster productive uses, ADELE is conducting ana- Ministry of Water and Energy, has been developing and lytics to get a better understanding of the potential deploying programs and projects with the support of for a new appliance result-based financing scheme. development partners like the World Bank, the African This scheme would complement the ongoing project Development Bank and its Sustainable Energy Fund activities and maximize the transformational eco- for Africa (AfDB SEFA), the Deutsche Gesellschaft für nomic impact in the communities targeted with mini Internationale Zusammenarbeit (GIZ), the Rockefeller grid rollouts. Under the utility-led modality of ADELE, Foundation, the Foreign, Commonwealth and Develop- the EEU intends to build in a number of activities to ment Office, the Ikea Foundation, the Rocky Mountain improve appliance availability and appliance financing Institute, and many others. for the communities expected to be electrified by EEU One important example is the $500 million Access with mini grids. Such activities are expected to range to Distributed Electricity and Lighting in Ethio- from community engagement and appliance demon- pia (ADELE) project, financed by the World Bank. strations to connecting communities to microfinance Approved in 2021, ADELE is a comprehensive, nation- institutions, nongovernmental organizations, appli- al-level rural electrification program powered through ance distributors, and so on. the grid, mini grids, and stand-alone solar solutions. The Ministry of Water and Energy, in partnership with So far, $270 million had been committed to the mini the Agricultural Transformation Agency and with grid program alone. The mini grid component under extensive support from the Rockefeller Foundation ADELE is being implemented by the EEU and covers and AfDB SEFA, is currently piloting nine mini grid sites two key modalities for rolling out greenfield mini grids. with major agricultural and irrigation potential, as part As a substantial scale-up of the utility-led model, the of the Rockefeller Foundation-supported Distributed EEU expects to deploy about 200 mini grids over the Renewable Energy Agriculture Modalities (DREAM) coming years through various engineering, procure- project. The pilot is intended to test a grant-supported, ment, and construction (EPC) and operation and private-sector-led modality, with results-based financ- maintenance modalities; it is also supported by the ing from Rockefeller and concessional financing launch of a large-scale, private-sector-led, perfor- expected from AfDB SEFA. mance-based grant program. Finally, the Rockefeller Foundation is preparing a Pro- Through ADELE and other programs, the Government ductive Use Appliance Financing Facility (integrating of Ethiopia has chosen to center all its mini grid and off- grid, mini grid, and off-grid electrification), with sup- grid rollout activities around income-generating uses port from the Collaborative Labeling and Appliance of electricity: identifying, prioritizing and stimulating Standards Program (CLASP) and Nithio. The facility them. In practical terms, this means that it has been should make income-generating appliances more assembling a holistic and up-to-date map of every site affordable and accessible. across the country with the most agricultural and pro- Source: ESMAP and the World Bank ADELE team. MINI GRIDS FOR HALF A BILLION PEOPLE    133 BOX 3.3 LESSONS FROM A UTILITY-NGO PARTNERSHIP IN INDONESIA Indonesia’s Rural Grid Electrification projects in the in leadership, proximity to larger markets, landholding 1990s, funded by the World Bank, also promoted pro- and crop patterns, and existing nonfarm income-gen- ductive uses. The lessons learned then are relevant erating activities, and the NGOs would need to craft for mini grid development today (Finucane, Besnard, their marketing to fit the specifics of the individual and Golumbeanu 2021). They show how partnerships businesses and their contexts. between energy providers (in Indonesia the national For the NGOs, the task was to market and sell an utility PLN) and local nongovernmental organizations already designed, deployed, and priced service (rural (NGOs) built an ecosystem that boosted rural electrifi- business services) in ways that would motivate pur- cation through jobs, income, and productivity. Fieldwork chase decisions in the different rural and business helped to tout the potential role of NGOs as channels contexts. The role of the utility, PLN, was to manage to promote PLN service. NGOs in rural Indonesia had the program, including: (1) NGO contracting, supervi- impressive operational capabilities, experience in the sion and payments; (2) village selections, which were marketing of changes, and experience with rural cot- expected to be recently electrified communities; (3) tage and small businesses and the poorest households. target setting (by village, NGO, and key metrics); (4) With the support of NGOs, project stakeholders held vetting of NGO marketing materials for accuracy; and consultations in order to better understand the char- (5) performance and impact monitoring and report- acteristics of current and potential customers and the ing. An effective outreach program, such as Rural reasons behind the slow take-up of grid services and Business Services, can improve load use and generate design-marketing campaigns. economic activity and employment. To run the campaigns, PLN contracted experienced Although the grid-based activity implies an exten- NGOs, skilled in outreach to families and community sive service area and varied clientele, the concept of groups in rural literacy, health, nutrition, and microen- customer-responsive service that makes good use of terprise development. The goal was to determine strat- capacity is transferable to other supply situations. The egies, marketing mixes, and communication methods success of this approach stemmed mostly from the suitable for the different village contexts, and conduct a holistic, opportunistic, context-specific design pro- series of time-limited marketing campaigns. The com- cess, a process that outsourced marketing to a local plexities of each village would be different, for instance, NGO as a possible market entry point. Source: ESMAP analysis of World Bank project documentation. 134   MINI GRIDS FOR HALF A BILLION PEOPLE REFERENCES IEG (Independent Evaluation Group). 2008. The Welfare Impact of Rural Electrification: A Reassessment of the Costs and Benefits. Washing- Alibaba. 2022. “Alibaba Online Store.” Accessed 2022. https://www. ton, DC: World Bank. alibaba.com/. ILO (International Labour Organization). 2014. “Evolution of Informal Baer, Tobias, Goland Tony, and Robert Schiff. 2013. “New Credit-Risk Employment in the Dominican Republic.” Notes on Formalization Models for the Unbanked.” McKinsey & Company. https:/ /www. Series, International Labour Organization, Geneva. https://www.ilo. mckinsey.com/business-functions/risk/our-insights/new-credit- org/americas/sala-de-prensa/WCMS_245893/lang--en/index.htm. risk-models-for-the-unbanked. Kes, Aslihan, and Hema Swaminathan. 2006. “Gender and Time Poverty Banerjee, Sudeshna Ghosh, Kabir Malik, Andrew Tipping, Juliette in Sub-Saharan Africa.” In Gender, Time Use, and Poverty in Sub-Sa- Besnard, and John Nash. 2017. Double Dividend: Power and Agri- haran Africa, edited by Mark C. Blackden and Quentin Wodon. World culture Nexus in Sub-Saharan Africa. Washington, DC: World Bank. Bank Discussion Paper Series No. 73. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/26383. Mlinda. 2021. “Renewable Energy for Rural Communities India.” https:// Bhatia, Mikul, and Niki Angelou. 2015. Beyond Connections: Energy www.mlinda.org/democratising-energy-supply/. Access Redefined. Energy Sector Management Assistance Program Mukasa, A. N., and A. O. Salami. 2016. “Gender Productivity Differen- (ESMAP) Technical Report 008/15. Washington, DC: World Bank. tials among Smallholder Farmers in Africa: A Cross-Country Com- https://openknowledge.worldbank.org/handle/10986/24368. parison.” Working Paper Series No. 231, African Development Bank, Blodgett, Courtney, Peter Dauenhauer, Henry Louie, and Lauren Kick- Abidjan, Côte d’Ivoire. ham. 2017. “Accuracy of Energy-Use Surveys in Predicting Rural NREL (National Renewable Energy Laboratory) and E4I (Energy4Im- Mini-Grid User Consumption.” Energy for Sustainable Development pact). 2018. Productive Uses of Energy in African Micro-Grids: Tech- 41: (December): 88–105. https://www.sciencedirect.com/science/ nical and Business Considerations. Washington, DC: NREL. article/pii/S0973082617304350. ODI (Overseas Development Institute), GOGLA, Practical Action, and Cheney, Catherine. 2016. “How Alternative Credit Scoring Is Transform- Solar Aid. 2016. Accelerating Access to Electricity in Africa with Off- ing Lending in the Developing World.” Devex. https:/ /www.devex. Grid Solar: The Impact of Solar Household Solutions. London: ODI. com/news/how-alternative-credit-scoring-is-transforming-lend- https://cdn.odi.org/media/documents/10229.pdf. ing-in-the-developing-world-88487. Plutshack, V., J. Free, and R. Fetter. 2020. “Taking Electrification on CrossBoundary. 2020. “Study Design: Appliance Financing 3.0 Ener- the Road: Exploring the Impact of the Electric Farm Equipment gy-Efficient Productive Use.” https://www.crossboundary.com/wp- Roadshow.” Published online by the Rhodes Information Initiative content/uploads/2020/02/CrossBoundary-Innovation-Lab- at Duke University. Accessed June 12, 2022. https:/ /bigdata.duke. Study-Design-Appliance-Financing-3.0-Energy-Efficient- edu/projects/taking-electrification-road-exploring-impact-elec- PU-Anonymized-25-Feb-2020.pdf. tric-farm-equipment-roadshow. De Gouvello, Christophe, and Laurent Durix. 2008. Maximizing the Ramachandran, V., M. Shah, and T. Moss. 2018. “How Do African Firms Productive Uses of Electricity to Increase the Impact of Rural Elec- Respond to Unreliable Power? Exploring Firm Heterogeneity Using trification Programs: An Operational Methodology. ESMAP Formal K-Means Clustering.”CGD Working Paper 493, Center for Global Devel- Report 332/08, Energy Sector Management Assistance Program. opment,Washington,DC.https:/ /www.cgdev.org/sites/default/files/ Washington, DC: World Bank. https:/ /openknowledge.worldbank. how-do-african-firms-respond-unreliable-power-exploring-firm- org/handle/10986/17538. heterogeneity-using-k-means.pdf. Duke University. 2022. “Come Join the Electric Circus! What the US RMI (Rocky Mountain Institute). 2018. Closing the Circuit: Stimulating Rural Electrification Story Can Teach Us Today.” Website published End-Use Demand for Rural Electrification. Golden, CO: RMI. by the James E. Rogers Energy Access Project at Duke University. Accessed June 12, 2022. https:/ /energyaccess.duke.edu/come-join- Rogers, Everett. 2003. Diffusion of Innovations. 5th ed. New York: Simon the-electric-circus/. and Schuster. ISBN 978-0-7432-5823-4. Factor[e] Ventures. 2020. “Looking Beyond Appliances: Systemic USAID (U.S. Agency for International Development). 2018. “How Can Barriers to Minigrid Demand Stimulation.” https:/ /www.factore. Productive Uses Enhance the Economics of Mini-Grids?” USAID, com/systemic-barriers-to-stimulating-electricity-demand-in-afri- Washington, DC. https:/ /www.usaid.gov/energy/mini-grids/eco- can-off-grid-energy. nomics/productive-use/enhancement/. Finucane, James, Juliette Besnard, and Raluca Golumbeanu. 2021.  Village Infrastructure Angels. 2019. Fixed Amount Award (FAA): PAEGC “Raising Rural Productive Uses of Electricity: A Case Study of a Suc- Innovator Annual Report. Washington, DC: USAID. https:/ /pdf.usaid. cessful Utility-NGO Partnership in Indonesia.” Live Wire 2021/119,  gov/pdf_docs/PA00WGSS.pdf. World Bank, Washington, DC. https:/ /openknowledge.worldbank. World Bank. 2000. Implementation Completion Report on a Loan in org/handle/10986/36670. the Amount of US$398 Million to the Government of Indonesia for GIZ (Gesellschaft für Internationale Zusammenarbeit), BMZ (Federal the Second Rural Electrification Project. Washington, DC: World Ministry for Economic Cooperation and Development), ESMAP Bank. https://documents1.worldbank.org/curated/en/1208214680 (Energy Sector Management Assistance Program), and AEI (Africa 49833272/pdf/multi-page.pdf. Electrification Initiative). 2013. Productive Use of Energy—PRO- World Bank. 2020. Women, Business, and the Law. Washington, DC: DUSE: Measuring Impacts of Electrification on Small and Micro-En- World Bank. doi:10.1596/978-1-4648-1532-4; https://wbl.world- terprises in Sub-Saharan Africa. Eschborn and Bonn: GIZ. bank.org/. IDS (Institute of Development Studies) and GIZ (Gesellschaft für Internationale Zusammenarbeit). 2019. Unlocking the Benefits of Productive Uses of Energy for Women in Ghana, Tanzania and Myan- mar. Research report RA6. ENERGIA. https:/ /www.energia.org/ NOTES assets/2019/03/RA6-Unlocking-the-benefits-of-productive-us- es-of-energy.pdf. 1. Global Findex database: https://globalfindex.worldbank.org/. MINI GRIDS FOR HALF A BILLION PEOPLE    135 CHAPTER 4 ENGAGING COMMUNITIES AS VALUED CUSTOMERS CHAPTER OVERVIEW This chapter focuses on the role of community engagement in mini grid projects and the impact that inadequate or insufficient community engagement can have on their sustainability. It begins by underscoring the value of con- tinuous community engagement through every phase of the mini grid project—starting at the early design and planning stages, through financing, procurement, operation, and maintenance. It next delves into concrete steps that a mini grid developer can take at each project phase to engage the community. The chapter concludes with examples of innovative ways to scale up community engagement across multiple projects at the national and inter- national levels. WHY IS COMMUNITY ENGAGEMENT mann (2010) reveal that rural electrification programs ben- efit greatly from local participation. Involving communities IMPORTANT? from the start can help improve the design (Peru, Vietnam), As shown in previous chapters, mini grids offer a least-cost avoid disputes and gain local support (Bangladesh), mobi- option for providing reliable, affordable electricity to mil- lize contributions in cash or in kind (Nepal, Thailand), and lions of people now living without access, people who would increase local ownership, thus contributing to operational otherwise have to wait years for the main grid to arrive. sustainability. In line with this, a study covering mini grid projects of the Energy and Environment Partnership (EEP) However, bottlenecks must be cleared before mini grids Trust Fund in 13 countries in Sub-Saharan Africa concluded can take off on a large scale. Some, such as access to that building strong relationships in the community is crit- finance, workable regulations, and enabling business envi- ical to the financial sustainability of mini grids (EEP 2018). ronments, are systemic and best addressed at the national and international levels. Others are highly localized, such Community engagement is beneficial at all phases of a as the specific socioeconomic, cultural, and environmental mini grid project, as discussed below. characteristics of each community to be electrified with a DESIGN AND PLANNING: mini grid. A review of relevant literature and interviews with • As the primary customer of the mini grid, the local com- mini grid developers1 demonstrates that, to identify the munity should be the first to be consulted and engaged best technical solution for a given site and ensure its long- if the mini grid developer is to obtain the “social license” term sustainability, it is essential to continuously engage (that is, buy-in and acceptance from the community) to with the local community and determine the optimal fit operate the system. between the community and the new mini grid. • The local community has the best understanding of Existing evidence points to several reasons why community the surrounding conditions and resources (Mishra and engagement is essential for the successful implementation Sarangi 2016), which can help the mini grid developer of mini grid projects—and rural electrification more broadly. choose the optimal technology mix and operating Case studies reviewed by Crousillat, Hamilton, and Ant- model. 136   MINI GRIDS FOR HALF A BILLION PEOPLE Existing evidence highlights reasons why The central role of community engagement community engagement is essential for the in ensuring the long-term sustainability of successful implementation of mini grid projects: mini grids is further illustrated by mini grid failures improving the design (for example, in Peru, Viet- linked to insufficient local community involvement. nam); avoiding disputes and gaining local support Overlaying the common causes of mini grid failure (Bangladesh); mobilizing cash or in-kind contribu- with the typical stages of mini grid development tions (Nepal, Thailand); increasing local ownership indicates that (1) more than half of the causes of and operational sustainability; and ensuring finan- failure involve inadequate community engagement, cial sustainability (for example, in 13 Sub-Saharan and (2) community engagement plays a particularly African countries). decisive role in the stages that precede the com- mencement of operations. • During this early stage, community engagement will • In the case of a project backed by a public-private part- ease communication with local authorities, identifica- nership, transparency in the tender procedure will also tion of reliable and capable local (technical and sales) raise the community’s trust in the project. staff, and identification of prospective customers • In some instances, mini grids may also recruit commu- (including anchor clients and businesses). nity members to install the systems. • Early community engagement will facilitate land acqui- OPERATIONS AND MAINTENANCE: sition or right of use for the project, assist in obtaining good socioeconomic and cultural information from pro- • Community engagement will improve the communica- spective customers to underpin the business case of the tion between operator and customer, creating trust and project, and inform the community as well as the project mutual understanding about system repairs and other developer about expectations, requirements, roles, and matters. responsibilities. • Assuming that sufficient capacity has been developed • During this phase, a mini grid developer will use its con- within the community, the participation of locals in the tact with the community to accurately assess customer setup, maintenance, and repair of systems has been demand2 (including willingness and ability to pay), anecdotally shown to reduce the frequency of repairs compare the likely demand with prospective revenue (Fahey and others 2014) and increase the likelihood of streams (tariffs and subsidies), and reach conclusions more judicious use of systems by the community.3 about the project’s viability. • At this stage, the developer can also set up training • In addition, during this phase, community engage- sessions to stimulate demand and drive consumption ment will enable the developer to gauge residential among productive users. and business energy needs and expectations, and • In addition, mini grid developers often recruit local vil- to recognize the roles of men and women as energy lage agents to assist in sales and ensure rapid responses consumers and stakeholders. Raising awareness and to customer complaints and concerns. engaging in open dialogue will allow communities to The central role of community engagement in ensuring make well-informed decisions about their mini grid the long-term sustainability of mini grids is further illus- options, improving the quality of “service fit” (in terms trated by mini grid failures linked to insufficient local com- of situation, costs, payment modalities, business munity involvement. An extensive study of mini grids in Fiji case) and protecting the project’s viability by making found many of them failing as a result of various technical customer satisfaction more likely. It will also assist in and nontechnical issues (Dutt and Macgill 2013). Overlay- market segmentation. ing the causes of failure identified in the analysis with the FINANCING, PROCUREMENT, AND CONSTRUCTION: typical stages of mini grid development (as shown in fig- • The information obtained during the design and plan- ure 4.1) indicates that (1) more than half of the identified ning phase will yield data useful in developing the proj- causes of failure have a community engagement element ect’s tariff structure and improving its risk profile for (shown in the boxes outlined in red), and (2) community investors. engagement plays a particularly decisive role before oper- ation commences. MINI GRIDS FOR HALF A BILLION PEOPLE    137 In another comprehensive analysis, the main risks of mini Community engagement can thus be considered an essen- grid deployment were evaluated in terms of the potential tial component of the mini grid process during the design, impacts of the risk and the probability of the risk occur- implementation, operation, and maintenance phases. ring (Manetsgruber and others 2015). Proper community Based on existing evidence and interviews with mini grid engagement can significantly limit the exposure of mini developers, the next section will explore the principal steps grids to some of these risks (table 4.1). that a mini grid developer can take during each phase to engage the local community. FIGURE 4.1 • Typical issues hindering the mini grid development process Financial/ Implementation/ Operation Design Planning procurement construction Inadequate load Failure to consider local Inadequate and Inadequate transfer of skills to assessment needs and expectations inappropriate tariff local population (system upkeep, of the system structure productive use, payment, etc.) Inadequate risk analysis Inadequate technical evaluation of system performance Inadequate evaluation of project sustainability Component failure Inadequate evaluation of commercial and economic viability Inadequate technical evaluation of system performance Lack of consultation between suppliers and end users (services, land acquisition, etc.) Poor maintenance Inadequate evaluation Lack of budget to maintain the system of renewable source General lack of community participation Source: ESMAP analysis of Dutt and MacGill (2013). TABLE 4.1 • Potential for community engagement to limit mini grid risks Risk Community engagement aspect Impact Probability Nonpayment of electricity bills as a result of either Inadequate assessment of customers during High Medium inability or unwillingness to pay planning, limited social consensus regarding grid services, inadequate services Unpredictable electricity demand, negatively Poor feasibility assessment that fails to address Medium to Situational affecting project sizing and cost structure “A-B-C” considerations a; insufficient promotion of high services to productive users Insufficient social acceptance leading to poor Unfavorable public opinion, lack of transparency, Moderate Low to embedding of the project in the sociocultural context insufficient involvement and development of local medium (for example, social services, agricultural activities) capacity Theft (of materials for which there is a secondary Inability to create a “social compact”; insufficient Moderate Medium to market) and vandalism (particularly in cases of understanding of local power structures low conflict of interest among stakeholders) Defective operation resulting from Insufficient understanding of “A-B-C” expectations, Low to Medium miscommunication between business and needs, and limitations; insufficient understanding medium customer; conflicts of interest of local power structures Unfavorable mini grid regulations do not permit Insufficient understanding of “A-B-C” expectations, High High attractive tariffs (for either supplier or consumer) or needs, and limitations; insufficient consumer competition organization Policy and planning fail to sufficiently stimulate Insufficient market segmentation, limited High High productive and income-generating activities, productive-use awareness, and insufficient access undermining the viability of the investment to appliances for productive uses Source: ESMAP analysis of Manetsgruber and others (2015). In mini grid market segmentation, A-B-C refers to anchor, business, and community customers. a.  138   MINI GRIDS FOR HALF A BILLION PEOPLE COMMUNITY ENGAGEMENT DESIGN AND PLANNING PHASE THROUGHOUT THE MINI GRID During the design and planning phase, community engage- ment can PROJECT CYCLE • Establish a relationship between local authorities and Examples from the field indicate that successful com- the project developer munity engagement begins with raising awareness in the • Identify partners, local stakeholders and authorities, potentially connected community. It continues during and local staff adoption and productive operation, as satisfied custom- ers promote the technology to their neighbors and friends. • Improve community confidence in the project developer Each community requires a flexible approach, with a clear or service provider understanding of the local socioeconomic and cultural • Provide insight into local expectations of the technology characteristics, and a potential tailoring of the promotional and services tools, materials, and channels.4 The typical mini grid project • Improve the assessment of demand and load cycle, outlined in figure 4.2, maps these steps. • Create awareness of the technology and the potential uses of electricity, and develop a portfolio of prospective customers Successful community engagement begins • Assist with segmentation of the market into A-B-C cus- with raising awareness in the potentially tomers connected community. It continues during adop- • Profile (and segment) potential customers or arche- tion and productive operation, as satisfied cus- types and forecast their ability to pay (ATP) and willing- tomers promote the technology to their neighbors ness to pay (WTP). and friends. Each community requires a flexible In this initial phase, community engagement is preceded by approach, with a clear understanding of the local a preliminary market assessment, and public data are ana- socioeconomic and cultural characteristics, and a lyzed by the developer to identify potential mini grid sites. If potential tailoring of the promotional tools, materi- the developer is new to the country, this exercise is also likely als, and channels. to include an evaluation of national regulatory, policy, politi- cal, economic, financial, and environmental considerations. FIGURE 4.2 • Typical mini grid project cycle DESIGN AND 1 PLANNING OPERATION & PROMOTION / 5 MAINTENANCE 2 INFORMATION FINANCING / 4 IMPLEMENTATION / 3 PROCUREMENT CONSTRUCTION MINI GRIDS FOR HALF A BILLION PEOPLE    139 When a site or a group of potential sites has been identified, Customer profile the developer will typically conduct a site visit to assess Where possible, the information gathered during an initial conditions on the ground, accompanied by a brief survey survey of potential customers is supplemented with the of residents and contacts with the local government. This findings of a domestic energy baseline study and, in a later step—outreach to the local authorities and, in particu- stage, user surveys. The surveys inform a detailed customer lar, village chiefs—is emphasized by developers as key to profile that ensures prospective customers are effectively demonstrating respect and securing the favorable atten- served. In practice, these profiles are further differenti- tion of the village residents.5 Once communication with the ated—by type of farming, type of enterprise, and so on. local village representatives is underway, the developer can The customer profile prepared by the developer based go on to sign a memorandum of understanding with the on the initial data gathering can be used to inform a gen- local authorities, discuss land acquisition or rights of use, der-specific approach to connecting customers. In such and pursue any subnational government approvals that an approach, the cultural and socioeconomic nuances and may be required. potential barriers to women’s participation are taken into Following this, the developer will typically dispatch a full account. customer survey team to carry out household-size surveys using tablets and standardized survey instruments. For Segmenting the market into A-B-C customers larger developers that operate in several markets, these An additional approach to segmenting the market further— surveys are often adapted to the local context.6 The objec- into A-B-C customers—may enhance the robustness and tive is to gain a deep understanding of how to satisfy cus- viability of a mini grid. The A-B-C approach is a business tomers’ needs. Good surveys will help developers gain an approach rather than a community engagement method- understanding of six areas that ultimately affect the satis- ology, but it can offer added value in helping the mini grid faction of their customers. developer identify A, B, and C client segments through proper community engagement. Customer satisfaction is directly influenced by four con- siderations, all of which can be addressed in the survey The A-B-C business model (figure 4.3) was first developed questions: by companies such as OMC Power in India, which built mini grid companies around anchor clients (GIZ 2014). • Livelihood: The potential role of mini grid electricity in Under the model, the supply of electricity from the mini domestic or commercial activities and the effect the grid is prioritized for an anchor load customer, typically services may have on employment, entrepreneurship, a commercial or industrial user, followed by businesses and quality of life (shops, small enterprises, and so forth), and then house- • Influences on the mini grid’s activities, including sea- holds. The A-B-C model stipulates that the size of the sonal labor migration; national and regional media; mar- generation unit is typically determined primarily by the ket effects of nearby major towns (particularly if those demand needs of the anchor load (Bhati and Singh 2018), towns are connected to the national grid); and existing with the aim of securing stable and predictable revenues. marketing campaigns, including those developed by the OMC and other mini grid operators in India have used government telecommunications towers close to population centers • Aspirations of prospective consumers, including educa- to scale up mini grid systems and expand into surround- tion, careers, and communication • Consumption: Need-based or additional con- FIGURE 4.3 • The A-B-C model sumption, return on investment, purchasing behavior. Community: Community Low electricity demand, mostly for These considerations are in turn influenced by lighting, mobile phone charging and household appliances • Community: The role of the family, gender, and community institutions Business Businesses: Higher electricity demand for • Infrastructure: Energy and appliance mar- productive use kets; available technology, the existing infra- Anchor structure for power and information and Anchor: Financially sound, guarantees electricity communications technology, and the avail- purchase, secures commercial operation ability of finance and services (health care, education). Source: GIZ 2014. 140   MINI GRIDS FOR HALF A BILLION PEOPLE ing businesses and households. Agroprocessing and min- consume the lion’s share of the energy generated during ing activities have also frequently been used in Africa to the day, whereas household consumption tends to inten- provide reliable loads. sify during the hours before and after work. A study of mini grid projects in 13 countries in Sub-Saha- PROMOTION AND INFORMATION-SHARING ran Africa (EEP 2018) finds that “the most financially sus- PHASE tainable mini grids use an A-B-C strategy: first, identify and During the promotion and information-sharing phase, negotiate an agreement with an anchor load client (often in community engagement can agroprocessing); then identify, or help develop, small local businesses; and only last target domestic consumers.” Or, • Improve the information on mini grid services shared as Kennas and Barnett (2000) put it in their microhydro with the community assessment: “It is easier to make a profitable micro-hydro • Fine-tune customer segmentation plant socially beneficial than to make a socially beneficial plant profitable.” • Leverage women’s social and trust networks to spread information Growing evidence (discussed in detail in chapter 3) sug- gests the need to increase the prioritization of small and • Increase consumers’“energy education” on matters such medium-size enterprises—the B of the A-B-C model—as as understanding energy bills, performing maintenance, income streams for the mini grid. This is largely because handling grievances, and maximizing safety and health securing anchor clients can be a challenge, particularly • Select, record, and share early successful adoption in rural economies based on subsistence agriculture and stories. small-scale artisan networks. In addition, anchor clients The objective of this phase is to increase awareness of may have unrealistic expectations about tariffs, and oper- the technology, its services, and the facilities offered by ators may become overreliant or even dependent on them, the developer. Community engagement activities in this threatening their financial position. phase traditionally target fairs, radio, word of mouth, local Some developers have been combining multiple business extension services, construction activities, and—increas- clients to act as the anchor. In this approach, the mini ingly—social media. Marketing hubs are an additional grid’s generation assets are sited near a cluster of small engagement “tool,” as described below. In addition, through businesses that may be housed in the same building. billboards, social media, radio, and television, projects can This approach allows for mitigation of commercial risks target family members who migrated to urban centers, as and diversification of the mini grid customer base. A good they are often interested in helping parents gain access to example is the community engagement, load acquisition electricity. and micro-enterprise development approach (CELAMeD), To build up their presence in the community after the ini- which has been promoted by Smart Power India and imple- tial site visits and meetings, mini grid developers often mented by several energy service companies in India— cohost workshops for community members with mem- notably TARA Urja. CELAMeD is discussed in further detail bers of relevant cooperatives, nongovernmental organi- in the next section. zations (NGOs), microfinance institutions, and financial Including both anchor and business clients will enable the institution workshops for community members. These project developer to connect households at affordable workshops can be used to build awareness of (1) oppor- rates, because these clients (1) may cross-subsidize con- tunities to connect income-generating and household nection and consumption fees for households and (2) will appliances and machines to the mini grid; (2) agricultural applications of mini grid electricity; (3) financing available for the purchase of household, agricultural, and commer- Segmenting the customer base into anchor, cial appliances and machines; and (4) entrepreneurship business, and community customers can and business training. enhance the robustness and viability of a mini grid. The A-B-C approach is a business technique rather Establishing marketing hubs than a community engagement methodology, but it For the community engagement process to be success- can offer added value in helping the mini grid devel- ful, the underlying communication platform must have oper identify client segments through proper com- sufficient reach and trust within the community. Some munity engagement. established mini grid companies have developed commu- nity engagement practices as a core capability. However, new entrants to the market are likely to have insufficient MINI GRIDS FOR HALF A BILLION PEOPLE    141 resources and skills to properly develop a robust commu- Information should also be provided to local leaders, staff nity engagement process. of value chain organizations, and financial institutions, as they are often consulted about innovative technologies and Marketing hubs are one solution that can help mini grid practices. startups engage with their communities. These local plat- forms strengthen consumer confidence and allow for com- Engaging women as leaders of promotional and bined community engagement efforts between the mini information-sharing activities grid developer and organizations over the entire project Energy service companies are increasingly engaging cycle. Strong candidates for marketing hubs are entities women in promoting their service. Women are well aware that perform local or regional network functions, exhibit of the challenges that other women face, such as unique entrepreneurial drive, and have a natural affinity with mini time constraints, household responsibilities, and energy grid or off-grid electricity services. needs. Their access to other potential female customers Examples include savings and credit cooperatives, dairy may be significantly less constrained by social or cultural or agricultural cooperatives, agricultural input dealers, norms than male sales agents in certain cultural contexts. and rural development NGOs. The project developer may However, despite women’s potential to build robust distri- support the marketing hubs with capacity building, a ded- bution networks for energy solutions, especially in rural icated liaison or extension officer, and incentives for sales areas, they are still underrepresented in the sector. and extension activities. One successful example of engaging women in promot- The advantages of such hubs include lower operating costs ing energy access is Solar Sister, an African initiative that (promotion, training, extension), access to a more targeted recruits, trains, and supports female entrepreneurs to market, and knowledge sharing. In addition, hubs often serve as last-mile distributors of clean energy products, perform initial client eligibility screenings and may serve such as solar lights, mobile phone chargers, and clean as a physical location where microfinance institutions and cookstoves. By 2012, Solar Sister had empowered 2,000 other local financiers can meet their potential (or current) female entrepreneurs in Uganda, Nigeria, and Tanzania, customers. The hub-hosting entity, on the other hand, ben- who in turn provided solar and clean cooking solutions to efits from being able to offer an additional service to its more than 370,000 beneficiaries (Arc Finance 2012). members. FINANCING AND PROCUREMENT PHASE Experience with marketing hubs from other sectors During the financing phase, community engagement can, shows that a hub’s activities must fit into its regular on the demand side: agenda, and the hub should not be converted into a “sales machine” for mini grid developers, as this erodes • Assist in ATP/WTP assessments and the development trust within the local community. When set up properly, of a viable demand forecast marketing hubs can instead develop into “marketing • Organize in-kind contributions (for example, construc- beachheads” (Moore 1999), preparing the market for tion, at both individual and community levels) uptake by early adopters. • Improve consumer access to energy and credit for pro- Sharing information with prospective customers ductive-use appliances. After prospective customers have a basic awareness of the technology, they will require more detailed information to properly assess its usefulness to them. It is important to Energy service companies are relying ensure that the information provided is correct, unbiased, increasingly on women to promote their and comprehensive (addressing the merits, costs, and lim- service. Women are close to their female custom- itations of the technology). ers, can more easily tap into social networks, and are well aware of the challenges that other women The information will have to be provided by well-quali- face, such as unique time constraints, household fied project staff. At this stage, information can be shared responsibilities, and energy needs. Their access to either individually or in smaller functional groups (for other potential female customers may be signifi- example, hosted by local authorities, microfinance institu- cantly less constrained by social or cultural norms tions, NGOs, or other community stakeholders). In addi- than male sales agents in certain cultural contexts. tion, exchange visits to existing installations, where peers can share their experience, can be powerful sources of information. 142   MINI GRIDS FOR HALF A BILLION PEOPLE On the supply side: • Facilitate the registration and connection process • Improve the risk perspective of the project (detailed • Build consumers’ capacity to contribute to O&M and portfolio, WTP/ATP assessment, A-B-C segmentation, safety reliable load assessment). • Stimulate community dialogue on additional or improv- Adequate financing is key to the sustainability of a mini ed services grid. Customer financing plays an important role in this • Encourage productive uses of energy area for households and enterprises alike. It allows house- holds to spread the costs of the connection over time and • Mediate disagreements among consumers, project per- to finance their domestic and productive appliances. Inno- sonnel, and local authorities vative credit arrangements, such as pay-as-you-go and • When local community members are engaged as part of lease-to-own, are showing their potential in the scale-up of the O&M team for the mini grid, activities will focus first mini grid uptake and in enabling customers to gain access on ensuring that local staff have a proper understand- to appliances and equipment. ing of how to operate and maintain the mini grid system safely and in a way that ensures good customer service. Mini grid developers may offer such financing options Training sessions are usually held before and immedi- to customers directly, often through the support of bilat- ately after the installation of the system. eral or multilateral development partners. Support from the Rockefeller Foundation, for example, allows Nigeria’s Many households and businesses will also require ongoing Havenhill Synergy to offer household and productive-use support to make the most of the electricity provided by the appliance financing to customers with 20 percent down mini grid. From a commercial perspective, activities during and a 12-month payback period.7 the O&M phase focus on signing up new customers and ensuring growth in energy consumption to achieve the tar- In addition, the developer’s staff and partners may help gets set to ensure financial solvency. farmers and businesses during the loan application pro- cess and facilitate the relationship between the appliance One example of such demand stimulation is the intensive supplier, the lender, and the customer. customer awareness campaign launched in 2017 by the Bangladesh Infrastructure Development Company Lim- IMPLEMENTATION AND CONSTRUCTION PHASE ited (IDCOL) to support customer uptake and demand During the implementation and construction phase, com- growth for 20 solar mini grids. The campaigns combined munity engagement can customer training with public events, such as folk songs, shows, and street theater. The comprehensive efforts have • Involve future stakeholders and recruit local human resulted in a five-fold increase in customers. Based on the resources for construction and installation jobs success of the campaign, IDCOL is now developing train- • Use the construction and installation processes as ing content focused on women, specifically highlighting opportunities to promote the new technology to the the benefits of social facilities, street lighting, schools and surrounding community clinics, rickshaw charging stations, and so forth, to enable local entrepreneurs to see the broader picture and make • The construction phase provides an excellent opportu- linkages between community benefits and the sustainabil- nity for the developer to connect with prospective cus- ity of their own businesses. A more detailed example of a tomers. The developer, who may already be joined by the government-supported community engagement initiative local sales team, local leaders (including from marketing is described on the companion website to this handbook: hubs), and possibly representatives from the national or www.esmap.org/mini_grids_for_half_a_billion_people. local government, can invite local households and busi- nesses to the construction site for the commissioning Another community engagement tool that developers of the mini grid, thus allowing local residents to observe can leverage during the O&M phase is customer service the system, ask questions, and sign up for service. If centers (CSCs) (SNV 2008). These centers can be both local community members were involved in building the virtual—with customers accessing the CSC by phone or system, they can be trained to engage with the audience online—and in person. The primary tasks of the CSC are on the expected benefits and purposes of the electricity to track mini grid functionality, address customer issues, produced by the system. and both track and increase overall customer satisfaction. CSCs will flag issues and relay them to the developers—for OPERATIONS AND MAINTENANCE PHASE example, a string of customer complaints about the tariff During the operations and maintenance phase, community may result in the developer holding a workshop to explain engagement can tariff options. CSCs can also relay customer questions and MINI GRIDS FOR HALF A BILLION PEOPLE    143 issues to other relevant entities. For example, the CSC can IMPORTANT GENDER ASPECTS OF connect customers with local suppliers of household and productive-use appliances and can route customer con- THE COMMUNITY ENGAGEMENT cerns about lack of financing for income-generating appli- PROCESS ances to a local microfinance institution. Women are increasingly important customers for mini CSCs may be established in-house by the developer or grids, both at the household and enterprise levels, particu- operated by a community organization or business. In larly in Sub-Saharan Africa, where female entrepreneurship addition, and increasingly, CSCs play a role in capturing rates are estimated to be the highest in the world (GEM the potential demand for mini grid electricity that exists 2017). Inclusive community engagement during A-B-C cli- in the community. The CSC will follow up with prospective ent scoping can identify these key clients. customers identified in earlier phases of the mini grid’s To gain insight into women’s roles and energy needs in a development. given community, it is important to understand the under- SUMMARY OF COMMUNITY ENGAGEMENT OVER lying gender norms and the financial, educational, and THE PROJECT CYCLE social patterns that may hinder their access to and full par- ticipation in the mini grid project cycle. Table 4.2 tracks the community engagement objectives, channels, targets, actors, and activities over the mini grid’s Men and women often have different experiences and entire project cycle. perceptions of community assets, such as mini grids, and TABLE 4.2 • Key community engagement activities over the project cycle Design and Promotion and Financing and Implementation Registration Operations and planning information procurement and construction and connection maintenance Obtain prospective Increase community- Establish a clear Construct a robust, Secure a viable Ensure reliability of user and situational level awareness of financial model to appropriately sized, customer base services, maximizing information energy access issues determine ideal and scalable mini with (fairly) benefits of investment Customer expectations Provide relevant public/private-sector grid secured power Ensure functionality and information (for investment needs Include provisions load domestic and productive Current energy use example, potential Prepare a project for potential future Identify anchor benefits (potentially Geographic conditions productive uses) and document showing expansion (for customers offering household OBJECTIVES Prepare a mini grid raise active interest to financial structuring, example, modular in advance and productive-use design that is: promote early buy-in risk assessment construction) and secure appliances for sale or Socially inclusive (in its and sign-up and mitigation, Install simple control connections lease) tariff structure and in Assist prospective financing partners, systems to manage Strengthen and meeting expectations) customers in their and adherence to demand levels incentivize local service registration decision regulations capacity Technically sound (load distribution and likely with specific, detailed Carry out procurement Establish a cost-effective future demand) information (signup, based on proper maintenance program, tariff, payment tender documents and with local service Financially viable (tariff methods, etc.) a transparent process covers cost) providers engaged Community Community members Financial institutions, Experienced service Households Households TARGET GROUPS Local leaders and Productive users investors, local and providers Local businesses Local businesses and authorities national government Local businesses and and startups startups agencies laborers Financial institutions Operating Operating companies Local and national companies government agencies Meetings with Campaign Meetings with local Community ICT for ICT for connections local institutions, and national finance meetings connections CHANNELS Local radio, ESCO office stakeholders, and newspaper, social institutions and local Social media, community government CSC media, brochures user video Community meetings CSC Marketing hubs Project developer, ESCO Project developer, Project developer, Project developer, ESCO manager ESCO manager and staff, CE manager and ESCO manager, CE ESCO manager, local ESCO manager, local and staff, CE technical staff ACTORS team team banks technicians and manager and CE manager Local authorities, lead Local media laborers team community members continued 144   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 4.2, continued Design and Promotion and Financing and Implementation Registration Operations and planning information procurement and construction and connection maintenance Carry out a community Carry out a promotion Finalize WTP/ATP Recruit local human Continue Continue community scan to assess WTP/ and information assessments resources community dialogue to monitor ATP and expectations campaign regarding Finalize market Engage with future dialogue to customer satisfaction of potential customers, mini grid services segmentation stakeholders stimulate Promote PUE extension particularly if the A-B-C Increase consumers’ (A-B-C customers) (A-B-C clients, local inclusion approach is followed Prepare impact “energy education” in Finalize tariff structure authorities) Carry out a PUE documentation Ensure women’s understanding their Offer gender- inventory (including video) Mobilize women’s COMMUNITY ENGAGEMENT participation in all bill, maintenance, differentiated user Carry out a design and planning grievance handling, finance networks Carry out O&M training to make available training on O&M community activities safety aspects, and and productive-use service inventory Mediate conflicts as health benefits investment financing needed Engage on local permits for female-run opportunities and approvals Finalize customer households and Develop a client portfolio segmentation businesses (A-B-C segmentation) Prepare and present Inventory local human impact videos resources Attract, organize, and Engage on land include women in acquisition or right of promotional and sales use activities Mobilize women’s networks to disseminate mini grid information Note: The term ESCO (energy service company) is analogous to the term “developer” in the chapter text. A-B-C = anchor-business-community; CE = community engagement; CSC = customer service center; ICT = information and communications technolo- gy; O&M = operations and maintenance; PUE = productive uses of energy; WTP/ATP = willingness to pay/ability to pay. their benefits are not equally distributed. In many con- texts, women are excluded from community consultations Men and women often have different expe- because of cultural norms whereby men assume leader- riences and perceptions of community ship, women are uncomfortable speaking in front of men, assets, such as mini grids, and their benefits are not or household and care responsibilities constrain women’s equally distributed. In many contexts, women are time and ability to attend. Working to address these barri- excluded from community consultations because ers is key to ensuring a more equitable division of benefits of cultural norms whereby men assume leader- and viable loads for developers. In the case of communi- ship, women are uncomfortable speaking in front ty-owned mini grids in India, enabling women’s participa- of men, or household and care responsibilities con- tion has improved governance processes and equity (Katre, strain women’s time and ability to attend. Working Tozzi, and Bhattacharyya 2019). Mini grid operators in India to address these barriers is key to ensuring a more keenly distinguish gender aspects in customer behavior equitable division of benefits and viable loads for and provide loans to support productive uses of mini grid mini grid developers. electricity for women’s income-generating activities (Katre, Tozzi, and Bhattacharyya 2019). Accounting for the gender aspects of the community more likely to gain employment outside the home following engagement process is important because mini grids can electrification (Smith 2000). The time savings delivered by transform female clients’ lives, as discussed below. electric power and the ability to perform domestic chores Drudgery and time savings. Mini grids can significantly in the evening frees women to accept paid work. Findings reduce women’s drudgery and save them time, particu- from electrification programs in Bangladesh, however, sug- larly in female-dominated labor-intensive agricultural and gest that the increase in time devoted to wage labor may food processing activities, through the use of electrical not result in a decrease in unpaid work in the home, causing appliances such as water pumps, grinders, mills, blenders, women to work longer hours overall (Magis 2010). refrigeration, and, in a few cases, electric stoves. Health and well-being. Electrification can reduce fertil- Employment. Studies in South Africa, Nicaragua, and ity levels. Greater exposure to television often improves Guatemala show that women are 9–23 percentage points access to information and may depict new norms, such as MINI GRIDS FOR HALF A BILLION PEOPLE    145 ble with conducting high-quality community engagement activities? The following two examples are innovations Incorporating gender in community engage- from the frontier of community engagement efforts—the ment is important because mini grids can first at the national level in India; the second at the develop- transform women’s lives by reducing drudgery, sav- er’s level, also in India. ing time, opening opportunities for employment, bettering health and well-being, and improving COUNTRY-LEVEL PROGRAM: SMART POWER intergenerational outcomes for families. INDIA Smart Power India was launched by the Rockefeller Foun- dation in 2015 as part of its $75 million Smart Power for family planning and smaller family size (Buckley 2012). In Rural Development initiative, focusing on the states of Indonesia, exposure to television increased the use of mod- Uttar Pradesh, Bihar, and Jharkhand. With a goal of bring- ern contraception by approximately 12 percent (Wiemann, ing electricity to a thousand villages and a million Indians Rolland, and Glania 2014). Anecdotal evidence from West (Muther 2016), the program has supported the installation Africa suggests that electrifying clinics for lighting and of more than 100 mini grids, affecting more than 40,000 refrigeration of medications has an especially beneficial people (Rockefeller Foundation 2017). impact on maternal health. The program’s approach was to establish Smart Power Gender norms and women’s agency. Evidence from several India as an India-based, Indian-led organization that would states in India highlights the role television plays in decreas- play an intermediary role among key stakeholder part- ing domestic violence and preference for sons. Women also ners—the private sector, national and local governments, report increased autonomy, measured by such factors as community organizations, and developers such as OMC, the ability to go out without permission and participation in Husk Power, TARA Urja, and DESI Power—thereby develop- household decision making (Wiemann, Rolland, and Glania ing a harmonious ecosystem to drive sustainable mini grid 2014). Mini grid projects can also shape new community installations at scale through comprehensive community decision-making and leadership models. If local electrifi- engagement (Rockefeller Foundation 2017). cation committees give women equal opportunities to run As part of its larger design, Smart Power India has used for key positions, for example, their voice in decision-mak- the CELAMeD approach, which involves close engagement ing may be enhanced. Another example is provided by the with local communities and partners to not only secure Nigerian Rural Electrification Agency, which requires that strong buy-in and sustained demand for electricity, but women be included in local decision-making bodies set up also to stimulate economic development in the villages. for the implementation of mini grids. The agency also hosts Through work with partners such as the Society for Tech- women-only consultations. nology and Action for Rural Development (TARA) and Intergenerational effects on education and health. Some People’s Action for National Integration, developers have evidence indicates that the improved lighting provided by developed communication and marketing strategies to electrification can promote more hours of study among inform consumers about the benefits of renewable energy children. In Bangladesh, years of schooling for both boys and enable rural businesses to emerge and expand (Smart and girls rose after electrification (Magis 2010). Reports Power India 2017). from West Africa also suggest that lighting and refrigera- The aim of the CELAMeD approach is to mobilize com- tion in clinics improve child health. Both effects shape inter- munities around the mini grid, helping developers exploit generational outcomes for families. latent demand and induce new businesses to use its elec- tricity (Smart Power India 2016). Some community engagement activities begin as soon REMOVING BARRIERS TO SCALE as villages are identified by the developer. These include THROUGH INNOVATIONS IN focus group discussions, door-to-door contacts, street COMMUNITY ENGAGEMENT plays, fairs, and community meetings. In the first four to six weeks following installation of the system, the developer’s One of the primary challenges of community engage- sales team works with the CELAMeD team to sign up cus- ment is that it requires boots on the ground—a significant, tomers (Smart Power India 2017). Initially, the vast majority human presence in the community served by the mini grid. of the connections are for households and small shops, pri- But developers are increasingly building large portfolios of marily for small load needs, such as lighting and cell phone mini grids to achieve economies of scale. Is this compati- charging. A subsequent effort is made to engage existing 146   MINI GRIDS FOR HALF A BILLION PEOPLE businesses to persuade them to switch from self-genera- on Mini Grids has piloted a novel approach to community tion using diesel with the prospect of gaining expanded ser- engagement. Called Mini Grid Stories, this pilot supported vices at lower cost. the creation of a digital community for mini grid stakehold- ers. Piloted in India with support from Quicksand Design One of the key principles advocated by the Smart Power Studio, the program used a YouTube-like online video plat- India program is that village residents are treated not as form and accompanying smartphone application to eas- “consumers,” but rather as “producers,” with needs beyond ily create, upload, download, and share videos. Together, simple lighting solutions (Smart Power India 2016). With they were intended to enable developers, households, this approach, the program works to create a thriving businesses, microfinance institutions, NGOs, and other ecosystem by catalyzing entrepreneurship through pro- groups to share customers’ experiences as a means of motional events, training, and capacity building. During community engagement across the development phases the first month, the developer’s team typically focuses of a mini grid project. on preplanning activities, such as site scoping, identifica- tion of focus areas, and preparation of tailored enterprise In this pilot program, mini grid developers, their small packages. It spends the next three months identifying busi- business and household customers, and other commu- nesses and entrepreneurs—both existing and potential— nity members used the smartphone application to record for contacts and promotional activities. One example of the short videos. The application guided users as they created latter is the Expansion Mela, an event carried out locally to their mini grid story—reporting, for example, the impact ramp up adoption of appliances (Smart Power India 2016). the electricity from the mini grid has had on their lives. The event promotes on-the-spot registration for electricity Once uploaded to the online platform, the video could be services for interested participants by offering discounted viewed online or downloaded for viewing offline and could tariff packages and lottery tickets for additional subsidies. be used by mini grid developers, marketing hubs, and CSCs to showcase the transformative potential of mini grids to To support new entrepreneurs, TARA Urja uses Start and prospective clients. Improve Your Business training programs, which have been certified by the International Labour Organization. The The approach was inspired by the success of a similar pro- programs are focused on identifying entrepreneurs with gram focused on peer-to-peer learning for farmers set up high chances of success; they also have a specific focus by Digital Green, a global development organization. Digi- on women. The programs include Generate Your Business tal Green’s agricultural web platform amplifies the existing training for novice entrepreneurs looking to flesh out a informal networks of farmers, extension providers, and business idea and Improve Your Business training for exist- markets with digital tools that transform development from ing businesses looking to scale up. the bottom up by raising local voices. An ongoing two-year randomized control trial conducted by the International By 2016, TARA Urja had trained more than 140 potential and Food Policy Research Institute in Ethiopia has yielded some existing entrepreneurs, of whom 30 percent were female; promising preliminary results: 40 percent had set up new businesses, and 20 percent had expanded their existing businesses (Smart Power India • Digital Green’s video-enabled approach reaches 24 per- 2016). As part of the program, TARA Urja offers support for cent more farmers than the Ministry of Agriculture’s the purchase of productive equipment, often customizing conventional approach and results in 35 percent higher tariff packages to bundle in equipment costs. By 2016, the uptake of promoted practices. program had enabled 30 existing entrepreneurs (includ- • The inclusive approach increases women’s access to ing small computer shops, printing businesses, grocery extension information by 20–25 percent. shops, mobile repair shops, and photo studios) to expand • Public extension agents who use the video approach their businesses, reporting overall increases in incomes of make a greater effort to visit farms, inspect technology approximately 15–20 percent (Smart Power India 2016). use, and provide follow-up advice to farmers than those With this program, the developer saw returns on its invest- who use the conventional extension approach. ment within approximately five to six months (Smart Power India 2016). A separate controlled evaluation found the approach to be 10 times more cost-efficient than traditional community ICT FOR COMMUNITY ENGAGEMENT AT THE engagement services on a cost-per-adoption basis. Key DEVELOPER LEVEL: A PILOT VIDEO HUB success factors of this approach include highly localized DELIVERING MINI GRID STORIES content, human mediation to reinforce key messages, and As part of its efforts to mainstream new solutions and best capacity building that strengthens service provision. practices into programs to scale up mini grids, the Energy Translating this innovative approach into the mini grid sec- Sector Management Assistance Program’s Global Facility tor, such as ESMAP’s Mini Grid Stories pilot, is likely to have MINI GRIDS FOR HALF A BILLION PEOPLE    147 similarly transformative impacts if implemented at scale. SAS), Trivandrum, Kerala, India, August 23–24. https://ieeexplore. ieee.org/document/6629898/. A detailed description of the Digital Green initiative is pro- EEP (Energy and Environment Partnership Trust Fund). 2018. Oppor- vided on the companion website of this handbook: www. tunities and Challenges in the Mini Grid Sector in Africa; Lessons esmap.org/mini_grids_for_half_a_billion_people. Learned from the EEP Portfolio. South Africa: EEP. https://eepafrica. org/wp-content/uploads/2019/11/EEP_MiniGrids_Study_Digi- talVersion.pdf. Fahey, Alison, C. C. Huang, Chikara Onda, Hana Scheetz Freymiller, Jac- CONCLUSION queline Wong, Nathan Ratledge, Simonne Li, Stephanie Speirs, Ste- phen Moilanen, and Venu Aggarwal. 2014. Rural Energy Alternatives Insufficient community engagement has been identified as in India: Opportunities in Financing and Community Engagement one of the major limitations of many energy access proj- for Renewable Energy Microgrid Projects. Woodrow Wilson School ects (in particular mini grid projects), as the work requires of Public Policy. Princeton, NJ: Princeton University Press. http:/ / wws.princeton.edu/sites/default/files/content/591f%20Rural%20 a substantial commitment to educate, train, and build the Energy%20Alternatives%20in%20India.pdf. trust of communities (Valencia and Caspary 2008). Even GIZ (Gesellschaft für Internationale Zusammenarbeit). 2014. “The when communities are involved in day-to-day operations ABC-Model: Anchor Customers as Core Clients for Mini Grids in and management of mini grid systems, insufficient capac- Emerging Economies.” https:/ /www.german-energy-solutions.de/ ity building and training may inhibit a project’s sustainabil- GES/Redaktion/DE/Publikationen/Praesentationen/2015/2015- ity and impact. 03-19-iv-mini-grids-05-giz.pdf?__blob=publicationFile&v=7. GEM (Global Entrepreneurship Monitor). 2017. Women’s Entrepre- This chapter explored the importance of involving local neurship 2016/2017 Report. https:/ /www.gemconsortium.org/ stakeholders in all phases of setting up and operating a report/49860. mini grid. The chapter highlighted how local stakeholders Katre, Aparna, Arianna Tozzi, and Subhes Bhattacharyya. 2019. and authorities can help assess electricity needs, support “Sustainability of Community-Owned Mini-Grids: Evidence from project monitoring, help organize the community, develop India.”  Energy, Sustainability and Society 9 (1): 2. https://link. springer.com/article/10.1186/s13705-018-0185-9. local productive enterprises or value-added activities, stim- Khennas, Smail, and Andrew Barnett. 2000. “Best Practices for Sus- ulate demand, and enable women to reap equal benefits tainable Development of Micro-Hydro Projects.” ESMAP Technical from the new source of electricity. Based on examples from Paper 006, World Bank, Washington, DC. https:/ /openknowledge. the field, the chapter showed that when communities feel worldbank.org/handle/10986/20314?show=full. a sense of ownership and all local stakeholders are suffi- Magis, Risten. 2010. “Community Resilience: An Indicator of Social ciently involved in setting up, maintaining, and repairing the Sustainability.”  Society & Natural Resources 23 (5): 40–416. Doi: 1 new systems, more customers tend to become connected, 0.1080/08941920903305674. https:/ /www.tandfonline.com/doi/ the level of consumption tends to increase, and the mini abs/10.1080/08941920903305674?journalCode=usnr20. grid’s O&M costs tend to be lower. Manetsgruber, David, Bernard Wagemann, Bozhil Kondev, and Katrin Dziergwa. 2015. Risk Management for Mini Grids: A New Approach to Guide Mini Grid Deployment. Brussels: Alliance for Rural Electrifi- cation. https://www.ruralelec.org/sites/default/files/risk_manage- REFERENCES ment_for_mini-grids_2015_final_web_0.pdf. Mishra, Arabinda, and Gopal Krishna Sarangi. 2016. “Off-Grid Energy Arc Finance. 2012. “Solar Sister’s Energy Consignment Model.” Development in India: An Approach towards Sustainability.” Eco- https://docplayer.net/13062422-Solar-sister-s-energy-consign- nomic and Political Weekly 51 (22, May). https://www.epw.in/jour- ment-model.html nal/2016/22/special-articles/grid-energy-development-india.html. Bhati, Priyavrat, and Mandvi Singh. 2018. Mini-Grids in Uttar Pradesh: Moore, Geoffrey. 1999. Crossing the Chasm. Marketing and Selling High- Policy Lessons. New Delhi: Centre for Science and Environment. Tech Products to Mainstream Customers. New York: Harper Collins http:/ /cdn.cseindia.org/attachments/0.67714200_1530692543_ Publishers. https://books.google.com/books/about/Crossing_the_ Mini-grids-in-Uttar-Pradesh_Status-and-Policy-Evaluation_CSE. Chasm.html?id=cj7EQgAACAAJ. pdf. Muther, Kyle. 2016. Smart Power for Rural Development: Creating a Buckley, Andrew. 2012. “Best Practice Community Engagement for In- Sustainable Market Solution to Energy Poverty. Shared Value Ini- frastructure Projects: Building Community Ties That Dig Deeper.” tiative. https://dokumen.tips/documents/smart-power-for-rural- Public Infrastructure Bulletin 1 (8): article 3.Crousillat, Enrique, Rich- development-sun-connect-news-2019-04-24-smart-power-for. ard Hamilton, and Pedro Antmann. 2010. “Addressing the Electri- html?page=17 cityAccessGap.”WorldBank,Washington,DC.https:/ /openknowledge. Rockefeller Foundation. 2017. “Smart Power for Rural Development worldbank.org/bitstream/handle/10986/12530/690620ESW0 Brochure: Transforming Lives through Energy Access.” https://www. whit0s0Gap0Final0Version.pdf?sequence=1&isAllowed=y. rockefellerfoundation.org/report/smart-power-rural-development- https://siteresources.worldbank.org/EXTESC/Resources/Address- brochure/. ing_the_Electricity_Access_Gap.pdf Smart Power India. 2016. Smart Power Connect Magazine 1 (1, July). Dutt, Pranesh Kumar, and Iain MacGill. 2013. “Addressing Some Issues https://smartpowerindia.org/wp-content/uploads/2021/07/ to Hybrid Mini Grid Failures in Fiji.” Presented at the IEEE Global smartpowerindia_magazine_july_2016.pdf. Humanitarian Technology Conference: South Asia Satellite (GHTC- 14 8   MINI GRIDS FOR HALF A BILLION PEOPLE Smart Power India. 2017. Smart Power Connect Magazine 2 (1, May). tion of an adequate demand assessment for appropriate sizing of https://smartpowerindia.org/wp-content/uploads/2021/07/ a mini grid system. The Powerhive team highlighted the challenges smartpowerindia_magazine_may_2017.pdf. of customers who have not previously used electricity in accurately Smith, J. 2000. Solar-Based Rural Electrification and Microenterprise translating their needs and expectations into monthly expenditure Development in LatinAmerica:AGenderAnalysis.Golden,CO: National and realistic consumption plans. Overreliance on such data can sub- Renewable Energy Laboratory. https://www.nrel.gov/docs/fy01osti/ sequently lead to substantial under- and oversizing of systems. At 28995.pdf. the same time, the Powerhive team emphasized the value of other demographic and socioeconomic data that should be gathered SNV (Netherlands Development Organization). 2008. Building Viable across the community from early stages to serve as prediction DomesticBiogasProgrammes;SuccessFactorsinSectorDevelopment. markers for electricity use (for example, number of residents in a https://cleanenergysolutions.org/es/resources/building-viable- household, number of homes in a compound, sources of income, domestic-biogas-programmes-success-factors-sector-develop- number of acres of farmed land, types of crops, number of children ment. in a school). Such data can be further validated through comparison Valencia, Adriana, and Georg Caspary. 2008. “Barriers to Successful with other sites that may be operated by the same mini grid devel- Implementation of Renewables-Based Rural Electrification.” Briefing oper in or near the area. Paper 7/2008, German Development Institute, Bonn. https:/ /www. 3. One example from the mini grid run by Gram Oorja in the village of die-gdi.de/uploads/media/BP_7.2008_Valencia.Caspary.pdf. Darewadi in the Indian state of Maharashtra shows that, with the Wiemann, Marcus, Simon Rolland, and Guido Glania. 2011. Hybrid local community having an active stake in the mini grid, its members Mini-Grids for Rural Electrification. Brussels: Alliance for Rural Elec- self-moderate their electricity use during the monsoon season to trification (reprint). https://www.ruralelec.org/sites/default/files/ ensure its continued functioning (Fahey and others 2014). hybrid_mini-grids_for_rural_electrification_2014.pdf. 4. The importance of tailoring the community engagement approach to each local context was particularly emphasized in an interview by Havenhill Synergy Ltd., a Nigerian mini grid developer operating sev- NOTES eral solar-hybrid mini grids in the Kwali and Kuje local government areas of Nigeria. 1. For this chapter, ESMAP consulted with the teams of Havenhill 5. The importance of securing the approval of the local chiefs at the Synergy Ltd., Powerhive, PowerGen International, and ENGIE Pow- initial design and planning stage of the mini grid project was high- erCorner. lighted in an interview by ENGIE PowerCorner. 2. In one of the interviews carried out for this chapter, a mini grid devel- 6. Based on an interview discussion with ENGIE PowerCorner. oper—Powerhive—noted the importance of not relying too heavily on community-sourced data on expected demand for the prepara- 7. Based on an interview discussion with Havenhill Synergy Ltd. MINI GRIDS FOR HALF A BILLION PEOPLE    149 CHAPTER 5 DELIVERING SERVICES THROUGH LOCAL AND INTERNATIONAL ENTERPRISES AND UTILITIES CHAPTER OVERVIEW This chapter first provides context for the rise of private-sector participation in the mini grid market, the motivation behind that participation, and a high-level overview of the various types of companies involved. It then presents a preliminary assessment of the profit at stake for private-sector players along the value chain, as well as for select mini grid operators. Third, the chapter outlines the private-sector players that are active in various stages of the mini grid industry’s value chain, and how they generate revenue and manage costs. Finally, the chapter highlights the ways companies across the value chain can collaborate. Throughout this chapter, results from the World Bank’s recent surveys of mini grid operators in Cambodia, Myanmar, and Nepal are provided to further illustrate concepts and provide concrete data from operating mini grids. Small and medium enterprises (SMEs) play a major role in large firms; instead, they rely on internal funds, or cash most economies, particularly in developing countries.1 They from friends and family, to launch and initially run their represent about 90 percent of businesses and more than 50 enterprises. About half of formal SMEs do not have access percent of employment worldwide. Formal SMEs contribute to formal credit. The financing gap is even larger when up to 40 percent of national income (measured as gross micro and informal enterprises are taken into account domestic product, GDP) in emerging economies. These (World Bank 2022). numbers are much higher when informal SMEs are included. A key priority is to improve SMEs’ access to finance and find Mini grid companies are a small yet important group of innovative solutions to unlock sources of capital, for exam- SMEs. As we describe later in this chapter, the Energy Sector ple, through SME lines of credit, partial credit guarantee Management Assistance Program (ESMAP) has identified schemes, and early stage innovation finance. Good prac- 168 private-sector mini grid developers (registered compa- tices also show the necessity of financial sector assess- nies) currently active in countries with low rates of electricity ments to determine where regulation and policy might access. All of them are SMEs (World Bank 2022). be improved; identify an enabling environment, design, According to World Bank estimates, 600 million jobs will be and setup for credit guarantee schemes; improve credit needed by 2030 to absorb the growing global workforce, infrastructure; and deploy innovative SME finance such as which makes SME development a high priority for many e-lending platforms, use of alternative data for credit deci- governments around the world. In emerging markets, most sioning, and supply chain financing (World Bank 2022). formal jobs are generated by SMEs, which create 7 out of In the past, small, local companies—often relying on inter- 10 jobs. nal or philanthropic funding—would design, build, and However, access to finance is a key constraint to SME operate mini grids. Increasingly, however, private compa- growth, including for mini grid companies. It is the sec- nies and private capital are participating across the mini ond-most frequently cited obstacle to growing their busi- grid value chain. Their participation brings the benefits of nesses in emerging markets and developing countries. more advanced technology, access to lower-cost capital, SMEs are less likely to be able to obtain bank loans than and a focus on innovation, replicability, and scale. Today, 150   MINI GRIDS FOR HALF A BILLION PEOPLE the mini grid space is seeing participation not just from SECOND-GENERATION MINI GRIDS local enterprises, but also from international companies As focus shifted from the developed markets to lower-in- providing tens of thousands of families with an income. come countries, the original notion of using mini grids as Scaling up the mini grid sector to achieve the Sustainable starting points for electrification re-emerged. But as the Development Goal 7 (SDG 7) targets will require merging technology supporting larger generation and distribution the advantages of large international companies that have had matured by the time these countries began to elec- the scale to manufacture standardized mini grid compo- trify, many urban areas in these lower-income countries nents and systems, with those of “on-the-ground” compa- could simply begin rolling out large-scale electric grids. nies that can engage the local communities being provided However, for reasons spanning both social and economic electricity from mini grids. factors, many rural areas were left unconnected to these growing grids. Although some of the largest developers of mini grids today are national utility companies, it is the private sector This gave rise to the second generation of mini grids: small- that will drive exponential growth in the industry, particu- scale power systems often deployed in remote or rural larly as many utility companies in electricity-access-deficit areas with little industrial activity or low population densi- countries are struggling financially.2 Indeed, private-sector ties. To date, these mini grids have been built and operated developers have the top five largest portfolios of mini grids by local communities and local entrepreneurs. If and when planned for construction in the next five to seven years, total- the main grid is extended to these rural areas, these mini ing more than 23,000 mini grids. This represents 76 percent grid installations are generally abandoned, such as in the of all currently planned mini grids, according to ESMAP’s case of Sri Lanka and Indonesia. But not all second-genera- database of more than 50,000 installed and planned mini tion mini grids become stranded assets when the main grid grids around the world. Are we doing enough to support arrives; in Cambodia, many mini grids converted to small all 160+ private-sector developers? Do we consider wom- distribution companies, using the existing mini grid foot- en-specific needs? Do we offer service providers the right print (Tenenbaum and others 2018). tools to increase their productivity and build back stronger? The World Bank recently undertook large-scale, nationally These need to be top-priority questions for government offi- representative surveys of mini grid operators in three coun- cials, financiers, and development partners as they launch tries—Cambodia, Myanmar, and Nepal—to understand the or scale up national and regional mini grid programs. market landscape of their second-generation mini grid operations. For more than three decades, mini grids have provided AN EVOLVING TECHNOLOGY FROM electricity service in these three countries to those with- FIRST-GENERATION TO THIRD- out access to grid-based electricity, particularly in remote GENERATION MINI GRIDS rural areas. Mini grid operators, on average, started oper- ating 12 years ago in Myanmar, 14 years ago in Cambodia, Mini grids are not a new concept; indeed, they have played and 10 years in Nepal. Almost half of all mini grids sur- a critical part in the electrification of many countries. veyed had been in operation for more than 10 years. The earliest mini grid operation, in Myanmar, dates to the late FIRST- AND SECOND-GENERATION MINI GRIDS 1950s. The number of mini grid operations in these three More than 100 years ago, mini grids served as a starting countries soared in the mid-2000s. Nearly a decade later, point for electrification in countries as diverse as China, the pace of growth has slowed, but the number of mini Sweden, and the United States. The first modern elec- grids continues to grow. tric power plant, Thomas Edison’s Pearl Street Station in Manhattan, New York, began operating in 1882 and served THIRD-GENERATION MINI GRIDS only 80 customers using thermal power. Generally, these Over the past few years, a new, third, generation of mini first-generation mini grids were located either in dense grid technologies and business models has emerged. urban environments or in areas where the cost of sup- As noted in the overview of this report, these mini grids ply was low. As the technology improved and demand differ from the second generation in several important increased through the Industrial Revolution, larger gener- ways. For example, third-generation mini grids use more ators and distribution systems could be built. Eventually, modular technologies, such as solar photovoltaic (PV)- these larger systems connected to what became the main based hybrid plants, which are combined with mobile- grid known today. The smaller mini grids that had been the based pay-as-you-go billing and internet-based remote starting point for this evolution were either integrated into monitoring. the expanding main grid or simply abandoned. MINI GRIDS FOR HALF A BILLION PEOPLE    151 costs running as much as $2,500 per connection, it would take more than $600 billion to provide a grid connection to The characteristics of third-generation mini every person without electricity (Attia and Shirley 2018). For grids—in particular, their use of new technol- private companies that develop or operate mini grid technol- ogies and market mechanisms alongside improved ogies, such a prohibitive cost provides a unique opportunity regulations specific to mini grids—are creating to offer a more cost-effective alternative solution to grid-level environments that attract local and international service. Indeed, as noted in the overview, average connec- industry players to invest in portfolios of mini grids in tion costs for mini grids are currently less than $2,000 per countries with low levels of electricity access. customer and are expected to continue to fall through 2030. With an addressable market of 167,000 load centers for mini grids in energy-access-deficit countries, representing Third-generation mini grids are increasingly being sup- about 86 million connections by 2030, this amounts to an ported, developed, and deployed by a variety of private- estimated investment opportunity of approximately $105 sector companies, including large local companies and billion by 2030. Mini grid electricity sales could grow from multi-national operators. These companies use new mar- an estimated $500 million per year in 2021 to $16 billion ket mechanisms, such as public-private partnerships, as per year by 2030, assuming 9 million connections in 2021 seen in Kenya and Sierra Leone. Such new operating par- and 86 million connections in 2030.3 Meanwhile, ESMAP adigms are coupled with a more targeted regulatory envi- estimates that the addressable market for productive use ronment, particularly when mini grids must contend with appliances and machines is approximately 3 million units by the need to connect to the main grid or operate in parallel 2030 (assuming 15 appliances per mini grid and 200,000 with the main grid (Shakti Foundation 2017). new mini grids). With an average cost of $1,200 per appli- ance, this represents a $3.6 billion per year market, and a $3 When taken together, these factors provide conditions billion end-use finance opportunity by 2030. that lead to greater participation of the private sector in the development and operation of mini grids. Interna- The impact of mini grids on the private sector is not lim- tional technology providers use global supply networks to ited to companies directly involved in their development offer standardized, economical mini grid systems. Large or operation. By some estimates, not having an adequate local and international operators, by using advanced mini supply of electricity can result in enterprises losing more grid designs and relying on clear regulatory systems, are than 30 percent of potential sales (Ramachandran, Shah, having discussions with private investors to support the and Moss 2018). Often, an enterprise must rely on its own deployment of portfolios of mini grids. Public-private part- electricity sources; globally, the expenditure on backup nerships and targeted regulations are providing greater diesel generation for individual homes and businesses certainty to investors, thereby creating an environment likely exceeds $40 billion annually (Velamala 2016; Orlandi that allows operators to raise equity and debt to supple- 2017). Offsetting such expenditures by delivering reliable ment grant capital. electricity to commercial customers through a mini grid can further bolster the mini grid’s economic viability. The mini grid sector complements the vertically integrated THE BUSINESS CASE FOR PRIVATE- utilities and distribution companies as well as the solar companies that sell stand-alone solar systems. SECTOR PARTICIPATION IN THIRD-GENERATION MINI GRIDS For electric utilities, mini grids can reduce the financial bur- den of costly utility connections for rural customers and More than 800 million people today lack access to electric- provide such customers with reliable electricity. The unfor- ity, while nearly 3 billion lack access to clean cooking. By tunate rural reality is that (1) utility connection rates remain some estimates, this segment of the population spends as low; (2) utilities have struggled to recover costs when they much as $37 billion per year on kerosene and biomass for rapidly increase electrification rates; and (3) even when lighting and cooking fuel; converting this use to sustainable customers are connected, reliability is often a problem. alternatives in line with the universal electrification ambi- Data on electrification rates by electric utilities in Africa are tion of SDG 7 could require more than $50 billion to be surprisingly sparse. The most up-to-date publicly available invested annually through 2030 (Asmus 2013; IEA 2017). data on utility connections in Africa are from the African For lower-income countries that must invest in expanding Development Bank’s Africa Infrastructure Knowledge Pro- electricity infrastructure to unelectrified areas, third-gen- gram Portal and only cover the period from 2004 to 2014 eration mini grids will play a critical role in providing a more (African Development Bank 2016). Initial analyses of these economical alternative to grid extension. With grid extension data are shown in table 5.1. 152   MINI GRIDS FOR HALF A BILLION PEOPLE Table 5.1 reveals several important points about utility elec- the decade leading up to 2014, the pace of utility electrifica- trification rates in Africa. First, we lack consistent, recent tion was just barely keeping up with population growth, con- data. The most recent publicly available comprehensive data necting an additional 1.2 percent of the population per year cover only up to 2012, with one country covered through on average. In some countries, we can see that population 2014. Because data are scarce and outdated, it is not pos- growth was actually outpacing utility connections (Comoros, sible to draw conclusions about utility connection rates over Guinea, Madagascar, and Togo). In terms of annual connec- the past several years. What we can say, however, is that in tion rates, the fastest-growing utilities are in Africa: TABLE 5.1 • Utility connection rates, 2004–14 Utility connections as percentage of total number of households Avg. annual change 2004 2006 2009 2008 2005 2007 2010 2014 2013 2012 2011 Country Utility Benin SBEE 19.0 19.9                   0.8 Burkina Faso Sonabel 10.7 11.3     12.4 13.2 13.7 14.7 15.6     0.6 Burundi REGIDESO   2.3 2.3 2.3 2.6 2.6 3.2 3.2 3.1 0.2 Cameroon AES SONEL 14.7 14.8   15.2 15.9 16.6 17.4 18.1 18.2     0.5 Cabo Verde Electra 71.3 76.0                   4.7 Central African Republic ENERCA             3.1 3.1 3.1 0.0 Chad STEE 1.1 1.2                   0.1 Comoros Ma Mwe   20.6 20.1 19.6 19.2 18.7           -0.5 Congo, Rep. SNEL 44.1     42.9 41.9 40.4 40.2 42.8 50.0     0.3 Côte d’Ivoire CIE 23.3 24.3   25.3 25.9 26.4 26.8 26.8 26.2     0.4 Djibouti EDD   21.4 22.4 23.1 23.6 24.2           0.7 Ethiopia EEU       5.5 6.8 7.0 6.9 7.3 7.1 0.3 Gambia, The NAWEC   29.3 34.3 37.7               4.2 Ghana ECG       23.5 25.5 28.4 29.9 32.4   2.2 Guinea EDG             9.0 7.0 7.6 8 8 -0.2 Kenya KPLC       8.7 9.7 11.3 12.7 14.8   1.5 Lesotho LEC                 29.0   Madagascar JIRAMA 10.9 10.8   10.2 10.0 10.0 9.9         -0.2 Malawi Escom       4.5 4.7 5.0 5.5 5.2 5.4 0.2 Mali EDM 6.8 7.2 7.6 8.0 8.2 8.5 9.0 9.7 10.5     0.4 Morocco ONE   50.8 55.7 59.9 62.6 66.1           3.8 Mozambique EDM 5.8 6.7   10.3 12.0 14.0 15.9         1.7 Namibia NORED 6.0 7.0                   1.0 Niger NIGELEC 5.2 5.4   5.7 5.9 6.7 6.7 6.8       0.2 Rwanda RECO       2.8 4.0 5.4 7.0 9.0   1.5 Senegal Senelec 44.1 46.7   52.6 55.2 58.0 60.0 59.8 61.0     2.2 Seychelles Public Utilities Company             17.0   17.3 0.1 South Africa Capetown and ESKOM   41.7           97.3       9.3 Tanzania TANESCO       7.9 8.4 8.8 9.3 9.8 10.5 0.5 Togo CEET   14.2 13.8 13.4 13.1 12.7           -0.4 Uganda UMEME           4.6 5.4 5.9 5.7 0.4 Zambia ZESCO             10.9 12.1 11.7 0.4 Source: African Development Bank 2016 and ESMAP analysis. High Low MINI GRIDS FOR HALF A BILLION PEOPLE    153 • ESKOM and Capetown were each year, together, con- access is less than 5 percent over this period (Balabanyan necting 9.3 percent of South Africa’s population. and others 2021). • Electra in Cabo Verde: 4.7 percent. Across Sub-Saharan Africa, households and small busi- • NAWEC in The Gambia: 4.2 percent. nesses have outages lasting for hours. In some countries— including Burundi, Ghana, Guinea, Liberia, Nigeria, and • ONE, in Morocco: 3.8 percent. Zimbabwe—more than half of households connected to the Still, in nearly every country large portions of the popula- main grid reported receiving electricity less than half of the tion remained unelectrified by the main utility companies. time (Blimpo and Cosgrove-Davies 2019). Disaggregated In 2012, the most recent year for which data were available data from diagnostic survey reports carried out by ESMAP for many countries, only three countries in Sub-Saharan in a range of countries, based on the Multi-Tier Framework, Africa saw more than half their households served by util- provide additional evidence of this lack of reliability, both ities: SNEL in the Republic of Congo, Senelec in Senegal, in the Sub-Saharan region and beyond. The report from and the combined utilities in South Africa. Rwanda indicates 97 percent of grid-connected house- holds experience more than four electricity disruptions a According to a World Bank study (Balabanyan and others week (Koo and others 2018). The Ethiopia report shows 2021), while generation capacity grew from 63 gigawatts that 57.6 percent of grid-connected households face 4–14 (GW) in 2000 to 106 GW in 2017, electricity access in outages a week, and 2.8 percent face more than 14 outages Sub-Saharan Africa remains mixed: 12 countries represent- a week (Padam and others 2018). The report from Cam- ing 14 utilities4 have low access rates (less than 33 percent of bodia indicates that 69.3 percent of grid-connected house- the population), 22 countries representing 46 utilities have holds face frequent, unpredictable power outages, and 9.9 mid-level access (33 to 67 percent), and 11 countries rep- percent of all grid-connected customers receive less than resenting 16 utilities have high access (above 67 percent). four hours of service per day (Dave and others 2018). These World Bank figures are roughly corroborated by find- ings from an International Energy Agency global database Utility information, while limited, corroborates this survey (IEA 2022) on electricity access rates. The IEA indicates an information. In Sub-Saharan Africa only about a third of the overall electrification rate for Sub-Saharan Africa of 49 per- vertically integrated utilities reported the System Average cent, but in rural areas this falls to 28 percent of the popu- Interruption Duration Index (SAIDI) and System Average lation. Many countries are substantially lower than this. For Interruption Frequency Index (SAIFI) in 2018, and only 5 out example, in Central Africa, Chad has an 8 percent overall of 21 distribution companies reported SAIDI and SAIFI. Of electricity rate, while less than 1 percent of the rural pop- those that did, the median reported SAIDI and SAIFI in 2018 ulation has electricity access; the Democratic Republic of were 51.6 hours and 24.7 interruptions per year, respec- Congo has similar access rates (9 percent overall, less than tively. These are high by international standards. In order to 1 percent rural). In East Africa, Djibouti’s access rates are 42 receive any points under the scoring methodology used by percent overall and less than 1 percent in rural areas, while the World Bank’s Doing Business indicators, the maximum in Burundi the overall rate is 10 percent and the rural is also SAIDI and SAIFI is 12—equivalent to a one-hour-long outage less than 1 percent. In West Africa, Mauritania has 47 per- each month (Balabanyan and others 2021). These SAIDI cent overall, and less than 1 percent rural; Sierra Leone has and SAIFI are corroborated by survey data: in some coun- 22 percent overall and less than 1 percent rural. tries—including Burundi, Ghana, Guinea, Liberia, Nigeria, and Zimbabwe—more than half of connected households As low as these figures are, they overestimate electricity reported receiving electricity less than 50 percent of the access from utility connections because they include off- time in 2014 (Blimpo and Cosgrove-Davies 2019). Informa- grid sources such as solar home systems (SHSs), diesel tion on load factors for these utilities was not available. generators, and mini grids (IEA 2020). Mini grids also complement the expansion of basic electric- In Sub-Saharan Africa the median time to connect cus- ity services like solar lanterns and SHSs by providing next- tomers steadily improved in recent years, dropping from step options as rural communities grow and needs increase, 117 days in 2012 to 90 days in 2018. But with high costs of and by providing electric service to towns that complement connecting new rural customers ($2,500 per connection, household-scale systems in surrounding areas. as already mentioned), utilities have found it difficult to cover costs, particularly those that have attempted rapid As of June 2022, over 100 companies were members of expansion of service. In countries where on-grid electric- the Global Association for the Off-grid Solar Energy Indus- ity access improved by more than 5 percent from 2015 to try (GOGLA). Globally, there were 261 quality-verified solar 2018, 69 percent of utilities failed to recover their costs. lanterns and SHSs from 67 brands listed by VeraSol (the Conversely, utilities recover their costs in 53 percent of sector’s quality assurance program) in 2021, up from 201 the countries where the increase in on-grid electricity products from 51 brands in 2019. SHSs are increasingly 154   MINI GRIDS FOR HALF A BILLION PEOPLE sold with appliances, for which there is also a sizable mar- Operators can be local or international, while facilitators ket of quality-tested products, including 152 TVs from 88 generally are international entities. Local companies are brands, 131 fans from 86 brands, and 94 refrigerators from generally SMEs whose operations are geographically 52 brands. Solar off-grid systems range from 0.5 watt to restricted to an area that could be a district, a state, or power a light bulb, up to several hundred watts to power even an entire country. These entities are generally staffed efficient DC (direct current) appliances such as TVs, fans, with local resources from the same region where the com- refrigerators, and productive-use equipment. Global mar- pany operates. Their products and services are developed ket turnover is now $2.1 billion annually. locally, tailored to the needs of the region they serve. International operators, or companies that operate mini Based on conservative life span assumptions, GOGLA esti- grids in more than one country, tend to set up in-coun- mates that in Sub-Saharan Africa, 64.5 million people are try operations similar to local companies in each region using an off-grid solar product. Using the multitier frame- they operate, relying on local resources to engage with work, 33 million people are in Tier 1 and 9.7 million in Tier the communities they seek to serve, while maintaining a 2. Globally, there are 34.6 megawatts of newly installed headquarters function that engages in activities common capacity through off-grid solar products. When customers across geographies, such as raising capital to deploy mini have the opportunity to shift to mini grids, these household grid equipment. solar products could continue to provide supplemental energy services or be passed on through family networks Facilitating organizations generally do not engage with to areas with no electrical services (Galan 2022). In 2012, the community receiving electricity directly; rather, they equity and debt investment in the sector totaled $17 mil- engage with the local and international companies that will lion; in the past decade, that grew over 25-fold to $447 mil- operate the mini grids. International companies are suited lion (GOGLA 2022). to serving such a role. Their operations across geographies allow them to develop wide-scale supply chains, source PRIVATE-SECTOR SEGMENTATION technology components more cost efficiently, and place Operators and facilitators manufacturing facilities in low-cost regions. These com- The private sector’s role in the mini grid space spans from panies generally have established processes for quality the development of mini grid technologies to their instal- assurance and after-sales service. Local and international lation, operation, and servicing. Broadly, two types of pri- companies have their respective advantages; therefore, vate-sector entities participate in the mini grid space: success in scaling the mini grid space will require engage- operators and facilitators. Operators include energy ser- ment across different types of organizations. Box 5.1 offers vice companies (ESCOs), utilities, and independent power an example of such a partnership. producers (IPPs). Facilitators include original equipment manufacturers (OEMs), engineering, procurement, and construction companies (EPCs), system integrators (SIs), BOX 5.1 and financiers. Operators assume primary responsibility for the life cycle of a mini grid or a portfolio of mini grids, GREEN VILLAGE ELECTRIC such as identifying a mini grid location, engaging with the PARTNERS WITH SCHNEIDER community, and operating the mini grid. Facilitator organi- ELECTRIC zations help operators design, build, and manage mini grids and include original equipment providers and financiers. Green Village Electric (GVE), a local Nigerian enter- prise focused on mini grid development, has part- nered with Schneider Electric, a large multinational power electronics manufacturer, to develop sev- Broadly, two types of private-sector entities eral mini grids. For example, GVE and Schneider participate in the mini grid space: operators partnered on a mini grid for the villages of Bisanti, and facilitators. Operators assume primary respon- Kolwa, and Onono-anam, developing a 24-kilowatt sibility for the life cycle of a mini grid or a portfolio solar mini grid with prepay meters. GVE engaged of mini grids, such as identifying the mini grid loca- with the community and developed the project, tion(s), engaging with the community, and operat- while Schneider provided the solar hybrid system ing the mini grid(s). Facilitator organizations help based upon each site’s unique needs. The net operators design, build, and manage mini grids; result was power for 250 homes, four classrooms, such companies include original equipment provid- small businesses, and a health center (Schneider ers and financiers. Electric 2018). MINI GRIDS FOR HALF A BILLION PEOPLE    155 During the early development of the electricity sector, when Table 5.2 highlights the different categories of local and first-generation mini grids were prevalent, their design, international private-sector players, the stage of the value manufacturing, and operation were often handled by a sin- chain in which they participate, and what levers each type gle company. When Thomas Edison first began operating of player can use to accelerate how mini grids achieve scale. Pearl Street Station, the world’s first mini grid, the Edison In addition to operators and facilitators, a few technology Illuminating Company took responsibility for designing and companies provide services to the mini grid industry. This operating the power plant, signing up customers, and sell- niche category consists of pure software developers that ing electricity. When second-generation mini grids began provide software to operate, monitor, and manage mini deployment to remote areas, the power plant components grids remotely, or facilitate the project development pro- were manufactured by different companies, while another cess of mini grids. For example, AMMP Technologies and local company or the community itself would operate and Infinite Fingers conduct remote monitoring; Odyssey and maintain the system. This is the case for the mini grids sur- Village Data Analytics facilitate project development and veyed in Cambodia, Myanmar, and Nepal. Mini grids are preparation. mostly owned by private entities or individuals in Myanmar (82 percent, with the remaining mini grids owned by com- Mini grid developers / operators / ESCOs munities or the government) and Cambodia (99 percent), This first category of private-sector participants is catego- while virtually all mini grids in Nepal (95 percent) are owned rized as operators, or entities that operate mini grids to sell by the local community. electricity directly to end customers. These SMEs focus With third-generation mini grids, more private-sector com- primarily on developing and operating mini grids, and also panies are participating across the value chain. Whereas may provide additional goods and services. Many ESCOs second-generation mini grids have been primarily hydro received their start during the rollout of second-generation or diesel based, with third-generation mini grids, other mini grids, when local companies were required to operate manufacturers, such as solar or battery companies, are mini grids that were targeted for rural areas. ESCOs are becoming involved. Whereas a good portion of second-gen- often staffed using local resources; therefore, they often are eration mini grids have been operated by communities or in a better position to engage with communities that can nongovernmental organizations, third-generation grids benefit from mini grid deployment. are increasingly run by profit-seeking companies. Indeed, Once a site has been identified, ESCOs may either directly across the value chain described earlier, private-sector or through another provider survey the village to develop companies are performing one or more of the steps to a demand profile, work with facilitator organizations to design, develop, deploy, and operate mini grids. design and install the mini grid, and then oversee its oper- TABLE 5.2 • Categories of local and international private-sector players Primary strategies to: Type Category Value chain segment Generate revenues Manage costs Operators Mini grid Market assessment, permitting Sell electricity, other Local staffing, bulk procurement, developers and financing, grid design and products, or services digital automation, mobile procurement, O&M, after-sales payment, prepaid meters, replicable processes Utilities Grid design and procurement, O&M Sell electricity Facilitators Original equipment Component manufacturing, O&M Sell equipment for mini Low-cost manufacturing, mass manufacturers grids customized packages, digital (OEMs) automation Engineering, Grid design and procurement, Procure, design, and Local staffing, mass customized procurement, integration and installation install mini grids processes, replicable processes and construction companies (EPCs) System integrators Grid design and procurement, (SIs) integration and installation Financiers Permitting and financing Finance operators, Data analytics, standardized projects, additional offerings goods, or services Source: ESMAP analysis. O&M = operation and maintenance. 156   MINI GRIDS FOR HALF A BILLION PEOPLE ation. Given the close relationship that ESCOs form with the energy access market. These are registered compa- their customers, they are also in an ideal position to pursue nies that have built mini grids, are planning mini grids, the “after sales” portion of the value chain, using their elec- are members of a mini grid industry association, and/or tricity delivery service to support selling other products. have secured investment for mini grids. These developers operate at least 2,100 mini grids in energy-access-deficit Mini grid ESCOs generally can be categorized as local enti- countries, serving approximately 5 million people. Of these ties, or international entities with local operating units. Local developers, 115 are active in Africa, operating at least 900 entities may be small companies, such as those that have mini grids and serving approximately 2 million people. As been operating second-generation mini grids, or larger, of early 2022, 41 of these developers were members of the more established companies that operate several mini Africa Minigrid Developers Association (AMDA). A further grids, among other business lines, such as Muhanya Solar 50 developers are active in Asia, operating approximately in Zambia and Nayo Tropical Technologies in Nigeria. Inter- 1,200 mini grids that serve around 2 million people. national companies may have their headquarters in a loca- tion outside the countries where they operate; each local The majority of active mini grid developers are taking a unit assumes the same structure as a local entity. Sample portfolio approach to building mini grids. Of the 167 active mini grid developers are listed in table 5.3. developers that we identified, 93 had built 3 or more mini grids. The average portfolio size is 23 mini grids per devel- Status of mini grid operators globally and in oper. The top three largest portfolios of installed mini grids Sub-Saharan Africa are in Asia (BRAC in Afghanistan, Husk Power in India, and In addition to a global database of 50,000 installed and Tata Power Renewable Microgrids in India), followed by planned mini grids around the world, discussed in the over- MeshPower in Rwanda and NS Resif in Senegal. view, ESMAP has compiled a database of more than 180 There are also clear signs that the private-sector mini grid investment deals in mini grid companies active in energy-ac- industry is growing. As mentioned in the overview of this cess-deficit countries, using publicly available sources. handbook, the largest portfolios of planned mini grids Combined, these databases give us an unprecedented view are all being developed by private-sector developers, and of mini grid developers targeting energy access. these portfolios are an order of magnitude larger than The mini grid industry is starting from a solid foundation existing portfolios. The five companies with the largest of already active private-sector developers. We have iden- portfolios of planned mini grids—OMC Power, Tata Power tified 168 active mini grid developer companies targeting Renewable Microgrids, Husk Power, Engie Energy Access, TABLE 5.3 • Sample mini grid developers Number of Company Countries mini grids Description OMC Power India > 200 OMC Power is a rural utility company in India that builds, owns, and operates solar hybrid mini grids. They target areas with a telecommunications tower or other anchor client and build the mini grid to serve the surrounding community. Website: https://www.omcpower.com/page/whatwedo. Tata Power India > 100 Tata Power Renewable Microgrid builds, owns, and operates solar hybrid mini grids in India. Renewable It is a wholly owned subsidiary of Tata Power, a multibillion dollar utility company in India. Microgrid Website: https://www.tata.com/newsroom/business/rural-india-solar-microgrids-tata- power. Redavia Tanzania > 10 Redavia uses a lease structure to deploy solar systems that can be integrated with diesel generators. The containerized units reduce investor risk, as they can be redeployed. The company has secured more than $20 million in investment. Website: https://www.redaviasolar.com/. Jumeme Tanzania > 20 Jumeme builds, owns, and operates solar hybrid mini grids in Tanzania, with a strong focus on tailored solutions for productive uses of electricity. Website: http://www.jumeme.com/. Havenhill Nigeria > 20 Havenhill Synergy builds, owns, and operates solar hybrid mini grids in Nigeria. The Synergy company provides other renewable energy services as well, including solar water pumping Ltd. and energy audits. Website: https://havenhillsynergy.com/. Earthspark Haiti < 10 Earthspark has focused entirely on Haiti through its local company, Eneji Pwop. The company builds, owns, and operates solar mini grids under a concession framework. https://www.earthsparkinternational.org/. Source: ESMAP analysis. MINI GRIDS FOR HALF A BILLION PEOPLE    157 and Renewvia—have collectively made public commit- TABLE 5.4 • Profit potential of mini grid operators ments to build more than 22,000 mini grids. In addition, given certain tariffs and costs of service the industry is professionalizing as it matures: AMDA, for Profit component 2022 2030 example, grew its membership from just 17 members in 2019 to 41 in early 2022. Average tariff (US$/kWh) 0.41 0.23 Cost of service (US$/kWh) 0.38 0.20 Investors have taken note of this strong base of active mini Profit on mini grids deployed 66 754 grid developers, and the pace of investment has acceler- in that year (US$, millions) ated over the past decade. Between 2010 and 2022, 188 Profit across all mini grids 290 3,324 investment deals were made between investors and mini deployed through that year grid companies, totaling more than $560 million. The top (US$, millions) five developers in terms of investment capital raised are Source: ESMAP analysis. Husk Power ($80 million), OMC Power ($65 million), Engie kWh = kilowatt-hour. Energy Access ($60 million), PowerHive ($52 million), and PowerGen ($46 million). Of the $302 million of investment the largest number of new mini grids are required in order capital raised by these five developers since 2010, $206 to achieve universal access by 2030. million was raised since 2019 alone. And looking across the full data set of investment deals, the 2018–22 period saw While operators’ potential profit, under the assumption that an average of 28 deals per year, compared with just 3 deals universal access will be achieved, is quite large, the finan- per year between 2010 and 2014. cial results of mini grid operators indicate that profitability remains difficult to achieve. Table 5.5 highlights the audited Profitability of mini grid developers financial results of select mini grid energy service compa- The profit potential for operators is significant, with the nies (ESCOs 15) from 2018; for comparison, the audited right enabling environment and if universal access to elec- financial results of three multinational IPPs are also shown. tricity is achieved by 2030. The primary mechanism oper- Analyzing the current financial health of the ESCOs, par- ators use to generate a profit is charging a tariff for each ticularly in relation to the more established IPPs, leads to unit of electricity higher than the fixed and variable costs several interesting conclusions. of generating that electricity. We note that in many cases there is a “viability gap” between what the developer must Profits are constrained. Many of the selected ESCOs charge to be profitable, and what the customers are able exhibit large negative profit margins, due in part to the high to pay in terms of disposable household income. Perfor- capital expenditure and other early investments needed to mance grants would be needed to bridge this viability gap, establish operations, while not yet having sufficient time as we discuss in more detail in chapter 6. or capability to drive greater revenue from those invest- ments. As indicated by the comparison companies, ESCOs The profit potential of mini grid operators, given assumed should target a 1–10 percent net profit range as operations tariffs and costs of service, is summarized in table 5.4.5 become more established. To this end, ESCOs might adopt It is important to note that financial support packages, digital solutions and focus on economies of scale to allow including performance grants from governments and for the costs of overheads and assets to be spread among development partners, will be needed to unlock this profit many units sold. potential, particularly over the next few years, to set the market on the trajectory of rapid scale-up. Public funds Greater focus on cost containment is needed. Several of enabled high-income countries to achieve universal elec- the ESCOs have higher administrative costs, as a percent- tricity access; the same will be true for electricity-ac- age of revenue, when compared with the IPPs. To address cess-deficit countries today. this, ESCOs should focus on economies of scale, standard- ized hardware, greater use of local resources, and digital As table 5.4 shows, the profit potential for mini grid opera- tools to lower personnel costs and improve productivity. tors is expected to increase over the next decade, even as tariffs decline. This analysis indicates a profit potential that Consider expanding into energy services. The ESCOs con- could exceed $3.3 billion annually across all third-genera- sidered have low asset turnover, which encapsulates how tion mini grids deployed globally through 2030, under the efficiently a company can monetize its asset base; they scenario in which mini grids serve 490 million people by also exhibit low or negative returns on assets. Part of this 2030. These tariffs are reflective of the low end of mini grid finding is driven by the deployment of relatively expensive tariffs today, resulting in a conservative estimate for mar- mini grid infrastructure into areas where the customer mix gins. More than 70 percent of operators’ expected profit may be primarily residential, thus leading to lower tariff rev- potential is concentrated in Africa, also the region where enue. To offset that, ESCOs should use the installed asset 158   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 5.5 • Mini grid developers and large-scale IPPs: A comparison of audited financial results, 2018 Financial category ESCO 1 ESCO 2 ESCO 3 ESCO 4 ESCO 5 IPP 1 IPP 2 IPP 3 Revenue (US$, thousands) 46 85,050 1,564 40 145 73,900,000 43,736,000 10,530,000 COGS (US$, thousands) 297 21,882 578 36 181 41,700,000 33,845,000 8,043,000 Gross profit (US$, thousands) –251 63,168 985 4 –36 32,100,000 9,892,000 2,487,000 Operating expense 1,505 83,270 Unknown 958 Unknown 10,963,000 2,551,000 8,043,000 (US$, thousands) Personnel (US$, thousands) 1,273 13,141 248 147 127 11,600,000 3,420,000 216,000 EBIT (US$, thousands) –3,029 –33,243 1,195 –1,100 –126 5,365,000 1,862,000 2,372,000 Interest expense (US$, N/A 205 219 19 23 988,000 1,429,000 1,170,000 thousands) Tax (US$, thousands) 344 N/A 304 N/A N/A 483,545 500,000 990,000 Net income (US$, thousands) –2,600 –33,448 672 –11,100 –$148 2,216,000 4,750,000 –148,000 Net profit (% of revenue) –5,454 -39 43 –2,744 –102 2 10 –11 SG&A as % of revenue 2,700 15 16 370 88 16 8 2 Asset turnover ratio 0.01 0.15 0.12 0.01 1.88 0.43 0.69 0.32 Return on assets (%) –65 –6 5 –32 –191 1 7 –4 Current ratio 7.14 0.81 1.82 0.32 0.04 1.06 1.12 1.06 Source: ESMAP analysis. Note: All values are in thousands of US dollars. Revenue is from business operations only; it excludes revenue identified as “other” and other extraor- dinary items. Comparable companies are large multinational energy service providers, and are used only for the purpose of providing a benchmark for the financial ratios. COGS refer to cost of goods sold, the variable cost to sell electricity. SG&A refers to all selling, general, and administrative overhead costs; it excludes personnel costs, which are shown in the table separately. ESCO = energy service company; IPP = independent power producer. base and established customer relationships to provide other services that would generate additional revenue. Cumulative profit potential for mini grid ESCOs also can consider identifying sites with a greater developers could exceed $3 billion by 2030 mix of commercial activity, thus allowing for a wider range if the SDG 7 target is achieved. However, an analysis of tariff values and structures. of the financial results of existing mini grid develop- Capital constraints vary. The current ratio encapsulates a ers suggests that profitability is difficult to achieve. company’s ability to address short-term liabilities based on As a result, financial support packages, including its current asset base; a number higher than 1 indicates a subsidies from governments and development company’s ability to successfully address short-term liabil- partners, will be needed to unlock the profit poten- ities. Select ESCOs have large current ratios, while others tial and subsequent scalability for mini grid compa- are more constrained. This situation implies that certain nies, particularly over the next few years. ESCOs are in a position to scale quickly, while others may need to focus first on their existing installations, and find additional ways to generate more revenue before accruing additional liabilities to fuel expansion plans. the same will be true for electricity-access-deficit coun- tries today. Financial support packages are needed to achieve scale. Profitability for existing mini grid companies remains dif- Another way to facilitate growth and economies of scale ficult to achieve. As a result, financial support packages, is through partnerships between local and international including subsidies from governments and development industry players. Organizations can act as conveners, pub- partners, will be needed to unlock the profit potential and lishing information or organizing events that enable knowl- subsequent scalability for mini grid companies, particu- edge sharing and networking among various private-sector larly over the next few years. Public funds enabled high- entities. Groups such as the Africa Minigrid Developers income countries to achieve universal electricity access; Association (box 5.2), the Alliance for Rural Electrification, MINI GRIDS FOR HALF A BILLION PEOPLE    159 BOX 5.2 AFRICA MINIGRID DEVELOPERS ASSOCIATION The Africa Minigrid Developers Association (AMDA) grid sector and act as a unified focal point for stake- was founded in 2017 to act as a trade association for holders to engage the sector. mini grid developers across Africa. AMDA currently has • Industry intelligence: Provide a platform that more than 40 members, with chapters in Kenya, Nige- enables transparency in industry performance ria, and Tanzania. a through comprehensive market data and analyt- AMDA has established the following governing princi- ics to establish, evaluate, and promote key finan- ples: cial, business, and policy solutions to overcome the sector’s major barriers to growth. • Advocacy: Collaborate with industry, policy makers, government authorities, donors, and other stake- To become a member of AMDA, a company must be holders to advocate for optimal policies and efficient a for-profit entity, have developed or be developing capital deployment that will benefit the mini grid at least one alternating-current mini grid, and have sector and the people it serves. some type of nonconcessional investment. In addition, member companies must be willing to share data on • Coordination: Serve as the voice of the mini grid 40 key performance indicators, such as cost, quality, development industry in Africa to promote the and reliability. growth and sustainable development of the mini Source: ESMAP analysis. a. More information about AMDA is available at http://africamda.org/. as well as companies such as Odyssey, can provide a valu- investment. Regardless of how tariffs are set, they generally able platform for the exchange of information about poten- fall into three categories: flat tariffs, volumetric charges or a tial mini grid projects and equipment specifications or combination of these. requirements; they can also serve as a forum for the private • Flat tariffs are standardized fees that a customer pays sector to engage with regulators and other public officials. monthly to the mini grid operator. When Husk Power National and regional utilities Systems first started, one of its initial tariffs was a flat 50 rupees a month (approximately $1/month), which was These traditional energy service providers are tasked with used to support a mini grid that could power two 15-watt distributing and selling electricity; they may also own gen- lights at a customer’s house or business. In this case, eration assets. In certain instances, these utilities may be customers were assured that by paying a fixed amount directed to provide electricity to remote areas under a uni- each month, a certain amount of electricity would be versal access mandate. In other instances, they may view available. However, with a flat-tariff design, customers providing electricity service through mini grids as a growth cannot dynamically adjust how much electricity they use. opportunity, to access a new customer base. Either way, they may decide to invest in mini grids directly, or may • Volumetric charges are paid according to the quantity of enter into an agreement with ESCOs, such as through a electricity used, whether on a per kilowatt-hour (kWh) distribution franchisee arrangement. Sample national util- basis (for energy), a per-kW basis (for power), or both. ity projects servicing mini grids are listed in table 5.6. One way to make volumetric rates more accessible is to tie them to appliance use; for example, a solar PV mini Generating revenues for mini grid developers and grid operator in the Indian state of Odisha charges for utilities each hour of television watched. Volumetric rates can One of the primary means developers and utilities use to also be designed to vary based upon time of day or avail- generate revenue is by selling electricity to existing and ability of supply. If a renewable energy source is abundant new customers. Mini grid tariffs are often, but not always, during certain hours of the day, tariff levels can be kept regulated. As discussed in chapter 9, allowing developers low, to incentivize customers to use low-cost energy. In to set a tariff using a “willing-buyer, willing-seller” approach the evening, when a more expensive diesel backup gen- can be a powerful enabler for scaling up private sector erator must be run, customers can be charged a higher tariff to reflect the higher cost of service. 160   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 5.6 • Sample utility mini grid projects Number of Company Country mini grids Description of sample project Horizon Power Australia > 30 Horizon Power currently services one of the largest remote microgrid portfolios in the world. The Onslow microgrid will be set up as a solar hybrid, which will make it one of Australia’s largest distributed microgrids. Northwest Canada < 20 Northwest Territories Power Corporation installed a 104 kW solar array to Territories supplement existing diesel generation, which is enough to power approximately Power 17 households. The installation is expected to reduce CO2 emissions by 76 Corporation tonnes per year. TANESCO Tanzania > 20 Of TANESCO’s 26 operational mini grids, 19 are powered by fossil fuel and 7 are hydro. Source: Horizon Power n.d.; Northwest Territories Power Corporation 2018, n.d.; Odamo and others 2017. CO2 = carbon dioxide; kW = kilowatt. BOX 5.3 UTILITY-LED ROLLOUT OF MINI GRIDS ON THE NATIONAL SCALE: CASE STUDY FROM ETHIOPIA With a population of nearly 120 million and an elec- The component has earmarked $217 million for the trification rate hovering close to 50 percent, Ethiopia scale-up of the utility-led mini grid model, primarily has the second-highest electricity access deficit on through the deployment of several modalities of var- the Sub-Saharan continent, in terms of total popu- iously bundled engineering, procurement, and con- lation without access to electricity, outpaced only struction, plus short- and long-term operation and by Nigeria. Faced with this significant challenge, and maintenance contracts. Prior to ADELE, such mod- focused on an ambitious goal of reaching universal els have already been successfully tested across the electrification by 2030, in 2017 the government of country through various donor-supported programs, Ethiopia issued a National Electrification Program such as the 12 mini grids deployed by the EEU with (NEP), followed by an even more comprehensive NEP financing from the World Bank’s Ethiopia Electrifica- 2.0 in 2019. NEP 2.0 is an integrated, national-level, tion Program (ELEAP), as well as the 25 systems cur- data-driven plan that combines the fast-paced grid rently under implementation by the utility with funding connection rollout of the earlier NEP—aiming at a support from the African Development Bank. 65 percent connection rate by 2025—with a comple- mentary off-grid and mini grid access program tar- ADELE’s Component 2 equally targets private-sector- geting the remaining 35 percent of the population in led modalities—$53 million under the project is harder-to-reach and more remote areas. dedicated to the launch of a national-level perfor- mance-based grant program. The program, also In alignment with the goals and vision of NEP 2.0, on implemented through the off-grid unit within the EEU, March 29, 2021, the World Bank approved a compre- will offer viability gap financing to help close the gap hensive new project: Access to Distributed Electric- between the cost to the developers of constructing and ity and Lighting in Ethiopia (ADELE). At $500 million, operating the systems, and the affordability of the local ADELE is the largest energy access program on the communities. The program aims to engage both local continent to date. and international developers. At the time of the prepa- Component 2 of ADELE, “Mini Grids for Rural Economic ration of this Handbook, the private-sector-focused Development,” is a $270 million commitment to roll out program is in an active design stage, with the EEU mini grids at scale across the country, both through working together with the World Bank and the Ministry public- and private-sector-led modalities. The compo- of Water and Energy of Ethiopia to refine the planned nent is implemented by the Ethiopian Electric Utility implementation model, building on lessons learned (EEU), through a dedicated off-grid unit. from experiences within and outside the country. continued MINI GRIDS FOR HALF A BILLION PEOPLE    161 BOX 5.3, continued In parallel to the preparation of the ADELE project, as ant example to note in this context is the Distributed part of a holistic approach aimed at targeting nation- Renewable Energy Agriculture Modalities (DREAM) wide mini grid scale-up along key frontiers (in align- program, currently being implemented by the coun- ment with the GFMG 10 building blocks philosophy), try’s Ministry of Water and Energy in partnership Ethiopia’s Government worked with the World Bank, the with Ethiopia’s Agricultural Transformation Agency, United States Agency for International Development, the Rockefeller Foundation, and African Develop- and other key development partners to develop new ment Bank’s Sustainable Energy Fund for Africa. The mini grid regulations. The resulting mini grid directive, program aims to have a transformational impact by which was issued by the Ethiopian Energy Authority in leveraging the Agricultural Transformation Agency’s December 2020, is a comprehensive, streamlined, tai- ongoing work in supporting household farmer-based lored policy document that provides detailed guidance Agricultural Commercial Clusters, through facilitating to the sector on minimum technical and performance access to reliable and affordable solar mini grid power standards, as well as licensing and tariff setting. for large-scale cluster irrigation farming. The program is intended to be implemented in partnership with pri- In addition to the ADELE project, a number of other key vate-sector mini grid developers. It is currently rolling donor partners and stakeholders are actively growing out a proof-of-concept pilot at nine sites, with a tender their presence in the mini grid space in Ethiopia, rec- underway, with viability gap financing and conces- ognizing the significant opportunity for impact across sional debt offered to developers interested in bidding the country and working on mobilizing local and inter- on the sites. national private sector and financiers. One import- • Combined flat and volumetric tariffs: Some developers customers. Nepali and Cambodia operators charge all cus- charge a flat fee for all energy and/or power consumed tomers very similar tariffs, according to their consumption up to a certain threshold, above which the customer (figure 5.2). pays a volumetric tariff according to how much electric- Operators provide a wide range of flat-fee tariffs for res- ity is consumed. idential, business, and public customers: $0.34–$5.32/ Results from the World Bank’s operator surveys provide month in Myanmar and $0.94–$5.71/month in Nepal. Busi- some examples of how these types of tariffs are imple- ness customers are charged apparently much higher than mented. Operators in Myanmar are more likely to charge residential and public consumers in Myanmar and Nepal, most customers a flat fee on a regular basis than they are to indicating that serving more business consumers could charge customers by electricity consumption. By contrast, effectively increase revenue. The flat-fee tariffs for public some Nepali operators adopt a flat fee with a consump- customers to consume mini grid electricity are markedly tion-based one, while a large majority of Cambodian oper- low—only $0.34/month in Myanmar and $1.03/month in ators charge consumers according to their consumption. Nepal (figure 5.3). In Myanmar and Nepal, among all classes of customers, Regardless of how the mini grid operator designs tar- operators have a greater tendency to charge business cus- iffs, money can be collected in two ways: before electric- tomers by consumption rather than a weekly or monthly ity is delivered (that is, prepaid) and after electricity has flat rate. Public customers may use electricity by paying a been consumed (that is, credit or postpay). Certain tariff flat fee or even obtain free electricity (figure 5.1). approaches naturally lend themselves to a prepay model, If charging according to consumption, Nepali operators such as the flat tariff. Although utilities have traditionally offer low tariffs in the range of $0.07–$0.08/kWh to all used credit to charge customers volumetric rates, many consumers, while the tariff range for Myanmar mini grid mini grid operators use prepay models, as they have sev- customers is $0.10–$0.34/kWh, and for Cambodia it eral benefits over credit mechanisms. With prepay electric- is $0.23–$0.24/kWh. Mini grid operations in Myanmar ity, the risk of noncollection is reduced, and it eliminates the provide a distinctly lower price to their public customers, costs associated with sending staff out to collect payment including schools, health clinics, and government buildings, (box 5.4). Prepay electricity also providers the customer on average $0.2/kWh lower than residential and business greater transparency regarding electricity use and costs 162   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 5.1 • Tariff-charging type by customer class 17% 14% 16% 11% 37% 38% 8% 4% 5% 67% 60% 33% 80% 40% 78% 81% 83% 63% 30% 32% 33% 20% 22% 20% 7% Resident Resident Resident Business Public Business Public Business Public MYANMAR NEPAL CAMBODIA By kWh (%) kWh+ at fee (%) Flat fee (%) Free Source: ESMAP analysis. kWh = kilowatt-hour. FIGURE 5.2 • Tariffs by customer class ($/kWh, charged by consumption) 0.4 0.34 0.30 0.3 0.24 0.4 0.23 0.24 ($/kWh) 0.34 0.2 0.30 0.3 0.24 0.10 0.23 0.24 ($/kWh) 0.07 0.08 0.08 0.1 0.2 0 0.10 0.08 0.08 0.1 0.07 Myanmar Nepal Cambodia 0 Residential Business Public Myanmar Nepal Cambodia Source: ESMAP analysis. Residential Business Public 6 5.71 5.32 FIGURE 5.3 • Tariffs by customer class ($/month, charged in flat fee) 5 ($/month) 4 5.71 6 5.32 3 2.07 5 2 ($/month) 4 0.94 1.03 0.82 1 0.34 3 2.07 0 2 Myanmar 0.94 Nepal 1.03 Cambodia 0.82 1 0.34 0 Myanmar Business Residential Public Nepal Cambodia Residential Business Public Source: ESMAP analysis. — = no responses to the survey for this category of customer. MINI GRIDS FOR HALF A BILLION PEOPLE    163 (Jack and Smith 2017). Where mobile phone coverage is cable will be critical. By developing standard business pro- available, mobile payments can also facilitate payment in cesses for site assessment, supplier engagement, mini grid both prepay and credit schemes. operation, and maintenance and repair procedures, mini grid operators are in a better position to efficiently manage Regardless of whether a mini grid operator uses a prepay multiple sites. Similarly, if a mini grid operator were to tar- or credit payment scheme, smart meters can help facilitate get similar customer demographics, such as communities payment and collection. Companies such as Sparkmeter with the presence of a telecommunications anchor tenant, have developed meters geared toward off-grid mini grid then the operator could reduce the amount of variability, deployments. These meters can remotely connect or dis- and resulting inefficiency, in its operations. connect customers, manage customer billing information, integrate with mobile payment plans, and help operators In the after-sale segment of the value chain, another way track and limit electricity use in real time. to increase revenue is to provide additional goods and ser- vices, particularly those that rely on electricity. This provides An important note on tariff levels: mini grid operators need the added benefit of opening up new revenue streams for to be able to charge tariffs at a rate that allows for viable the operator; it uses the relationship the operator already operation, and that is commensurate with a community’s has with the end customer, and addresses additional needs ability to pay. Where there is a gap between the commer- or aspirations that the end customer may have. cially viable tariff and customers’ ability to pay, subsidies will be needed. Dictating tariffs that are below the cost of As an example, the Rockefeller Foundation and CrossBound- service in the absence of subsidies will prevent money from ary are running a pilot in Sub-Saharan Africa through their being reinvested into operating the system or into sub- joint Mini Grid Innovation Lab, providing capital to mini grid sequent expansion. When service is provided for the first companies, who in turn can provide financing to customers time, customers still have a choice; they might use solar for equipment purchases to increase the productive uses lanterns, install SHSs, or even continue with kerosene lan- of mini grid electricity. (More information regarding the terns. Therefore, to be commercially viable the mini grid methods and benefits of increasing demand, particularly operator needs to offer rates that compete with alternate through productive uses, can be found in chapter 3.) technologies, and reliable service that will attract cus- tomers to sign up for a connection. Chapter 9 provides a Reducing costs for mini grid developers and utilities detailed discussion of regulating retail mini grid tariffs. In addition to identifying ways the operator can generate additional revenue, it is important to manage costs, as As previously noted, if an operator is to pursue multiple greater profitability is driven both by higher revenue as well mini grid sites, designing business processes that are repli- as by reducing capital and operating costs. Mini grid oper- ators such as ESCOs and utilities can apply several strate- gies to manage their costs. BOX 5.4 Local staffing. By employing staff who work or live close PREPAY IS NOT JUST FOR MINI GRIDS— to the mini grid facility, operators can reduce the cost of ESKOM’S “POWER FOR ALL” SCHEME transporting personnel to the site, as well as avoid salary adjustments that result from living in higher-cost areas. In In 1988, the South African power utility company addition, having local staff has the added benefits of mak- Eskom launched the “Power for All” initiative, ing an operator more responsive to end customers and intended to supply electricity directly to domestic addressing potential operational issues. customers. At the time, Eskom’s primary custom- Procurement. When developing and operating multiple ers were mines and municipalities. When the con- mini grid sites, it is important not only to engage multiple cept was introduced, many of the target customers suppliers, but also to buy equipment for more than one mini were rural, and did not have a bank account or a grid site if possible. Negotiating with multiple suppliers with fixed address. Eskom sought to develop a scheme similar technical specifications can help drive down the ini- that would target these customers with the lowest tial equipment cost. Ordering multiple units and utilizing overhead, since many of them would be low-use volume discounts can help reduce capital expenditures. customers (Eskom n.d.). A recent study in Cape Town found that the investment returns to the util- Even if an operator is executing a single project, receiving ity were 10 percentage points higher in a prepay bids from facilitating organizations, such as equipment scheme when compared with a postpay scheme manufacturers, can help reduce initial capital expenditures. (Jack and Smith 2016, 2017). On the online mini grid management platform Odyssey, mini grid operators post project specifications for bidders. 164   MINI GRIDS FOR HALF A BILLION PEOPLE Equipment providers, financiers, and others can then offer their services to help complete the project. Mini grid operators make their money by Prepaid metering. As noted earlier, by having customers selling electricity to customers. Some earn pay for electricity up front (rather than metering and then additional revenue by selling appliances and elec- billing after the electricity has been used), operators can tro-mechanical equipment to increase demand for often reduce the amount of staff and overhead dedicated electricity. Strategies they deploy to manage their to billing and payment collection. With prepaid systems, costs include using local staffing, procuring for a the customer pays up front for a certain amount of electric- portfolio of mini grids, collecting tariffs through pre- ity, and once that has been used, the customer must pre- pay schemes, hedging against fuel cost increases, pay for another allotment. Prepaid metering is not without and deploying digital automation in their operation challenges, however. Issues to address include the need for and maintenance activities. cellular data coverage at the mini grid site and methods to protect against customer tampering. Fuel hedging. For operators using a fuel-based generator, MINI GRID FACILITATORS adding a fuel hedge contract may be valuable. Fuel hedg- In this second category of private-sector participants, facil- ing involves entering into a contract with a fuel supplier for itators are not selling electricity directly but are engaged in a certain amount of fuel at a certain price. In environments facilitating the mini grid sector. where fuel prices are volatile or are expected to rise due to increased demand or a removal of a subsidy, entering Engineering, procurement, and construction into such a contract can make costs more certain and also companies and system integrators maybe lower them, relative to a spot market price (Villadsen These companies focus on designing, installing, and com- 2017). However, before entering into a hedge contract, the missioning generation or distribution assets. They often operator should analyze historical and future price trends design power systems, procure and assemble various to avoid entering into a long-term agreement when spot fuel subsystems, transport equipment to the site, and oversee prices are below historical averages. In addition, hedging installation and commissioning. International EPCs and SIs and locking in fuel costs will require explaining to users why often have extensive experience operating in remote areas, their bills do not go down when fuel prices fall, and hedges although they tend to focus on large-scale industrial oper- may not be available in some markets. In these cases, fuel ations, such as extractive mines. These large EPCs have spot prices with a fuel surcharge may be more practical. also, at times, pursued rural electrification programs. For Digital automation. Digital technologies, such as mobile example, when Botswana Power Corporation issued a ten- payment or remote monitoring, can help reduce opera- der to build 20 solar mini grids, large EPCs such as Japan’s tional costs, and thus directly improve profitability. As noted Marubeni, China Harbor Engineering Company, and China earlier, using prepay meters with mobile payment can help Mechanical Engineering Company all expressed interest reduce overhead by eliminating the need for staff to collect (Benza 2017). Local EPCs may lack the scale to work on cash or to pursue customers for payment. Remote-moni- large industrial projects, but often have the requisite local toring tools can help monitor different mini grid equipment, knowledge and ability to profitably assist with small-scale ensure that it is functional, and proactively address any power installations. EPCs usually employ local resources issues that may arise, reducing overall maintenance costs to assist with projects; moreover, their capacity to design (Aaron 2014). power systems can be transferred to assisting with mini grid development. For example, using digital twin technology, a virtual model of a product, such as a solar hybrid system, allows for the EPCs that tend to specialize in mini grids are sometimes analysis of operational data to address issues before they called SIs. These companies buy different solution com- occur (Marr 2017). This might increase the availability of ponents from OEMs and package them into a complete operators’ assets by up to 15 percent and reduce main- solution. For example, an SI might purchase solar panels, tenance costs by up to 25 percent (Bradbury and others batteries, a generator, and power electronics, and bundle 2018). Mini grid equipment providers, such as SMA Solar these components into a containerized solution. SIs may Technology through its Sunny Portal system, allow for be independent, or they may affiliate with an OEM, acting operators to monitor a variety of parameters in real time as a local channel partner. SIs are generally local entities, to assess system performance. The Sunny Portal system and therefore can provide needed fabrication and assem- is the largest solar monitoring platform, with more than bly facilities close to project sites, whereas EPCs are typi- 20 GW of capacity monitored in real time, including both cally international firms. on-grid and off-grid installations (SMA Solar 2019). MINI GRIDS FOR HALF A BILLION PEOPLE    165 TABLE 5.7 • Sample mini grid experience of EPC companies Location of Company headquarters Mini grid experience Hatch Canada In Canada, nearly 300 remote communities do not have access to the grid. Generally, these communities have relied on diesel generation. Hatch, a global EPC, has developed its own microgrid controller called HµGrid, which integrates renewables with diesel generation to improve power quality and reliability, and to lower cost as a result of lower diesel consumption. Clarke Energy United Kingdom In February 2018, Clarke developed and installed a hybrid in Nigeria that used 4 megawatts of natural gas engines with 250 kilovolt-ampere of energy storage, and could operate either while connected to the grid or in islanded operation. Clarke integrated GE engines with a FlexGen energy storage system. Sterling & India Sterling & Wilson is an EPC executing projects of various types— including solar, cogeneration, Wilson and diesel generation—in more than 40 countries. In April 2018, the company established a business unit dedicated to developing hybrid systems. In May 2018, the company won an order for three hybrid systems, including one with 17 megawatt-hours of energy storage, across three sites in western Africa. Source: Sedighy 2017; Clarke Energy 2018; Kenning 2018. EPC = engineering, procurement, and construction company. TABLE 5.8 • Sample system integrators with technologies in mini grids Number of Company Country mini grids Sample project Tiger Power Belgium > 10 In November 2018, Tiger Power signed an agreement with the Ugandan rural electrification agency to develop three mini grids using its own containerized solar and storage system, backed up by a hydrogen generator, to provide power to 3,000 households. Winch Energy United > 30 In 2016, Winch Energy installed a 17 kW and 30 kW containerized solar and storage Kingdom system in the village of Nimjat in Mauritania. It is providing electricity to a school, dispensary, mosque, streetlights, and 70 households. Nayo Nigeria > 10 Nayo Tropical Technology sells individual mini grid components, and also integrates them into containerized systems. In January 2019, NTT won a N96.3 million (approximately $260,000) contract to power the Kare and Dadin Kowa communities with a 90 kW system. Sources: ARE 2018; Okafor 2019; Winch Energy 2019. kW = kilowatt. Sample EPC experiences with mini grids are presented in As noted earlier, the primary components of a solar hybrid table 5.7. Table 5.8 describes sample SI projects related to mini grid include solar panels, an energy storage solution, mini grids. a generator, and power electronics. Table 5.9 lists select OEMs that have deployed their technologies in mini grids. Original equipment manufacturers OEMs develop subcomponents of or complete mini grid Generating revenues for mini grid facilitators systems. While they may have a business dedicated to Broadly, facilitators have three ways to increase revenue providing equipment for mini grids, generally OEMs view while participating in the mini grid space: sell more of their mini grids as simply another channel or application to sell existing products or services, expand their scope to include their equipment. By pursuing mini grids as another chan- new products or services, or develop products or services nel, OEMs can improve their business operations, such adjacent to mini grid operations. As the number of mini as by increasing factory or channel use. Often OEMs will grids expands, it may be viable to explore all three options. establish dealers or will partner with local entities to pro- Perhaps the easiest option to manage would be for a facil- vide sales and after-sales support to customers. While not itator to simply sell more of its existing mini-grid-focused a necessity, OEMs generally develop and sell technology products or services. To do this, it would need to rely on the with some type of proprietary feature; this enables them to growth of mini grid deployments. This would necessitate provide a unique value proposition and differentiate them- working across multiple geographies; it might also require selves from the competition. working closely with regulators and other stakeholders to 166   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 5.9 • Sample original equipment manufacturers with technologies in mini grids Equipment Location of Number of mini Company provided headquarters Company description grids Jinko Solar Solar panels China 100 GW of solar panels to over 160 countries. 107 JA Solar Solar panels China 95 GW of solar panels to 135 countries and regions. 46 Canadian Solar Solar panels Canada Over 70 GW of products to over 160 countries. 29 Huawei Batteries—lithium China Large information and communications technology Battery: 85 ion LFP and infrastructure and smart devices company Battery inverter: battery inverters operating in over 170 countries and regions. 87 Alpha ESS Batteries—lithium Australia Energy storage solution provider, 1 GWh annual 38 ion LFP production capacity, working in 60 countries. Hoppecke Batteries—lead acid Germany Battery manufacturer focusing on power supply 20 backup, renewable energy, and motive power. SMA Power electronics Germany Manufacturer of solar and storage inverters. 45 Victron Power electronics Netherlands Manufacturer of solar and storage inverters. 32 Schneider Power electronics France Manufacturer of low- and medium-voltage industrial 29 Electric products. For mini grids, Schneider provides inverters and controls. Source: ESMAP analysis. GW = gigawatt; GWh = gigawatt-hour; kW = kilowatt; LFP = lithium ferrophosphate. establish rules to enable operators to identify and develop discrete solar space, particularly those who manufacture mini grid sites. For example, ABB has provided equipment stand-alone solar systems, have begun to diversify their to mini grid operators in India as well as Africa, using its local offerings. For example, SolarKiosk designed an E-HUBB, a presence and engineering capability to provide products stand-alone solar and storage solution that, in addition to that are tailored to those markets’ unique requirements. generating electricity 24/7, can be customized for various commercial applications, including as a small commercial A second way to increase revenue is to provide additional store, a health center, or even a movie theater. scope to a mini grid product or service. This allows the facil- itator to capture additional “wallet share.” For example, ven- Reducing costs for mini grid facilitators dors of individual components can expand their focus to To seek maximum profitability across the mini grid value system integration, providing packaged mini grid solutions chain, facilitators will need not only to grow revenue, but that can be deployed easily. Schneider Electric recently also to better manage operational costs. Costs can be bet- introduced containerized solar and battery systems, thus ter managed generally by employing mass customization, expanding beyond individual components. For EPCs, such replicability, and digital tools. a move might involve expanding into product development and testing. Financiers might move from project financing Mass customization. A key strategy in other industries, into corporate financing, providing debt to project vehicles from apparel to automobiles, mass customization is the as well as corporate expansion. To offer another product process of combining elements of mass production with or service allows companies to establish new partners and those of bespoke tailoring (The Economist 2009). As in develop new competencies. But expanding into adjacent mass production, the process uses a few standard plat- parts of the value chain is best done only after carefully forms that are common across products; this allows large studying customers’ needs and economic decision points investments to be amortized across a greater number of (Zook and Allen 2003). units. Once these standard components are produced, the remaining pieces of the product can be tailored for The third way to increase revenue is to expand into spaces individual applications. This allows companies to respond adjacent to the mini grid value chain. Just as mini grid oper- to unique customer requirements without developing ators are expanding their offerings into appliances and bespoke products. In the case of mini grids, OEMs that are other goods, so too can facilitators provide products to mini producing equipment should consider mass producing grid customers, perhaps even utilizing the customer rela- components that can apply to multiple applications, but tionship that the operator has established. While mini grid then customize them to suit the mini grid space’s unique facilitators have yet to embrace this strategy, many in the requirements. MINI GRIDS FOR HALF A BILLION PEOPLE    167 For example, a solar inverter manufacturer might develop of production for a manufactured good, the cost declines a component that can be installed in multiple products by a fixed percentage. Solar panels are an example of this: or applications, and yet use it in a unique offering for every time their production doubled, their cost declined mini grids. Outback Power manufactures individual solar by 18–35 percent (d’Avack 2010). By engaging with a vari- and battery inverters; in addition to offering these as ety of successful and high-potential mini grid developers, stand-alone products, the company integrates them into instead of targeting only the largest ones, facilitators can packaged offerings with batteries to deploy in off-grid improve the economics of their product or service offering environments. The package itself is scalable based on to the mini grid industry and achieve greater profitability as the project’s power requirements. This allows Outback the market expands. to amortize the fixed investment of building a factory to One important note is that for mass customization and develop inverter components across multiple applica- replicability to apply to mini grids, companies that supply tions, from small-scale off-grid systems to large-scale technologies and services to the mini grid industry will utility inverters. Thus, while performing minor tasks that need as large a market as possible. This means standard- customize the unit for off-grid duty, the OEM is simulta- izing technical specifications for mini grid equipment neously responding to the unique requirements of the across countries. The greater the discrepancy between mini grid operator. Operators too will benefit from such regulations across regions, the more that customization an arrangement: they will receive a product designed and and a bespoke approach will be required, thus adding tested for a variety of environments, and therefore proven cost and reducing the ability of facilitator organizations effective, yet tailored to their requirements. to scale. Financiers might also consider a mass customized ap- Digital tools. Using digital tools can help facilitator orga- proach to reduce overhead, while still responding to a nizations standardize their processes around mini grid mini grid operator’s unique needs, by establishing a set development and allow for a more efficient operational of standardized, yet flexible, processes or products. When cost position. Using remote monitoring and diagnostics Facebook, Microsoft, and Allotrope Partners established technologies, OEMs can analyze a mini grid’s performance the Microgrid Investment Accelerator in the spring of 2017, remotely, diagnose any issues, and dispatch personnel its goal was to use a blended capital fund to make grant, with the right tools and parts. Using such digital capa- equity, and debt investments available at a project or cor- bilities can avoid costly field visits, minimize unplanned porate level. Through this one fund, it could access various downtime, and even allow for future product revisions to investment instruments and utilize data analytics to offer be based on operational data. For EPCs, using digital tools a customized financing package based upon a mini grid such as geospatial analysis can help with process repli- operator’s maturity level. cability. Financiers can also use digital tools to help drive Replicability. For equipment manufacturers, it is relatively transparency in the projects and companies they finance. clear how mass customization as a concept can apply; In addition to using tools such as Odyssey to identify however, it is not as clear for companies that focus on projects, financiers can analyze operational data to gain process, such as EPCs. Just as mass customization rests greater insight into projects, benchmark performance on the ability to mass produce key pieces of a product, against other projects, and work with mini grid operators so too should process companies such as EPCs focus to implement best practices. on replicability. Developing standard processes for site preparation, distribution planning, staffing requirements, and packaging for transport will help reduce variability Mini grid facilitators make their money by and, therefore, cost. Using digital inputs and algorithms to selling products and services to companies process certain tasks quickly, automation can also play a that participate in the mini grid industry value chain, role. For example, geospatial analysis using satellite image particularly operators. They generally have three processing can aid in the planning and layout of distribu- ways to increase revenue: sell more of their existing tion grids. products or services, expand into new products or Given the current stage of the mini grid market, and the rel- services, or develop products or services adjacent atively few installed projects, it may be difficult to establish to mini grid operation. In parallel, facilitators gen- tested processes for mass customization and replicability. erally use three strategies to manage their costs: But the key will be to invest, execute a variety of projects mass customization of their products, replicability (of which some may not be profitable), and use that expe- of their services and processes, and the use of digi- rience to develop product strategies and implementation tal tools to increase efficiency. methodologies that can drive down cost. For every doubling 168   MINI GRIDS FOR HALF A BILLION PEOPLE FACILITATING COLLABORATION BETWEEN AND AMONG LOCAL AND There is a natural delineation between the steps of the value chain best served by INTERNATIONAL PRIVATE-SECTOR local entities, and those that are best served by ENTITIES international entities. Local entities are best posi- tioned to focus on aspects of the value chain that As the mini grid sector matures, and both international require knowledge of local rules and regulations, or and local private-sector entities begin to specialize in cer- that require coordination with the customer being tain aspects of the value chain, establishing partnerships served by the mini grid. International companies are among entities will be critical. There is a natural delineation best suited to elements of the value chain that do between (1) the steps of the value chain that are best served not require local context; they are ideally positioned by local entities, and (2) those best served by international to perform tasks that can be replicated across geo- entities. Facilitating the deployment of mini grids at scale graphic boundaries. will require collaboration between both types of entities; indeed, a partnership that uses the unique strengths of local and international industry is the only way to maximize overall value chain profitability. Facilitating deployment of mini grids at As noted earlier, local entities are best positioned to focus scale requires collaboration between local on those aspects of the value chain that require knowledge and international entities; indeed, such a partner- of local rules and regulations, or that require coordination ship that uses the unique strengths of local and with customers being served. These local entities might be international industry is the only way to maximize independent companies or affiliated with a larger interna- overall value chain profitability. Industry associ- tional company. ations play an important role in facilitating deal International companies are best suited to elements of making between local and international companies, the value chain that do not require local context; they are and helping the local industry players speak with a ideally positioned to perform tasks that can be replicated unified voice to both policy makers and the broader across geographic boundaries. For example, international marketplace alike. OEMs can deliver packaged mini grid equipment to oper- ators across countries. For the operator, working with an international company provides assurance of a competi- means that mini grid components and systems should be tively sourced and manufactured product, with the appro- procured from wherever they are cheapest to manufac- priate service and warranty provisions. For the OEM, being ture. Much of the mini grid value chain must be executed able to sell components or solutions across geographies by local industry; therefore, it is reasonable to consider would enable amortizing fixed costs, such as capital expen- removing barriers related to international industry partic- ditures on more units, reducing the overhead ascribed to ipation in equipment provision. each individual unit. Collaboration between local and international industry Table 5.10 outlines four types of partnerships among local does not simply have to be within the bounds of the mini and international industry, along with the benefits each grid space. For facilitators, such as OEMs, that need to party receives. develop a local presence to provide after-sales support and generate additional leads, there may be value in part- Though beneficial, facilitating partnerships between dif- nering with other companies that have already developed ferent local and international industry participants does such a presence. Likewise, non–mini grid companies may have certain policy implications, particularly related to derive value from working closely with mini grid providers, trade and local content requirements. While there may be particularly if their products require electricity. For opera- strong justifications for imposing tariffs or local content tors, working with companies that sell goods or services to requirements on certain mini grid components—whereby the same end customers may increase demand and help a certain portion of the mini grid’s bill of materials must improve the customer experience through a bundled prod- be manufactured locally—when viewed through the nar- uct or service offering. row lens of mini grid viability, these policies are typically counterproductive. Mini grid viability is influenced to a Productive uses of electricity are another area where col- significant degree by the ability to procure and install laboration between local and international industry can mini grid systems as cheaply as possible; this necessarily drive the mini grid sector to scale. For mini grid operators, MINI GRIDS FOR HALF A BILLION PEOPLE    169 TABLE 5.10 • Benefits of local and international partnerships Benefit to local industry partnering with international Benefit to international industry Structure Description entity partnering with local entity Buyer–supplier Local buyer company Access to technically proven Additional revenue from new purchases goods from technology with support for customer segment one or more international warranty, service and repair, etc. More product volume reduces supplier companies Competitively priced components amortized fixed cost, improves manufacturing use Channel partner or The local company will act Access to technically proven Lower-cost mechanism to establish distributor as the local agent for an technology with support for presence in new markets or remote international company, warranty, service and repair, etc. locations (for example, lower base selling goods and services Branded, differentiated technology costs from not having full-time staff) to customers improves competitiveness vs. Can be more responsive or tailored to other local companies (especially if local needs exclusive) Cross-channel partner The local company will act Access to proven technology that Access to new distribution partners as the local agent for an can generate revenues adjacent to to increase product sales volume international company in a core business Access to new customer segments different industry or market Increased customer awareness, Increased customer awareness sector than the one in which brand recognition the local company operates Access to other complementary Access to other partners with other products or services to generate complementary services revenue adjacent to core Equity or debt The local company receives Access to lower cost of capital Access to investments that can investment an investment from an Access to more financial products, generate higher return international entity earlier-stage capital Long-term growth potential Source: ESMAP analysis. identifying and facilitating productive uses is a key element COMPONENT MANUFACTURING of generating more revenue. One way to achieve this is For mini grids, three general categories of technology are taking an ecosystem approach to productive-use develop- needed: generation (which also includes storage), power ment. Companies that can provide goods and services that electronics, and distribution. Generation technologies can generate productive uses can align commercial efforts either produce energy or, in the case of energy storage, with mini grid operators, as they are ultimately seeking can save produced energy for later use. In the case of solar to address the same customer base. This “channel align- hybrid systems, the key components include solar arrays, ment” can help reduce costs, as it streamlines commercial energy storage, and backup diesel generation. Power elec- resources needed, while improving the end customer expe- tronics help combine multiple generation sources into a rience while providing bundled services (for example, sell- single flow of power, while distribution technology trans- ing electricity and refrigeration). ports the power from the plant to the end customer. Rather than being manufactured specifically for mini grids, these necessary components often are manufactured for gener- LOCAL AND INTERNATIONAL al-purpose energy applications that are then adapted for INDUSTRY PLAYERS ACROSS THE mini grid use. MINI GRID INDUSTRY VALUE CHAIN MARKET ASSESSMENT The mini grid industry value chain consists of a series of This step involves identifying sites that are suitable for mini activities, each with one or more participating private-sec- grid deployment. Generally, mini grid sites that have some tor entities, that support mini grid design, deployment, and combination of commercial and residential loads will be operations, as indicated in figure 5.4. more viable than those that are strictly residential. Com- mercial customers have higher energy needs and greater ability to pay; they also tend to cluster geographically (IFC 170   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 5.4 • Mini grid industry value chain Develop and Obtain operating Containerize where Perform repairs and manufacture mini concession, acquire applicable, transport other preventive grid components capital for project to site, install and maintenance development commission system Component Market Permitting & Grid design & Integration & Operation & After sales manufacturing assessment financing procurement installation maintenance Identify portfolio of Use survey data Connect and Identify and pursue villages, prioritize to specify system disconnect adjacent businesses, deployment, initial configuration customers, adjust enable additional site surveys tariffs, enroll new productive uses customers Source: ESMAP analysis. 2017). Moreover, ideal mini grid sites are typically far from a mini grid operator survey conducted in Cambodia indi- the main grid or have an unreliable and intermittent supply cated that mini grid operators in the country were serving of grid power. mostly residential customers (94 percent) (EDC 2001). Today, 84.6 percent of operators simultaneously serve Geospatial analysis can accelerate the identification of mul- residential, public, and commercial customers, while 68.1 tiple potential mini grid deployment sites, and assist with percent of Nepali and 11.1 percent of Myanmar operators prioritizing those sites best suited for development first. serve these customer groups (figure 5.5). The presence of The cost of acquiring relevant data continues to decline. productive uses and anchor customers, such as small busi- (Chapter 2 provides a detailed discussion of how geospatial nesses, could diversify the load profile of mini grid opera- analysis can help develop portfolios of economically viable tions. In Myanmar, peak demand for residential customers mini grids.) is concentrated during the evening, while commercial and One key determinant of the mini grid’s viability is the mix public customers may regularly consume electricity during of end customers. Generally, mini grids were primarily for the day, based on Cambodia’s peak-hour profile. This could residential customers who reside too far from the exist- allow nonresidential customers to function as important ing grid for service to be economically viable. A mini grid daytime loads or as anchor customers. Thus, having non- that serves both residential and commercial customers residential customers could enable a higher and more can achieve greater use, in part because commercial loads balanced distribution of the load, hence increasing the utili- tend to be active during the day, when solar or solar hybrid zation factor of the plant. mini grids can provide power from the cheapest source Mini grid operators in Cambodia deliver electricity to the they have—solar PV. In contrast, residential loads peak largest number of residential, business, and public cus- during evening hours or at night, when these mini grids tomers, followed by Nepal and Myanmar operators. The have to rely on more expensive diesel or battery backup mini grid operations in Nepal and Myanmar typically serve power. Larger commercial customers may even enter into one village of roughly 200–300 households; Cambodian a power purchase agreement with the mini grid operator; operators serve a cluster of villages containing, on average, these agreements can offer the operator a guaranteed rev- 3,842 households. In particular, the Cambodian operations enue, and therefore make it easier to finance the mini grid’s with distribution licenses cover approximately 4,182 house- development (IFC 2017). holds, more than five times the number of households Second-generation mini grids, while primarily serving served by consolidated operations. The number of busi- residential customers, also serve commercial and other ness customers also varies from one country to another: customers. Out of three countries in a World Bank survey, a typical mini grid operator in Myanmar and Nepal serve mini grid operations in Cambodia serve the most diverse about 5 and 12 business customers, respectively, while set of customers, including not only residential but also the typical Cambodian mini grid operator with distribution commercial and public clients. Until recently, however, mini licenses serves around 133 business customers, covering grid operations in Cambodia were quite different. In 2001, many more than local consolidated operators. MINI GRIDS FOR HALF A BILLION PEOPLE    17 1 FIGURE 5.5 • Customer types served by mini grids 1.9% N=104 Cambodia 7.7% 5.8% 84.6% N=407 Nepal 7.1% 16.5% 8.4% 68.1% 1.6% N=805 Myanmar 31.7% 55.6% 11.1% Residents Resident + public Resident + business Resident + public + business Source: ESMAP analysis. PERMITTING AND FINANCING Mini grid developers are increasingly using a portfolio approach; this reduces the risk of any one project by devel- If after the market assessment phase, a decision is made oping multiple projects in parallel. Assuming each mini grid to further pursue mini grid deployment, the necessary in the portfolio exhibits different characteristics, financiers regulatory review must be conducted to receive approval may feel more comfortable providing funds to the entire to develop the site. This may involve the operator sub- portfolio than to individual projects. If a particular project mitting a proposal to the rural electrification authority or does not meet expectations, others may provide high- notifying the distribution company of the intent to pro- er-than-expected returns to compensate. (Chapter 6 pro- ceed with a mini grid installation at that site, along with vides a more in-depth discussion on financing.) the proposed system configuration. While each mini grid likely may require separate approval, it may be more effi- The second-generation mini grids surveyed by the World cient from both the mini grid developer’s and the regu- Bank employed a diversified approach to raising capital. lator’s perspectives to discuss approval on multiple mini Mini grids in Nepal are financed by a more diversified grids in parallel. funding portfolio than those in Cambodia or Myanmar, which show relatively high reliance on a single type, Also at this stage, a mini grid operator should make a pre- either an equity or loan. In Nepal, almost 90 percent of liminary determination as to the financial viability of the operators received both a subsidy from the government mini grid or portfolio of mini grids. This includes assessing or organizations and a contribution from community or the revenue potential (a product of the number of custom- private developers, while this share was only 4 percent ers, the demand, and the tariff rate), as well as working with in Cambodia and 6 percent in Myanmar. Apart from sub- facilitating organizations to estimate the costs of the mini sidy-based grants and equity investments, bank loans grid equipment and financing, along with other expendi- are another important funding type for Cambodia and tures, such as fuel and administrative overhead. A more Nepal; of projects in these countries, 46 and 22 percent, accurate estimate of capital expenditure can be derived respectively, have ever applied for loans from local banks. during the design phase of the mini grid. However, at this Though lending money could increase the risks of capi- stage, a determination can be made as to whether it would tal management, loans are of great importance for some be worthwhile to invest additional money and resources Cambodian operators to cover the cost of large mini grid into the design of the mini grid. operations. In Myanmar, since the business scale is com- In addition, operators may need to demonstrate viability, paratively much smaller, operators were likely to start the including both the number of interested customers as well business with funding from the community, family loans, as their ability to pay for the electricity. Operators will often or their own savings. ask customers to prepay a certain amount for an electric- The main funding type for mini grid operations varies ity connection, mostly to ensure that customers are indeed widely across the three countries. Equity is the main one able to pay for ongoing electricity provision. For example, in Myanmar, with an average share of 37 percent for each to validate customers’ interest in receiving electricity, Husk operation. Though most Nepali operations receive both Power in India asks them to provide a deposit for three subsidy and community-based funding, subsidies from the months’ supply. government and other funding agencies (grants) generally 17 2   MINI GRIDS FOR HALF A BILLION PEOPLE account for a higher percentage than contributions from of the operations in Myanmar and Nepal, respectively, have the community and private developers (equity). In Cam- imposed the capacity limit (which is, on average, around bodia, overall, a considerable amount of funding is contrib- 100 watts) to end users by limiting the number and type of uted by loans from banks, relatives, or friends (figure 5.6). services they can use. GRID DESIGN AND PROCUREMENT Mini grid operations in Nepal rely mainly on hydropower, which is affordable and locally available, while most of Once the site has been identified and assessed and the the mini grid operations in Cambodia Nepal, carried out requisite approvals to proceed are in place, the next steps by consolidated licensees, rely on diesel for electricity center on finalizing the system design: identifying gener- generation. Solar PV technology was introduced into die- ation sources, modeling the design, validating the distri- sel-based mini grid operation in Myanmar only recently—in bution layout, issuing permits, and ensuring regulatory 2017, according to the survey results. In Myanmar, 7 per- compliance. cent of mini grid operations, particularly in Shan state, use hydropower to generate electricity. Mini grids using hydro Identifying generation sources resources in Myanmar, with an average capacity of 30 kW Depending on the location of the mini grid, there may be (median value), have a relatively larger capacity than die- one or more suitable options for electricity generation. sel-based operations, which have an average capacity of 19 Solar and solar hybrid mini grids account for 80 percent of kW (median value). the planned mini grids in ESMAP’s database of more than 26,000 installed and planned mini grids around the world. Modeling the design Hydropower remains an important generation technology To determine the appropriate mini grid size and finalize for mini grid sites located close to a river, in which case a the overall business case, it is necessary to model the mini run-of-river mini hydro system is often viable. grid’s operation to estimate fixed and variable operational In the World Bank’s survey of second-generation mini grid costs. Tools such as the HOMER Energy mini grid modeler operators in Cambodia, Myanmar, and Nepal, most opera- can use inputs from the site assessment phase, such as tors used diesel or hydropower as the primary generation the demand profile, to run an hourly dispatch model that source. Cambodia’s average generation capacity per mini will estimate the configuration and operating costs of the grid is larger than it is in Myanmar and Nepal, reflecting the mini grid. difference in the number of customers that the operators serve in each country. The median value of the genera- Validating the distribution layout tion capacity is 19 kW and 23 kW for Myanmar and Nepal, Another critical element of the mini grid design is the lay- respectively, while the average generation capacity in Cam- out of the distribution system, which can add significant bodia is about 250 kW. About 50 percent and 28 percent up-front costs to the mini grid. Factors that affect the cost FIGURE 5.6 • Average contribution share of each disclosed funding type 40% 37% 30% 23% 24% 20% 10% 10% 8% 7% 1% 3% 0% 0% Myanmar Nepal Cambodia Grant Equity Loans Source: ESMAP analysis. Note: Many developers did not disclose the sources of all of their funding; as a result, totals in each country do not sum to 100 percent. Funding sources are generally divided into three types: grants, equity, and loans. Grants include government sub- sidies and other subsidies from nonprofit organizations or funding agencies (for example, the Rural Electrification Fund and Reimbursable Advisory Service in Cambodia). Equity includes community contributions, investment from private developers or banks, as well as company-owned resources. Loans include those from banks, relatives, friends, and others. MINI GRIDS FOR HALF A BILLION PEOPLE    17 3 FIGURE 5.7 • Generation source of the mini grid mar use nonstandard poles for distribution, which may 2% pose a hazard and cause unreliable electricity service with 100% 3% 10% 7% significant voltage fluctuation. 1% 100% 75% INTEGRATION AND INSTALLATION 50% Once the design of the power plant and distribution sys- 90% 87% tem has been finalized, the various power plant compo- 25% nents and distribution infrastructure can be procured. Facilitating organizations, such as equipment manufactur- 0% ers and system integrators, can play a role in assembling N=805 N=407 N=10 the mini grid according to the operator’s requirements, determined through the prior steps. As these companies Myanmar Nepal Cambodia have experience helping to configure equipment in other applications, they may be able to partner with mini grid Diesel Solar Hydro Biomass Mix operators to ensure that the system is designed to meet the expected load. Source: ESMAP analysis. System integrators can develop prefabricated contain- erized mini grids, especially for solar hybrid mini grids. include the number of houses served, the distance over These shipping containers may house much of the nec- which electricity must be distributed, and the capacity of essary technology, including generators, batteries, and the distribution lines. Geospatial analysis can help deter- power electronics. In some instances, the solar panel array mine the optimal layout, considering physical barriers and may even be attached to the container. For example, Tiger specific customer locations, as discussed in more detail in Power offers a containerized solar and storage system that chapter 2. contains the solar, batteries, and power electronics needed to distribute electricity. Once a distribution layout has been designed, there still remains selection of the materials used to implement One important consideration when procuring mini grid the distribution system—namely, the type and size of the components or completing integrated systems is the poles and wires. Here, and with other mini grid compo- benefit of volume purchases. Component manufactur- nents, we see a wide variety of technologies and materials ers, system integrators, and EPCs are likely to offer some used around the world. This is partly the result of the dif- form of discount for purchasing more than one unit. Vol- ferent standards and specifications used across countries. ume purchases allow for greater resource use, potentially According to the results of the World Bank’s survey, oper- lower variable cost, and an ability to amortize fixed costs ators in Cambodia, Myanmar, and Nepal prefer different over more units. For mini grid operators and others pur- conductor materials. In Cambodia and Nepal, aluminum chasing components or systems, procuring for multiple is the most prevalent conductor material for the mini grid mini grid sites provides greater negotiating leverage when transmission system, adopted by more than 95 percent of interacting with suppliers. operators, although aluminum is still an important material Once the system has been designed and various compo- in Myanmar, the percentage of operators using it is much nents have been procured, the next step in the process lower than in the other two countries; instead, copper is involves preparing the mini grid site to receive the mini used for about 80 percent of operations. grid system components and finalizing the installation of Distribution network poles made of insulated, qualified, the system. Facilitating organizations, such as EPCs, play and standardized materials can effectively improve the a key role here as well. These companies can help clear quality and safety of electricity service. In Cambodia, the the land to receive the mini grid, arrange for transporting material for an electric pole is standardized: all of the the system to the site, and perform all the installation and operators reported that they use concrete for an electric wiring according to the system’s design. Unlike the prior pole. In contrast to Cambodia, it is difficult to find a stan- step related to the design and procurement of the sys- dard material for an electric pole for the mini grid opera- tem, this step in the process will rely almost entirely on tions in Myanmar and Nepal. About half of the operators local resources. To minimize the amount of field resources in Nepal use steel poles, while the rest rely on wood poles. required, as well as to reduce the amount of time dedi- Operators in Myanmar use nonstandard wood (50 per- cated to installation, mini grids can also be preconfigured cent) or bamboo (2 percent) poles, or both (22 percent). into transportable units. In general, around 87 percent of the operations in Myan- 174   MINI GRIDS FOR HALF A BILLION PEOPLE OPERATIONS AND MAINTENANCE If an operational issue is caused by a faulty part, finding a suitable replacement part quickly becomes critical. Mini Once the mini grid has been installed and commissioned, grid equipment manufacturers may be able to provide a operation can commence. Mini grids can be operated either replacement part, particularly if that part is still under war- locally or remotely. Local control requires an operator turn- ranty. It may be prudent, however, to create an inventory of ing on the unit and manually switching between generation the parts that are most prone to failure so operators can sources. With remote operation, the different generation resolve the issue quickly. To do so, it is important to discuss sources automatically turn on or off based on an algorithm with the equipment manufacturers which critical compo- that is designed to always dispatch the lowest-cost gen- nents would be best to inventory, then procure and store eration source first. Many second-generation mini grids those parts, and develop a process to ensure they do not require a local operator to monitor the system and make degrade while in storage (Dyess 2017). adjustments as demand changes. But with third-genera- tion mini grids supplied by international equipment manu- Respondents to the World Bank’s survey of second-gen- facturers, much of the control work is automated through eration mini grids documented difficulty in providing year- digital controls. This digital system can also send real-time round availability, overcoming resource shortages, natural operation data back to a monitoring system, using cellular calamities, or system breakdowns: 37, 27, and 50 percent connectivity, to allow technicians not located near the mini of mini grid operators in Cambodia, Myanmar, and Nepal, grid to monitor its operation. respectively, have experienced operational difficulties in supplying electricity to customers. In Cambodia, threats to Proper mini grid maintenance boosts system availabil- service reliability may include system damage from inclem- ity and reliability. Proper maintenance procedures are ent weather and shortages in resources and finances. founded upon understanding how the system is perform- Operations in Myanmar are greatly affected by resource ing, identifying issues when they occur, and having the shortage, mini grid system breakdown, and high diesel fuel right personnel and parts available to address the issue. prices. For Nepali operators, natural calamities, such as Having enough skilled technicians available is one of the floods and landslides, are the primary concern; they are biggest challenges facing operators. Husk Power, a mini also bothered by hydro shortages and finance shortage grid operator in India, has designed its own training pro- problems. Thus, these multiple challenges could lead to rel- grams, educating more than 100 technicians each quar- atively unreliable mini grid operation. Comparison between ter to supply the needed repair services on the company’s typical and worst months can indicate the stability with mini grids. (Training and skills gaps in the mini grid sector which the operators are able to provide electric services to are discussed in chapter 7.) end users throughout a year (figure 5.8). FIGURE 5.8 • Causes behind operators’ worst months of delivering electricity Responses to the survey question: “Are there certain months/seasons every year when the mini grid has difficulties in supplying electricity services to its customers?” Myanmar Nepal Cambodia (N = 805) (N = 407) (N = 92) Yes Yes 27% 37% No Yes No 50% No 50% 37% 63% Of developers who responded yes . . . Of developers who responded yes . . . Of developers who responded yes . . . 36.4%  Network line was broken 86.1% N  atural calamities like floods 44.1%  Heavy rains damage the and landslides system 36.3% S  hortage of nonmonetary resources 45.5% S  hortage of nonmonetary 29.4%  Shortage of nonmonetary resources (e.g., inadequate resources 18.2%  Resources are expensive water in case of hydro system) 23.5%  Shortage of finances 14.4%  Shortage of finances Source: ESMAP analysis. Note: The three most common causes behind service delivery challenges are shown. Respondents often gave multiple causes; as a result, percentages do not add to 100%. MINI GRIDS FOR HALF A BILLION PEOPLE    175 Daily availability of 2nd generation mini grids hours during the worst months. In Cambodia, the operators Nearly 90 percent of Cambodian mini grid operations pro- can provide power for 22.7 hours during typical months and vide electricity for 24 hours a day, while less than 10 per- 22 hours during the worst months (figure 5.9). cent of operations in Nepal and Myanmar are able to do the same during typical months. Large seasonal fluctuations Evening availability of 2nd generation mini grids in daily availability between typical and worst months are Virtually all Cambodian and Nepali mini grid operators pro- found among mini grid operations in Nepal: 60 percent of vide four full hours of electricity during the evening; more them provide electricity for more than 23 hours a day during than three-quarters of mini grid operators in Myanmar typical months, but the share falls to 34 percent during the provide more than three hours of electricity supply during worst months. On average, mini grid operators in Nepal pro- the evening between 6 and 10 p.m. A comparison between vide electricity for 17.1 hours a day during typical months daily availability and evening availability shows that most and for 13.4 hours during the worst months. In Myanmar, mini grid operators in Myanmar and Nepal focus more on the numbers are 5.2 hours during typical months and 4.5 electricity supply during the evening hours (figure 5.10). FIGURE 5.9 • Tiers of daily availability in typical and worst months Daily availability (typical months) N=91 Cambodia 11% 89% 2% 2% N=407 Nepal 38% 49% 9% 1% 1% N=805 Myanmar 67% 21% 4% 8% Daily availability (worst months) N=34 Cambodia 15% 6% 79% 2% 1% N=202 Nepal 9% 54% 33% 1% N=805 Myanmar 71% 18% 5% 5% 0–4 hours 4–8 hours 8–16 hours 16–23 hours 23–24 hours 24 hours Source: ESMAP analysis. FIGURE 5.10 • Tiers of evening availability in typical (left) and worst (right) months 6% N = 34 N = 91 Cambodia 100% Cambodia 94% 1% 2% N = 202 N = 805 N = 407 Nepal 98% Nepal 9% 25% 64% 3% N = 805 Myanmar 22% 45% 32% Myanmar 29% 35% 27% 6% 0–1 hours 1–2 hours 2–3 hours 3–4 hours 4 hours Source: ESMAP analysis. 176   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 5.11 • Daily peak hour profile 100% Percentage of peak load 80% 60% 40% 20% 0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Myanmar (N=764) Nepal (N=407) Cambodia (N=93) Source: ESMAP analysis. Evening hours between 6 and 10 p.m. are the peak hours for all three countries (figure 5.11). Apart from these, peak hours The mini grid value chain consists of a series can also arise during the day. More operators in Cambodia of activities that support the design, deploy- than the other two countries reported having daytime peak ment, and operations of mini grids. Both local and hours in the morning and afternoon. This may be because international companies participate. Upstream activ- they serve a relatively more diverse customer base. ities—those that occur before construction—include component manufacturing, market assessment, Reliability permitting and financing, and design and procure- The majority of mini grids in Myanmar, Nepal, and Cambo- ment. Downstream activities, which occur mostly on dia provide electricity at relatively high levels of reliability— site, include integration and installation, operations fewer than four disruptions per week totaling no more than and maintenance, and after-sales service. two hours of outage (figure 5.12). During the worst months, mini grid operators in all three countries have seen more unplanned outages, some lasting a day or more, while ser- vice reliability has plunged. For example, the mini grid operator MeshPower in Rwanda offers its customers three price levels for electricity; each AFTER SALE level comes with a different bundle of appliances, which may From the site identification process through operations include television sets, stereos, and fans. PowerGen Renew- and maintenance, mini grid operators build a relationship able in Kenya provides refrigerators and freezers to com- with the end customer. Operators can provide these cus- mercial customers, allowing them to provide cold beverages tomers with additional goods or services—for example, by or keep produce longer while generating needed electricity offering residential or productive-use appliances or by pro- demand to improve mini grid use (Edwards 2017). viding a small business a microloan to procure equipment. PROFIT POTENTIAL OF THE SOLAR MINI GRID In some instances, regulators, such as those in Tanzania, VALUE CHAIN actively encourage this type of activity as a way to further enhance cost recovery. Such an approach allows a mini grid ESMAP has developed a process for evaluating the private operator to act more as a bundled service provider, rather sector’s potential profit if mini grids are deployed so as to than simply an electricity company. achieve universal access. It considers the relative profit potential among various stakeholders along the mini grid Bundling such offerings or providing them side by side value chain. For the private sector to accelerate its invest- generates several benefits to the operator as well as the ment in mini grids, a sufficient profit margin is needed end customer. For the operator, providing and supporting along the entire chain. end users that consume more electricity help with the mini grid’s economic viability. For customers, a bundled offer- Figure 5.13 depicts this profit potential if SDG 7 is achieved ing of an appliance or commercial product with electricity by 2030. It does not show yearly revenue, which would makes the mini grid’s value proposition more accessible, as correspond to the total capital expenditure for the mini customers are in the market for energy services, not just grids deployed in a particular year, but rather the gross the electricity itself (USG 2016). MINI GRIDS FOR HALF A BILLION PEOPLE    17 7 FIGURE 5.12 • Tiers of reliability in typical and worst months Daily availability (typical months) N = 84 Cambodia 18% 31% 51% 1% N = 338 Nepal 7% 92% N = 805 Myanmar 15% 84% Daily availability (worst months) N = 84 Cambodia 21% 45% 33% 2% N = 338 Nepal 43% 55% N = 805 Myanmar 19% 22% 59% > 14 disruptions per week Max 14 disruptions per week Max 3 disruptions per week of total duration < 2 hours Source: ESMAP analysis. FIGURE 5.13 • Profit potential for facilitator organizations across the value chain, 2019 and 2030 2022 Value chain pro t 2030 Value chain pro t Total pro t: $790m Total pro t: $5.8b Project Project management, 2% management, 2% Balance of Balance of system, 13% Solar PV, 20% system, 13% Solar PV, 24% Battery Battery storage, 24% Logistics and storage, installation, 16% 21% Logistics and installation, 22% Distribution Distribution and meters, 20% and meters, 22% Source: ESMAP analysis. Note: These figures use the 21,557 mini grids installed globally in 2021 and 217,000 in 2030, in line with ESMAP’s data and analysis presented in the overview. The figures exclude taxes and financing. Balance of system includes such components as inverters, breakers, gensets, and converters. Estimates of the cost of service are derived from the HOMER analyses presented in chapter 1. 178   MINI GRIDS FOR HALF A BILLION PEOPLE profit after all variable production and manufacturing SUMMARY costs are taken into consideration (costs for taxes, trans- port, and financing have not been accounted for). The fig- Mini grids can play a pivotal role in the transformation of ure also does not show net income, that is, income after all the energy industry. In addition to providing reliable access other overhead, such as personnel costs, are accounted to electricity, mini grids can help improve power quality and for. Detailed assumptions on the cost and manufacturing resiliency. Fast-growing countries will need to invest signif- margins are documented on the companion website of icant resources in expanding and strengthening the elec- this handbook: www.esmap.org/mini_grids_for_half_a_ tricity infrastructure and should consider facilitating mini billion_people. grid deployments as a complement to grid extension. The mini grid industry offers significant profit potential to The private sector will be instrumental in ensuring the ongo- private-sector equipment and service suppliers, particu- ing viability of the mini grid sector. As the market transi- larly for solar PV, batteries, and power electronics. ESMAP tions from first- and second-generation to third-generation projects the annual profit potential across the value chain mini grids, more local and international industry players will at almost $5.8 billion by 2030 (figure 5.13).6 Regionally, the begin focusing on developing, building, and operating mini largest share (74 percent) of this will be in Africa, where the grids. The market for third-generation mini grids is nascent; most new mini grids are needed to achieve universal access however, to achieve universal access, the size of the poten- to electricity by 2030, followed by South Asia (13 percent of tial market certainly can provide the opportunity for profit. the total profit potential). The estimated revenue for private-sector participants is nearly $100 billion, with a cumulative profit opportunity of To derive these figures, ESMAP used current and projected nearly $6 billion over the next 10 years. component costs (as presented in chapter 1). When com- bined with the forecast for mini grid deployments, the total Deploying and operating mini grids involve a variety of expected revenue for the different component suppliers aspects, such as manufacturing mini grid components, was estimated. To determine each component supplier’s integrating them into a complete system, identifying and expected profit margin, ESMAP analyzed recent audited validating mini grid sites, designing the grid, procuring the financial statements; conservative estimates for profit relevant components, installing and operating the grid, and margin were then applied to the total revenue derived ear- further enabling productive uses of electricity. A variety lier. The largest profit centers for mini grid components will of private-sector entities compete in these aspects of the be solar PV, battery storage, and distribution infrastructure value chain: ESCOs and IPPs identify sites and procure and and technologies such as smart meters. As the costs of operate mini grids; OEMs and SIs develop individual mini solar PV and battery storage continue to fall, the fraction grid components and assemble them into prepacked units; of energy produced by solar PV and batteries will approach and EPCs install and commission the system, and in certain 100 percent, resulting in the profit potential for diesel drop- cases also can assist with system design, including distri- ping to nearly zero over the next decade. bution line layout. Private-sector equipment providers have an opportunity to Successfully scaling mini grids to achieve universal access generate additional revenue and profit growth in a market will require partnership among local and international enti- segment that has not traditionally played a meaningful role ties. Those aspects of design and operations that require in creating value. an in-depth understanding of local environments or require resources near the mini grid will naturally be aligned with the strengths of local industry. Those aspects that rely on scale, such as manufacturing, are naturally aligned with The largest profit centers along the mini grid international companies. The viability of mini grids will rely industry value chain by 2030 will be solar on both local and international companies maximizing PV, battery storage, and distribution infrastructure profit, through investing and pursuing actions that maxi- and technology (for example, smart meters). As mize revenue while minimizing costs. the costs of solar PV and battery storage continue To that end, removing certain market or regulatory obsta- to fall, the fraction of energy produced by solar PV cles, such as local content requirements or other import and batteries will approach 100 percent, resulting in restrictions, will allow private-sector players to minimize the profit potential for diesel dropping to nearly zero cost. Allowing mini grid operators to charge tariffs com- over the next decade. mensurate with the community’s willingness to pay, and using their customer relationships to provide additional services, will enhance their ability to generate revenue. MINI GRIDS FOR HALF A BILLION PEOPLE    17 9 Implementing digital technology will be key to driving more Dave, Rutu, Sandra Keller, Bryan Bonsuk Koo, Gina Fleurantin, Elisa Por- efficient operations across the value chain. In addition, facil- tale, and Dana Rysankova. 2018. Cambodia—Beyond Connections: Energy Access Diagnostic Report Based on the Multi-Tier Framework. itating organizations are critical actors in seeking a market Washington, DC: World Bank. https:/ /openknowledge.worldbank. environment favorable to mini grids, and across geographic org/handle/10986/29512. boundaries. Ultimately, achieving universal access through Dyess, Nicole. 2017. “Four Best Practices in Spare Parts Management.” mini grids will be achieved only if the private-sector partic- Plant Engineering, October 27. https:/ /www.plantengineering.com/ ipants active across the mini grid value chain can operate articles/four-best-practices-in-spare-parts-management/. profitably. EDC (Enterprise Development Cambodia). 2001. Survey of 45 Cambo- dian Rural Electricity Enterprises. Phnom Penh: EDC. Edwards, Isabelle. 2017. “Company Profiles: Mini-Grid Developers in Emerging Markets.” Bloomberg New Energy Finance. REFERENCES Eskom. N.d. “Prepayment Electricity.” Eskom, Government of Zambia. Aaron, Marty. 2014. “Remote Energy Monitoring Improves Plant Per- https://www.prepayment.eskom.co.za/. formance, Reduces Downtime.” Plant Engineering. https:/ /www. Galan, Johanna Christine. 2022. Personal Communication, June 10, plantengineering.com/articles/remote-energy-monitoring-im- 2022. proves-plant-performance-reduces-downtime/. Horizon Power. N.d. “Onslow Distributed Energy Resource Microg- African Development Bank. 2016. “Energy Utilities Database, 2014.” rid.” https://www.horizonpower.com.au/your-community/getting- Africa Infrastructure Knowledge Program Portal, January 23. future-ready/onslow-distributed-energy-resources-manage- https:/ /infrastructureafrica.opendataforafrica.org/ewicrwc/ener- ment-system-derms/ gy-utilities-database-2014. IEA (International Energy Agency). 2017. Energy Access Outlook 2017: ARE (Alliance for Rural Electrification). 2018. “ARE Member Tiger From Poverty to Prosperity. Paris: IEA. https://www.iea.org/reports/ Power Seals Deal to Run World’s First, Solar-Hydrogen Powered energy-access-outlook-2017 Mini-Grids in Uganda.” https:/ /www.ruralelec.org/news-from-are/ IEA. 2020. “Defining Energy Access: 2020 Methodology.” IEA, Paris. are-member-tiger-power-seals-deal-run-worlds-first-solar-hydro- https://www.iea.org/articles/defining-energy-access-2020-meth- gen-powered-mini-grids. odology. ARE. N.d. “Mlinda—Solar Mini-Grids for Off-Grid Rural Markets (India).” IEA. 2022. “Access to Electricity Data Set (SDG 7.1.1).” Excel database. https://www.ruralelec.org/project-case-studies/mlinda-solar-mini- https://trackingsdg7.esmap.org/downloads grids-grid-rural-markets-india. IFC (International Finance Corporation). 2017. From Gap to Opportunity: Asmus, Peter. 2013. “Power for the 20%.” Forbes, August 22. https:// Business Models for Scaling Up Energy Access. Washington, DC: IFC. www.forbes.com/sites/pikeresearch/2013/08/22/power-for-the- https://www.ifc.org/wps/wcm/connect/topics_ext_content/ifc_ 20-percent/#527e1e1a2ffb. external_corporate_site/sustainability-at-ifc/publications/publica- Attia, Benjamin, and Rebekah Shirley. 2018. “Distributed Models tions_report_gap-opportunity. for Grid Extension Could Save African Utilities Billions of Dollars.” Jack, Kelsey, and Grant Smith. 2016. “Charging Ahead: Prepaid Electric- Greentech Media, June 13. https:/ /www.greentechmedia.com/ ity Metering in South Africa.” Working paper E-89201-ZAF-2, Interna- articles/read/grid-extension-done-right-for-sub-saharan-af- tional Growth Centre, London. https:/ /www.theigc.org/wp-content/ ricas-utilities#gs.8vzp11. uploads/2017/07/Jack-Smith-2017-Working-paper.pdf. Balabanyan, Ani, Yadviga Semikolenova, Arun Singh, and Min A. Lee. Jack, Kelsey, and Grant Smith. 2017. “Prepaid Electricity: Better Service 2021. “Utility Performance and Behavior in Africa Today.” Technical Delivery for the Poor?” International Growth Centre, London. https:/ / Paper, World Bank, Washington, DC. https:/ /openknowledge.world- www.theigc.org/project/pre-paid-electricity-better-service-deliv- bank.org/handle/10986/36178. ery-for-the-poor/. Benza, Brian. 2017. “Asian Giants Bid for Rural Solar Power Tenders.” Kenning, Tom. 2018. “Sterling and Wilson to Build ‘Largest’ Battery MmegiOnline, July 12. https:/ /www.mmegi.bw/business/asian-gi- Energy Storage Project in Africa.” Energy Storage News, May 30. ants-bid-for-rural-solar-power-tenders/news. https://www.energy-storage.news/news/sterling-and-wilson-to- Blimpo, M., and Malcolm Cosgrove-Davies. 2019. Electricity Access in build-largest-battery-energy-storage-project-in-afri. Sub-Saharan Africa: Uptake, Reliability, and Complementary Factors Marr, Bernard. 2017. “What Is Digital Twin Technology—And Why Is It So for Economic Impact. Africa Development Forum Series. Washington, Important?” Forbes, March 6. https:/ /www.forbes.com/sites/ber- DC: World Bank. https://openknowledge.worldbank.org/bitstream/ nardmarr/2017/03/06/what-is-digital-twin-technology-and-why- handle/10986/31333/9781464813610.pdf?sequence=6&isAl- is-it-so-important/?sh=227641ed2e2a. lowed=y. Northwest Territories Power Corporation. 2018. “Adding Renewables to Bradbury, Steve, Brian Carpizo, Matt Gentzel, Drew Horah, and Joël Electricity Grids.” https://www.ntpc.com/energy-alternatives/cur- Thibert. 2018. “Digitally Enabled Reliability: Beyond Predictive Main- rent-alternative-energy-projects. tenance.” McKinsey and Company. https:/ /www.mckinsey.com/ business-functions/operations/our-insights/digitally-enabled-reli- Northwest Territories Power Corporation. N.d. “Fort Simpson Solar ability-beyond-predictive-maintenance. EnergyProject.”https://www.ntpc.com/energy-alternatives/current- alternative-energy-projects/fort-simpson-solar-energy-project. Clarke Energy. 2018. “OK Plast Hybrid Power Plant Solution, Nigeria.” https://www.clarke-energy.com/2018/ok-plast-hybrid-power- Odamo, Lily, Estomih Sawe, Mary Swai, Maneno Katyega, and Alli- plant-solution/. son Lee. 2017. Accelerating Mini-Grid Deployment in Sub-Saharan Africa: Lessons from Tanzania. Washington, DC: World Resources d’Avack, Francesco. 2010. “Maturing PV Technology: The Experience Institute and World Bank. https:/ /www.wri.org/publication/tanza- Curve Revisited.” Bloomberg New Energy Finance. nia-mini-grids. 180   MINI GRIDS FOR HALF A BILLION PEOPLE Okafor, Chineme. 2019. “43000 Rural Homes, Businesses to Bene- Winch Energy. 2019. “Renewable Energy and Utility Services for Off- fit from N1.95B REF.” This Day Live. https://www.thisdaylive.com/ Grid Communities in Africa and the Rest of the World.” https://www. index.php/2019/01/22/43000-rural-homes-businesses-to-bene- winchenergy.com/. fit-from-n1-95bn-ref/. World Bank. 2018. International Development Association Project Orlandi, Itamar. 2017. “Distributed Energy in Emerging Markets.” White Appraisal Document on a Proposed Credit in the Amount of SDR Paper, Bloomberg New Energy Finance. https:/ /data.bloomberglp. 243.4 Million to the Republic of Nigeria for the Nigeria Electrification com/bnef/sites/14/2017/11/BNEF-2017-11-21-Distributed-Energy- Project. Washington, DC: World Bank. http:/ /documents.worldbank. in-Emerging-Markets-White-Paper1.pdf. org/curated/en/367411530329645409/pdf/Nigeria-Electrifica- Philipp, Daniel. 2014. “Billing Models for Energy Services in Mini-Grids.” tion-PAD2524-06052018.pdf. Micro Energy International. https:/ /www.bmwk.de/EEE/Redaktion/ World Bank. 2022. “Small and Medium Enterprises (SMEs) Finance.” DE/Downloads/Publikationen/Praesentationen/2015-03-19-iv- https://www.worldbank.org/en/topic/smefinance. mini-grids-07-micro-energy.pdf?__blob=publicationFile&v=2. Zook, Chris, and James Allen. 2003. “Growth Outside the Core.” Har- Ramachandran, Vijaya, Manju Kedia Shah, and Todd Moss. 2018. “How vard Business Review, December. https://hbr.org/2003/12/growth- Do African Firms Respond to Unreliable Power?” Working Paper outside-the-core. 493, Center for Global Development, Washington, DC. https:/ / www.cgdev.org/publication/how-do-african-firms-respond-unreli- able-power-exploring-firm-heterogeneity-using-k-means. NOTES Rockefeller Foundation. 2019. “CrossBoundary Mini-Grid Facility Announces First Close with The Rockefeller Foundation and Ceniarth.” 1. There is no standard international definition of SMEs. They are https://www.rockefellerfoundation.org/about-us/news-media/ defined differently in the legislation across countries, in particular crossboundary-mini-grid-facility-announces-first-close-rockefel- because the dimension “small” and “medium” is relative to the size of ler-foundation-ceniarth/. the domestic economy. For statistical purposes, the Organisation for Schneider Electric. 2018. “Access to Energy: Schneider Electric Economic Co-operation and Development refers to SMEs as firms Solutions Panorama.” https:/ /download.schneider-electric.com/ employing up to 249 persons, with the following breakdown: micro (1 files?p_enDocType=Brochure&p_File_Name=Brochure+AC- to 9), small (10 to 49), and medium (50 to 249). This provides for the CESS+TO+ENERGY+2018_V3_BD.pdf&p_Doc_Ref=ACCESSTOEN- best comparability given the varying data collection practices across ERGY_BROCHURE. countries, noting that some countries use different conventions. The first four paragraphs in this chapter are drawn from the pages of the Sedighy, Mohammad. 2016. “Powering Remote Communities and World Bank’s website devoted to financing for SMEs: https:/ /www. Industrial Facilities.” Hatch. https://www.hatch.com/en/About-Us/ worldbank.org/en/topic/smefinance. Publications/Blogs/2016/11/Powering-remote-communi- ties-and-industrial-facilities. 2. RAO Energy in Russia, TANESCO in Tanzania, JIRAMA in Madagas- car, and KPLC in Kenya are utility companies that operate dozens of Selding, Peter. 2015. “Established Imagery Providers Face Changing mini grids nationwide. These mini grids are typically diesel powered Competitive Landscape.” SpaceNews, September 24. https:/ /space- (or, in the case of JIRAMA, hydro powered). They tend to be large, news.com/established-imagery-providers-face-changing-competi- typically on the order of several hundred kilowatts to a few mega- tive-landscape/. watts. Some utilities (in Niger, for example) have started to hybridize Shakti Foundation. 2017. Beyond Off-grid: Integrating Mini-Grids with their diesel systems with solar PV panels. India’s Evolving Electricity System. New Delhi: Shakti Foundation. 3. These estimates are in line with data from ESMAP’s global database https://shaktifoundation.in/wp-content/uploads/2017/09/Inte- of mini grid projects presented in the introduction, and the Global grating-Mini-grids-with-Indias-Evolving-Electricity-System.pdf. Electrification Platform analysis presented in chapter 2. SMA Solar. 2019. “Sunny Portal.” https://www.sma.de/en/products/ 4. There are a total of 76 utilities in Sub-Saharan Africa, including trans- monitoring-control/sunny-portal.html. mission and generation utilities of which 61 are vertically integrated T&D India. 2018. “Tata Power DDL Inaugurates Its Second Microg- utilities, or companies that have a distribution function. rid Project.” http://www.tndindia.com/tata-power-ddl-inaugu- 5. The number of mini grids that are modeled to be deployed in 2022 is rates-second-microgrid-project/. 6,298, with a cumulative deployment of 27,771. By 2030, more than Tenenbaum, Bernard W., Chris Greacen, and Dipti Mulrajsinh Vaghela. 49,000 mini grids need to be deployed annually. This is the pace of 2018. Mini Grids and the Arrival of the Main Grid: Lessons from development required to reach 490 million people by 2030, through Cambodia, Sri Lanka, and Indonesia. ESMAP Technical Report No. more than 217,000 mini grids. 013/18. Washington, DC: World Bank. https:/ /openknowledge. 6. Rather than provide a definitive number, this analysis is designed worldbank.org/bitstream/handle/10986/29018/134326.pdf?se- to understand the relative profit potential among different mini grid quence=6&isAllowed=y. value chain stakeholders. Such an analysis can be used to determine The Economist. 2009. “Mass Customisation: Combining Elements of the viability of establishing business lines focused on the mini grid Mass Production with Those of Bespoke Tailoring.” The Economist, market. The data reflect the profit potential after all variable produc- October 22. https:/ /www.economist.com/news/2009/10/22/ tion and manufacturing costs are taken into consideration. Detailed mass-customisation. assumptions and methodology are documented in the companion Velamala, Mohith. 2016. “India’s Diesel Addition Is a Target for Solar and website to this handbook: www.esmap.org/mini_grids_for_half_a_ Battery.” Bloomberg New Energy Finance. billion_people. Villadsen, Bente. 2017. “Should Regulated Utilities Hedge Fuel Cost and If So, How?” The Brattle Group. https:/ /surfa.memberclicks.net/ assets/articlesfromconference/should_regulated_utilities_hedge_ fuel_cost_and_if_so%20-%20how_villadsen_04-2017.pdf. MINI GRIDS FOR HALF A BILLION PEOPLE    181 CHAPTER 6 FINANCING SOLAR MINI GRID PORTFOLIOS AND END-USE APPLIANCES CHAPTER OVERVIEW This chapter lays out the main barriers that private-sector mini grid developers face when trying to finance their mini grid projects and portfolios, the types of financing mechanisms that are available to them, and how different financing mechanisms address the different barriers. The chapter also presents details of the World Bank’s mini grid investment portfolio. The fundamental message of this chapter is that governments and their development partners should prepare a package of financial support for mini grid developers. The elements of this financial package are access to equity, debt, subsidies, and risk-sharing mechanisms. The nature and role of these elements will differ with the type of developer being supported. The economic and social case for promoting mini grids energy mini grids, which tend to have a high share of ini- has already been laid out. Many of the Bank’s client coun- tial capital costs, because there are no major fuel costs tries are now looking to private firms to develop mini grids. (though there are some costs for supplemental diesel Some countries will opt for publicly funded mini grids, and generation). The payback period for mini grids is long, some may use a combination of private- and public-sec- as most generation and distribution assets have lives of tor developers. While the focus of this chapter is access to 10–15 years. For example, using costing data from chap- finance for private developers, publicly funded mini grids ter 1, a hybrid solar diesel 120-kilowatt (kW) mini grid that are also considered. could serve 500 households in Sub-Saharan Africa would have initial capital costs of around $500,000 and variable The fundamental message of this chapter is that govern- costs of about $45,000 a year. ments and their development partners should prepare a package of financial support for mini grid developers. The It is important that these developers be able to attract sig- elements of this financial package are access to equity, nificant funding, so that mini grids are developed rapidly on debt, subsidies, and risk-sharing mechanisms. The nature a large enough scale to provide electricity quickly to large and role of these elements will differ with the type of devel- numbers of unserved people. oper being supported. From a financial viewpoint, there are two broad categories of mini grid developers: larger international and local firms that already have significant access to equity and debt, WHAT ARE THE FINANCIAL NEEDS OF and smaller, mostly local firms that do not have such ready MINI GRID DEVELOPERS? access. Since the financial needs of the two categories are different, the support package for the two will be corre- All mini grid developers need to finance both preinvest- spondingly different. ment and initial capital costs. Preinvestment costs include It is important to note that female-led enterprises and proj- site visits; feasibility studies; discussions with communi- ect developers may require an expanded support package, ties, government officials, and other interested partners; as women may face additional barriers in accessing finance and other elements of due diligence. Initial capital costs (box 6.1). are very important for mini grids, particularly renewable 182   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 6.1 WOMEN’S LIMITED ACCESS TO FINANCE Female developers face particularly significant chal- even after controlling for individual characteristics, such lenges in accessing finance. Women own and lead as income, education, and age (Demirguc-Kunt, Klap- roughly 6.6 million formal small and medium enter- per, and Singer 2013). Drivers can include collateral con- prises and 39 million microbusinesses, forming about straints caused by a lack of land ownership, among other 28 percent of business establishments in developing aspects. Further, across regions, legal barriers inhibit countries. However, women-owned small and medium women’s rights to inheritance, immovable assets, and enterprises face a significant credit gap (IFC 2017). collateral, which in turn affects their access to finance Women are 15 percent less likely than men to have a (World Bank 2020). Women also face discriminatory bank account, and significantly lag behind men in sav- social and cultural norms and gender biases, which ing and borrowing through formal financial institutions, make them less credible to potential investors. Different financing packages are required Female-led enterprises and project devel- for different types of mini grid develop- opers may require an expanded support ers, consisting of a combination of debt, equity, package, as women often face additional barriers subsidy, and risk-sharing mechanisms. Larger to accessing finance. Women own and lead roughly international and local firms tend to already have 28 percent of business establishments in develop- significant access to equity and debt, while smaller, ing countries, are 15 percent less likely than men to mostly local firms usually do not have ready access have a bank account, and significantly lag behind to equity and debt. men in saving and borrowing through formal finan- cial institutions. WHAT TYPES OF FINANCE ARE MINI GRID DEBT AND EQUITY AVAILABLE? INVESTORS The funds provided by commercial entities can be clas- While governments and their development partners, along- sified as project finance and corporate finance. In non- side philanthropic entities and other nongovernmental recourse project finance, the returns to the financiers organizations, will continue to play an important role in come only from the project’s revenues; financiers cannot providing grants and risk mitigation mechanisms to mini get their returns from anything else the company does or grid developers, private-sector investment will be needed owns. Project finance is normally available only to large to achieve the exponential growth necessary to bring mini companies that have predictable revenues and low risks. grid electricity to half a billion people by 2030. Indeed, a However, some of the larger mini grid developers may be lack of access to affordable financing has been highlighted able to access project finance funds, and smaller-scale as a barrier to growth by mini grid developers. A reported developers may also be able to access project financing if 37 percent of mini grid developers in Cambodia, Myan- they aggregate individual mini grids into a portfolio. mar, and Nepal have said that access to capital is a severe or very severe constraint, and two-thirds rely on their own In corporate finance, also called balance sheet finance, pools of capital to fund projects, according to World Bank the returns to private financiers come from all of the surveys of mini grid developers in these countries. company’s assets, not just the revenues from the project. Smaller mini grid developers are likely to get only corpo- Positive momentum is underway, however, and both the rate finance funds. pace and scale of private-sector investment in mini grids are increasing. The Energy Sector Management Assis- Private finance can be in the form of debt or equity. In addi- tance Program (ESMAP) identified 188 unique investment tion, mini grid developers can access direct and indirect deals between investors and active private sector mini grid subsidies. The sources of these types of finance are shown developers between 2010 and 2022, totaling $557 million in table 6.1. MINI GRIDS FOR HALF A BILLION PEOPLE    183 TABLE 6.1 • Types and sources of commercial finance Type of finance Source of finance Debt, including mezzanine, which is a loan Local banks, impact investors that can convert later to equity Equity Impact investors, own funds from project developers and associates, commercial investors Direct and indirect (risk-sharing) subsidies Governments, development partners, nonprofit organizations Source: ESMAP analysis. FIGURE 6.1 • Cumulative private-sector investment in mini grid companies 600 500 Millions USD 400 300 200 100 0 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Source: ESMAP analysis of 188 deals between investors and mini grid companies between 2010 and 2022, using publicly available data. in 2022 dollars, adjusted for inflation (figure 6.1). Active a marked slowdown in dealmaking during the COVID-19 developers are those that are registered companies and pandemic (figure 6.2). have built mini grids, are planning mini grids, are members The top 20 investors in terms of total investment in mini of a mini grid industry association, and/or have secured grid companies account for 54 percent of the cumulative investment for mini grids. total of $557 million (figure 6.3). None of the current top The deals were completed between 54 developers and 100 20 investors had made any investments in mini grids prior investors, and averaged $3 million per deal. The largest deal to 2015. was a $41 million equity investment by the Chubu Electric The largest investors are a mix of private companies and Power Company in the Indian mini grid developer OMC organizations and funds supported by governments. Three Power in 2022. It is important to note that these are the large electric power companies feature among the top 20: totals for investment deals for which we were able to find the Chubu Electric Power Company (Japan), Enel Green publicly available data. As a result, these totals may under- Power (Italy), and Engie (France). Multinational corpora- represent the full scale of investment in private sector mini tions in the top 20 include Mitsui & Co. (Japan), Bank of grid developers. Indeed, we were able to find information America (US), Microsoft through their Climate Innovation for only 54 of the 177 active mini grid developers, indicating Fund (US), and Shell through Shell Technology Ventures that the investment totals reported here may be just half of LLC (UK). Several government-supported funds and enti- the real amount. ties are among the top 20 investors as well: ElectriFi, an The number of deals between investors and private sec- impact investment fund supported by the European Union; tor developers grew from an average of just 3 per year in InfraCo Africa, a private investment firm supported by the 2010–14, to an average of 28 per year in 2017–21. The aver- UK government; REPP—the Renewable Energy Perfor- age deal size also grew, from less than $1 million per deal mance Platform, an impact investment fund supported by in 2010–14, to more than $2 million per deal in 2017–21. the UK government and the European Union; the Develop- Almost half—22 out of 54—developers received their first ment Finance Corporation, a US government–supported investment in just the past three years. However, there was investment organization; the European Investment Bank; 184   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 6.2 • Annual number of deals between FIGURE 6.3 • Top 20 investors in mini grid companies investors and mini grid companies, 2010–22 by cumulative investment 60 Chubu Electric Power Co., Inc. ElectriFI 50 InfraCo Africa Number of deals 40 Mitsui & Co., Ltd REPP 30 Enel Green Power 20 Development Finance Corporation (DFC) European Investment Bank 10 Engie International Finance Corporation (IFC) 0 2010 2012 2014 2016 2018 2020 2022 Shell Technology Ventures LLC Green Climate Fund Source: ESMAP analysis of 188 deals between investors and mini grid companies between 2010 and 2022, using publicly available data. FMO Claritas Capital FMO, the Dutch development bank; and Norfund, an invest- Norfund ment fund supported by the Norwegian government. Two Chapel Hill Denham specialized private funds are also among the top 20: Infra- Infracredit credit, an impact investment fund in Nigeria; and Omid- ARCH Emerging Markets Partners yar Network, an impact investor in the United States. Two Bank of America multilateral development institutions (the International Finance Corporation and the Green Climate Fund) as well Microsoft Climate Innovation Fund as four private investment firms (CrossBoundary, ARCH 0 10 20 30 40 50 Emerging Markets Partners, Claritas Capital, and Chapel Millions (USD) Hill Denham) make up the rest of the top 20 investors. 2015 2016 2017 2018 2019 2020 Different categories of investors emerge when they are cat- 2021 2022 egorized by their average deal size, and how many invest- Source: ESMAP analysis of 188 deals between investors and mini grid ments they have made. Table 6.2 presents some of these companies between 2010 and 2022, using publicly available data. categories. Four of the top five-most-active investors are government- combination of private investment houses (for example, supported impact investment funds or entities: ElectriFi, Chapel Hill Denham) and government-supported funds REPP, the Energy and Environment Partnership Trust Fund (for example, Norfund), although two large companies also (EEP Africa), and InfraCo Africa. All On, the third most active feature in this list: Microsoft and Bank of America. investor, is an impact investment fund based in Nigeria and The last category of investors highlighted in table 6.2 con- supported by Shell. The top investors seeking big deals are sists of active mini grid developers (having made at least for the most part large companies (for example, Chubu, two investments in mini grid companies since 2010) with Enel, and Mitsui) and international investment institutions average deal sizes of less than $5 million. This category (for example, the Development Finance Corporation, Euro- comprises various types of investors, but they all focus on pean Investment Bank, International Finance Corporation, impact. and Green Climate Fund). Only one is a stand-alone private investment house: Claritas Capital. As a final insight gleaned from the database of investment deals in mini grid companies, the number of equity deals The list of investors funding early stage start-ups are has declined, and the number of debt deals has risen (fig- grouped together because their average deal size is greater ure 6.4). than $5 million, indicating they are looking for well-estab- lished companies, but less than $10 million, indicating The downward trend in equity combined with the upward that they are looking to catalyze the scale-up of small but trend of debt could be an indication of the maturation of well-established developers. These investors include a the mini grid industry. As companies develop more exten- MINI GRIDS FOR HALF A BILLION PEOPLE    185 TABLE 6.2 • Categories of investors in mini grid companies, 2010–22 Investors seeking large Investors funding Investors making small but Most active investors deal sizes early-stage scale-up targeted investments (Top investors with at least (Top investors with an average (Top investors with average (Top investors with two or more three deals, ranked by deal size larger than $10 million, deal sizes of $5–$10 million, deals and average deal sizes under number of deals) ranked by average deal size) ranked by average deal size) $5 million, ranked by deal size) ElectriFI Chubu Electric Power Co., Inc. Norfund Shell Technology Ventures LLC Renewable Energy Enel Green Power Chapel Hill Denham Omidyar Network Performance Platform Development Finance Infracredit FMO (REPP) Corporation ARCH Emerging Markets REPP All On European Investment Bank Partners Engie Energy and Environment International Finance Bank of America Caterpillar Ventures Partnership Trust Fund Corporation Microsoft Climate Innovation Tao Capital (EEP Africa) Mitsui & Co., Ltd Fund InfraCo Africa Total Energy Ventures Green Climate Fund Development Bank of Central US African Development Proparco Claritas Capital African States (BDEAC) Foundation CRE Venture Capital ElectriFI Engie Acumen Swedfund FMO (Dutch Development Bamboo Capital Partners Beyond the Grid Fund for Bank) Zambia Oikocredit CrossBoundary Energy InfraCo Africa All On Access Nigeria Infrastructure Debt Energy Access Ventures Shell Technology Ventures Fund SunFunder LLC Pi Investments EEP Africa Omidyar Network SunFunder Facility for Energy Inclusion Off- Grid Energy Access Fund (African Acumen Development Bank) Energy Access Ventures GReeN Investor Group Charm Impact Ashden Trust Source: ESMAP analysis of 188 deals between investors and mini grid companies from 2010 to 2022, using publicly available data. FIGURE 6.4 • Trends in debt and equity 70% 60% Share of all deals 50% 40% 30% 20% 10% 0% 2015 2016 2017 2018 2019 2020 2021 2022 Share of all deals that are debt Share of all deals that are equity Share of all deals with undisclosed type Source: ESMAP analysis of 188 deals between investors and mini grid companies between 2010 and 2022, using publicly available data. 186   MINI GRIDS FOR HALF A BILLION PEOPLE would cover approximately 30 percent of the $8.9 billion total planned investment in mini grids globally, accord- Both the pace and scale of investment in ing to ESMAP’s analysis of funding commitments from mini grids are increasing. ESMAP has iden- donors and development partners. tified 188 unique investment deals in active pri- vate-sector mini grid developers between 2010 and Meanwhile, the World Bank has committed more than $1.4 2022, totaling $557 million in 2022 dollars (adjusted billion to mini grids through at least 2027. At the end of for inflation). The deals were completed between 54 June 2022, its mini grid portfolio consisted of 41 projects developers and 100 investors, and averaged $3 mil- approved by the World Bank board, covering 31 countries lion per deal. The number of deals between investors (table 6.3). The investment plans of this portfolio project and private-sector developers grew from an average the deployment of more than 3,000 mini grids by 2027, of just 3 per year from 2010 to 2014, to an average of with the expectation of bringing electricity to more than 13 28 per year from 2017 to 2021. The largest investors million people. Even more important than the direct impact are a mix of private companies and organizations is the expected “crowding in” of private-sector develop- and funds supported by governments. Four of the ment and investment, which is expected to leverage $1 top five most active investors are government-sup- billion of cofinancing from private-sector, government, and ported impact investment funds or entities. development partners. With strategic market-enabling work supported by ESMAP (for example, geospatial planning, feasibility studies, and rapid response support to project implementation, among sive track records, they become more attractive (less risky) other types of support), the portfolio is expected to cata- investment opportunities for debt investors because the lyze private-sector engagement in mini grid markets world- developers have a proven cashflow with which to repay wide. A recent example of this work in action is the flagship their debt. But the data also demonstrate an upward trend “piloting at scale” project in Nigeria, where the World Bank’s in more undisclosed types of deals—perhaps another sign financial commitment is crowding in investment from the that the mini grid industry is maturing. As more sophisti- private sector: $150 million in International Development cated investors participate and as more complex deals are Association funding for the project’s mini grid component brokered, more and more of the details of individual trans- will bring in more than $230 million in private-sector invest- actions may remain confidential—particularly for invest- ment in mini grids. ment entities that are not supported by governments or other donors. When investing in mini grids, the World Bank projects employ various financing mechanisms. The following list highlights some examples: DEVELOPMENT PARTNERS’ • The Electricity and Water Access and Governance Proj- INVESTMENT IN MINI GRIDS ect in the Democratic Republic of Congo has the largest expected coinvestment from the private sector in the Development partners, including the World Bank, have World Bank’s current portfolio. The World Bank team increased financing for mini grids, from millions of dollars supported the government of the Democratic Republic in the 2000s to billions of dollars in 2018. A group of 15 of Congo to develop the project in close collaboration major international donors and development partners, with the International Finance Corporation’s Scaling including the World Bank, has collectively committed Mini Grids program. This program is providing signifi- more than $2.6 billion just to mini grid investment (that is, cant upstream support to develop strong contractual excluding funding for technical assistance and research). frameworks that are expected to attract $300 million in Six of these organizations have investment commit- cofinancing from the private sector for the construction ments for mini grids that total more than $100 million: and operation of several large mini grids under a con- the Agence Française de Développement, the African cession contract, with each mini grid serving tens of Development Bank, the Global Environment Facility, the thousands of customers. The large investment required German Society for International Cooperation, KfW, the for each of these mini grids makes them attractive to Islamic Development Bank, and the World Bank. Moving infrastructure project investors, who, in contrast to the from commitments to disbursements and investments, more venture capital–oriented investors described in however, will require sustained efforts and close collabo- the previous section, seek large transactions backed by ration with governments and the private sector. ESMAP a strong contractual framework between the mini grid estimates that all donor funds committed to mini grids developer and the government. MINI GRIDS FOR HALF A BILLION PEOPLE    187 TABLE 6.3 • World Bank mini grid investment portfolio as of June 30, 2022 Project details Mini grid investment (US$, millions) Expected mini grid results Approval Closing World Mini Country Project name date date Bank Cofinancing Total grids Connections People Afghanistan Herat Electrification 13-Jun-17 31-Jan-23 1.0 — 1.0 4 444 2,220 Project Argentina Renewable Energy for 07-Apr-15 30-Jun-22 6.9 1.2 8.1 14 720 3,600 Rural Areas Project Bangladesh Rural Electrification 20-Sep-12 18-Dec-23 22.6 17.6 40.2 30 15,600 78,000 and Renewable Energy Development II (RERED II) Project Burkina Faso Electricity Sector 30-Jul-13 31-Dec-22 16.6 — 16.6 18 11,135 55,675 Support Project Burkina Faso Solar Energy and 21-Jun-21 31-Dec-28 40.0 25.0 65.0 180 78,000 390,000 Access Project Burundi Solar Energy in Local 28-Feb-20 01-Mar-26 37.0 — 37.0 45 31,000 155,000 Communities Cameroon Rural Electricity Access 13-Dec-18 30-Jun-25 19.0 — 19.0 12 3,000 15,000 Project for Underserved Regions Central African Water and Electricity 17-Jan-18 30-Jun-22 5.2 — 5.2 2 1,200 6,000 Republic Upgrading Project Central African Electricity Sector 03-Jun-22 31-Dec-27 15.0 - 15.0 5 20,100 100,500 Republic Strengthening and Access Project Chad Energy Access Scale Up 24-Mar-22 30-Jun-27 100.0 50.0 150.0 25 105,000 525,000 Project Congo, Electricity Access & 04-May-17 31-Oct-23 37.0 0.5 37.5 9 5,500 27,500 Dem. Rep. Services Expansion (EASE) Congo, Electricity & Water 31-Mar-22 30-Sep-29 263.5 300.0 563.5 123 353,000 1,765,000 Dem. Rep. Access and Governance Project Ethiopia Ethiopia Electrification 01-Mar-18 07-Jul-23 7.5 — 7.5 12 6,240 31,200 Program (ELEAP) Ethiopia Access to Distributed 29-Mar-21 31-Mar-27 270.0 50.0 320.0 600 240,000 1,248,000 Electricity and Lighting in Ethiopia Gabon Access to Basic 17-Sep-15 30-Nov-22 8.6 — 8.6 2 1,100 5,500 Services in Rural Areas and Capacity Building Project Guinea Electricity Access Scale 15-Feb-19 31-Dec-23 3.5 3.5 7.0 10 10,000 50,000 Up Project Haiti Renewable Energy 25-Oct-17 31-Dec-24 15.1 24.0 39.1 43 74,490 100,000 for All Haiti Modern Energy Services 25-Oct-17 30-Apr-28 5.0 12.5 17.5 n.a. n.a. n.a. for All (CTF DPSP) Kenya KE Electricity 31-Mar-15 31-Dec-22 10.0 11.7 21.7 7 3,640 18,200 modernization project Kenya Off-grid Solar Access 26-Jul-17 30-Jun-23 40.0 125.0 165.0 137 27,000 135,000 Project for Underserved Counties Lesotho Renewable Energy and 30-Jan-20 31-Jan-27 20.0 10.0 30.0 40 4,800 24,000 Energy Access Project Liberia Renewable Energy 11-Jan-16 31-Dec-23 20.4 2.0 22.4 1 10,000 50,000 Access Project 188   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 6.3, continued Project details Mini grid investment (US$, millions) Expected mini grid results Approval Closing World Mini Country Project name date date Bank Cofinancing Total grids Connections People Liberia Electricity Sector 12-Mar-21 30-Jun-26 4.0 1.0 5.0 39 6,500 32,500 Strengthening and Access Project (LESSAP) Malawi Electricity Access 20-Jun-19 30-Jun-24 10.0 — 10.0 10 5,200 26,000 Project Mali Rural Electrification 11-Dec-13 31-Mar-23 51.0 4.9 55.9 42 9,770 48,850 Hybrid System Project Mozambique Energy for All 28-Mar-19 31-Dec-23 10.0 — 10.0 6 4,000 20,000 (ProEnergia) Nepal Business Models for 31-Jan-19 30-Apr-23 5.6 9.6 15.2 28 27,109 126,000 Private Sector–Led Mini Grid Energy Access Project Niger Solar Electricity Access 07-Jun-17 31-Jan-24 34.1 — 34.1 41 16,500 82,500 Project Niger Accelerating Electricity 10-Dec-21 31-May-27 40.0 — 40.0 80 15,750 90,860 Access Project (Haské) Nigeria Electrification Project 27-Jun-18 31-Oct-23 150.0 250.0 400.0 850 330,000 1,650,000 Rwanda Renewable Energy Fund 20-Jun-17 30-Sep-23 5.0 20.0 25.0 10 26,000 105,000 Sierra Leone Energy Sector Utility 18-Dec-13 31-Dec-22 1.0 — 1.0 — n.a. n.a. Reform Project Additional Financing Sierra Leone Renewable Energy 28-Jan-21 31-Dec-25 6.0 — 6.0 10 5,200 26,000 Development and Energy Access Project Solomon Electricity Access and 05-Jul-18 31-May-23 11.5 — 11.5 5 1,500 7,500 Islands Renewable Energy Expansion Project (Phase II) Somalia Somali Electricity 21-Dec-18 30-Jun-22 1.0 — 1.0 — — — Access Project Somalia Somali Electricity 08-Dec-21 31-Dec-26 95.0 — 95.0 382 198,848 5,714,000 Sector Recovery Project Tajikistan Rural Electrification 11-Jul-19 31-Dec-25 23.8 — 23.8 35 2,401 6,694 Project Tanzania Rural Electrification 21-Jun-16 31-Jul-23 32.0 120.0 152.0 288 197,635 988,175 Expansion Project Vanuatu Rural Electrification 31-May-17 30-Jun-22 6.8 4.9 11.7 5 550 2,750 Project Stage II Yemen Integrated Urban 30-Apr-18 30-Dec-22 4.3 — 4.3 1 520 2,600 Services Emergency Project Zambia Electricity Services 27-Jun-17 31-Aug-23 5.9 — 5.9 5 2,200 11,000 Access Project  Total approved World Bank projects 1,460.9 1,043.4 2,504.2 3,156 1,887,846 13,753,834  Total pipeline World Bank projects (estimated) 200.0 100.0 300.0 300 150,000 750,000  Total approved + pipeline World Bank projects 1,660.9 1,143.4 2,804.2 3,456 2,037,846 14,503,834 Source: ESMAP analysis. — = 0. n.a. = not applicable. MINI GRIDS FOR HALF A BILLION PEOPLE    189 • In the Nigeria Electrification Project, the Bank is pursu- Though already significant, the World Bank’s investments in ing top-down and bottom-up approaches simultane- mini grids are designed to act as a magnet for private-sec- ously. Under the top-down approach, private-sector tor investment to support exponential growth in mini grids developers bid on the minimum capital cost subsidy to reach universal access. required to deliver electricity in areas preselected by the Rural Electrification Agency, given specific regula- tory guidance on tariffs. Geospatial planning and socio- BARRIERS FACED BY PRIVATE economic analysis helped identify sites that would be suitable for mini grids. Under the bottom-up approach, DEVELOPERS IN ACCESSING private-sector mini grid developers can select their COMMERCIAL FINANCE own sites and apply for performance-based grants on the basis of new customer connections. MINI GRIDS ARE PERCEIVED AS RISKY • In Kenya, up to $80 million financing of the Off-grid Many potential private financiers consider mini grids too Solar Access Project for Underserved Counties pro- new and hence too risky to invest in. Most local equity and gram, is allocated for private-sector developers to bid debt financiers lack experience providing funds for proj- on Engineering-Procurement-Construction contracts ects with such long payback periods (typically between 7 with the Kenya Power and Lighting Company (KPLC) and 15 years). The absence of relevant financial experience and the Rural Electrification and Renewable Energy increases the risk perception. In addition, most countries Corporation (REREC). The 137 sites are divided over have had limited operational experience with mini grids, thirteen lots in 14 low-income counties. Bidders are which are seen as a new type of investment, increasing the asked to install the solar mini grids including KPLC risk perception. certified prepaid/smart meters as well as provide for a Where operational experience has been gained, it has been social infrastructure project as selected by the county. for only a few years, making it difficult to prudently assess In addition, the Government is considering expanding the risks mini grids may face in later years. For example, it the performance grant program for solar systems and is likely that the main grid will eventually reach the areas cookstoves, to private sector-led solar mini grid proj- where the mini grid is located. What will happen then? Sev- ects as well. eral countries have designed regulatory financial schemes • The recently completed National Electrification Proj- under which the main grid takes over at least some of the ect in Myanmar used a tripartite agreement between mini grid’s assets, with compensation. Lack of experience developers, the community, and the government entity with this compensation scheme creates a significant risk responsible for implementing the World Bank project. for investors. The agreement sets out the terms that developers must adhere to as they build, own, operate, and main- DEMAND UNCERTAINTY REMAINS HIGH tain the mini grid. In addition, the agreement stipulates Developers estimate future demand for their electricity a minimum equity investment from both the developer services not only to appropriately design their mini grids and the community, as well as the capital cost subsidy but also to attract external financing. Demand for mini grid amount to be paid by the implementing agency. The electricity varies widely between mini grids in the same capital cost subsidies were technology neutral. country, and between countries, so estimating demand remains challenging for developers. As a result, concerted efforts are still needed to stimulate income-generating uses of electricity during the daytime. In parallel, as we Development partners including the World discuss in this chapter, financing mechanisms can help Bank have committed more than $2.6 billion de-risk some of the demand uncertainty. to mini grids over the next several years. The World Bank’s portfolio of mini grid investment commit- CONSUMERS ARE UNABLE TO PAY THE FULL ments includes $1.4 billion in approved funding for COSTS OF SUPPLY AND LACK CREDIT HISTORY 41 projects. This investment is expected to mobilize Many potential consumers are unable to pay the full costs an additional $1 billion in government, private-sec- of electricity from the mini grid. This implies that, in the tor, and development partner funding for mini grids absence of subsidies, the mini grid would realize losses, over the next five to seven years. which is not acceptable to private developers. Some countries allow mini grids to charge cost-reflective tariffs. These tariffs create their own problems, as poorer poten- 190   MINI GRIDS FOR HALF A BILLION PEOPLE tial consumers may choose to not connect to the mini grid can take to reduce uncertainty and create a more favorable because they are unable to pay the tariff. This would tend regulatory environment for mini grids. In parallel, however, to leave the poorer people without electricity, even though policy decisions such as national electrification plans (see electricity is available where they live. Furthermore, many chapter 2) and the roles and responsibilities of the national potential mini grid customers do not have a credit history, utility (see chapter 8) will also affect the mini grid sector. which is a key risk and potentially a deterrent for institu- The ability of mini grid developers to attract external financ- tional investors (especially debt). ing can depend significantly on the political and regulatory environment in which they intend to operate as businesses. Even households that do connect to the mini grid may end up consuming little electricity because of the high tariff. FUTURE MACROECONOMIC CONDITIONS ARE Minimal use will make it more difficult for the mini grid to UNCERTAIN recover its costs. Uncertainty about future macroeconomic conditions dis- The mini grid’s revenues may be variable over time in areas courages investment. One major macroeconomic variable where agriculture is a major source of livelihood, as the is the exchange rate. It is important because at least some fluctuations in agricultural output and incomes may lead to of the equipment used by the mini grid, such as solar pan- corresponding fluctuations in electricity use and revenues. els, will be imported. If the exchange rate depreciates sig- This variability may also be a barrier. nificantly and unexpectedly, the local costs of this imported equipment will rise, reducing the financial viability of new EQUITY IS INADEQUATE investments in mini grids. In extreme cases, the govern- Impact investors and commercial investors, as well as ment may impose foreign exchange controls, reducing the local and national banks, have developed equity, debt, and amount of foreign exchange available for importing mini blended finance options to help developers scale up their grid equipment. mini grid business. While this has driven some innovation In extreme cases, when the currency is depreciating, the in financing mechanisms for mini grids, their combined government may impose foreign exchange controls, and investment falls far short by an order of magnitude of what hike up the custom duties or charges for importing mini is needed to achieve the Sustainable Development Goal 7 grid equipment. For example, in 2015, Nigeria imposed objective by 2030. foreign exchange controls, which caused concerns about THE TENORS OF ASSETS AND LIABILITIES ARE limiting the import of solar panels. We note that this sce- MISMATCHED Mini grid investments (and other infrastructure investments Private-sector mini grid developers face seeking long-term asset financing in low-income countries seven main barriers to accessing finance: more generally) often create an unacceptable asset-lia- (1) many potential private financiers consider mini bility mismatch for local debt financiers. These financiers grids too new and hence too risky as investments; rarely have access to long-term funds, as they raise their (2) uncertainties around demand for the mini grid’s funds from term deposits of a limited number of years. If electricity services creates a risk for developers local debt financiers provide long-term loans (their assets) and financiers alike; (3) many potential consumers while their deposits (their liabilities) are for shorter periods, are unable to pay the full costs of the mini grid’s this is an unacceptable asset-liability mismatch. When this electricity; (4) unlike larger project developers, mismatch occurs, it makes it difficult for these financiers to local for-profit and community developers often do provide long-term loans to mini grid developers. not have enough of their own financial resources to POLITICAL AND REGULATORY RISKS CREATE meet the equity requirements imposed by conven- UNCERTAINTY tional lenders; (5) mini grid investments often cre- ate an unacceptable asset-liability mismatch for In a handful of countries around the world, mini grids oper- local debt financiers; (6) political and regulatory ate under well-established regulatory regimes that, while risks are prevalent in many countries with large perhaps not perfect, offer guidance to and reduce uncer- populations that do not have access to electricity; tainty for private-sector developers. However, in many and (7) uncertainty about future macroeconomic populous countries where many lack access to electricity, conditions—particularly exchange rates, interest regulatory and political risks create uncertainty for develop- rates, and the economic growth rate—discourages ers. Chapter 9 discusses in more detail the key regulatory investment. issues that need to be addressed in any regulatory frame- work for mini grids, and other steps regulatory agencies MINI GRIDS FOR HALF A BILLION PEOPLE    191 nario would also affect foreign investors’ ability to take prof- Developers should be able to demonstrate that their busi- its and repay their foreign lenders, which are longer-term ness plans have some desired financial features (guaran- issues than importing equipment. tees, collateral), including the following: Another relevant variable is the interest rate on local loans. • Adequate returns on investment, including in scenarios Any unexpected hike in interest rates would increase the that reflect some potential adverse external events. financial costs of loans to cover capital and operational costs. • Construction performance. Mini grid developers should While such an increase would not affect fixed-rate loans that be able to demonstrate that they have the skills and were already extended, it would affect future refinancing or resources to construct the mini grid on time and within the rollover of loans that have a shorter tenor than the devel- budget, including by hiring a reliable engineering, pro- oper wants. For example, a developer may want a 10-year curement, and construction contractor. loan to finance capital costs. The lender might offer only a 6-year loan with payments calculated as if it were a 10-year • Operational performance. Developers should be able loan. At the end of six years, the developer would still owe to demonstrate that they have the skills and resources part of the loan. It could be rolled over into a fresh four-year to operate the mini grid. Penalties could be imposed loan, but the old interest rate would no longer apply. The new for failure to generate and distribute electricity as pro- loan would be at the interest rate then prevailing. posed. A third relevant macroeconomic variable is the country’s • Maintenance fund. Developers should establish and economic growth rate, particularly in rural areas. Significant maintain a reserve fund for planned and unplanned economic growth would tend to increase the mini grid’s maintenance during the operational phase of the proj- revenues and improve its financial viability. A national or ect. This fund should be built into the project cost and regional recession or even slow growth leads to a “demand topped up from project revenues as needed. risk”—that is, a situation where demand would not be • Debt-service reserve fund and ratio. The financial plan adequate to allow the mini grid to turn a profit. This would should include a debt-service reserve fund that allows adversely affect the operation of mini grids and reduce the developers to continue to service the debt in the event of incentives to invest in them. a temporary shortfall in project revenues. The debt-ser- vice ratio should be satisfactory to lenders. • Financial support from the government or international OVERCOMING THESE BARRIERS agencies. Developers should be able to demonstrate that they will be supported by the government or inter- Mini grid developers and the government must take national agencies. This will give some confidence to actions to mitigate these barriers, with the support of potential equity and debt financiers. development partners and other interested groups. The extent to which these barriers can be mitigated will GOVERNMENT AND DEVELOPMENT PARTNER depend on local conditions. ACTIONS The government and development partners will need to DEVELOPER ACTIONS develop a support package containing four elements: debt Developers are at the center of the financing package. They facilitation, subsidies, equity facilitation, and risk-sharing need to understand what kind of business plan financiers and other risk-mitigation interventions. Table 6.4 summa- will view as bankable. In some cases, they may need assis- rizes how the support elements can be used to mitigate the tance from experts in developing such plans. barriers. Developers need to understand what kind of To be viable, mini grid projects and devel- business plan financiers will view as bank- opers need a complete financial package able, and they should be able to demonstrate that consisting of debt, equity, grants, and risk-sharing their business plans have some desired financial mechanisms. Government support for private-sec- features (guarantees, collateral), including ade- tor mini grids should therefore actively facilitate quate returns on investment, construction and oper- debt and equity investment and risk sharing through ational performance, maintenance and debt-service private-sector investors, alongside results-based reserve funds, and additional financial support from grants paid directly to the mini grid developers. the government or development partners. 192   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 6.4 • Overcoming barriers to investments in mini grids Risk-sharing and other Barriers Debt Subsidies Equity interventions Mini grids are Help potential lenders Use preinvestment subsidies Help local developers Develop risk-sharing viewed as too assess mini grid finances to reduce risk by facilitating prepare bankable instruments risky Help local developers adequate due diligence business plans Provide strong national prepare bankable Use capital cost—not output- commitment to mini grids, business plans based—subsidies to reduce to reduce risk perception risk exposure Demand Provide subordinated Reduce capital costs to a level — Offer first-loss guarantee uncertainty concessional loans so that brings the mini grid’s or minimum revenue that commercial lenders cost of electricity to below the guarantee mechanisms that can be repaid first customers’ ability to pay cover a period of expected Offer grace periods revenues covering the first year or two of operations to enable the mini grid to reach its expected demand Customers Develop new financial Use well-designed subsidies — Reduce costs by cannot pay full instruments that reduce Results-based subsidies may formulating large-scale costs borrowing costs be the most suitable for larger programs firms that have good access to Allow entities to operate debt and equity multiple mini grids Enter as guarantors, being liable in case of default payment by mini grid customers Developers lack — Use capital cost subsidies to Facilitate “patient Provide financial and equity resources partly finance initial capital capital,” which can technical training to costs of smaller firms take later returns from potential developers investment Develop new financial instruments that facilitate external equity Assets and Increase the term of Use capital cost—not output- — Reduce concerns about liabilities are liability (deposits) by based—subsidies, to reduce long-term loans through mismatched providing funds, as in debt requirements of smaller schemes such as “future for local debt World Bank credit lines firms, which tends to assure liquidity support” financiers local financiers Political and Develop mechanisms — — Develop risk-sharing regulatory risks that facilitate local instruments that protect currency debt developers for regulatory or political changes Macroeconomic Develop risk-sharing Provide short-term subsidies — Develop risk-sharing uncertainty instruments for specific to tide over adversely affected instruments for specific events firms, to avoid disrupting events electricity provision Provide currency-hedging instruments Develop contracts and agreements that allow developers to pay investors and suppliers in foreign currencies Source: ESMAP analysis. — = not applicable. MINI GRIDS FOR HALF A BILLION PEOPLE    193 DEBT FACILITATION • To the extent possible, interventions should not distort financial markets. In particular, the commercial risk for While it is common for commercial debt to be the largest the debt should remain with the private financiers, not part of the financing package for a typical investment proj- transferred to the government or some other official ect, most mini grid projects to date have been financed agency. Further, the interest rates should remain close mostly with equity and grants with a smaller portion of to the local market rates. Although high interest rates debt. Indeed, the International Finance Corporation has impose a financial burden on the mini grid borrowers, been tracking several financing deals with private-sector it is recommended to keep debt and subsidy separate. mini grid developers and found that debt tended to be a Instead of implicit subsidies through low-rate loans, the relatively small portion of the overall financing package. subsidy should be explicit and provided separately. One of the main reasons for this is that commercial debt— particularly from local financiers—is not readily available • The government should engage financial experts to help because of the local mini grid sector’s lack of a track record, mini grid developers develop their business plans. It or, if it is available, the terms are not financially attractive. may also be useful to assist private financiers with com- Typical debt terms for mini grid developers tend to be mercial appraisal of mini grid projects if the financiers between 9 and 12 percent if the debt is in US dollars, with a request this assistance. much broader and higher range for debt in local currency. • The government should have the option of including (We note that high interest rates are an artifact of the coun- nonprice factors in facilitating debt, as has been done in try risk and other commercial risks, where private-sector South Africa (box 6.2) for renewable energy projects. In lenders need to ensure an appropriate risk-return ratio.) particular, the social and environmental aspects of mini As a result, in most countries, the government will need grid projects could be more heavily weighted in bid eval- to induce the flow of debt to mini grids, from commercial uations, to ensure that winning firms—to whom debt investors and banks, multilateral agencies, private finance would be channeled—have these important nonprice initiatives, local pension funds, and others. The following factors built into their business plans. four principles should guide the design of instruments that assist lenders who participate in financing mini grids: Up-front credit lines One common intervention in government-supported proj- • If private financiers need their own long-term funds ects is the creation of an up-front credit line, whereby the before they can provide long-tenor loans to mini grid government makes its funds available to participating developers, the government should be willing to make financial institutions (PFIs) at the time the loan is issued to long-term funds available to them. This intervention the mini grid developer. The typical sequence of an up-front would mitigate or even eliminate the asset-liquidity mis- credit line is as follows: match. • The government gets funds from development partners, such as the World Bank. BOX 6.2 NONPRICE FACTORS IN RENEWABLE ENERGY PROJECTS Public and private procurement of goods and services In South Africa’s power sector, bidding documents can act as a lever to incorporate diversity measures under the Renewable Energy Independent Power Pro- and enhance corporate social responsibility. For exam- curement Program include nonprice bid-evaluation ple, South Africa has a preferential procurement policy factors. The bidding documents outline requirements framework and regulations for considering nonprice for job growth promotion, domestic industrialization, factors for 10–20 percent of its bid scoring. In 2011, community development, black economic empow- these regulations were amended to reflect the Broad- erment, and women-owned vendor expenditure. The Based Black Economic Empowerment (BBBEE) Act of socioeconomic requirements go beyond the custom- 2003, which ties award points to BBBEE status levels, ary nonprice criteria in the government’s preferential adding verification and remedies for fraudulent repre- procurement policy—accounting for 30 percent of total sentation. bid value. Source: ESMAP analysis. 194   MINI GRIDS FOR HALF A BILLION PEOPLE • The government selects local lenders who are inter- Market-conforming interest rates ested in financing mini grids and meet some operational While the ideal design of an up-front credit line would call financial criteria. for interest rates that are close to the prevailing interest rates in the country, in many cases, mini grid developers • The government on-lends these funds, usually through find these interest rates too high and ask for “soft loans” a financial intermediary, to participating PFIs in local at reduced interest rates. An example from Bangladesh of currency. The loans to the PFIs have a long tenor. The loans below market rate is offered in box 6.3. interest rates on the loans are similar to the prevailing local wholesale interest rates. They are adjusted period- Such interest rate reductions should be implemented with ically to reflect current market conditions. caution because of the problems they can create. In par- • The loans from PFIs to mini grid developers are denom- ticular, reduced-rate loans take the credit offered to mini inated and repaid in local currency. The interest rates grids out of the system of normal loans. Mainstreaming are close to the prevailing market rates. The main differ- is needed to provide debt finance to mini grid developers ence is that the loan tenor is longer than is locally com- at scale. The closer the mini grid loans are to conventional mon. The PFIs assume the commercial risk for the loan, debt finance, the greater the likelihood that debt finance to repaying the government even if the mini grid developer mini grids will grow. fails to repay the loan. One way to address the problem of high rates is to blend • The PFIs have to put some of their own money into the concessional debt along with commercial debt in the loan. The government provides only a certain portion same financing package. This can help reduce the amount (such as 80 percent) of the total loan amount to the mini of up-front subsidies required while keeping tariff levels grid developers. affordable, particularly in countries with a nascent mini grid market where capital costs are high. The need for and design of an up-front credit line requires careful assessment of local conditions and the interests of potential private finance initiatives. The World Bank’s Nige- ria Electrification Project does not include a credit line or similar facility. Instead, it was agreed that a detailed assess- BOX 6.3 ment of the financial conditions and the experience of mini A FINANCIAL SUPPORT PROGRAM grid developers would be carried out about 18 months after IN BANGLADESH the project became operational. This assessment will then recommend appropriate financial interventions, which In Bangladesh, mini grids are developed by pri- could include an up-front credit line, for promoting mini vate investors who may apply to the Infrastructure grids in Nigeria. Development Company Limited (IDCOL) for funds. IDCOL does due diligence on a proposal. To begin, it consults with the Rural Electrification Board to check on when the main grid is likely to serve the One common debt facilitation intervention proposed site. IDCOL also undertakes reviews of in government-supported projects is an other factors affecting project viability, including up-front credit line. This is a type of multi-tier lend- potential customers’ willingness to pay and the ing mechanism in which the government receives validity of cost estimates. funds from development partners such as the World Bank, selects qualified participating finan- IDCOL finances and supports renewable energy cial institutions (PFIs) interested in financing mini and energy access projects in Bangladesh using a grids, and on-lends the development partner funds variety of financial tools. It has provided $27 mil- to the PFIs in local currency using loans with a long lion to support 27 mini grids totaling 5 megawatts tenor and market-conforming interest rates. The (peak). PFIs then lend to mini grid developers using loans IDCOL provides a grant of 50 percent of the capital denominated and repaid in local currency, with mar- costs and loans for another 30 percent. The loan is ket-conforming interest rates and loan tenors that for 10 years, with a two-year grace period. The inter- are longer than what is locally common. The PFIs est rate is 6 percent per year, which is below market assume the commercial risk for the loan, repaying interest rates. the government even if the mini grid developer fails to repay the loan. Source: ESMAP analysis MINI GRIDS FOR HALF A BILLION PEOPLE    195 Governments considering up-front credit In markets with no shortage of liquidity when lines as a mechanism to make it easier for governments are set to deploy a debt facili- developers to secure private-sector loans should tation mechanism, up-front credit lines may not be ensure that the credit line’s interest rates are close to the most appropriate option. A better approach the prevailing interest rates in the country. Reduced- would be to use a financial instrument that induces rate loans take the credit offered to mini grids out of private lenders to make longer-term loans without the system of normal loans. However, mainstream- increasing liquidity in the short term. One way to do ing is needed to provide debt finance to mini grid this is to set up a financing facility that makes funds developers at scale. The closer the mini grid loans available to developers in the years after the com- are to conventional debt finance, the greater the mercial loan tenor expires. likelihood that debt finance to mini grids will grow. Another way to address the problem of high interest rates is Emerging debt options to increase capital cost subsidies, so that the total amount Significant technical innovation in the financial sector has of the loan is reduced. Reduced-interest rate loans are an recently led to new ways for debt funds to flow to borrow- indirect, implicit subsidy for mini grid developers; it is more ers. Funds for mini grids could come from worldwide inves- transparent and sustainable to provide this additional sub- tors, typically affecting investors interested in promoting sidy explicitly outside the loan. renewable energy or rural electrification. A third option is to introduce some risk-sharing guarantees, Some of the new financial instruments were recently intro- under which an independent third party shares the com- duced in Asia and Africa. Although they are still in the early mercial risk with private finance initiatives, as discussed stages of development, they have the potential to raise sig- later in this chapter. nificant funds for mini grids in the future. Convertible notes. While not a new debt product per se, Alternatives to up-front credit lines convertible notes are new to the mini grid sector. A con- One drawback of an up-front credit line is that it introduces vertible note is a loan that can be converted into company additional funds into the financial system in a country that equity at some later point. The loan has a lower interest rate often does not have a liquidity shortage. The fact that an than a conventional loan. The lender can convert the loan up-front credit line adds liquidity in the short term even into equity under terms favorable to the lender. though there is no shortage of liquidity indicates that it is an overly broad instrument that may distort credit markets. The main advantage of convertible notes is that they allow A better approach would be to use a financial instrument early investors to provide mini grid developers with quick, that induces private initiatives to make longer-term loans cheap capital that they can use to launch new companies without increasing liquidity in the short term. and execute their growth plans. These early investors will likely participate only if they can sell their converted equity One alternative is to set up a financing facility that could shares reasonably quickly at attractive prices. Once the make funds available in the years after the normal tenor mini grid company is well established, some people must of commercial loans runs out. Say, for example, a mini be willing to buy the shares of these early investors. If the grid developer wants a 12-year loan, but the private lender early investors are non-nationals, they must be confident may be willing to lend money from its own funds for only that they will be able to readily convert the money from the five years. The alternative facility could kick in after five sale of their shares into foreign currency. years. It directly addresses the issue of longer-term loans without adding liquidity to a financial system that may not Convertible notes are suitable for early-stage mini grid have a liquidity shortage. Another advantage of this mech- developers that have significant growth and profitability anism is that the total external funds required would likely potential. be smaller than the amount required by a conventional Peer-to-peer business lending. Peer-to-peer business up-front credit line. (P2PB) instruments are suitable for companies that have This alternative facility is more complex than an up-front a track record of revenues. In a P2PB scheme, a mini grid credit line, however. Although it is easily understood by developer borrows, without collateral, from a group of indi- potential private finance initiatives, its longer-term horizon viduals or institutional lenders, using an online P2PB lend- and complexity make it more difficult to design and imple- ing platform. The lending platform itself does not provide ment. It has not yet been used for financing mini grids. the loan; it is just an intermediary that matches borrowers and lenders. 196   MINI GRIDS FOR HALF A BILLION PEOPLE It is important to induce more developers to consider more mini grid sites—that is, to cast a wide net. One way to do so is Two debt financing innovations have to meet these costs with preinvestment subsidies. The main emerged recently for private-sector mini elements of preinvestment subsidies include the following: grids: convertible notes, in which a loan can be con- verted into equity on terms that are favorable to the • General market and resource assessments, including lender, thereby reducing the loan’s interest rate; and geospatial planning (see chapter 2). Governments peer-to-peer business lending, in which a mini grid should undertake these assessments for potential mini developer borrows, without collateral, from a group grids and provide the information free to all interested of individuals or institutional lenders, using an online developers. lending platform. • Prefeasibility and feasibility studies. The costs for these studies should be shared with shortlisted firms for proj- ects that pass basic screening criteria set by the govern- One major advantage of the P2PB system is that it is ment agency responsible for promoting mini grids. quick, as limited due diligence is conducted. However, the • Technical and financial information assistance. This cost of the loan is likely to be high. If the lenders are for- assistance should be provided free to local developers, eigners, they must be confident that they will be able to at least in the early years, so that they can develop bank- readily convert the money from repayments into foreign able business plans. currency. CAPITAL COST SUBSIDIES SUBSIDIES Capital cost subsidies should be designed so that the finan- Subsidies should be part of the financial package avail- cial package—consisting of equity, debt, and subsidies—is able to mini grid developers and customers. The subsidy sufficient to finance all capital costs. If all capital costs can- amounts and design should respond to the needs of the not be covered, the mini grid will not be constructed. following three groups: Capital cost subsidies can be preset as a share of “reason- • Households and firms, which want an affordable com- able” capital costs in contexts where the tariff is not pre- bination of initial (connection and internal wiring) costs defined, or is based on lowest-subsidy bids in cases where and usage charges the tariff is predefined. The preset option tends to be easier • Mini grid developers, who are looking for financial via- for both administrators and mini grid developers, but may bility lead to higher-than-necessary subsidy levels. • Subsidy fund providers, who want subsidies to be prop- In recent years, many governments have provided out- erly targeted, easy to administer, and not be a financial put- or results-based capital cost subsidies. These should burden be applied to mini grids with caution, as relying exclusively on final output makes it difficult for developers to finance Subsidies for mini grids come in two broad categories. The the initial, up-front capital costs of mini grids, particularly if first category consists of subsidies that help the develop- affordable debt is not readily available to them. As a conse- ers and customers finance the costs of project prepara- quence, it is reasonable to designate some “intermediate” tion, construction, and operations. The second category results as a basis for some of the subsidy payments, such reduces the financial costs of mini grid developers. as when goods are purchased and when they arrive on site. First-category subsidies: Helping developers and Performance-based subsidies of capital expenditure customers finance their costs ESMAP used HOMER® Pro (Hybrid Optimization of Mul- PREINVESTMENT SUBSIDIES tiple Energy Resources) to model an optimized “best- in-class” mini grid based on component costs and load Mini grid developers face certain expenses before they magnitude averaged from three high-performing mini grids decide to undertake their investments, such as market in Nigeria, Myanmar, and Ethiopia. The software was also assessments and prefeasibility and feasibility studies. used to model a “best-in-class 2030” mini grid reflecting Developers may be hesitant to incur these costs because equipment cost reductions expected in 2030 (see chapter they are not sure that their firm will actually build a mini 1 for details). The modeling software was used to assess grid at the site being considered. As a result, the number of the impacts of reducing initial capital expenditure (CAPEX) mini grids may be smaller than needed for scale-up. costs by 40 and 60 percent under different productive-use and component cost scenarios. The results of this analysis are presented in table 6.5. MINI GRIDS FOR HALF A BILLION PEOPLE    197 TABLE 6.5 • Impact of performance-based subsidies of capital expenses on the levelized cost of energy of a well-designed mini grid While the combined impact on LCOE of grants and productive uses is typically greater than Performance-based either approach on its own, when used together, their Load grants as share of factor capital expenditure 2021 2030 cumulative impact is lower than the sum of expected (%) (%) (US$/kWh) (US$/kWh) impacts from each intervention when OPEX costs are 22 0 0.38 0.30 relatively large with respect to CAPEX. 22 40 0.28 0.20 22 60 0.23 0.15 40 0 0.28 0.21 costs that must be covered by electricity sales. A second 40 40 0.22 0.16 reason has to do with diesel generators. While higher asset 40 60 0.19 0.12 utilization lowers the portion of LCOE associated with 80 0 0.23 0.17 investment costs of equipment, it can raise the portion of 80 40 0.20 0.14 LCOE associated with operational expenditure (OPEX), 80 60 0.19 0.13 especially if that OPEX involves burning expensive diesel Source: ESMAP analysis. fuel. Because OPEX accounts for a larger portion of LCOE Note: Data on the levelized cost of electricity are for a well-designed so- than CAPEX in a subsidized system, the higher OPEX costs lar-hybrid mini grid with 231 kilowatts of firm power output serving around can outweigh the increased capital utilization benefits that 800 customers. would otherwise lower LCOE. kWh = kilowatt-hour. CONNECTION COST SUBSIDIES Table 6.5 shows how a 40 percent capital cost grant low- Poorer households will find it difficult to pay the up-front ers the LCOE from $0.38/kWh to $0.28/kWh in a scenario connection and internal wiring costs needed to connect to with low productive use. In a scenario where productive use a mini grid. A special fund can subsidize connection costs raises the mini grid’s load factor to 40 percent, the same 40 for these households. percent capital cost grant reduces the LCOE from $0.28/ kWh to $0.22/kWh. The subsidy calculation should use a broad definition of connection cost, which includes all the costs to connect, Interactions between performance-based capital cost including internal house wiring. It is possible that these subsidies and productive use costs will vary according to the level/tier of service, as CAPEX subsidies and increased load factors through defined by the Multi-Tier Framework. productive use both act to lower LCOE. But when used together, their interaction diminishes their cumulative Low-interest loans may be useful to help households finance impact. In an unsubsidized mini grid, the impact of shifting their internal wiring and connection costs. Box 6.4 presents from a 22 percent load factor to a load factor of 40 percent an example from the Lao People’s Democratic Republic greatly lowers LCOE. In the case of the best-in-class mini (Lao PDR) of an interest-free loan program, designed with grid modeled above, a shift from 22 to 40 percent load fac- a gender focus, available to low-income households to help tor lowers LCOE by $0.10/kWh from $0.38/kWh to $0.28/ them cover the costs of connecting to the mini grid. kWh when the CAPEX subsidy is zero (see table 6.5). With Where households are responsible for part of the total a higher capital subsidy, however, the LCOE-lowering ben- connection costs, should the subsidy be paid to the house- efits of a higher load factor are diluted. In the case of a 60 hold or the developer? This depends on the specific cir- percent subsidy, the LCOE of the same mini grid drops only cumstances of each case. It is typically most convenient $0.04/kWh, from $0.23/kWh to $0.19/kWh, when the to make a single subsidy payment to the developer, which load factor increases from 22 to 40 percent. would then be responsible for paying any contractor who The erosion of LCOE benefits in combinations of subsi- does the internal household wiring. dies and productive use arises because the mini grid must While it is generally desirable that subsidies decline over deliver more electricity to meet load. One consequence of time, it is unlikely that the household’s share of the costs can more electricity consumed means that the batteries must be increased. One possible exception occurs in locations cycle more electricity, and therefore require more frequent where the income levels of poor households have risen. replacement. CAPEX subsidy only covers equipment costs when the mini grid is built but does not cover equipment It may also be useful to pay some subsidies after connec- replacements. More frequent equipment replacements tion is made, to increase the number of households with through heavy use increases the overall (unsubsidized) access to electricity. 198   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 6.4 Customer connection schemes need to pay attention to equity issues related to gender. A REVOLVING FUND FOR CONSUMER If connection charges are too expensive for house- FINANCE IN LAO PDR holds headed by women, or if connecting to the mini grid requires a land title or proof of collateral, the The “Power to the Poor” (P2P) program of the Lao benefits of mini grid electricity will not reach male PDR government is a targeted, subsidized, afford- and female customers equally. able, and sustainable financing mechanism for connection and indoor wiring for the poorest rural households. It is designed with a gender focus to provide interest-free credit that allows the poorest rural households, which cannot otherwise afford Four types of subsidies can help mini grid to pay the entire up-front costs of connection and developers and customers finance their internal wiring, to access the main electricity grid costs: (1) preinvestment subsidies, such as market for basic service. The monthly payments for both and resource assessments, including geospatial the credit ($75 per household) and electricity planning (see chapter 2), prefeasibility and feasi- consumption are designed to be about the same bility studies, and technical assistance; (2) capital as the cost for lighting via candles, diesel lamps, cost subsidies, typically either preset as a share of or car batteries, as used before electrification. “reasonable” capital costs or based on lowest-sub- The pilot results indicate an increase (from 63 to sidy bids; (3) connection cost subsidies, paid either 90 percent) in the connection rate of households to the developer or to customers, either as grants headed by women, which was attributed to the or as concessionary loans; and (4) usage subsidies, P2P program’s design. including those built into the tariff structure, such as lifeline tariffs, and those paid to customers to subsi- dize the purchase of energy-efficient appliances and Finally, in designing consumer connection schemes, it is electromechanical equipment. important to pay attention to equity issues related to gen- der. For example, are the connection charges too expensive for households headed by women? Do connection require- watt-hours used per month. In Uganda, for example, the ments such as land titles prohibit women from gaining a charge for the first 15 kWh is U Sh 250/kWh (about $0.07/ connection in their name? If the answer to either of these kWh); the rate thereafter is nearly three times higher. questions is yes, then the benefits of mini grid electricity will be unequally distributed across the mini grid’s potential Providing usage subsidies to customers may reduce their customer base. incentives to limit their energy consumption, leading to the purchase of energy-inefficient appliances. To prevent USAGE SUBSIDIES this from happening, governments can consider providing In some situations, usage subsidies may be built into the subsidies for the most efficient appliances and end-use tariff, particularly if the mini grids adopt the national tariff equipment. that is applied to the main grid. Although mini grid custom- ers may initially agree to pay higher tariffs, they may later Second-category subsidies: Reducing the financial change their minds—possibly as a result of statements by costs of mini grid developers local or national-level politicians about fair electricity pric- ABSORPTION OF SOME COSTS BY THE GOVERNMENT ing—and ask to pay a lower tariff. Such a change would Governments can reduce some of the capital costs of threaten the financial viability of the private investment in developers by owning part of the mini grid. For example, the the mini grid and deter future private investments in mini government could pay for the distribution system and allow grids. To prevent this from happening, the government the mini grid developer to use it for a fee. The fee could be must be prepared to increase the subsidies provided to the subsidized (such as a zero fee in the initial years). private developers. Alternatively, the mini grid developer could build the distri- Regardless of which tariff scheme is applied, it is important bution system and then sell it to the government once it is to ensure that lifeline rates are provided for poorer house- ready for use.1 Under such an arrangement, the mini grid holds. Lifeline rates charge a low fee for the first few kilo- developer would not need long-term capital to finance the MINI GRIDS FOR HALF A BILLION PEOPLE    199 costs of the distribution system. After the sale, the mini grid developer would lease back the system from the govern- Two types of subsidies can lower the finan- ment. Another advantage of this system is that the arrival cial costs mini grid developers face: (1) the of the main grid would create less of a financial transaction government can absorb some of the costs, typically for the mini grid developer. The developer’s lease would end by owning some of the mini grid’s physical assets with the arrival of the main grid. At that point, the primary or by providing developers with tax breaks, import physical assets the developer would be concerned about tax waivers, and tax holidays; and (2) wage, employ- would be the generation system, which is much easier to ment, and training subsidies can increase employ- sell. The developer would, however, still need to repay its ment in rural areas and boost productive uses of creditors if any debts were outstanding. mini grid electricity. For example, in Nigeria, GIZ supports mini grids through the split-asset model.2 The split is between the generation and distribution assets. The entrepreneur finances and REDUCING SUBSIDIES OVER TIME owns the generation facility. The GIZ grant is used to build the distribution grid, which is owned by the state govern- It is important that a subsidy scheme has an exit, or taper, ment. This government leases the distribution grid to the policy. The need for subsidies should diminish over time as entrepreneur, with the lease payment possibly subsidized. a result of the following factors: The United Nations Office of Project Services is planning a • Experience should allow financiers to assess risks more similar approach in Sierra Leone. accurately, reducing the risk perception. Thus, the avail- The disadvantage of this method is that the government ability of debt and equity finance is likely to grow. needs to finance the costs of distribution systems. Part of • The need for preinvestment subsidies should fall as mini the reason for asking private firms to develop mini grids grid developers become active in particular markets. is because governments do not have adequate funds to • Costs are likely to fall as firms gain experience and the finance a rapid scaling up of mini grids. But if the govern- scale of the industry increases. ment can raise these funds from development partners, or through a development impact bond scheme,3 it may be However, as mini grids move into poorer, more remote feasible to move forward. areas, affordability may decline, making it more difficult to taper subsidies. A second way governments could absorb some of the finan- cial costs is through such schemes as tax breaks, import- FACILITATING EQUITY tax waivers, and tax holidays. Although these schemes are Mini grid developers invest some of their own money as financially helpful, they suffer from two major disadvan- equity. Project developers also look for equity from other tages where the aim is to scale up mini grids rapidly. First, sources. these schemes are subject to change and amendment later on because of competing priorities in a country’s bud- Over the past few years, infrastructure assets have become get. And second, these implicit subsidies make it difficult a more attractive asset class for international investors, to know how much support is being provided to mini grid driving competition and hence shrinking investor returns developers. on equity and debt margins. However, compared with sev- eral other infrastructure assets, mini grids are relatively WAGE, EMPLOYMENT, AND TRAINING SUBSIDIES small deals with high levels of risk. Such characteristics Modern mini grids support firms and entrepreneurs lead to higher expected returns on equity (RoE), typically engaged in productive uses of electricity, which in turn in the range of 18–22 percent for commercial investors and creates jobs. At the same time, governments are increas- around 15 percent for impact investors. But the RoE also ingly looking to create rural employment opportunities, depends on a number of other factors, such as the overall particularly when investment and new technologies are risk profile of specific mini grid transactions, contractual being introduced on a large scale. Subsidy schemes that and regulatory frameworks, and potential risk mitigants. incentivize or support new employment created by mini Reducing the RoE—and attracting a higher volume of debt— grids are therefore of interest. The design of such sub- on mini grid projects is achievable. Most mini grid projects sidies would have to draw on broad-based schemes for are relatively small, so they do not necessarily benefit from rural development, as no such subsidies have been pro- economies of scale; competition among investors is rather vided yet for mini grids. limited for a variety of reasons; and high up-front CAPEX subsidies are needed to make project economics work. 200   MINI GRIDS FOR HALF A BILLION PEOPLE Subsidy schemes should have an exit, or Equity usually takes the form of common taper, policy as a country’s experience with stock (issued to founders, family, friends, mini grids increases, more private sector develop- and early employees) and preferred stock (issued to ers become active in a country, and mini grid com- financiers outside the developer’s internal group of ponent costs fall. Also, subsidy policy should be supporters). Holders of preferred stock have prior- location specific, allowing mini grid developers to ity over holders of common stock in getting returns move into low-income and remote areas. on investment, so negotiations between common stockholders and potential investors can often be intense; if the negotiations break down, the devel- oper is unable to raise adequate equity. As a result, Developing mini grids through a portfolio approach would governments should consider developing standard improve the bankability and risk-allocation framework, rep- financial models to facilitate these negotiations, licability, and pace of their implementation through some particularly when the developer is local and the level of standardization. This could also help expand the potential investors are international. market potential and increase the size of individual trans- actions. In turn, larger and more bankable projects combined with The connection between the developer and investors is the appropriate risk-mitigation framework, regulatory through online platforms that have been established to regime, public-sector intervention, and subsidies will help channel equity into a variety of investments in developing attract more private investors and lenders, in turn contrib- countries, not just mini grids. uting to lower costs of capital and debt margins. Mini grids appear to have little experience with this model. STRUCTURING EQUITY INVESTMENTS But given its major potential in the scaling up of mini grid development programs, it would be useful for development Common stock is the equity issued to founders, family, partners and their financial experts to help interested gov- friends, and early employees. It is usually issued to people ernments make it easier for equity crowdfunders to finance who invest their time, effort, and money in the earliest days mini grids. of a start-up. For financiers who are outside the developer’s group of supporters, preferred stock is a suitable option. Mezzanine finance. Mezzanine finance is a hybrid of equity The most important feature of preferred stock is that its and debt. It counts as equity when debt lenders consider holders come ahead of common stock investors in getting the debt-equity ratio for financing the capital costs of a returns on investment. mini grid, because payments to mezzanine financiers are made after project operating costs, conventional (senior) The terms of preferred stock may be subject to intense debt, and required reserve balances are determined. There negotiation between common stockholders and potential appears to be no experience with this type of financing for investors who want to become preferred stock investors. It mini grids in developing countries. can be helpful if there is a government-approved standard financial model that would be the starting point of discus- Socially oriented capital. These funds come from investors sions between investors and mini grid developers. Simpli- who are interested in supporting mini grids in a commercial fied models are used in Tanzania and Haiti, among other manner but who also recognize the social rationale for mini countries, and a detailed standardized financial model grids. One way for this social recognition is to accept a lower is automatically produced by Odyssey’s online platform, than commercial return. In other words, the investor expects based on developers’ inputs (see chapter 2 for a discussion to get a financial return on the investment but will accept a on Odyssey). Such models would be particularly helpful lower financial return in recognition of the project’s social when the developer is local and the potential investors are returns. This is called “social impact investment.” Other foreigners. investors may want a commercial return, but are willing to wait for a longer period. This is called “patient capital.” INNOVATIONS IN EQUITY FINANCING FOR THE MINI GRID SECTOR The “AssetCo” Model. Under this model, an investor (the Asset Company, or AssetCo) agrees to purchase a portfolio Various innovations in equity financing can help bring of mini grids from a developer once certain milestones are affordable equity investment to the mini grid sector. achieved, such as obtaining all necessary licenses, success- Equity crowdfunding. Equity crowdfunding is a way of ful commissioning, and serving a minimum number of cus- allowing a large number of people to invest in a mini grid. tomers for a prespecified amount of time. The mini grids are MINI GRIDS FOR HALF A BILLION PEOPLE    201 • Site and portfolio selection: The rural customer base has low and episodic income (for example, seasonal income Three innovations in equity investment for for predominantly agricultural economies), leading to mini grids are emerging: (1) equity crowd- affordability issues and potential renegotiation requests. funding, in which developers offer small amounts In this context, site selection is key to improve the level of equity to a large number of investors through of strategic cross-subsidization and hence overall credit an online platform; (2) mezzanine finance, such as and project economics. Measures for achieving such convertible notes, in which a lender has the right to objectives include (1) mixing commercial and industrial convert its loan to an equity stake in the company; customers with residential loads, (2) blending nearby and (3) socially oriented equity investment, where greenfield projects with robust brownfield projects, and investors take an equity stake in a company but (3) aggregating new projects together. accept a lower financial return in exchange for a demonstrable social impact. • Rigorous demand assessment: Future electricity up- take is uncertain, given the myriad factors that could affect it (such as level of income, access to appliances for mini grid users, and access to microfinance) and the typically built and operated by the same developer, though inclusion of biases in initial assessments. While elec- in some cases the AssetCo will contract with two different tricity uptake would remain a source of uncertainty, companies – one for construction and initial operations, and it can be mitigated through (1) rigorous methodology one for long-term O&M. The AssetCo, meanwhile, owns the regarding the assessments, (2) a level of standardiza- portfolio of projects. A primary advantage of this model is tion to allow comparability between projects, and (3) that it enables the AssetCo to tap into low-cost, longer-term a postimplementation review to better understand the debt that is not typically available to a mini grid company. factors responsible for the uptake and bias. It can also be further enhanced by creating a larger pool of data to MITIGATING RISK look back on. Given the perceived riskiness of the mini grid business as an • Generation sizing and modular expansion: Mini grids investment opportunity, several mechanisms are designed have traditionally suffered from oversizing generation to mitigate some of the demand and investment risks that components as a result of various factors, including international investors might face. misleading demand estimates or misaligned incentives and grant mechanisms. This oversizing has led to the While the existence of latent demand for electricity in economic and financial underperformance of proj- unserved and underserved communities is well accepted, ects. Sizing a mini grid generation component for the the expected demand and associated ability to pay is demand expected in the first years of operation with rather uncertain. In addition, the steady state growth in built-in modularity allows mini grid developers to meet electricity demand and corresponding payments is likely demand growth over time, while curtailing early finan- to be highly dependent on the pace at which value-added cial losses from underutilization after commissioning. products and services, which in turn generate additional need for electricity, become available. This uncertainty is • Load factor optimization strategy through stimulation likely to have an adverse effect on the cost of capital and of productive uses: The appropriate development of capital structure, further increasing the need for subsidies productive uses will augment revenues by optimizing for commercial viability and to limit the kinds of parties will- the mini grid’s load management. Providing incentives ing to provide equity and debt financing. On the flip side, in to scale up access to appliances and microfinance to the rare cases when demand rises dramatically, exceeding mini grid users is key to boosting demand, and these initial installed capacity, the developer may ask for (scarce) value-added services can even be added up front in the additional capital for capacity expansion in order to meet mini grid’s value proposition to the community. the requisite service levels. The need to develop adequate demand risk–mitigation instruments is thus acute. Mitigating risk through risk allocation Just as risk allocation mitigates demand risk through mini Mitigating risk through mini grid design grid planning, design, and operations, there are several As highlighted in chapters 1–4, there are ways to design options that lessen the risk of demand uncertainty and and plan mini grids and engage with customers that can make the mini grid business a more attractive investment. help mitigate against the risk of uncertain demand for elec- In risk-sharing schemes, a third party shares some of the tricity. These include the following: risks with debt providers and equity investors. The third party could be a development partner or an interested 202   MINI GRIDS FOR HALF A BILLION PEOPLE independent foundation or fund. It is common to charge a shadow toll rate was not flat but, rather, varied with the vol- fee for the risk-mitigation mechanism, though it could be at ume of traffic relative to the user category. Bidders for the a subsidized rate, at least initially. UK projects were asked to bid up to four separate toll rates for four bands of traffic volumes. The main restriction was These schemes ease the financial risks private finance ini- that the toll rate for the uppermost band of traffic volumes tiatives and equity investors face in their fear of high losses. was to be set at zero, effectively capping the government’s The losses may be from a single mini grid developer or from financial exposure to paying tolls to the concessionaire, a pool of mini grids, with pooling generally considered a even if traffic volumes were well above expectations. The more workable option. Some risk-sharing mechanisms are lowest band was set at a level that would cover operating discussed below. costs and debt service. Bands 2 and 3 would offer a return Minimum revenue guarantee (MRG). An MRG issued to equity, and Band 4 set the maximum toll that the gov- in favor of the mini grid developer can partly reduce the ernment would have to pay, thus capping the government’s demand risk and would benefit both the lenders (principal liability to pay shadow tolls. and interest) and equity indirectly. The instrument would Loss-sharing mechanisms. These are generally of three be sized to cover a percentage of annual projected revenue, types. and if the demand does not materialize, then the guarantee would be called. In this case, the demand risk is distributed • In a first-loss guarantee scheme, the third party agrees between the developer and the government or a third party, to bear the first tranche of loss. If losses extemd beyond with payouts ultimately benefiting the entire financial equi- the first tranche, the private finance initiative has to bear librium of the project. Replenishment and renewal modali- them. The definition of where the first tranche ends var- ties would need to be in place to ensure the sustainability of ies, depending on local financial conditions. First-loss such a mechanism. schemes are a useful approach from an investment portfolio perspective—for them to work, the investor Chilean highway concessions (between 1992 and 2004) must have sufficient volume or deal flow to spread risk were among the pioneers integrating MRG mechanisms across a large investment base. to provide short-term liquidity essential to cover debt ser- vice and other financial obligations. For these projects, the • In a pari passu guarantee scheme, the guarantor and Chilean government was offering an MRG as high as 80–85 the private finance initiative share losses proportion- percent of expected revenues, with the government paying ally. The private finance initiative shares the loss right the MRG if traffic fell under a given year’s guarantee level. from the start. This scheme may therefore be less Developers would pay a guarantee fee of 0.75 percent of attractive to private financiers than is the first-loss the MRG amount, also contributing to the scheme. MRGs scheme. were also weighted toward the early years of concessions • In a last-loss guarantee scheme, the private finance ini- to reflect lenders’ liquidity concerns. tiative bears the first tranche of losses, with the guaran- Shadow tolls. Similar to MRGs, these mechanisms were tor absorbing any further losses. This scheme may be developed for toll roads and entail the government paying the least attractive to private financiers. the concessionaire service payments calculated based on Benefit sharing. In some cases, demand may increase way the number of users within a set time frame. In other words, beyond base case projections. Though rare, this case may the concessionaire collects tolls from the government lead to the developer realizing excess returns. This would rather than actual infrastructure users. In principle, shadow offer an opportunity to implement a benefit-sharing mech- toll structures enable the government and concessionaire to anism if the mini grid was developed as part of a public-pri- share the demand risk without affecting end users’ service vate partnerships model or was supported by government fees and without imposing real tolls on road users. Shadow grants. Depending on the tariff regulation, such mecha- toll advantages include: (1) minimizing demand risk and nisms can transfer benefits directly to end users through hence making it easier for private investment partners to tariff reductions or to the government, possibly contribut- find more advantageous financing; (2) if structured prop- ing to a demand risk–mitigation instrument for additional erly, reducing the effect of lower-than-expected demand; projects. (3) capping the public sector’s exposure, thereby eliminat- ing the risk of superprofitability by the concessionaire; and Subordinated loans. Risk-mitigation instruments can also (4) capturing the profit-seeking motives of the private sec- be designed specifically for debt providers. Subordinated tor, often resulting in capital construction cost savings. concessional loans can be added to the debt structure to ensure that commercial lenders are repaid in priority if the This mechanism was used in early road projects in the demand, revenues, and hence debt-coverage ratios are United Kingdom, and it had a distinguishing feature: the lower than anticipated. In such an event, the developer will MINI GRIDS FOR HALF A BILLION PEOPLE    203 be able to defer or even write off its debt repayment to the mechanism, in which the borrower takes the risk of deval- subordinated debt portion. uation up to a certain point, with the lender absorbing the loss beyond that. Currency-hedging instruments can be Mitigating foreign exchange risk quite costly, though, and help drive up overall investment The future value of the local currency in terms of foreign costs for the mini grid. Finally, any contracts or license currency is uncertain, and presents a substantial risk to mini agreements that the developer enters into with the gov- grid developers. During the planning phase, a sharp decline ernment should explicitly state that the developer can pay in the value of the local currency would disrupt the financial its investors and suppliers in any currency and make these plans and projections of finance providers as well as mini payments internationally. grid developers. During the construction and operations phases, currency risks stemming from volatile exchange rates and devaluation of the local currency are also pres- CONCLUSIONS, AND GENERAL ent. The up-front investment typically requires hard cur- rency to purchase the equipment (solar panels, cables, and RECOMMENDATIONS FOR so forth), though revenues are earned in local currency. In GOVERNMENTS addition, financing is often in hard currency, rather than local currency. In some countries, developers can index First, it is essential that the government, together with their tariffs to hard currency if they show evidence that its development partners, takes the steps necessary to they have creditors and investors to repay in hard currency. make sure that potential mini grid developers are able to This can minimize the exchange rate risk, but increases the access a financial package that makes it possible for them risk that the tariffs will no longer be affordable for custom- to finance their projects, particularly the capital costs, and ers. Currency risks are significant in some countries, where earn an acceptable rate of return on their investments. This local currency has been devalued by 25 percent or more financial package will comprise equity, debt, subsidies, and over a one-year period. This ultimately means that there is risk-sharing mechanisms. 25 percent less revenue in terms of hard currency if devel- Second, the elements of the financial package should opers are not able to increase their tariffs accordingly. be in sync with the stage of development of the mini grid One instrument to mitigate currency risk is a foreign program. In the kickstart phase, the focus should be on exchange hedge, which allows companies or lenders to mit- formulating a simple package that is easy to administer. igate foreign exchange risk by locking in an exchange rate Although ideally it should include equity, debt, subsidies, for a transaction that will occur in the future. This lock-in and risk-sharing instruments, the choice should be based requires the payment of a premium by the party looking on the level of financial development of potential financiers to mitigate the risk. Another instrument is a quasi-hedge and developers. The focus should be on making it easy for firms to get going, rather than using the most efficient financial instruments. For example, capital cost subsidies paid in tranches linked Demand and foreign exchange are the two to construction results may be more practical than subsi- categories of risk that can be mitigated to dies linked to results after the mini grid is commissioned make mini grid investments more attractive to debt (for example, actual connections). This is because sub- and equity investors. To mitigate the demand risk— sidies paid later increase the up-front debt and equity in addition to measures taken during the planning, needs, as capital cost subsidies are a substitute for debt design, and operations of the mini grid—financial and equity. It may take significant time to develop and mechanisms include minimum revenue guarantees, implement some of the newer financial interventions that shadow tolls, loss sharing, and subordinated debt. have not been used for mini grids. Relying heavily on them Foreign exchange risks can be mitigated by curren- increases the risk of delays in getting the program going. cy-hedging instruments and explicit language in The planning and development of the newer financial inter- contracts and license agreements that the devel- ventions should start as soon as possible, however, as these oper can pay its investors and suppliers in foreign interventions will be critical in the scale-up phase. currency and can make these payments interna- tionally. While there have been few significant expe- Third, financial packages should consider support to poor- riences with these interventions in the mini grid er households, so that they can connect to the mini grid sector, they would be helpful in a large-scale mini and use electricity in their homes. Financial assistance to grid development plan. pay for the initial costs can be a combination of subsidies and loans. 204   MINI GRIDS FOR HALF A BILLION PEOPLE REFERENCES NOTES Demirguc-Kunt, Asli, L. Klapper, and D. Singer. 2013. “Financial Inclusion 1. This approach is being adopted as part of the World Bank–sup- and Legal Discrimination against Women: Evidence from Developing ported Off-grid Solar Access Project for Underserved Counties in Countries.”Policy ResearchWorking Paper 6416,World Bank,Washing- Kenya. This project will construct mini grids in 14 counties with low ton, DC. https:/ /openknowledge.worldbank.org/handle/10986/ electrification rates. The private sector will bid on Engineering-Pro- 15553. curement-Construction contracts with the Kenya Power and Light- IFC (International Finance Corporation). 2017. “MSME Finance Gap: ing Company (KPLC) and the Rural Electrification and Renewable Assessment of the Shortfalls and Opportunities in Financing Energy Corporation (REREC). Bidders are asked to install the solar Micro, Small and Medium Enterprises in Emerging Markets.” IFC, mini grids including KPLC certified prepaid/smart meters as well as Washington, DC. https:/ /openknowledge.worldbank.org/han- provide for a social infrastructure project as selected by the county. dle/10986/28881. 2. This model is related to GIZ’s procurement rules, which do not per- World Bank. 2020. Women, Business, and the Law. Washington, DC: mit performance grants. World Bank. https://wbl.worldbank.org/. 3. A development impact bond is a performance-based investment instrument intended to finance development outcomes in develop- ing countries. MINI GRIDS FOR HALF A BILLION PEOPLE    205 CHAPTER 7 ATTRACTING EXCEPTIONAL TALENT AND SCALING UP SKILLS DEVELOPMENT CHAPTER OVERVIEW This chapter highlights the training and skills needed to connect half a billion people to mini grids by 2030. It surveys the types of training and skill-building initiatives needed in the mini grid sector, identifying skills gaps that can hamper successful scaling. The chapter concludes with suggestions for implementing effective training and skills-building programs and offers selected examples of capacity-enhancing efforts undertaken in several countries. A database of training and skills-development programs relevant to mini grid industry stakeholders is available on the companion website to this handbook: www.esmap.org/mini_grids_for_half_a_billion_people. Scaling up mini grid deployment to connect 490 million tive training and skill-building programs, with examples of people by 2030 will be possible only if human capital keeps capacity-enhancing efforts undertaken in several countries. pace with financial capital. However, recent mini grid proj- A database of training and skill-development programs rel- ects have revealed a gap in skills that impedes the scale-up evant to industry stakeholders, as well as organizations of mini grids (Energy 4 Impact and INENSUS 2018). Fur- that implement training programs, is available on the com- thermore, most of the second-generation mini grids built panion website of this handbook: www.esmap.org/mini_ to date have been done on a project-by-project basis, which grids_for_half_a_billion_people. does not incentivize sustainable skills development in the The following definitions are important for the purposes of country of implementation. After a project is completed, this chapter: any skills gained tend to dissipate, and training often must start over again with the next project (IEA 2003). • Skills building and capacity building are often used inter- changeably, and in this chapter they both refer to devel- Strategic, sustainable, and needs-based skill-building and oping the knowledge, know-how, and capabilities of an training initiatives are critical to ensure the long-term scal- organization or an individual. ability and sustainability of mini grids. To develop mini grids using a portfolio approach training and skills should be built • Training is defined as the process that targets special- in from the outset. ized groups to develop specific skills and capacities for certain activities or tasks (Energy 4 Impact and INEN- In this chapter, we illustrate how skills gaps have been SUS 2018), taking into account the interests of all stake- closed, enabling successful electrification through mini holders, including the local community, government and grids. We also illustrate some key limitations and inadequa- utility officials, project developers, and operators. cies in how the interventions were designed and carried out, as learning from experience can improve interven- This chapter is intended for policy makers, regulators, tions elsewhere and avoid damaging the reputation of development partners, and project developers who seek mini grid development programs (GIZ 2017b). The chapter to understand how their role in electrification can be concludes with recommendations on how to set up effec- enhanced by building skills and capacity. 206   MINI GRIDS FOR HALF A BILLION PEOPLE MINI GRID DEVELOPMENT REQUIRES Generation and distribution systems make up the largest portion of construction. Installing those systems requires DISTINCTIVE SKILLS AND CAPACITY site visits by the design team to tailor system design to Scaling up mini grid deployments successfully requires that local conditions (for example, ground conditions, wind actors along the industry value chain possess certain skills strength, and lighting incidence). Customer connections, attuned to the unique characteristics of mini grids—tech- indoor electrical installations, and metering also need to be nical, legal, financial, and political, and others (Ochs and considered by designers of the mini grid system, particu- Indriunaite 2019). As mini grid networks grow in scale, the larly when the system must comply with local codes and need for specific skills will only grow. Consider the following standards. Specialized installation tools and manuals in the skills required during three broad phases of mini grid devel- local language are also needed. opment—preinvestment, construction, and operations and Finally, community engagement is a crucial part of the con- maintenance (O&M). struction phase so as to build local ownership and avoid PREINVESTMENT PHASE future conflicts. Developers and project teams must either possess the skills to engage with the community or hire The preinvestment phase requires skills to assess the fea- trained local organizations to deliver awareness programs sibility of the project, decide on a corporate structure, and on tariffs and system use. hire local staff. Feasibility studies are critical to ensure the technical and financial viability of a project in a specific PROJECT OPERATIONS AND MAINTENANCE market (GMG n.d.[d]). Such studies cover site selection, O&M is a critical phase for the sustainability and reliabil- demand assessment, technical system design and system ity of a portfolio of mini grid systems. An operator should sizing, distribution network mapping, business models, incorporate operational spending and maintenance plans financial modeling, capital raising, community engage- as part of overall project development and be able to carry ment, and legal compliance (AfDB 2016). out O&M tasks (e-MPF 2014) across the portfolio. Ideally, Distinctive skills for the preinvestment stage are also such tasks should be set well in advance of the start of required at the policy level. To achieve electrification goals operations: even the best-designed mini grids can fail if through mini grids, civil servants may be involved in devel- O&M is not properly prepared and executed (GMG n.d.[b]). oping national electrification plans and in working with The O&M phase requires many skills, such as O&M pro- project developers. However, civil servants responsible for cess management and software; marketing and customer giving such support may themselves suffer from a lack of service; metering; demand-side management; demand skills and experience in the mini grid market, preventing stimulation (including micro enterprise development); them from achieving their intended purpose (IRENA 2017). performance monitoring and evaluation; and enterprise This challenge is pertinent at all staff levels, including policy management (Energy 4 Impact and INENSUS 2018). O&M making, standard setting, company management, finance can be managed by the utility or contracted out to local and accountancy, project management, and engineering (GMG n.d.[d]). PROJECT CONSTRUCTION Each phase of mini grid development Project construction involves contracting, procurement, requires specific skills. The preinvestment installation, commissioning, community engagement, and phase requires skills related to site selection (includ- project management. Foundational principles of the proj- ing through geospatial analysis), demand assess- ect construction phase are to ensure safety and efficiency, ment, technical system design, distribution net- provide adequate power to meet local demand, and build in work mapping, business models, financial mod- scalability (GMG n.d.[e]). Developers and government staff eling, capital formation, community engagement, play an important role in ensuring that mini grids are built and legal compliance, among others. Project con- to code and comply with national and local rules and regu- struction skills include contracting, procurement, lations (USAID 2018f). installation, commissioning, community engage- ment, and project management. Some of the key A robust mini grid system is built with certified compo- skills for the O&M phase are process management, nents. Similarly, the installation should be conducted under financial and technical software, marketing and the supervision of certified and experienced technicians. customer service, metering, demand-side man- Implementing portfolios of technically sound, safe, and agement, demand stimulation, and performance affordable mini grids requires that everyone in the project monitoring and evaluation. unit have requisite skills: developers, technicians, suppliers, and electricians. MINI GRIDS FOR HALF A BILLION PEOPLE    207 O&M companies (GMG n.d.[b]). The staff responsible for fying gaps early in the development process, developers O&M must be properly trained, have detailed procedures and their partners can create a plan to address the gaps in place, and have access to higher-level technical special- among key stakeholders and help mitigate potential risks ists as required. The local staff require training on customer in the project. contract signing, revenue collection, cost reporting, instal- For example, DESI Power in Bihar, India, conducts a capac- lation and quality control of customer connections, and ity needs assessment based on market surveys that inform internal wiring. targeted training activities to develop local entrepreneurial In addition, the staff responsible for managing the mini grid and small business capacity. By enhancing the skills and system must possess knowledge and skills in customer knowledge of local entrepreneurs about how to run busi- service, the monitoring software or platform, data manage- nesses that use mini grid electricity, DESI Power overcomes ment, productive uses of electricity, and demand stimula- a key barrier: low or nonexistent demand in poor communi- tion. The training needs may vary, depending on whether ties (USAID 2018a). the developers are local or international. Capacity needs can vary depending on the project’s own- While some engineering and technical skills are shared ership model. For example, in some community-based more broadly with the renewable energy sector; some are models, the developer will depend on local expertise to specific to mini grid projects. Developers need to under- operate and maintain the mini grid system as well as con- stand national policy and regulations to meet legal require- duct business operations, such as tariff collection, admin- ments for their projects, as these are typically very specific istrative duties, and customer relations (USAID 2018e). to the mini grid sector. The technical knowledge and skills In these cases, local technicians may need to know how for system design are also very specific to the sector, as to install distribution infrastructure and how to perform they involve generation and transmission network devel- various O&M activities, so capacity gaps in local technical opment and standards. Business models, tariff design, expertise become particularly important. Capacity needs community engagement, and demand stimulation are also assessments will therefore vary depending on the specific very specific to the sector, given that the life span of a mini needs of the project, but they nevertheless generally follow grid project is 15–20 years. The mini grid will have to be the steps outlined in table 7.1. expanded if the load or customer base increases, and the Capacity needs assessments demand an honest assess- battery bank needs to be replaced every 8–9 years based ment of the gaps between existing capacity, in terms of on the life span of the system. Mini grid development also human resources and institutions, and what is needed to requires the technical skills involved in connecting to the effectively implement scalable and sustainable projects main grid, when and if it arrives. (Pew Charitable Trusts 2014). Capacity needs can be assessed using mixed methods IDENTIFYING SKILLS GAPS applied to existing data or freshly collected quantitative and qualitative data. Information can be gathered using key A shortfall in the skills required for mini grid develop- informant interviews, focused group discussions, and sur- ment can impede the scale-up of mini grids. Capacity veys. Interviews can target specific project personnel, such needs assessments are important tools to identify skills as developers, energy-access country coordinators, proj- gaps along the value chain. They reveal gaps in key areas, ect managers, or policy shapers in the responsible govern- including technical expertise, management skills, insti- ment ministries (Kang’ethe and others 2017). If capacity tutional capacity, policy, knowledge, partnership, and needs are not assessed, developers may face unexpected implementation (Pew Charitable Trusts 2014). By identi- challenges that delay project development. TABLE 7.1 • Conducting a project- or portfolio-level capacity needs assessment Step 2: Determine the project’s Step 3: Assess existing Step 4: Identify Step 1: Identify key actors or portfolio’s capacity needs capacity capacity gaps Who will build, own, operate, What technical, financial, managerial, What technical, financial, Which of the required and maintain the mini grid(s)? and other capacities will actors managerial, and other capacities are in short Who are the customers, need, and at what level of expertise? capacities do actors already supply or lacking potential investors, upstream Key areas to consider include have, and at what level of altogether? Are levels of suppliers, and relevant policy, knowledge, partnership, and expertise? expertise sufficient? government entities? implementation, among others. Source: Adapted from Energy 4 Impact and Inensus (2018). 208   MINI GRIDS FOR HALF A BILLION PEOPLE users and grid extension (Pedersen 2017). Its diesel off-grid power plants operate at a loss, and the reconciling of off- Capacity needs assessments are a critical grid electrification planning with grid extension plans has early step in designing effective training not always been easy (EUEI PDF, n.d.[a]). Another import- and skill-building initiatives. They reveal gaps in key ant skills gap for utilities is how to manage what happens areas, including technical expertise, management when the main grid arrives in the service area of a pri- skills, institutional capacity, policy frameworks, vate-sector mini grid—including the technical and financial partnerships, knowledge, and know-how. Needs capabilities to interconnect with the mini grid or purchase assessments generally follow a four-step process: the mini grid’s eligible assets. (See chapter 9 for a detailed (1) identify key actors, (2) determine the project’s discussion of options when the main grid arrives.) or portfolio’s capacity needs, (3) assess existing capacity, and (4) identify capacity gaps. They can BANKS AND FINANCIAL INSTITUTIONS be carried out using existing data or data collected from interviews, group discussions, and surveys. Local banks can be reluctant to lend to mini grid projects because of a lack of experience in appraising them (UNDP 2018b). As a result, they often consider financing mini grids to be too risky (Odarno and others 2017). Common gaps in the skills and expertise of key stakehold- ers can lead to considerable delays or missed opportunities Financial institutions at the national level, such as national to rapidly deploy mini grids, particularly when the mini grids development banks, may not be aware of opportunities to are to be deployed in portfolios instead of one-off projects. invest in mini grid projects. Even if a financing facility or To rapidly scale up mini grid development at the national credit line within a bank is created, a lack of skills in con- level, gaps in skills and knowledge should be identified for ducting due diligence and project appraisal can prevent all stakeholders in the value chain—developers, utilities, funding from flowing to developers. In Zambia, for example, banks and financial institutions, systems engineers, regula- training in financial evaluation and project management led tors, policy makers, suppliers, and local communities. The the Development Bank of Zambia to increase financing for following sections present some of the more common skills mini grid projects. In this case, bank staff received on-the- gaps for a selection of these key stakeholders. job training by being involved in the appraisal of the three pilot mini grid projects (Draeck and Kottász 2017). PROJECT DEVELOPERS Many small-scale and local developers are missing core ENGINEERS skills in project management, risk assessment, and O&M, Mini grids require technical knowledge for system design, particularly when it comes to owning and operating a port- construction, and O&M, and many projects have failed folio of mini grids. For example, most mini grid developers because of poor technical design, improper sizing, and in Tanzania that obtained local finance were foreign-owned poor maintenance. In many countries, local engineers may businesses. Commercial banks cited the poor quality of lack experience designing and sizing mini grids to the proj- documentation submitted by local developers as one rea- ect-specific context, yet because of limited local resources son for not extending credit to local developers (Odarno and capacity, they are often made responsible for the design and others 2017). International developers, by contrast, and maintenance of mini grids (Inversin 2000). Meanwhile, may have more capacity to raise finance and operate port- training local engineers to conduct O&M can add significant folios of mini grids but may lack knowledge and skills about costs for developers, particularly when they need a relatively how to operate in the local context (Energy 4 Impact and large number of people with the right technical capabilities INENSUS 2018). All developers—both local and interna- to install, operate, and maintain a portfolio of mini grids. tional—may also lack the skills needed to integrate produc- tive uses across a portfolio of mini grids. SUPPLIERS Like any business, mini grids depend on a well-function- UTILITIES ing supply chain—not just for their equipment but also for In countries where national utilities own or operate mini household and productive-use appliances for their cus- grids, the utility often lacks the expertise needed for mini tomers. Suppliers of generic electrical equipment are often grid–specific project development, such as community unfamiliar with certain mini grid–specific technologies, engagement, planning using geographic information such as smart meters and next-generation power electron- system (GIS) mapping, and payment collection (USAID ics, and may not have the capacity to import such technol- 2018c). For example, even though the Kenya Power and ogies. This can reduce competition in the market for mini Lighting Company (KPLC) is mandated with implementing grid components and limit the availability of locally tailored mini grids in Kenya, its focus has remained on industrial hardware (UNDP 2018a, 2018b). When this happens, mini MINI GRIDS FOR HALF A BILLION PEOPLE    209 grids can be put at risk: component failure or damage can tion that energy-access projects can have a higher socio- lead to long periods of system downtime when replace- economic impact by including productive uses (Cabraal, ment parts are unavailable, which can be compounded by Barnes, and Agarwal 2005). Unfortunately, different stake- a lack of local expertise and technical capacity (Ricardo holders in the mini grid ecosystem often lack the knowl- Energy & Environment 2016). edge or technical skills needed to integrate productive uses into the programs to scale up mini grids. Retailers and distributers of household and productive-use appliances may also lack the knowledge and capacity to reach rural populations, and they may not be familiar with highly efficient appliances or larger equipment that can run TRAINING AND SKILL-BUILDING on mini grid electricity. Filling these gaps can improve the INTERVENTIONS TO ADDRESS quality of available products and lead to better services. SKILLS GAPS POLICY MAKERS AND REGULATORS How do you empower thousands of people with the right Policy makers and regulators often do not have the prior skills in their specific domains to create an ecosystem experience and skill sets needed to develop and imple- capable of supporting hundreds or thousands of mini ment comprehensive policies, regulations, and techni- grids? In the following sections we present examples of cal standards to create an enabling environment for mini training programs and initiatives that answer this question grid scale-up. A lack of knowledge in GIS planning and in whole or in part—at the national level, at the project and resource assessment can prevent policy makers from portfolio levels, and at the community level. making data-driven and coordinated decisions on elec- NATIONAL-LEVEL TRAINING AND SKILL BUILDING trification planning. For example, a successful mini grids At the national level, training and skill-building interven- project implemented in India’s Sunderbans Delta by the tions should target government ministries, regulators, util- West Bengal Renewable Energy Development Agency in ities, rural electrification agencies, and other agencies that the late 1990s fell apart because of poor coordination influence the mini grid sector. (See chapter 8 for a discus- between the Ministry of Power and the Ministry of New and sion of the different government institutions that interact Renewable Energy. The conflicting and overlapping electri- with mini grids.) Most national-level initiatives are financed fication efforts of the two ministries resulted in the central and supported by development partners and international grid arriving at villages where mini grids already existed. In financial institutions. those villages, rather than being integrated and comple- menting each other, the central grid took over mini grids, Generally, training interventions at the national level cover rendering them obsolete (IT Power and AETS 2015). the development and implementation of policies and reg- ulations that support mini grids, the planning of mini grid LOCAL COMMUNITIES AND CUSTOMERS projects, and the monitoring and certification processes In community-based models, technical expertise may not required. These training needs are illustrated in figure 7.1, be available to install, maintain, and operate mini grid sys- along with the relevant stakeholders, and the benefits that tems. Community cooperatives responsible for managing training and skills building can provide. mini grids may lack the skills in business and accounting that are essential in collecting payments and maintaining Policy and regulation business operations. For example, between 1997 and 2012, Training sessions on policy and regulations should target more than 250 isolated community-owned micro hydro- regulators, ministries, rural electrification agencies, and power projects came on stream in Sri Lanka. The program any other authority that oversees mini grids and partic- was designed to create village-level mini grids that would ipates in national electrification planning. Training plans be owned and operated by community organizations, also need to be updated at regular intervals to reflect con- known as electricity consumer societies. However, the soci- ditions in the country and adapt to current targets and eties lacked the technical expertise to integrate the mini plans (IRENA 2018). Mini grids are most successful when grid systems into the national grid infrastructure. When the they have the support of the national government. Govern- main grid arrived, more than 100 of the isolated mini grids ments can mandate that mini grids be considered part of were closed, as just three societies converted their mini formal energy planning processes and help support them grids into main grid–connected small power producers with financial resources and incentives where required (Greacen 2017). (USAID 2018b). Lack of community engagement at the consumer level, Currently, development partners are mostly providing train- particularly around productive uses of electricity, can ing and skills building to policy makers and regulators. The also impede mini grid scale-up. There is a growing realiza- Energy Sector Management Assistance Program (ESMAP) 210   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 7.1 • National-level training needs and relevant stakeholders Outcomes Policy Regulations Planning Monitoring from trainings Training needs Training needs Training need: Training needs Key outcomers from policy • National electrification • Licensing and • GIS mapping • Stakeholder and regulatory level training • Business environment procurement • Resources assessment complaints and skill development • Mini grid–specific • Tariffs • Financial planning • Technical metrics • Mini grid development • Technical standards • Institutional and stake- • Regulation integrated into national Relevant stakeholders • Grid interconnections holder coordination compliance electrificationplan • Government ministries • Financial metrics • Enabling environment • Rural electrification Relevant stakeholders Relevant stakeholders through regulations agencies • Regulators • Regulators Relevant stakeholders and policies for mini • Government • Rural electrification • Regulators grid scale up ministries agencies • Rural electrification • Transition from • Utilities agencies project-level training • Banks to programmati approach Source: ESMAP analysis. GIS = geographic information system. provides direct support to regulators and ministries, work- ing alongside these entities to develop workable mini grid BOX 7.1 regulations and national electrification plans in more than a dozen countries. A good example is Economic Consult- POLICY AND REGULATORY TRAIN- ing Associates. With support from development partners, INGS PROVIDED BY ECONOMIC it has implemented successful training programs for policy CONSULTING ASSOCIATES makers and regulators (box 7.1). In recent years, Economic Consulting Associates • In Senegal, Power Africa focuses on capacity building, (a UK-based consulting company), has developed primarily for government entities (MEDER, Senelec, and policy and regulatory frameworks for mini grids in CSRE). Through the United States Energy Association, Rwanda, The Gambia, Ghana, Kenya, Mozambique, Power Africa has led training sessions with government and across the Southern African Development stakeholders (including the Agence Sénégalaise d’Elec- Community region. Each assignment has involved trification Rurale and the Agence Nationale pour les a training course on mini grid policy and regulation. Energies Renouvelables and Agence Sénégalaise d’Elec- Course participants are typically from the public trification Rurale). Power Africa also provides advisory sector—ministries, regulators, utilities, and rural support to help connect independent power producers energy agencies—but courses have also included to Senelec. private-sector representatives. Some courses in- • The German Agency for International Cooperation volve representatives of the regulators as trainers. (GIZ) provided technical support and training to the Nigerian Electricity Regulatory Commission in devel- Course content has included all key policy and oping mini grid regulations (Dalberg Global Develop- regulatory topics: licensing frameworks, business ment Advisors 2017). models, procurement, technical guidelines, and economic issues. Much of each course typically • In Zambia, the national utility ZESCO and Zambia’s Rural focuses on economic issues, including tariff calcu- Electrification Authority, along with other government lation and design, project financing, and subsidy staff, received training in sustainable energy regulation adjustments, with a series of practical examples and tariff setting. A training program was conducted and modelling exercises. The goals of the training to train technical staff at the Department of Energy, and skill-building workshops are to fully empower the Rural Electrification Authority, the Energy Regula- stakeholders to implement policy and regulations tion Board, and ZESCO to formulate and implement without the need for further external support, to renewable energy projects in the country. The project ensure effective collaboration between all sector has had a catalytic effect on the development of several stakeholders, and ultimately to ensure electricity is new mini grids in Zambia, particularly solar (Draeck and provided to those without access to it. Kottász 2017). Source: Economic Consulting Associates. MINI GRIDS FOR HALF A BILLION PEOPLE    211 Planning Monitoring Training on geospatial least-cost electrification planning Government entities tasked with monitoring mini grids’ (see chapter 2 on geospatial planning), resource assess- compliance with rules and regulations also require train- ment, and financial planning can help government agencies ing to avoid unnecessary red tape. Under the Promotion of make informed decisions on mini grid scale-up. Resource Solar-Hybrid Mini Grids (ProSolar) program in Kenya, GIZ planning and assessment can help to accurately pre- developed toolkits for solar photovoltaic (PV) and solar-hy- dict electricity demand, resource availability (solar, wind, brid mini grids and provided training to enable energy pol- hydro), and market pricing, thereby mitigating risk. icy makers and regulators to do a better job of monitoring mini grid systems (GIZ 2017a). The training sessions high- In planning and implementation, it is important to recog- lighted the importance of avoiding processes that were nize the interaction between electrification solutions (grid overly burdensome for private-sector developers. extension, mini grids, and stand-alone solar systems) and the rapidly evolving technologies and delivery models of National training programs off-grid renewables (efficiency improvements, innovations Governments may also elect to establish national training in metering, and end-user financing). Planning requires an and accreditation programs. By certifying mini grid courses assessment of the renewable resources available, for which and accrediting local training providers, governments can there are various private providers of resource assess- raise the quality of training and expand delivery of stan- ments, as well as public resource tools, especially for solar dardized courses through accredited local training institu- and wind resources (IRENA 2018). Some of these tools are tions. GIZ has helped develop solar PV and mini grid training listed in table 7.2. programs in Nigeria and Kenya. Under the Nigeria Energy Financial and business model planning can help policy Support Program, GIZ, in partnership with the United makers understand mini grids’ profit potential (EUEI PDF States Agency for International Development (USAID) and n.d.[b]). Training government staff in project planning Winrock International (Warren 2018), helped the Nigerian can help them make informed decisions about business government develop certification courses for solar and models (see Chapter 8 on mini grid delivery models), tar- hydro mini grid technicians, supervisors, and engineers. In iff schemes (see chapter 9 on mini grid regulations), and addition to compiling training manuals for trainers, GIZ and financing schemes (see chapter 6 on access to finance). its partners facilitated the training of trainers in 12 technical Table 7.3 presents some tools for financial planning. training institutions. TABLE 7.2 • Selection of tools for resource planning Application Tool Resource Private: 3TIER, AWS Truepower, Digital Engineering, Meteonorm, SolarGIS, and Windlogic assessment Public: Solar and Wind Energy Resource Assessment (SWERA), IRENA Global Atlas, National Aeronautics and Space Administration Earth Observing System (NASA EOS) Web, and Joint Research Centre (JRC) of the European Commission PV design PVSyst, PVWatts, PV Sol, and PVPlanner Wind design AWS Truepower’s Windographer, Wind Atlas Analysis and Application Program (WAsP) from Risø National Laboratory, and WindSim Source: Adapted from IRENA (2018). Note: The tools listed in this table provide general resource data of sufficient quality for national-level planning; however, wind and hydro typically require on-site verification. TABLE 7.3 • Sample tools for financial planning Tool Description RETScreen RETScreen is a software system for analyzing energy efficiency, renewable energy, cogeneration project feasibility, and energy performance. NREL CREST The National Renewable Energy Laboratory’s Cost of Renewable Energy Spreadsheet Tool (CREST) is an economic cash-flow model designed to allow policy makers, regulators, and the renewable energy community to assess project economics, design cost-based incentives (such as feed-in tariffs), and evaluate the impact of various state and federal support structures. CREST has separate tools for solar (photovoltaic and solar thermal), wind, geothermal, and anaerobic digestion technologies. Source: Adapted from GIZ (n.d.). 212   MINI GRIDS FOR HALF A BILLION PEOPLE Further, training interventions can be developed to train pol- Each phase of the project cycle can benefit from specific icy makers on implementing policies that promote socio- interventions. Figure 7.2 illustrates the main phases in economic development. This has been demonstrated by project and portfolio development, along with the training Smart Power India (SPI), a Rockefeller Foundation initiative needs and relevant stakeholders at each phase. that supports the government in developing policies and Each phase in the figure and its respective training and regulations to align the government, investors, and devel- capacity-building needs are described in the following opers. SPI also provides skill-building interventions to mini sections. grid companies (energy service companies) on key aspects of project development, such as financing, business mod- Site identification eling, site and cluster selection, procurement, and train- Developers, technical experts, and local communities can ing for technicians and customers. SPI’s programs have be trained to identify appropriate projects and sites for been very successful, leading to more than 178 renewable mini grid development. Some tools that can help in site energy mini grids with 6 megawatts of cumulative capacity. identification and development are described below. Their The electricity from these grids is transforming the lives of outputs depend on the reliability of the input data. Training more than 80,000 people by providing electricity for light- for field investigators can improve reliability (Gambino and ing and to power fans, electric pump sets, and appliances others 2019). and motors for productive uses (SPI 2019). • HOMER (Hybrid Optimization Model for Multiple Energy PROJECT-LEVEL TRAINING AND SKILLS BUILDING Resources) is a software program that facilitates the A comprehensive mini grid training curriculum should aim analysis of planning for distributed generation and mini to cover the whole project life cycle, from project develop- grid systems. HOMER Energy offers certified training ment and construction through operation (Energy 4 Impact on how to use its software for analysis. Participants and INENSUS 2018). The training and capacity-building learn how HOMER calculates the technical feasibility, needs vary by project and ownership model. At the project economic value, and other metrics of different designs level, interventions should target local communities, devel- through its powerful sensitivity analyses and its ability opers, system designers and engineers, suppliers, and to simulate and optimize thousands of systems designs installers. Assessments of capacity needs can help stake- in minutes.1 holders develop customized training at the project level. • Odyssey Energy Solutions is an online platform provid- Training materials should be interactive and appropriate for ing a comprehensive suite of tools to help project devel- adult learners. The materials used for training power plant opers design, build, and operate rigorous, systematic, operators, especially, should be practical and adapted to and data-driven microgrids.2 the project context as much as possible. All the materials, • Powerhive’s “Site Wizard for Analysis, Reconnaissance, including equipment catalogs and O&M manuals, should and Mapping” (SWARM) software enables project devel- be translated into the local language (ADB 2017). opers to remotely identify customers and site locations over broad regions. SWARM analyzes sites based on FIGURE 7.2 • Project- and portfolio-level training needs and relevant stakeholders Operaton & Outcomes Identification Feasibility Construction Maintenance from trainings Training needs Training needs Training needs Training needs Key outcomers from • Site identification • Technical/ • Technical design • O&M software project-level training and • Site selection technological assessment • Business models • Customer relations skill development • Socio-economic analysis • Procurement • Demand stimulation • Improved capacity and skills of Relevant stakeholders • Financial analysis • Construction technical experts • Developers • Demand estimation Relevant stakeholders • Adoption of appropriate • Technical experts Relevant stakeholders • Local technician and sustainable business • Local community Relevant stakeholders • Developers • Local community models • Developers • Technical experts • Trained and certified suppliers • Technical experts • Suppliers and technicians • Financial experts • Installers • Informed cusomers and • Local community • Local community increased ownership at the community level Source: ESMAP analysis. MINI GRIDS FOR HALF A BILLION PEOPLE    213 financial, technical, and geospatial data. It then calcu- • Financial analysis. Mini grid project developers often lates optimal mini grid locations, reticulation design, and lack experience in financial analysis, risk mitigation, estimated system size, resulting in a prioritized list of the and business plan development—and they may not be most viable sites. The data from the analysis are used able to afford outside financial professionals. Develop- to streamline the customer acquisition process, saving ers or their relevant staff members need training and time and radically lowering project development costs.3 skill-building support in financial modeling, capital accu- mulation, and proposal writing. Project developers and public agencies responsible for rural electrification planning can benefit from training and • Demand estimation and projections. A critical step in skill building in project identification. In Nigeria, the Rural any mini grid project is to gauge demand for electricity. Electrification Agency (REA) benefitted from technical This demand assessment should cover all households, support and skill-building interventions from development productive users, and social institutions in the service partners and organizations such as the Rocky Mountain area. The project developers must estimate project Institute. Topics included data on households, commer- costs, connection costs, and the electricity tariffs. To cial users, potential system size, peak load, and geospatial gather more reliable data, developers should set the analysis to identify sites appropriate for mini grids.4 The business model before conducting the surveys (Ener- REA is presently engaged in project and site identification gypedia 2016). for 250 sites throughout Nigeria. By 2030, it will help scale mini grids to more than 10,000 sites powering 14 percent of Construction the population (Ogunbiyi 2017). Training interventions aimed at imparting the skills and knowledge required to build and commission a mini grid Feasibility assessment typically assume a minimum level of prior knowledge of Once projects are identified, developers need specific skills electrical systems. As a result, such interventions can be and knowledge to assess feasibility. Geospatial analysis is built into, or alongside, existing courses for engineers and increasingly used for detailed preliminary studies before technicians. In some cases, depending on the developer’s survey teams are dispatched to the most promising sites business model, training on how to build and commission (see chapter 2 on geospatial planning). An example of a mini grid may also include training on how to operate and training for these activities at the national level comes from maintain the system. Training and skill-building interven- a project in Rwanda. The Private Sector Participation in the tions should, at a minimum, cover the following: Generation and Distribution of Electricity from Renewable • Technical design. A mini grid’s size dictates its maxi- Sources project provides training to mini grid developers mum power output. The generation system must have on assessing the feasibility of renewable energy mini grids, sufficient installed capacity to meet loads. To size the with the goal of increasing the bankability of their proj- system, planners must calculate load variations in half- ect proposals (BRD 2016). At the international level, the hour intervals and estimate future load growth. Estimat- Renewables Academy Online (RENAC n.d.) provides online ing and planning for current and future loads are critical training courses, resulting in professional certificates for PV steps, especially for financial viability. Developers can and PV-diesel–hybrid systems. estimate current loads by surveying and assessing cur- Training programs on feasibility assessment have four key rent and potential customers (USAID 2018d). Anticipat- components. ing future loads is more difficult. • Technical and technological assessment: Technical soft- • Procurement. All mini grid projects involve acquiring ware helps developers design systems to meet existing equipment and services from external sources, often via and projected demand. The key is to ensure that the a tendering or competitive bidding process. Developers design is carried out by technicians with skills and expe- may issue tenders for individual components of the mini rience in the relevant technology for planning, installa- grid and then install and commission the system them- tion, commissioning, and O&M. selves (GMG n.d.[c]). • Socioeconomic analyses: Socioeconomic data and • Installation. Mini grids should be operated under the analysis can provide quantitative and qualitative infor- supervision of certified and experienced technicians. mation that can be helpful in designing projects and It is also important before work begins that the tech- estimating current and future demand. Conducting nicians test delivered equipment and locally produced socioeconomic surveys helps developers gather infor- products (such as foundations, channels, and bricks) mation on mini grid customers, their willingness to pay, and provide safety training to all personnel, including and their current spending on alternative sources of local support staff. electricity. 214   MINI GRIDS FOR HALF A BILLION PEOPLE The purpose of technical professional development is training technicians. Husk’s predicament highlights the fre- to ensure that system designers are well trained on all quent absence of comprehensive training programs, which the available technologies, such as solar-hybrid, inverter places the burden of training technicians on developers.5 technologies, storage technologies, and grid connection methods. System designers should also consider available COMMUNITY INVOLVEMENT AND AWARENESS materials, capabilities, and know-how at the community Training programs for communities typically center on rais- level, which can reduce the implementation cost and, later, ing awareness about mini grid electricity and how it can be the cost of training end users. used for household activities and for income-generating System designers and engineers holding university de- activities. (See chapter 4 for a full discussion on commu- grees or higher training certificates are well positioned to nity engagement and training.) advance mini grid technology in the country. In addition to engineering training offered through universities and poly- Awareness of household uses of electricity technics, technology companies can provide training in Educational campaigns can inform local communities of their products and technologies. the range of uses of electricity, which may increase their willingness to pay. For example, sessions might focus on For example, to overcome the technical knowledge gap how electricity can improve agriculture and sanitation in the mini grid sector in Chad, 35 public and private through water pumping, improve health and education, stakeholders were trained in the use of HOMER software, and support the development of new businesses. Tar- described earlier. Training included a course for 44 peo- geted training helps local communities understand their ple on the management of mini grids based on renewable energy use and can move them to a higher tier of energy energy. Further training was carried out at project sites. Six consumption, including productive uses, as seen in the people per site conducted training sessions on managing PowerCorner project in Arusha, Tanzania. PowerCorner solar installations, maintaining and managing technical goes beyond providing electricity: it not only sells ener- teams, and financing mini grids. The training prioritized gy-efficient home appliances through loans to the villag- gender mainstreaming. ers but also invests in training customers on how to use the electricity efficiently (Lebleu 2018). Communitywide Operations and maintenance education campaigns can teach people to use electricity Training in operating and maintaining mini grid systems is safely and demonstrate how it can improve their well-be- essential to their long-term sustainability. Training and skill ing through education and health. Communities new to building can be done online, face to face, or as a combina- electricity may need instruction in the safe use of electri- tion of the two. Sessions may be provided by governments cal appliances and household wiring. (Nepal, Mali), private groups (such as Trama Techno- Ambiental, Mee Paynar, and the Institute of Electrical and Various engagement strategies are being tested to Electronics Engineers [IEEE]), universities (University of increase community involvement in the development Strathmore, Arizona State University), training institutes of mini grids. Quicksand is a user-fed, video-based digi- (Renewable Energy Solutions for Africa and the Economic tal platform that showcases different users’ stories to Community of West African States [ECOWAS] Centre for demonstrate the benefits of mini grids around the world Renewable Energy and Energy Efficiency), as well as some (Quicksand 2018). This digital platform’s goal is to prop- development partners (such as GIZ). agate the use of mini grid electricity in businesses and homes in communities that receive either irregular or no Online training platforms can provide technical training to grid electricity. Quicksand visited 10 communities served help solar companies and organizations manage their staff by mini grids to learn, share, and test its platform. The and assets. LEDsafari is a Swiss-based startup in clean project resulted in a smartphone app, a dedicated digital energy that develops digital products for such purposes. platform to host mini grid videos, and 22 films on indi- It has developed a multilingual online training platform viduals’ experiences with mini grids. Of the final 22 films called HelioLearn that provides technical and business produced, 11 were made by users with the developed app, courses on solar energy. HelioHealth is a cloud-based sen- with minimal help from Quicksand, illustrating the poten- sor for solar panels. Panel-level monitoring, combined with tial of this app-plus-website to scale rapidly with minimal advanced machine learning and data analysis, helps solar intervention (Quicksand 2018). companies track the performance of their devices and to diagnose faults (LEDsafari n.d.). Trama TechnoAmbiental employs a more hands-on, train- the-trainers approach aimed at communities that will be To address the lack of qualified candidates for its O&M team, served by mini grids, as described in box 7.2. Husk Power has recently established its own university for MINI GRIDS FOR HALF A BILLION PEOPLE    215 BOX 7.2 LESSONS FROM COMMUNITY TRAINING BY TRAMA TECHNOAMBIENTAL Trama TechnoAmbiental (TTA) demonstrates how The TTA training model is part of the mini grid project strong capacity building and engagement of the local itself and is included in the overall project financing. community from the initial phases of a mini grid project According to TTA, this model has been cost-effec- can lead to well-maintained and sustainable mini grid tive, has reinforced the governance structure within systems. Training and enhanced awareness among the the community, and has helped build local capacity. beneficiaries of a mini grid are important in minimiz- Further, local demand has continued to rise, indicat- ing the vandalism of assets and managing community ing that the consumption categories were correctly expectations. estimated and that end users are satisfied with their TTA trains in two stages. First, local actors receive a electricity services. more theoretical training, followed by a second, cus- TTA has also noted some limitations to its approach. tomized training that reflects the actual operation of Despite the training provided and the small local the relevant system. In the first stage, training is offered structure set up for O&M, if a major technical problem to (1) end users on the possibilities and limitations of occurs, the community is unlikely to be able to resolve the system and on the uses of electricity, (2) the entity the issue or find replacement components quickly. that will be responsible for operation and maintenance Although a local committee is responsible for ensur- (O&M), and (3) the local technicians who will perform ing that funds put aside are used as intended, there the O&M work. In the second stage, after about six is always the risk that funds for future replacements months, TTA visits the project, addresses any problems may be used for other needs that may arise in the that have arisen, and completes the training. community. Source: TTA 2019. Awareness of productive uses of electricity that more than 80 percent of the trainees started a busi- Providing access to electricity does not guarantee that ness in the community. New activities included bakeries, communities will move on to productive uses of energy, preparing broiler chickens, agricultural processing mills, as discussed in chapter 3. Encouraging productive uses sawmills, photo studios, and producing incense sticks requires a multifaceted approach, including strategies (Fishbein 2003). to create awareness of how mini grid electricity can help Mentoring is needed to help build the commercial and increase revenues for local entrepreneurs and small busi- technical skills of local entrepreneurs, to train them in nesses. Box 7.3 provides an example of such a training pro- using electrical appliances, and to help their businesses gram developed by the IEEE. navigate the challenges of early development. Mentor- Trainings targeted at productive-use customers help them ing could be provided by governments, nongovernmen- identify income-generating electric appliances, access tal organizations, or potentially developers themselves if capital to acquire the equipment, and use the mini grid– they had the necessary capacity and knowledge (Booth powered equipment to cut costs, grow the business, and and others 2018). diversify. For example, Energy 4 Impact offers entrepre- The UNDP experience shows that the training and skills neurs training in business management skills (such as development of different local stakeholders lead to not only record keeping, marketing, and customer care), appliance improved O&M but also increased productive use of energy use, and health and safety standards to customers of Mesh from mini grid systems. Targeted trainings on productive Power. Currently, MeshPower operates more than 80 solar uses and machinery played a critical role in community mini grids that supply electricity to more than 2,400 house- capacity building. In Sri Lanka, the mini grid community is holds and small businesses across Rwanda (MeshPower fully involved and has received training in mini grid O&M. 2022). (See chapter 3 for an in-depth look at productive All assets were transferred to a consumer committee to uses of mini grid electricity.) In another example, a United operate and maintain the system. Since the project was Nations Development Programme (UNDP) initiative in also servicing irrigation systems, the training also included Nepal provided skills training to households and reported 216   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 7.3 PRODUCTIVE-USE TRAINING FROM IEEE SMART VILLAGE IEEE Smart Village is an initiative of the Institute of FIGURE B7.3.1 • Future community entrepreneurs Electrical and Electronics Engineers (IEEE) Foun- and mini grid technicians participate in a dation. It aims to reach 50 million people by 2025 classroom discussion at a training program through electrification, community-based educa- offered by Igniting Africa tion, and sustainable enterprise centered on produc- tive uses of electricity. IEEE Smart Village supplies philanthropic venture funding, mini grid equipment, pro-bono consulting, and extensive education and training resources to launch and grow locally operated micro-utilities. Igniting Africa is an IEEE Smart Village program. Launched in Bamenda, Cameroon, it delivers commu- nity electrification through mini grids serving 22,500 people. The program provides numerous vocational training programs designed to demonstrate the scal- ability of technical and business education. A train- the-trainer methodology is combined with hands-on courses for youth who return to their home villages with Source: Used with permission from IEEE. seed funding to start their own businesses centered on mini grids and productive uses of energy. munity-based education and development of local Recognizing that electricity is an enabler and a means entrepreneurship. The array of microentrepreneurial toward economic empowerment (rather than an end activities targeted by IEEE-funded trainings includes in itself), IEEE Smart Village structures its programs artisan crafts, construction, electrical wiring, elec- not just to provide electricity, but also to drive com- tronics assembly, electric transportation, information technology services, retail services, sustainable agri- Source: IEEE. culture, and tourism. irrigation infrastructure monitoring and water use manage- members of the community management committee are ment. In India, 30 community members from two project women. Earthspark has also supported local female entre- sites received training to use the renewable energy gener- preneurs to start small businesses using power from the ated from micro-hydro mini grids for productive purposes. grid and has become a strong advocate of gender inclu- These participatory trainings included marketing strate- sion for grid operators based on this experience (ESMAP gies and business models (Draeck and Kottász 2017). 2017). The Global Environment Facility and United Nations Industrial Development Organization project in The Gam- Gender-related training and capacity building bia has also succeeded in building the knowledge and Mini grid projects have been successful in empowering in-depth technical capacity of the regulators and the utility women and supporting female-led businesses. Training and (NAWEC). The project had a training program that focused skills-building interventions can be developed specifically on women. As a result of this program’s success, the gov- to provide trainings on how to benefit women and should ernment decided to use 50 percent of its overall renew- also incorporate gender-sensitive training. For example, able energy fund for projects focused on skills building for the US-based nonprofit Earthspark owns and operates a women. mini grid in Les Anglais, Haiti, serving 449 homes and busi- Other initiatives that have been successful in incorporating nesses with affordable, reliable electricity. Earth Spark has and customizing gender-based skills building and training made a commitment to integrate gender equality in every support are the Barefoot College and Solar Sisters. Bare- aspect of its operations. Local women have been trained foot College stands out because it has recently announced and employed to install parts of the grid, and 4 of the 10 MINI GRIDS FOR HALF A BILLION PEOPLE    217 plans to create a network of more than 2,000 female ery allows for more flexibility in terms of when, where, and engineers and entrepreneurs and has affected 200,000 how different types of training are conducted. For exam- households. Together, Barefoot College and Solar Sisters ple, theoretical training can be offered remotely or online, will establish 100 “solar demo hubs” in rural areas, start- while practical work is done in the classroom or the field. ing in India’s Rajasthan state, with plans to expand to five The blended approach helps the training providers keep states in India and 10 countries internationally (Power for their costs down and widen the audience of potential stu- All 2017). A new study in Ghana by Power Africa and Black dents. The success of any program involving online content Star Energy has revealed the significant impact of elec- depends on the training providers and students having tricity access on women-owned businesses and incomes, access to power, computers, and the internet (Energy 4 including helping them move from extreme poverty to Impact and INENSUS 2018). Box 7.4 details the blended near-middle-class status, while allowing them to stay in approach taken by IEEE. their rural communities (Poindexter 2018). Standardized training methods and accrediting improve the quality of training and standardize training processes through accredited institutions. However, there is still no LESSONS LEARNED FROM EFFECTIVE clear path on how this certification should be done. Three TRAINING AND SKILLS-BUILDING options that have emerged from ESMAP’s conversations with mini grid sector stakeholders are: creating a nation- PROGRAMS ally recognized mini grid certificates similar to the ones for Skills-building programs are useful for a wide set of stake- solar engineers; introducing regional as well as national holders, including regulators, rural electrification agencies, certificates, although this assumes the relevant national policy makers, academic institutions, finance institutions, institutions can agree on common quality standards; and developers, engineers, technicians, and end users. Many creating regional training programs that can be delivered governments (for example, Kenya, Ghana, India, and through a common platform and integrated through local Nepal), private entities (such as Schneider Foundation, certification or standardized tests. SELCO, KITTEC, IEEE), and educational institutions (such In addition, training should not be considered a one-off as Strathmore University and Arizona State University) are event, and people trained under any mini grid project or providing training and skills-building programs and courses program need follow-up and refresher courses after the to mini grid stakeholders worldwide. training. For example, TTA, a global consulting and engi- There is currently no widely recognized mini grid indus- neering company,7 delivers training in two stages. First, try training standard. While nearly all vocational courses during the project preparation and implementation phases are locally certified, professional development courses of the project, training on system limitation and use is deliv- are often not certified and there is no regional certifica- ered to end users and training on O&M is provided to local tion system for mini grid training. Many mini grid training technicians. Second, after about six months, TTA visits the programs have a narrow focus on technical training for project, responds to any problems that have emerged, and engineers and do not cover other key factors for devel- completes the trainings based on the challenges identified oping mini grid projects (Energy 4 Impact and INENSUS in the first six months of operation (Wiemann, Rolland, and 2018). However, to build the skills and capacity of local Glania 2014). entrepreneurs in the off-grid and mini grid sector, the In another example where training and skills development SELCO Foundation has started an incubation center that are undertaken by the state agency, India’s Chhattisgarh uses SELCO’s shared resources, management expertise, State Renewable Energy Development Agency (CREDA) intellectual capital, and bottom-up learning to enhance runs an Installers Certification Programme, designed for the capacity and vision of potential local entrepreneurs. personnel who are specifically assigned to carry out the Their process follows identification, initiation, incubation, installation and commissioning of projects. CREDA also growth, and analysis for selected candidates.6 provides refresher-training programs every six months A review of training programs conducted by Energy 4 for technicians, operators, and Village Energy Committee Impact and INENSUS for the African Development Bank members. CREDA has electrified around 35,000 house- indicated that most were offered locally via face-to-face holds across more than 1,400 villages and hamlets with trainings (Energy 4 Impact and INENSUS 2018). Some low-capacity (1–6 kilowatts-peak) solar mini grids in Chhat- professional development training providers, however, use tisgarh (Palit and Sarangi 2014). a blend of face-to-face and online training. Blended deliv- 218   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 7.4 COMPONENTS OF IEEE SMART VILLAGE’S COMPREHENSIVE TRAINING PROGRAM Members of a newly electrified community frequently The majority of schools at all levels in IEEE-funded lack not only technical knowledge and business skills, communities lack critical education resources, includ- but also education fundamentals, including basic lit- ing electricity, internet, clean water, safe sanitation, eracy and numeracy. To address this need, the IEEE required textbooks, course materials, or access to Smart Village has developed an extensive set of online mandatory government curricula and exams. Access and classroom-based curricula and education pro- to electricity from mini grids enables the creation of grams spanning from kindergarten to retirement, digital classrooms with electronic copies of govern- covering technical skills, vocational training, and K–12 ment curricula, textbooks, and scripted daily classes education. to assist teachers with limited pedagogical proficiency. IEEE Smart Village entrepreneurs have implemented a To date, IEEE Smart Village training initiatives have range of innovative education technologies, including conducted eight comprehensive training programs in EmpowerSchool (developed by EmpowerPack Social seven countries, reaching more than 97,000 people, Purpose Corp.), Rachel PI (by World Possible), Blue including almost 12,000 students and youth. Box (by Worldreader), and TalkingBook (by Mavis Technical education delivered to communities is devel- Computel Ltd.). oped by local subject matter experts and tailored to All IEEE Smart Village community entrepreneurs the local environment, culture, and language. Technical actively participate in a collaborative network pro- training provided by IEEE community entrepreneurs moting open sharing of lessons learned, best prac- covers a broad range of topics, including installation, tices, training curricula, and business plans for both operation, and maintenance of mini grids; siting and mini grid construction and microbusinesses focused sizing of wind turbines, hydro turbines, and solar arrays; on productive uses of energy. This forum for open battery system configuration and maintenance; com- innovation and information exchange enables each missioning and use of customer billing and payment successive round of new projects to increase the systems; computer-aided design; assembly and repair replicability, scalability, and sustainability of mini grid of electronics; and development of computer software applications worldwide. Additionally, project manag- and applications. ers and community leaders have access to a prac- Vocational training programs are designed to maximize titioner-oriented Masters of Development Practice the opportunities for community members to engage online program through Regis University and the IEEE in new entrepreneurial activities that are made possible Smart Village global classroom at the Posner Center by access to electricity. The course offerings combine for International Development in Denver, Colorado, hands-on activities with basic business skills and fun- for 2 percent of the cost of a traditional degree. Praxis damental literacy, numeracy, and social skills. courses in monitoring and evaluation techniques are also provided to help accelerate collection of metrics Source: IEEE. and impact data. DATABASE OF TRAINING PROGRAMS companion website to this handbook: www.esmap.org/ mini_grids_for_half_a_billion_people. The following are ESMAP has developed a database of more than 50 train- some examples: ing courses and entities that provide trainings that are The National Power Training Institute of Nigeria (NAPTIN) relevant to the mini grid sector. These training programs is a federal government institution that reports to the are available all over the world—and many can be deliv- Federal Ministry of Power, Works and Housing and ered remotely. They target policy makers and regulators, operates nearly 10 regional training centers across the developers, engineers, and operators. Most of the training country. NAPTIN is also responsible for helping the min- programs provide a formal certificate upon satisfactory istry develop policy to build capacity in the power sector.8 completion of the course. This database is available on the MINI GRIDS FOR HALF A BILLION PEOPLE    219 NAPTIN increasingly seeks to take a private-sector-driven REFERENCES approach to its training and operations to best meet the ADB (Asian Development Bank). 2017. Developing Renewable demand of the privatized power sector (Ley, Gaines, and Energy Mini-Grids in Myanmar: A Guidebook. Manila, Philip- Ghatikar 2015). pines: Asian Development Bank. https:/ /www.adb.org/sites/ default/files/institutional-document/391606/developing-renew- The ECOWAS Certification for Sustainable Energy Skills able-mini-grids-myanmar-guidebook.pdf. Program has been established by the ECOWAS Centre for AfDB (African Development Bank). 2016. “Green Mini-Grids in Sub-Sa- Renewable Energy and Energy Efficiency in partnership haran Africa: Analysis of Barriers to Growth and the Potential Role with the International Renewable Energy Agency and GIZ. of the African Development Bank in Supporting the Sector.” Green The program aims to improve the technical competency of Mini-Grid Market Development Program Document Series No. 1, various renewable energy professions across the ECOWAS AfDB, Abidjan, Côte d’Ivoire, December. https:/ /energy4impact. member states. The scheme is being piloted in Ghana and org/sites/default/files/pdf_green_mini-grids_a4_version_finale. compressed.pdf. Senegal for certification of stand-alone solar PV techni- Booth, S., X. Li, I. Baring-Gould, D. Kollanyi, A. Bharadwaj, and P. cians who will be required to clear a written and practical Weston. 2018. Productive Uses of Energy in African Micro-Grids: examination based on a regionally harmonized job-task Technical and Business Considerations. Washington, DC: United analysis that details competencies for installation, main- States Agency for International Development; Golden, CO: National tenance, safety, and basic design of off-grid solar systems. Renewable Energy Laboratory. https:/ /www.nrel.gov/docs/fy18o- Subsequently, the certification will be expanded to other sti/71663.pdf. solar PV technician profiles, such as on-grid solar PV and BRD (Development Bank of Rwanda). 2016. “Terms of Reference for mini grids (IRENA 2019). Renewable Energy Financing Expert.” https://www.brd.rw/brd/ wp-content/uploads/2016/04/Energy_Dpt_REF.pdf. The Micro-Grid Academy (MGA) is a regional capaci- Cabraal, A., D. Barnes, and S. Agarwal. 2005. “Productive Uses of ty-building platform that provides theoretical and practical Energy for Rural Development.” Annual Review of Environment training on energy access and decentralized renewable and Resources 30: 117–44. https:/ /www.annualreviews.org/ doi/10.1146/annurev.energy.30.050504.144228. energy solutions to young East African and international technicians, entrepreneurs, and engineers. The project Dalberg Global Development Advisors. 2017. Improving Access to Electricity through Decentralised Renewable Energy: Policy Analy- aims to enhance access to energy in rural communities sis from India, Nigeria, Senegal and Uganda. https://dalberg.com/ fostering local enterprise and job creation. Located in the wp-content/uploads/2017/05/Dalberg-offgrid-policy.pdf. KPLC’s Institute of Energy Studies & Research in Nairobi, Draeck, M., and E. Kottász. 2017. “Renewable Energy-Based Mini-Grids: Kenya, the MGA is coordinated by the RES4Africa Founda- The UNIDO Experience.” Department of Policy, Research and Sta- tion, in partnership with Enel Foundation, the national Ken- tistics Working Paper 1/2017, United Nations Industrial Develop- yan utility KPLC, Strathmore University, AVSI Foundation, ment Organization, Vienna. https:/ /www.unido.org/sites/default/ files/2017-03/Minigrid_report_Jan2017.v19_FINAL29906_0.pdf. and St. Kizito Vocational Training Institute, and endorsed by the East Africa Center of Excellence for Renewable Energy e-MPF (European Microfinance Platform). 2014. “Small-Scale Mini-Grids: Product Catalogue—2014.” http:/ /www.e-mfp.eu/ and Energy Efficiency. To support the theoretical lectures sites/default/files/resources/2014/11/Small-scale%20Mini- with practical learning, a real 30-kilowatt hybrid mini grid Grids_2014_web.pdf. will be installed on site thanks to the contribution of RES4A- Energy 4 Impact and INENSUS. 2018. Mini-Grid Training Needs Assess- frica members.9 ment: Gap Analysis for Developers. http:/ /energyaccess.org/ wp-content/uploads/2018/11/AFDB_Minigrid_Developer_Train- Founded in 1994, Peru’s Center of Demonstration and ing_Needs_Assessment_September-2018.pdf. Qualification in Appropriate Technologies (CEDECAP) Energypedia. 2016. “Demand Assessment for Mini-Grids.” https://ener- aims to develop technical and managerial skills in leaders, gypedia.info/wiki/Demand_Assessment_for_Mini-grids. students, manufacturers, technicians, professionals, and Escobar, R., D. Vilar, E. Velo, L. Ferrer-Martí, and B. Domenech. 2012. officials of Latin America.10 The most noteworthy aspect of “Promoting and Improving Renewable Energy Projects through CEDECAP is the hands-on approach that the training pro- Local Capacity Development.” In Modeling and Optimization of vides in real facilities. Practice lessons are carried out in real Renewable Energy Systems, edited by Arzu Şencan, 147–70. Lon- don: IntechOpen. systems. As a result, students learn exactly how systems work and how they should be operated once installed in the ESMAP (Energy Sector Management Assistance Program). 2017. “Mini-Grids and Gender Equality: Inclusive Design, Better Devel- communities. Their methodology is to spend 70 percent of opment Outcomes. Key Issues, and Potential Actions.” World Bank, the training time in fields of expertise and 30 percent in the Washington, DC. https:/ /openknowledge.worldbank.org/han- classroom. The energy module comprises three environ- dle/10986/29050. ments with equipment in hydraulic, solar, and wind energy EUEI PDF (European Union Energy Initiative Partnership Dialogue (Escobar and others 2012). Facility). N.d.(a). “Mini-Grid Policy Toolkit–Case Study: Country: Kenya.” Bonn. http:/ /minigridpolicytoolkit.euei-pdf.org/system/ files_force/attachments/Mini-Grid%20Policy%20Toolkit%20 Case%20Study%20-%20Kenya60ab.pdf?download=1. 220   MINI GRIDS FOR HALF A BILLION PEOPLE EUEI PDF. N.d.(b). “The Full Financial Model for Mini-Grid Projects.” Afri- IT Power and AETS (Application Européenne de Technologies et de Ser- ca-EU Renewable Energy Cooperation Programme. http:/ /minigrid- vices). 2015. Evaluation of the Financial and Economic Combination policytoolkit.euei-pdf.org/system/files_force/RECP_Financial%20 of SHS and Mini-Grid Systems. Abidjan, Côte d’Ivoire: African Devel- Model_Explanation60ab.pdf?download=1. opment Bank. https:/ /www.african-ctc.net/fileadmin/uploads/ Fishbein, R. 2003. “Survey of Productive Uses of Electricity in Rural actc/Documents/GHANA_SE4All_evaluation_of_SHS_minigrids. Areas.” World Bank, Washington, DC. http:/ /www.martinot.info/ pdf. Fishbein_WB.pdf. Kang’ethe, E., L. Symekher, M. Lemma, P. Sambati, and I. Dror. 2017. Gambino, V., R. Del Citto, P. Cherubini, C. Tacconelli, A. Micangeli, and “Africa RISING Mali: Capacity Needs Assessment.” International R. Giglioli. 2019. “Methodology for the Energy Need Assessment to Institute of Tropical Agriculture, Ibadan, Oyo State, Nigeria. https:// Effectively Design and Deploy Mini-Grids for Rural Electrification.” cgspace.cgiar.org/bitstream/handle/10568/90398/cna_ar_Mali. Energies 12 (3): 574. https://www.mdpi.com/1996-1073/12/3/574/ pdf?isAllowed=y&sequence=1. htm. Lebleu, T. 2018. “PowerCorner, the Power of Sustainable Mini-Grids GIZ (Deutsche Gesellschaft für Internationale Zusammenarbeit). for Rural Villages.” Solar Impulse Foundation. https://solarimpulse. 2017a. “GIZ Portfolio: Renewable Energy and Hybrid Mini-Grid Sys- com/news/view/powercorner-the-power-of-sustainable-mini- tems.” Eschborn and Bonn. https:/ /energypedia.info/images/f/ff/ grids-for-rural-villages. GIZ_%282017%29_Mini-grid_Portfolio_Overview.pdf. LEDsafari. N.d. “Learning Management System for Solar Companies: GIZ. 2017b. “Nigerian Energy Support Programme II.” https://www.giz. Train and Track Your Customers/Staff.” Lausanne, Switzerland. de/en/worldwide/26374.html. http://ledsafari.com/wp-content/uploads/HelioLearn-presenta- tion_last-version.pdf. GIZ. N.d. “Tools for Mini-Grid Practitioners.” Clean Energy Solutions Center. https://cleanenergysolutions.org/sites/default/files/docu- Ley, K., J. Gaines, and A. Ghatikar. 2015. The Nigerian Energy Sector: An ments/mini-grid-tools.pdf. Overview with a Special Emphasis on Renewable Energy, Energy Effi- ciency and Rural Electrification. 2nd ed. Abuja, Nigeria: GIZ. https:// Greacen, C. 2017. “Mini-Grids and the Arrival of the National Grid: www.giz.de/en/downloads/giz2015-en-nigerian-energy-sector. Case Studies from Cambodia, Sri Lanka & Indonesia.” PowerPoint pdf. presentation at the Upscaling Mini-Grid Workshop in Naypyitaw, Myanmar, February 9. https:/ /www.esmap.org/sites/default/ MeshPower. 2022. “Our Installations”. https://www.meshpower.co.rw/ files/170206%20Greacen%20ESMAP%20mini-grid%20Myan- Ochs, A., and I. Indriunaite. 2019. Advancing Mini-Grids in Sierra Leone: mar%20grid%20arrival%20case%20studies%20technical%20 Exploring a Holistic Framework of Policies, Regulations, and Institu- breakout_optimized.pdf. tions. Clean Energy Solutions Center. https:/ /ledsgp.org/resource/ GMG. N.d.(b). “Operation and Maintenance.” https://greenminigrid. advancing-mini-grids-in-sierra-leone-exploring-a-holistic-frame- afdb.org/how-it-works/help-desk-developers-and-operators/ work-of-policies-regulations-and-institutions/. operation-and-maintenance Odarno, L., E. Sawe, M. Swai, M. Katyega, and A. Lee. 2017. Accelerating GMG. N.d.(c). “Procurement, Installation and Commissioning.” Mini-Grid Deployment in Sub-Saharan Africa: Lessons from Tanza- https://greenminigrid.afdb.org/how-it-works/help-desk-develop- nia. TaTEDO and World Resources Institute. https:/ /www.wri.org/ ers-and-operators/procurement-installation-and-commissioning. publication/tanzania-mini-grids. GMG. N.d.(d). “Setting Up a Mini-Grid Business.” https://greenminigrid. Ogunbiyi, D. 2017. “The Off-Grid Opportunity in Nigeria.” PowerPoint afdb.org/how-it-works/help-desk-developers-and-operators/set- presentation at the Upscaling Minigrids for Least Cost and Timely ting-mini-grid-business. Access to Electricity Action Learning Event in Abuja, Nigeria. https://www.esmap.org/sites/default/files/Presentations/REA_ GMG. N.d.(e). “Technical System Design.” https://greenminigrid.afdb. Damilola-Off-Grid%20Opportunity_03122017_web.pdf. org/how-it-works/help-desk-developers-and-operators/techni- cal-system-design. Palit, D., and G. Sarangi. 2014. Renewable Energy-Based Rural Elec- trification: The Mini-Grid Experience from India. New Delhi: The IEA (International Energy Agency). 2003. PV for Rural Electrification in Energy and Resources Institute for the Global Network on Energy Developing Countries: A Guide to Capacity Building Requirements. for Sustainable Development. https:/ /unepdtu.org/wp-content/ Report IEA-PVPS T9-03:2003. Paris: IEA. uploads/2014/09/renewable-energy-based-rural-electrifica- Inversin, A. 2000. Mini-Grid Design Manual. National Rural Electric tion-the-mini-grid-experience-from-india.pdf. Cooperative Association. https:/ /www.reseau-cicle.org/wp-con- Pedersen, M. 2017. “Rural Electrification through Private Models: The tent/uploads/riaed/pdf/Mini-Grid_Design_Manual-2.pdf. Case of Solar-Powered Mini-Grid Development in Kenya: Exploring IRENA (International Renewable Energy Agency). 2017. “Mini-Grid the Hybrid Nature of Private Business Models and the Interplay Project Guide Developed for Small Islands.” https://www.irena.org/ between New Players and Existing Structures in the Kenyan Rural newsroom/articles/2017/Sep/Mini-grid-project-guide-developed- Electrification Regime.” PhD Thesis, Technical University of Den- for-small-islands. mark, Copenhagen. http:/ /orbit.dtu.dk/files/134853297/PHD_ IRENA. 2018. Policies and Regulations for Renewable Energy Mini-Grids. FINAL_til_print_MBP_August_2017_inkl_For_bagside.pdf. Abu Dhabi: IRENA. https:/ /www.irena.org/-/media/Files/IRENA/ Pew Charitable Trusts. 2014. A Methodology for Capacity Needs Agency/Publication/2018/Oct/IRENA_mini-grid_policies_2018. Assessments towards Implementation of the Port State Measures pdf. Agreement. Washington, DC: Pew Charitable Trusts. https://www. IRENA. 2019. “Off-Grid Renewable Energy Solutions to Expand Electric- pewtrusts.org/-/media/legacy/uploadedfiles/peg/publications/ ity Access: An Opportunity Not to Be Missed.” IRENA, Abu Dhabi. report/cnareport2014updatev3pdf.pdf. https://www.irena.org/-/media/Files/IRENA/Agency/Publica- Poindexter, N. 2018. “Mini-Grids Boost Entrepreneurship; Up to 11x tion/2019/Jan/IRENA_Off-grid_RE_Access_2019.pdf. Income Rise.” Power for All blog. https:/ /www.powerforall.org/ insights/dre-technologies/mini-grids-boost-women-entrepreneur- ship-11x-income-rise. MINI GRIDS FOR HALF A BILLION PEOPLE    221 Power for All. 2017. “Stage Set for Women to Lead a Solar-Powered USAID. 2018d. “Key Steps in Mini-Grid Technical Design.” https://www. Gender Revolution in Rural India.” Medium. https://medium.com/ usaid.gov/energy/mini-grids/technical-design/key-steps/. energy-access-india/solar-powered-gender-revolution-taking- USAID. 2018e. “What Is a Community Needs Assessment?” https:// shape-in-india-645f155e096f. www.usaid.gov/energy/mini-grids/community/needs-assess- Quicksand. 2018. “Building a Digital Platform for Community Engage- ment. ment: Project Brief.” http://quicksand.co.in/work/building-a-digital- USAID. 2018f. “With What Technical Standards Should Mini-Grids platform-for-community-engagement. Comply?” https:/ /www.usaid.gov/energy/mini-grids/technical-de- RENAC (Renewables Academy Online). N.d. “Certified PV Professional: sign/standards. Online Training on Photovoltaic and PV-Diesel Hybrid Systems.” Warren, C. 2018. “Success Story: Creating a Professional Renew- RENAC, Berlin, Germany: Renewables Academy. https:/ /www. able Energy Industry in Nigeria through Training.” Winrock, renac.de/fileadmin/renac/media/pix/Trainings/Online_trainings/ Power Africa, and USAID. https://www.winrock.org/wp-content/ Brochure_Prof_PV.pdf. uploads/2018/02/Success-Story-Clean-Energy-Qualifica- Ricardo Energy & Environment. 2016. Recommendations for Accel- tions-in-Nigeria.pdf. erating Solar PV Mini Grids in India: Final Report. Didcot, United Wiemann, M., S. Rolland, and G. Glania. 2014. Hybrid Mini-Grids for Rural Kingdom: Ricardo Energy & Environment. https:/ /cdkn.org/ Electrification: Lessons Learned. Brussels: Alliance for Rural Elec- wp-content/uploads/2017/07/Up-Recommendations-for-acceler- trification. https:/ /www.ruralelec.org/sites/default/files/hybrid_ ating-solar-PV-mini-grids-in-India-1.pdf. mini-grids_for_rural_electrification_2014.pdf. SPI (Smart Power India). 2019. “Mini-Grid Model.” http://www.smart- powerindia.org/our-work/mini-grid-model/. TTA (Trama TecnoAmbiental). 2019. “Specialized Training and Educa- NOTES tion.” http://www.tta.com.es/en/specialized-training-and-educa- tion. 1. More information about the HOMER Energy training is available at UNDP (United Nations Development Programme). 2018a. “Derisk- https://www.homerenergy.com/services/training.html. ing Renewable Energy Investment: Mini-Grid LCOE Tool.” Mic- 2. More information is available on the Odyssey Energy Solutions at rosoft Excel spreadsheet. https://www.undp.org/publications/ https://www.odysseyenergysolutions.com/. drei-grid-electrification-2018. 3. More information about Powerhive is available at http://www.power- UNDP. 2018b. “Derisking Renewable Energy Investment (DREI): Der- hive.com/our-technology/. isking Matrix.” UNDP, Geneva. https:/ /www.undp.org/sites/g/ 4. Information about the mini grid tender as part of the National Elec- files/zskgke326/files/publications/DREI%20Minigrid%20Der- trification Project is available at http://rea.gov.ng/minigrids/. isking%20Table%20(Version%203.2,%20Jun%202018)%20 (FINAL).pdf. 5. More information about Husk Power is available at http://www.husk- powersystems.com. USAID (United States Agency for International Development). 2018a. “Biomass Gasification in India.” https:/ /www.usaid.gov/energy/ 6. More information about SELCO’s incubator is available at https:// mini-grids/case-studies/india-biomass/. www.selcofoundation.org/incubation-about/. USAID. 2018b. “How Should Governments Integrate Mini-Grids into 7. More information about TTA is available at http://tta.com.es/. National Electrification Planning?” https://www.usaid.gov/energy/ 8. NAPTIN’s 2016 training schedule includes a 15-day course focused mini-grids/policy/national-planning. on renewable energy (operation and maintenance of solar, wind, USAID. 2018c. “What Are the Capacity-Building Needs for Each Own- and hybrid systems). ership Model?” https:/ /www.usaid.gov/energy/mini-grids/owner- 9. More information about RES4Africa’s Micro-Grid Academy is at ship/capacity-building/. https://www.res4africa.org/micro-grid-academy/. 10. More information about Peru’s CEDECAP is available at https:// www.cedecapltda.cl/aula/. 222   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 8 DELIVERY MODELS AND SUPPORTING INSTITUTIONS CHAPTER OVERVIEW This chapter has two principal parts. The first describes models commonly used to deploy mini grids in low-in- come countries. In most cases, mini grids have been developed with support from and regulation by government institutions. Other entities have also provided support. The second part of the chapter follows from the first. It answers the question: What are the desirable features of an institutional framework to support mini grids? MINI GRID DELIVERY MODELS PUBLIC-PRIVATE MODELS Management and split-asset are two common public- The chief differences among the numerous delivery models private models. for mini grids lie in who finances, builds, and operates them (Tenenbaum, Greacen, and Vaghela 2018). The models dis- Under the management model, a government entity plans, cussed in this section are the build-own-operate model; the finances, and implements a mini grid up to the commission- public-private partnership model; the concession model; ing stage, with a private operator subsequently assuming utility models with and without private-sector involvement; responsibility. The operator first has to manage, maintain, and the cooperative model.1 Each is discussed in turn. and operate the entire mini grid, including generation and distribution. Second, it collects revenues from the mini THE BUILD-OWN-OPERATE MODEL grid’s customers. Contractual options can oblige the oper- In the build-own-operate model, a private mini grid devel- ator to assume responsibility from the governmental entity, oper carries out all the steps—designing, financing, and options that range from authorization arrangement to con- operating the mini grid. The developer might subcontract tracted operation, leasing contract, and, finally, transfer to some activities to third parties but retains control over full ownership (EUEI PDF 2014). A recent example is the and responsibility for the entire process. In many cases, Rural Renewable Energy Project implemented in Sierra the government or other entities provide subsidies to the Leone, with the assistance of the United Nations Office for developer, to customers, or both. In instances where the Project Services.2 model is implemented by a public institution, that insti- Under the split-asset model, a governmental institution tution is also responsible for program management and procures and owns the distribution assets of a mini grid, implementation (box 8.1 gives an example from Tanzania). while the developer owns the generation assets. This split In this model, because developers have full responsibility reduces investment costs for the developer, thus improv- for all aspects of project development, the role of govern- ing the mini grid’s financial viability from the developer’s ment institutions, apart from grant finance, is that of facil- perspective. This model has been used by GIZ (the German itator and enforcer, ensuring the developer’s compliance Society for International Cooperation) in Nigeria. with local policies and regulations. MINI GRIDS FOR HALF A BILLION PEOPLE    223 BOX 8.1 THE BUILD-OWN-OPERATE MODEL IN TANZANIA Tanzania recently implemented a build-own-operate to 1 megawatt) and certification (for sites below scheme managed by the country’s Rural Energy Agency 100kWp) to the mini grid developers’ sites. The local (REA) and financed by the U.K. Foreign, Common- national environment management councils issue wealth & Development Office (FCDO) and the Swed- environmental permits for the mini grids deployed. ish International Development Cooperation Agency Throughout the whole scheme, the Tanzanian Minis- (SIDA). FCDO and SIDA have handled the general over- try of Energy takes a back seat, providing guidance sight of the program. REA is the implementer, selecting and advice to the REA and other institutions only as the grant awardees, managing the distribution of the needed. grant, and monitoring implementation of the program. Mini grid developers identified and selected their own Tanzania’s Energy and Water Utilities Regulatory sites; REA gave them guidance on national electrifica- Authority ensures the mini grid developer’s compli- tion plans to ensure that they would not implement ance with local technical and tariff regulations and projects in areas where the national utility (TANESCO) issues licenses (for sites of 100 kilowatts-peak (kWp) would soon extend the main grid. Both models require careful thought regarding risk pro- CONCESSION MODEL files. Developers responsible for both the construction Under a concession model, the government issues a con- and operations and maintenance (O&M) of the assets cession to a developer providing particular contractual tend to take on more risk over the lifetime of the proj- rights to beneficial terms, such as monopoly status, prefer- ect. Government entities constructing the mini grid, and ential market access for a limited amount of time, or specif- then transfering some or all of it to an operator—or as ically designed tariffs that may differ from those set for the an alternative entering into a management contract with rest of the country (EUEI PDF 2014). Concession models the developer/operator—introduces risks. In response, have not attracted much private investment so far. This lack certain private-sector developers and operators may of investment, however, is not likely a function of the model seek to mitigate these risks because they did not oversee but instead more about risk allocation and mitigation. construction of the infrastructure. Without mitigation, operators might incur higher O&M costs resulting from Usually, the rural electrification agency plays a central role government-built infrastructure. in initiating mini grids under concession models, whereas the concession contract is usually signed by the conces- Moreover, recent experience with the IFC’s Scaling Mini sionaire and the ministry of energy, representing the gov- Grids program has identified a dearth of investment-ready ernment. contractual frameworks arising from public-private part- nerships (PPP). In fact, contracts between governments UTILITY MODELS WITH OR WITHOUT PRIVATE- and developers tend to offer scant protection for investors, SECTOR INVOLVEMENT often blocking large-scale investment: they are, in truth, In a utility model, the national utility usually shoulders not bankable for long-term infrastructure capital provid- all aspects of the mini grid, although the utility may ers. Moreover, government entities charged with managing tender out the engineering, delivery, installation, and the PPP can lack the capacity to develop and deploy the commissioning of all mini grid assets. In some cases, a pri- contracts. As a result, technical assistance is often needed vate-sector company must operate the mini grid for some to develop investment-ready contracts and to build robust time before handing it over to the utility. In other cases, capacity within the government entities that manage the the utility takes over operation right after commissioning. contracts. Kenya is using this model in developing 137 mini grids In general, the public-private model can be viewed as an under the World Bank-supported Off-grid Solar Access effective way to assign responsibilities and, in doing so, to Project for Underserved Counties program. Up to $80 boost capacities, especially since it allows financially con- million is allocated for private sector developers to bid on strained mini grid operators entry into the market. Engineering-Procurement-Construction contracts with 224   MINI GRIDS FOR HALF A BILLION PEOPLE the Kenya Power and Lighting Company (KPLC) and the and installation of the mini grid are often contracted out Rural Electrification and Renewable Energy Corporation to a third party, either from the community itself or via (REREC). The 137 sites are divided over thirteen lots in 14 a nongovernmental organization (NGO) or development low-income counties. Bidders are asked to install the solar partner. Through ongoing training, the cooperative is then mini grids including KPLC certified prepaid/smart meters capacitated to assume O&M for the mini grid (see also as well as provide for a social infrastructure project as chapter 7). selected by the county. Cooperative models have been implemented success- In Nigeria, some distribution companies (which are privat- fully in Burkina Faso, Indonesia, Peru, and the Philip- ized) are beginning to consider mini grids as a way to serve pines, among other countries. This model requires larger their rural areas. We know of one distribution company cooperatives that operate like a professional company in saying it has a potential portfolio of 340 sites with histor- that they possess adequate management and technical ical data on consumption patterns. This data could prove capacity. In the Philippines, some cooperatives resemble valuable in reducing demand risk. The exact construction professional public utilities in other industrial countries; and operations model is not yet defined in this case, but they run mini grids in parallel to their main grid business. this is an example of how the utility model in general can At this size, however, it looks rather more like a utility encompass private-sector distribution companies looking model than a typical cooperative model as described here. to engage private-sector mini grid developers. Cooperatives can operate under various institutional Government control of mini grid utility models is usually frameworks. They can be strongly government induced exercised by the energy ministry directly, often through its and influenced, as in Ethiopia, or independent, as in the direct control over the national utility. Therefore, in most Philippines. In the first scenario, the ministry of energy (or cases, no additional institutions are involved in overseeing a separate ministry for cooperatives) takes the lead from mini grids under this model. the government side. In the second scenario, the coop- eratives are subject to little supervision from the local COOPERATIVE MODEL government or, if several cooperatives are in one country, Under this model, one or more local communities finance they report to a dedicated authority. Cooperative-led mini and own mini grids. They procure most of the required grids can be fully funded by grants or can rely in whole or capital through grants. The motivation of the cooperative part on the cooperative’s own equity or debt members is to get access to electricity for themselves. In some cases, only members of the cooperative are con- COMPARISON OF VARIOUS MODELS nected to the mini grid. In other cases, the cooperative Figure 8.1 illustrates the different roles for developers and connects customers who are not members. Procurement utilities under the various models discussed above. To help decision-makers weigh the delivery model options when designing a mini grid program, we present in table 8.1 There are various delivery models for mini a comparative analysis. Please note that in many countries grids, where delivery model denotes who different models can be implemented in parallel. finances, builds, owns, and operates the mini grid. The five most common delivery models for mini grids are: (1) build-own-operate, where a private Each of the five main delivery models for sector entity raises money to carry out all of the mini grids has its own benefits and limita- activities; (2) public-private partnership model— tions. Important considerations for each model are including models where (a) the government sets up profitability for the private sector, government and the mini grid and hands it over to a private-sector local capacity, scalability, transaction costs, and operator, and (b) a split-asset model where the gov- availability of subsidies, among others. ernment and private sector co-own the mini grid; (3) a concession model, where the private sector enters into a contractual relationship with the government to deliver electricity in a specified area; (4) a utility model, where all or part of the mini grid ownership and operations is controlled by the national utility; and (5) a cooperative model, where local coopera- tives build, own, and operate mini grids. MINI GRIDS FOR HALF A BILLION PEOPLE    225 FIGURE 8.1 • Roles for developer and utility under different delivery models Build-Own-Operate and Concession Models Split Asset PPP Management PPP Utility Model Cooperative Model Customers Customers Customers Customers Customers Tariff Tariff Tariff O&M contract Tariff EPC contract Tariff EPC contract Developer Developer Utility Developer Utility Developer Utility Developer Cooperative Own Own Own Own Own Own Own Own Own Own Generation Distribution Generation Distribution Generation Distribution Generation Distribution Generation Distribution Developer finances and Developer finances and Developer operates mini Utility finances and Cooperative finances owns generation and owns generation and grid under a contract owns generation and and owns generation and distribution, and bills and bills customers; with the utility and distribution, and bills distribution, and bills customers directly. utility finances and owns collects tariffs on behalf customers directly. customers directly. Under concessions, distribution. of utility; utility finances Developer can build Developer can build ownership is for and owns generation and generation and/or generation and/or concession term only. distribution. distribution as an EPC. distribution as an EPC. Sources: IFC and ESMAP analysis. EPC = engineering, procurement, and construction company. TABLE 8.1 • Comparative analysis of mini grid delivery models Public Key funding Absolutely implementation requirement Benefits Limitations must-haves Deal breakers risks Build-own-operate model Low: CAPEX Developers having full Target beneficiaries Affordable financing Requirement to Lack of private-sector subsidies only control of the value chain may be suspicious of Results-based grants sell electricity at developer interest means they optimize cost private firms below-cost tariff level Lack of private-sector and quality for each step Ability for developers to without adequate Private firms will build portfolios instead of investor confidence Projects likely to be abandon the mini subsidy one-off projects Regulatory or policy financially viable with grid market if Heavy-handed, case- changes that make subsidies conditions change to Workable options for by-case regulation grid arrival private-sector Low transaction costs make it unprofitable operations unviable overall (for example, new, unworkable regulations) Highly scalable Contractual or split-asset PPP High: pays for Where governments Both PPP models: Contractual PPP: respect Contractual PPP: Procurement can be a all or most have significant financial high transaction costs, for legal contracts no government bottleneck to progress of mini grid and technical resources considering various Split-asset PPP: experience building If government builds CAPEX and to build mini grids, transactions between compensation mechanism mini grids some or all of the mini OPEX. these programs can be public and private for expropriation of Split-asset grid infrastructure, implemented quickly partners developer’s mini grid PPP: History of external quality control Responsibilities are Split-asset PPP: can assets uncompensated by an independent third effectively distributed lead to stop-and-start Both PPP models: expropriation of party may be necessary between financiers/ implementation issues procurement of private-sector assets investors and operators, if government-built private-sector developer by government with capacities aligned infrastructure and is competitive and with the tasks assumed private-sector–built transparent, and clearly infrastructure are not states requirements for developed in sync long-term O&M. continued 226   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 8.1, continued Public Key funding Absolutely implementation requirement Benefits Limitations must-haves Deal breakers risks Concession model Low: CAPEX Mini grid developer is May be difficult to Respect for legal contracts Main grid has Review and approval subsidies only assured that the main grid implement where the Clear provisions for taken over of concessionaire will not come into the area local community does key regulatory issues: concession areas activities can be a Developers have a clear not want to be served technical and service in the past without bottleneck to progress set of responsibilities. for a long time by mini standards, tariffs, and compensating Change in main grid grids, or in countries exclusivity developers expansion plans Can facilitate a portfolio with low levels of trust approach to project Requirement to New policies or in contract. sell electricity at development regulations conflict with below-cost tariff level existing concession without adequate agreements subsidy Different legal and regulatory requirements between concessions Utility model (with private sector) High: pays for Keeps overall responsibility Transaction costs may Source of long-term Financially insolvent Procurement of private all or most for electrification with one be high if contracting subsidy (or cross-subsidy) utility sector services can of mini grid entity. with different third if mini grid tariffs are not Noncompetitive or be a bottleneck to CAPEX and Brings in technology parties for various fully cost-reflective opaque procurement progress OPEX. innovation and process stages of mini grid of private-sector Insufficient long-term innovation from the private development. contractors allocation of staff and sector. Building and managing financial resources at Highly scalable if utility mini grids can the utility has high financial and exacerbate utility losses technical capacity. if mini grid tariffs are not fully cost-reflective. Utility model (without private sector) High: pays for Keeps responsibility for Building and managing Utility experience Financially insolvent Insufficient long-term all or most electrification with a single mini grids can operating isolated rural utility allocation of staff and of mini grid entity exacerbate utility losses mini grids financial resources at CAPEX and Utility can get economies if mini grid tariffs are Source of long-term the utility OPEX. of scale through bulk not fully cost-reflective subsidy (or cross-subsidy) purchases if mini grid tariffs are not Highly scalable if utility fully cost-reflective has high financial and technical capacity Cooperative model Low: CAPEX Local involvement leads Limited local capacities Cooperatives as a Tariffs are not fully Insufficient capabilities subsidies only to solutions that consider mean that capacity community organizational cost-reflective, and financial resources most of the locally relevant building and potentially structure have a strong including costs of local cooperatives to aspects and increases third-party oversight track record, even if not related to successfully operate sense of ownership of the need to be put in place yet in the mini grid sector depreciation, economically viable local community to ensure ongoing O&M replacement, repairs, mini grid businesses Local structures need and maintenance Progress generally to be created to limit tends to be slow unless the potential for social there are preexisting conflict as related to the well-managed mini grid cooperatives Not easily scalable or fast in deployment Source: ESMAP analysis. CAPEX = capital expenditure; O&M = operations and maintenance; OPEX = operational expenditure; PPP = private–public partnership. MINI GRIDS FOR HALF A BILLION PEOPLE    227 INSTITUTIONAL FRAMEWORK national technical and environmental standards. Government agencies may wish to assist smaller The term institutional framework broadly refers to the mini grid developers with compliance. various entities that, through their respective roles and • Confirming that mini grids are suitably sited (for responsibilities, shape the social, economic, and political example, by earmarking specific areas for mini grid environment within which a particular sector operates. development in the rural electrification agenda). THE DESIRABLE FEATURES OF A FRAMEWORK TO • Facilitating the flow of financing (equity, debt, and SUPPORT MINI GRIDS grants) to private-sector mini grids; public-sector Most countries depend on the main grid to increase elec- mini grids focus more narrowly on grants, as other tricity access. Thus, most countries have developed an funds would come from the national budget. institutional framework to support the main grid. In some • Making sure that enough skilled workers are on hand countries, an institutional framework for solar home for mini grid deployment. systems also exists; but these frameworks have not yet • Establishing and regulating tariff regimes as they emerged for mini grids. In most low-income countries, apply to mini grids. mini grids are still early in development stages and lack a comprehensive, supportive institutional framework. 3. The framework should be tailored to the country’s Below are five desirable attributes of institutional frame- electrification planning strategies, current and works that can support large-scale and rapid deployment future, which should include geospatial planning. of mini grids. We have discussed several delivery models above, and other relevant models might be available. In many Governments need to recognize mini grids as a desir- 1.  countries, achieving universal access to electricity able and viable electrification option. Any institutional will require both private-sector and government-led framework must be reinforced by national political and approaches to mini grid development. For this reason, social recognition that mini grids are indeed a viable the institutional framework should be able to support and desirable option to scale up electricity access. In different models—without implementing a multitude of practical terms, this means the role of mini grids should models that would create a fragmented mini grid mar- be recognized and delineated in high-level policy docu- ket. For example, in some countries where the national ments, such as national plans for electricity access. utility already operates mini grids, it may make sense For example, the Economic Community of West African to ensure that the institutional framework can accom- States’ 2015–20 Programme on Access to Sustainable modate a development approach led in some areas by Electricity Services notes the importance of government the national utility, and, in others, by an approach fully promotion of mini grids all the way to the level of rural led by the private sector. communities. Government promotion at the village level An important related element of the institutional frame- creates willingness to accept and embrace the arrival of work and its ability to accommodate the country’s elec- mini grids when they are being developed in their area trification planning strategy is the electrification plan (ECREEE 2015). itself. Geospatial analysis is a powerful tool for govern- 2. Government institutions must support mini grid devel- ments to use when developing their electrification plans opment through transparent actions and decisions. because it leads to better decision-making around both They should assign existing or newly created agencies the mini grid delivery model options and the institutional and make them responsible for achieving national elec- arrangement supporting these delivery models. Geo- trification targets through mini grids. Institutions should spatial planning is discussed in chapter 2. have explicit, transparent, and harmonized roles and The framework should be stable and should mini- 4. responsibilities, allowing mini grid developers to know mize duplication of oversight and conflicting rules. In and understand which actors are involved in which some countries with frameworks that already support capacity and at which point. In addition, it is particularly mini grid development, developers are wrangling with important to ensure transparency in the decision-mak- altered institutional frameworks—finding, for example, ing processes for institutions with direct jurisdiction over a newly created government entity with new jurisdiction mini grids. Institutions supporting mini grid development over mini grids. Instability inhibits private sector invest- have broad roles in the following areas: ment in mini grids. To the extent possible, governments • Ensuring that mini grid projects are in compliance should refrain from altering the institutional framework with the applicable rules and regulations, including other than to streamline or simplify. Additional com- 228   MINI GRIDS FOR HALF A BILLION PEOPLE plexity is not welcome. In addition, in many countries, Figure 8.2 presents an example of the institutional frame- developers struggle to navigate complex ecosystems work that affects mini grid developers and the mini grid of national and international institutions, which include sector. development banks and development partners, rural Other stakeholders have not been considered in this over- energy agencies, environmental protection agencies, view because of their comparatively minor impact on mini energy and finance ministries, national energy regula- grid delivery models, at least compared with national and tors and utilities (figure 8.2), each having an influence international institutions. These more minor stakeholders on the developers’ ability to start up and operate their include regional and local authorities and administrations, businesses. Navigating this web can be a daunting, the network of local organizations (including NGOs and time-consuming, and costly task for mini grid develop- other potential partners, such as suppliers of biomass), ers. As we discuss in chapter 10, it is therefore critical communities, and customers. Also, the institutions dis- to avoid duplication and conflicting rules through clear frameworks, relationships, and contact points. 5. Key institutions should be supported with ongoing capacity building. Finally, we note the importance of The five main characteristics of an insti- capacity building for each of the institutional actors. tutional framework that can support mini Particularly, although not exclusively, in countries with grids, given the diversity in potential mini grid limited mini grid experience and where government delivery models, are: (1) mini grids must be recog- and other local entities make up the institutional frame- nized by the government as a desirable and viable work (figure 8.2) will need technical support and train- electrification option; (2) government institutions ing to carry out their roles in a way that is conducive must support mini grid development through their to developing a mini grid sector that can grow at scale. actions and decisions; (3) the institutional frame- For example, regulatory agencies may need support in work needs to be tailored to the country’s current developing a light-handed regulatory framework that and future electrification planning strategies, and can help attract private-sector investment, while minis- should be aided by geospatial planning; (4) the tries of energy and planning may need technical assis- framework should be stable and should minimize tance to develop and implement national electrification duplication of oversight and conflicting roles; and plans that have been informed by robust and detailed (5), key institutions should be supported with ongo- geospatial analysis. Chapter 7 discusses capacity build- ing capacity building. ing within the mini grid institutional ecosystem. FIGURE 8.2 • Sample ecosystem of institutions affecting mini grid developers National institutions International institutions Rural Jointly develop Ministry of national programs Development Development electri cation energy bank partner agency Set locally relevant policies and Donors implement programs Ministry of Determines budgets nance and avails grants Institute research programs Mini grid Developer Determines international playing eld Regulator Collarborates with Determine Government local National institution in Environmental regulations developer utility agency base country Source: ESMAP analysis. MINI GRIDS FOR HALF A BILLION PEOPLE    229 played in figure 8.2 are not necessarily present in every country where mini grids are deployed—or are involved in The institutional framework for mini grids the deployment of mini grids at all. In some countries the contains various entities, including the regulator, for example, has no impact on the mini grid indus- ministries of energy, labor, finance, and the envi- try. One of these is Bangladesh, where regulation applies to ronment; the national utility; the energy regulator; installations with 6MW or greater of installed capacity. the rural electrification (or energy) agency; the At the moment, there is no standardized institutional environmental agency; the bureau of standards; framework for mini grid development. Such standard- energy associations; local development banks and ization may not be desirable, considering the various financial institutions; social agencies and NGOs; factors that go into the development of institutional and international development banks and develop- frameworks—not least of which are the wider governmen- ment partners. TABLE 8.2 • Roles of national and international institutions Institution Typical existing roles Potential mini grid roles National institutions Ministry of Design rural electrification targets, Transparently communicate electrification plans to integrate energy solutions energy strategy/vision, and mission providers, including mini grid developers in the absence of a rural electrification Design and administer national energy agency policy and planning Recognize and support mini grids as an option to increase electricity access Administer public resource allocation Initiate mini grid institutional framework Energy Issue, monitor, and ensure compliance Formulate and implement economic regulations regulator with local regulations (licensing, permit Formulate and implement technical regulations (for example, standards and requirements) interconnection requirements and scenarios), sometimes in collaboration with Mediate disputes and protect consumers the bureau of standards and rural electrification agency National Construct any national mini grids Understand and appreciate the role of mini grids in increasing electricity access utility Carry out grid-extension projects Develop mini grids under public-private model Rural Implement rural electrification agenda Develop mini grid electrification approach electrification Perform any regulatory tasks delegated Take overall responsibility for development of mini grids (or energy) to it agency Environmental Assist with formulating and Issue environmental rules relevant for mini grids agency implementing national environmental Issue permits as required regulations Monitor compliance with environmental standards Ensure mini grids meet minimum environmental standards and monitor compliance with environmental regulations Ministry of Provide rural electrification budget Provide guidance on national electricity tariffs and subsidies finance Assist with and coordinate grants Provide funds for fiscal incentives, including subsidies and concessionary loans for rural electrification Others Bureau of standards, which may set applicable technical standards Energy associations, which may provide some vocal support and advisory efforts on behalf of mini grid developers Ministry of labor, which may assume the role of ensuring local mini grid skills are developed Local development banks and financial institutions, which may finance mini grid projects Social agencies and NGOs, which may be particularly important for gender issues, community engagement, and promoting productive uses of electricity Other ministries, including the ministry of agriculture, health, and education, which may seek to increase mini grid implementation for the betterment of their respective objectives 230   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 8.2, continued Institution Typical existing roles Potential mini grid roles International institutions Development Design and implement programs in Possibly assist the government in designing and creating the financial support banks and coordination with national institutions, for mini grids development ensuring alignment of the programs with Address financing of national planning initiatives in the context of rural partners broader development targets electrification Provide finance for development Provide funds to support mini grid development programs Possibly design programs that they cofinance through a government agency Assist the government in building capacity for all relevant agencies Source: ESMAP analysis. tal landscape, the local cultural context and environment, Nigeria’s operational counterpart in the World Bank–sup- the legal framework, and socioeconomic characteristics ported National Electrification Program. (Deshmukh, Carvallo, and Gambhir 2013). Nevertheless, Other countries that do not have a rural electrification or certain roles are typical for most of the institutions, as pre- energy agency rely on different lead agencies. In Myanmar, sented in table 8.2. the lead agency is the Department of Rural Development, INSTITUTIONS INSIDE THE ENERGY SECTOR located within the Ministry of Agriculture, Livestock, and Irrigation. In Bangladesh, while a Rural Electrification Board Lead agency actually exists, it is only responsible for the main grid. The An energy ministry or a rural electrification agency is often lead agency for mini grids is the Infrastructure Develop- the lead for mini grids. Its responsibilities may vary accord- ment Company Limited, a government-owned company. ing to local circumstances, but under any scenario, the lead agency must have an in-depth understanding of all In Kenya, under a World Bank cofinanced project, mini grids aspects relating to mini grid development (technical, finan- will be developed jointly by the Rural Electrification Author- cial, social, gender, and community-based), in addition to ity and KPLC. pertinent regulations, procurement and labor rules, and dealings with development partners and other interested Energy regulator stakeholders. Many countries have an energy regulatory agency that oversees the main grid. If such an agency exists, it would If the lead agency for mini grids is a ministry, an internal have responsibility for mini grids also, as in Nigeria and unit is typically assigned: mini grids are not ready fits for Tanzania. The role of the mini grid regulator is discussed in a ministry’s broad responsibilities. This ministerial unit detail in chapter 9. should have adequate staff and capacity and flag relevant emerging issues for high government officials. If a lead Power utilities agency other than the ministry is appointed to handle mini Power utilities are often government owned in developing grids, it should be highly capable, managing the various countries. Overall, they need to understand the role and activities described above and be respected as the highest relevance of mini grids in increasing electricity access and authority on mini grids, even compared with the ministry of not hinder the development of mini grids. In addition, if the energy. The lead agency must enjoy direct communication national power utility takes ownership of mini grid distri- vis-à-vis other national institutions (such as the ministry bution facilities when the main grid reaches a previously of finance and the legislature). Where capacity is lacking, isolated mini grid, the distribution systems of the mini donors or development partners will need to invest to build grid will need to be built to grid standards for a successful country capacity in governance. integration (Kidenda 2018). Further, power utilities need to One example of a lead agency is Nigeria’s Rural Energy be aware of and understand the regulatory and financial Agency (REA). Established in 2005, the REA’s main func- issues that will affect their operations when the main grid tion is to facilitate the provision of modern energy services reaches the site of a mini grid. in rural areas by providing grants, subsidies, technical assistance, training, and capacity building to rural energy Energy associations project developers. Further, the REA works with the govern- Some countries may have an association of renewable ment as well as with development partners, the private sec- energy developers, or of solar dealers. The role and rele- tor, NGOs, and community-based organizations. The REA is vance of these associations should be considered in each MINI GRIDS FOR HALF A BILLION PEOPLE    231 country, with possibly some support provided to develop Permitting institutions their ability to facilitate mini grids. In Africa, the Africa These institutions handle a broad set of issues that arise Minigrid Developers Association (AMDA) promotes the over the use of land, water, and other natural resources. development of alternating current AC-powered mini grids One example would be the disposal of batteries by mini throughout Africa. AMDA has encouraged the development grids with solar backup. It is often the case that permitting of uniform mini grid policies and regulations across African agencies are not familiar with mini grids. As a consequence, countries so private developers do not have different reg- their procedures could become a major bottleneck for mini ulations in different countries, which is the current reality. grid developers. Hence, it is important for the lead govern- The Tanzania Renewable Energy Association and the Kenya ment agency to work closely with all the relevant permitting Renewable Energy Association also promote mini grids to institutions. the government on behalf of their members, and orga- nize training workshops, building further capacity for their Labor ministry members. Two issues are relevant to the labor ministry. First, many countries lack the skills needed for a rapid scale-up of mini INSTITUTIONS OUTSIDE THE ENERGY SECTOR grids. Hence, it is important to ensure that local skills are Environmental agencies or ministries developed whenever a new technology like mini grids is Mini grid projects often require environmental approval introduced. The labor ministry may be the right agency to from the relevant environmental agency or ministry. Carry- take on this responsibility. But it is more realistic to expect ing out the studies necessary to be granted approval can be that the relevant training will have to be performed by the a major cost to mini grid developers, delaying the mini grid mini grid developer (for example, Husk Power and Mera Gao development timeline. In Tanzania, for example, mini grid in India), because the developer will have direct and detailed developers were facing delays with obtaining the needed knowledge of the tasks and skills that will be needed. environmental clearances and were also having to cover Second, when mini grids operate at a small scale, it is the costs of an environmental consultant. In some cases, uncommon for labor issues to arise. However, when a sin- this cost approached a hefty $10,000, significant for some gle developer expands to operate several mini grids with developers, particularly local ones. From a timing perspec- a growing number of employees, employment issues are tive, some developers had to wait more than 10 months likely to arise. One frequently encountered issue is whether just to obtain the needed environmental clearance, creat- the mini grid operator’s staff will be employed as indepen- ing further delays and uncertainty in the mini grid project dent contractors or as employees. development cycle. Chapter 10 dedicates an entire section to how to reduce the red tape related to environmental Financial institutions clearances. These institutions are relevant for private and communi- ty-owned mini grids only. Some countries have a develop- Bureau of standards ment bank (such as Bangladesh and Sri Lanka), which may This agency would set many of the technical standards play a major role in financing mini grids, or even establish- for mini grids. In some cases, it may not have the capacity ing technical standards for equipment that the bank will to set some of the standards, particularly for mini grids finance. The bank, in turn, will require capacity building. whose distribution systems are not compatible with the Other relevant financial institutions are banks and nonbank main grid. In such cases, the bureau would have to work financial institutions comprising equity and debt investors. closely with the lead agency and the regulator to develop Some of these institutions may not have the capacity to a suitable, comprehensive set of standards that mini grids be involved with mini grids, and their capacities may have must follow. to be built over time. In particular, there may be a need to build the capacity of various agencies to understand how Finance ministry the financial risk of mini grids could be shared with inter- One important aspect of this ministry’s role would an ested parties. understanding of, and support for, any set of subsidies that are needed. Further, this ministry may also need to under- Social agencies and NGOs stand and approve of any financial interventions in the These agencies would also include community groups. debt and equity markets. Finally, it would be involved in any They can be instrumental in the acceptance and promotion taxes, import duties, and foreign exchange issues. Hence, it of mini grids in the target areas. They may be particularly is important that the ministry have the capacity needed to involved in gender-related issues, community engagement, support and facilitate mini grids. and the promotion of productive uses of electricity by entrepreneurs and small and mid-sized enterprises. 232   MINI GRIDS FOR HALF A BILLION PEOPLE INVESTORS’ PERSPECTIVE ON tional framework would go a long way toward providing investors with the confidence they would need to invest INSTITUTIONAL FRAMEWORKS in the mini grid sector. The presence of an indepen- Through conversations with private-sector investors and dent and competent energy sector regulator is another building on the International Finance Corporation’s exten- factor that can attract investors—the key words being sive experience investing in infrastructure projects around independent and competent. This would require that the world, we have identified three main barriers that the regulator be (1) protected from short-term political investors perceive in institutional frameworks that would interference, (2) appointed on the basis of professional severely constrain investment in mini grids. competence and integrity, (3) endowed with sufficient funding to hire personnel of high professional skills while • Low government capacity: Low institutional capac- enjoying budgetary independence from the government ity to run effective procurement, approve permits and budget, (4) empowered with a minimum set of compe- licenses in a transparent manner, or undertake any tencies, and (5) legally prevented from having personal other bureaucratic processes critical for mini grid interests in the regulated industry. development is a red flag for mini grid investors. Sim- • A national electrification plan that emphasizes mini ilarly, unstable government institutions with frequent grids: A national electrification plan that emphasizes turnovers within the different government agencies is mini grid deployment as a pillar of the government’s another signal that government capacity to support strategy to increase access to electricity sends one mini grids is inadequate. of the strongest signals possible to investors that the • Limited or nonexistent electrification planning: The government is committed to supporting mini grids. lack of an agreed and signed-off national electrification Updating this plan at predefined intervals (for example, plan is a major red flag for investors because it signals every three or five years), using geospatial analysis, and that the government has not identified a clear institu- ensuring that it is based on realistic cost assumptions tional framework within which mini grids would operate. and accurate socioeconomic data are ways to make the In addition, limited electrification planning capabilities, national electrification plan even stronger for attract- or national electrification plans based on unrealistic ing private-sector investment. (See chapter 2 for an assumptions about main grid expansion, impose major in-depth discussion on geospatial planning.) barriers to private-sector investment to the mini grid sector. • Unclear or overly complex institutional relationships: ROLE FOR DEVELOPMENT PARTNERS Complicated interfaces and allocations of responsi- bility for oversight of the mini grid sector at specific Development partners play key roles in institutional frame- government agencies and among agencies sends a works conducive to scaling up national mini grid sectors. strong signal to investors that the mini grid sector is The following are recommendations for development part- fraught with political risk. Similarly, different govern- ner organizations that have been identified in consultation mental entities having conflicting political motives with with mini grid sector stakeholders together with ESMAP’s respect to their interactions with mini grid developers experience with World Bank operations teams imple- and investors is another red flag that would dissuade menting mini grid projects over the past several years. To investment in the mini grid sector. Finally, institutional establish and support strong institutional frameworks, frameworks that prevent private-sector developers development partners can act as or assist with: from participating in the mini grid sector, or restrict • Advising on best practices on creating scale and eco- their participation to limited roles such as engineer- nomic sustainability—rapidly—in rural electrification ing, procurement, and construction contractors, would eliminate or severely reduce developers’ ability to • Fashioning institutional frameworks to support mini attract investment, respectively. grids, whether funded privately or publicly, in ways that consider a country’s resources, constraints, and prac- Balancing these three barriers are two “must haves” in an tices institutional framework for mini grids from an investor’s perspective. • Ensuring quick scale-up through capacity building among all the relevant agencies so mini grids can scale • Clear roles for the appropriate government agencies: up quickly The existence of a competent governmental body with • Designing the delivery model for mini grid scale-up full authority to implement a rural electrification strat- (countries may choose more than one model) egy embedded in a transparent and conducive institu- MINI GRIDS FOR HALF A BILLION PEOPLE    233 • Overseeing M&E programs to track deployment Rural-electrification-concessions-in-Africa-what-does-experi- ence-tell-us. • Securing external evaluations of institutional frame- Kidenda, John. 2018. “Mini-Grids on the Trajectory of Rural Electrifica- works and delivery models for mini grids in each tion in Africa.” Position paper, Africa Minigrid Developers Association. country. (Independent, outside groups of regional or https:/ /shellfoundation.org/learning/mini-grids-on-the-trajecto- ry-of-rural-electrification-in-africa-an-amda-position-paper/. international experts must do these assessments Tenenbaum, B., C. Greacen, and D. Vaghela. 2018. Mini Grids and the • Navigating the institutional framework to help them Arrival of the Main Grid: Lessons from Cambodia, Sri Lanka, and Indonesia. ESMAP Technical Report 013/18. Washington, DC: design bankable projects3 World Bank. https:/ /openknowledge.worldbank.org/bitstream/ • Cofinancing the above. handle/10986/29018/134326.pdf?sequence=6&isAllowed=y. World Bank. 2017. International Development Association Project Appraisal Document on a Proposed Credit in the Amount of EUR 133.8 Million (US$150 Million Equivalent) to the Republic of Kenya REFERENCES for an Off Grid Solar Access Project for Underserved Counties. Report No. PAD2008. Washington, DC: World Bank. Deshmukh, Ranjit, Juan Pablo Carvallo, and Ashwin Gambhir. 2013. Sustainable Development of Renewable Energy Mini-Grids for Energy Access: A Framework for Policy Design. Berkeley, CA: Law- rence Berkeley National Laboratory. https:/ /eta.lbl.gov/publica- NOTES tions/sustainable-development-renewable. ECREEE (ECOWAS Centre for Renewable Energy and Energy Effi- 1. Other models may also be useful for mini grids, and new models ciency). 2015. “ECOWAS Programme on Access to Sustainable may emerge over time. Electricity Services: 2015–2020.” Economic Community of West 2. Information about the UNOPS project is available at https://www. African States (ECOWAS). http://www.ecreee.org/project/epases- unops.org/news-and-stories/stories/access-to-energy-giving-sier- ecowas-programme-access-sustainable-electricity-services. ra-leone-the-power-to-change. EUEI PDF (European Union Energy Initiative Partnership Dialogue Facility). 2014. “Mini-Grid Policy Toolkit: Policy and Business 3. As part of the Scaling Up Renewable Energy Program, the govern- Frameworks for Successful Mini-Grid Roll-Outs.” http:/ /www. ment of Tanzania, together with multilateral development banks, minigridpolicytoolkit.euei-pdf.org/system/files_force/RECP_ developed a country investment plan with a component just for mini Minigrid_Policy_Toolkit_doublepage%20(pdf%2C%2012.5MB% grids. The implementation of this component was led by the Inter- 2C%20EN)_web60ab.pdf?download=1. national Finance Corporation and included a Transaction Advisory Hosier, D., M. Bazilian, T. Lemondzhava, K. Malik, M. Motohashi, and D. Services Facility (TASF). This TASF supported mini grid developers Vilar de Ferrenbach. 2017. Rural Electrification Concessions in Africa: to strengthen mini grid operating models and to increase their com- What Does Experience Tell Us? Washington, DC: World Bank. http:/ / mercial viability and, ultimately, bankability. documents.shihang.org/curated/zh/347141498584160513/ 234   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 9 ENACTING REGULATIONS AND POLICIES THAT EMPOWER MINI GRID COMPANIES AND CUSTOMERS CHAPTER OVERVIEW Key regulatory decisions for mini grids have a significant impact on whether and how the sector develops. The five key decision areas are market entry, retail tariffs, service standards, technical standards, and the relationship with the main grid. This chapter identifies how different countries have regulated—or not regulated— these areas, presents decision trees that can help regulators and policy makers think through these decisions, provides some “regulatory packages” that combine smart regulatory decisions given different starting conditions, and discusses two innovative regulatory mechanisms that can further incentivize private-sector investment in mini grids. A smart approach to regulating mini grids can enable This chapter also draws on a review of policy and regu- them to emerge in countries where they are a viable but latory documents in these countries; on the literature on untapped solution for electrification and to scale up in mini grid regulations; and on research in other countries, countries where they already exist. Regulators can benefit including Indonesia, Madagascar, and Sri Lanka. The anal- from understanding how other countries have handled reg- ysis focuses on privately owned and operated mini grids, ulatory decisions and determining whether similar action because they have more potential than community- or can work in their settings. utility-owned mini grids to attract the finance needed for rapid expansion of access to electricity and to operate To provide them with that information, this chapter sustainably. draws on a research project conducted by the World Bank’s Energy Sector Management Assistance Program The chapter provides practical guidance to regulators, (ESMAP), Castalia, and Ecoligo. The project included 11 rural electrification agencies, and project teams devel- field visits and 70 interviews, conducted between August oping mini grid programs. Because no single regulatory 2017 and September 2017, and produced a 300-page solution is optimal in all settings (and regulation has report; six country-specific case studies (on Bangladesh, costs as well as benefits), the chapter identifies multiple Cambodia, Uttar Pradesh [India], Kenya, Nigeria, and Tan- options and discusses the conditions under which they zania), which can be found on the companion website to are suitable. The decision trees provided are not exhaus- this book: www.esmap.org/mini_grids_for_half_a_bil- tive or prescriptive; they indicate the choices that might lion_people; and two articles on LiveWire, the World make sense in certain common scenarios. While the focus Bank’s series of online knowledge notes on development of this chapter is on regulatory decisions affecting mini issues in the energy and extractives sectors. In each coun- grids, general business and environmental regulations try, the research team interviewed regulators, developers, also touch on mini grids. These economywide regulations rural electrification agencies, and consumers, as well as and policies are discussed in chapter 10, Enabling Busi- utilities and development partners where appropriate. ness Environment. MINI GRIDS FOR HALF A BILLION PEOPLE    235 One limitation of this chapter is that by providing a “deep country-specific mini grid regulations and policies in more dive” on regulations, it does not engage much with broader than 50 countries and is a useful resource to complement conversations on all the policies that can affect mini grids. this chapter. The RISE program conducts its own analysis While policies signal the market that mini grids are an of policies and regulations to assign scores to each country accepted path to universal electrification, by themselves based on the comprehensiveness and quality of its policies they rarely provide enough clarity and guidance to private- and regulations. sector developers and investors need to risk investing in Of the 52 countries for which RISE has developed a score mini grids. Regulations, when done right, can complement for a mini grid regulations, 22 have scores of 70 or higher, policy by providing clarity and guidance on how mini grids out of 100 (figure 9.1). These countries have put in place can be deployed, thus incentivizing investment. regulatory and political frameworks that at least on paper Still, it is important to acknowledge the impact that pol- are conducive to private sector investment, and according icies can have on mini grids. Several countries have to analysis using ESMAP’s Global Electrification Platform integrated national electrification policies that explicitly (https:/ /electrifynow.energydata.info/) they collectively leverage mini grids as a part of the solution set—Ethiopia, have 224 million people who could be served at least Kenya, Rwanda are examples. These national electrifica- cost by mini grids. Another set of 15 countries with RISE tion policies send important signals to developers that scores between 50 and 70 have made strong progress to the government acknowledges the importance of mini strengthen their regulations over the past several years, but grids for achieving its electrification objectives. Other pol- still have work to do to put in place a political and regulatory icies that directly affect mini grids are related to renew- framework that is conducive to private sector investment. able energy and rural economic development, which can Collectively, these countries have 110 million people that incentivize developers to incorporate renewable energy would be best served at least cost by mini grids. Finally, a in their generation mix and seek out rural areas for mini group of 15 countries with RISE scores below 50 represent grid deployment, respectively. A third type of policy that high priority areas for improving policies and regulations affects mini grids creates key energy-sector institutions related to mini grids. These lowest-scoring countries col- and delegates responsibilities to them. Rural electrifica- lectively have 100 million people that would be best served tion agencies, such as those in Nigeria and Tanzania, and at least cost by mini grids. independent regulatory agencies, such as those in Haiti We then plotted each country’s RISE score against the and Rwanda, are often created by acts of parliament as number of planned mini grids in that country, using data part of energy-sector policy making. on planned mini grids in countries with a RISE score from the global database of mini grids that ESMAP compiled for this report (see the overview to this handbook). After THE IMPORTANCE OF WORKABLE removing outliers—following the statistical definition of REGULATIONS FOR SCALING UP outliers1—we found that the presence of a strong mini grid policy and regulatory environment correlates positively MINI GRID DEVELOPMENT with planning to build mini grids (figure 9.2). Developing a set of workable regulations can set a strong foundation upon which to scale up a mini grid market. At the same time, heavy-handed regulations, or an absence of mini grid–specific regulations that provide clarity to While improved policies and regulations to the private sector can hold back a market from its full support mini grids do not automatically lead potential. to more private-sector mini grid investment, we did find a statistical correlation between the number of While this may seem intuitive, it is also backed up by real- mini grids installed and planned in a country and world data. Using data sets that have never before been the quality of that country’s regulatory and policy compared, we found evidence for a positive correlation framework for mini grids. Implicit in this result is between the quality of a country’s mini grid–specific regu- that countries with higher RISE scores for their mini latory regime and the number of planned mini grids in that grid frameworks have a higher level of government country (figure 9.2). For data on the quality of mini grid reg- capacity to design and executive good policies and ulations, we used the World Bank’s Regulatory Indicators regulations to support mini grids. for Sustainable Energy (RISE), an online resource available at https:/ /rise.esmap.org/. This online resource tracks 236   MINI GRIDS FOR HALF A BILLION PEOPLE 2019 RISE Mini Grids Framework Score FIGURE 9.1 • Regulatory Indicators for (Score out of Sustainable Energy (RISE) scores for mini grid framework 100) Nigeria 100 Tanzania 100 Rwanda 100 Nicaragua 95 Haiti 88 Zambia 88 Zimbabwe 87 Benin 85 Philippines 83 Sierra Leone 83 Kenya 82 Bangladesh 80 India 78 Cambodia 78 Malawi 77 Nepal 75 Vanuatu 75 Countries with 73 RISE scores >70 Niger collectively have Myanmar 73 220 million people Uganda 73 that could be served Ethiopia 70 at least cost by mini grids.a Cameroon 70 Guatemala 65 Congo, Dem. Rep. 62 Solomon Islands 62 Central African Republic 62 Pakistan 60 Angola 60 Papua New Guinea 60 Lao PDR 57 Guinea 55 Honduras 55 Madagascar 52 Countries with RISE scores Ghana 52 50–70 collectively have Indonesia 52 110 million people that could be served at least Mauritania 52 cost by mini grids.a Eritrea 50 Mozambique 45 Liberia 45 Afghanistan 45 Burkina Faso 42 Burundi 42 South Africa 40 Sudan 37 Chad 35 Cong, Rep. 35 Mongolia 35 Somalia 25 Senegal 25 Togo 25 Countries with RISE scores < 50 collectively Yemen, Rep. 20 have 100 million people that could be served Mali 10 at least cost by mini grids.a 0 20 40 60 80 100 ESMAP Rise Score for Mini Grids Source: ESMAP 2022. Note: Scores are based on the comprehensiveness and quality of each country’s mini grid policies and regulations. a. Calculated from the Global Electrification Platform. MINI GRIDS FOR HALF A BILLION PEOPLE    237 FIGURE 9.2 • Correlation between mini grid policies and regulations and number of mini grids planned 3,500 3,000 Nigeria Number of Mini Grids Planned 2,500 2,000 Tanzania 1,500 Senegal Ethiopia 1,000 DRC Madagascar Myanmar 500 Mozambique Mali Burkina Faso 0 0 10 20 30 40 50 60 70 80 90 100 RISE Score for Mini Grids Framework (Size of bubble = mini grid market potential according to ESMAP's Global Electri cation Platform) Source: ESMAP analysis. Note: Countries with more than 100 planned mini grids have been identified for reference. DRC = Democratic Republic of Congo. Implicit in this result is that countries with higher RISE scores for their mini grid frameworks have a higher level Entities tasked with overseeing mini grids— of government capacity to design and execute good poli- whether a formal regulator or other entity— cies and regulations that support mini grids. We address will need to make decisions on five key areas: this element of government capacity throughout the market entry, tariffs, service standards, technical chapter. standards, and arrival of the main grid in the service area of a mini grid. FIVE KEY REGULATORY DECISIONS In this chapter, we focus on the key decisions a regulatory on how it makes those decisions (transparency is key) and agency tasked with mini grid oversight will need to make. on the independence with which it makes and implements This does not mean, however, that a formal regulator them (independence from political and private-sector needs to be in place to make these decisions. As we dis- pressure is key). cuss later in this chapter, regulatory decisions are often Mini grids often enter the market as small competitors with embodied in a legal contract between the developer and limited market power. Over time, they sometimes develop the government authority tasked with mini grid oversight. considerable market power in the local market for energy In addition, local governments, or a grant-giving entity, services. Smart regulation can maximize the chances of such as a rural electrification agency, may be appropriate early entry and growth while protecting against exploita- entities to make regulatory decisions about mini grids. tion of market power. As we discuss in chapter 10, however, it is important not to duplicate government oversight, which adds needless Each element of the regulatory framework—entry, tar- layers of bureaucracy—and hence costs and risks—to pri- iffs, service and technical standards, and what happens vate-sector investment in mini grids. when the main grid arrives—needs to be considered individually, but all need to work together as a coherent With that said, if a formal regulatory agency has been package. No element can therefore be designed without granted the authority to regulate mini grids, its effective- considering the others. ness and credibility will depend not only on the decisions it makes in the five areas outlined in this chapter but also 238   MINI GRIDS FOR HALF A BILLION PEOPLE REGULATING ENTRY cent of projects’ capital costs for solar and solar-diesel hybrid mini grids (Republic of Kenya and World Bank Entry into the electricity sector is commonly regulated, to 2016). Delays in the application procedures and the risk ensure the safety of operations, control the attributes of the that an application may not be approved can also deter companies that enter, and prevent multiple operators from developers. supplying different parts of an area that would be more effi- ciently served by one. Many jurisdictions have made it ille- Regulating entry through means other than permitting gal to supply electricity without a license or a permit. and licensing may be appropriate in certain circumstances (table 9.1). Such options include no regulation of entry and Obtaining a permit or a license can impose a significant registration. financial burden on developers, for whom administra- tive costs can represent a large share of project costs. In A comparison of entry regulations across five countries Kenya, for example, licensing costs can exceed 10 per- shows that developers generally offer less detail when TABLE 9.1 • Options for regulating entry Option Description Suitable where . . . Examples No regulation Leaves investors free to Government prioritizes rapid expansion of Cambodia (before 2001 in practice) of entry develop mini grids where electricity access through mini grids. Uttar Pradesh (most mini grids were not they want, without providing Government is willing not to regulate tariffs or registered as of 2017) any information to the standards, or it is willing and able to enforce government. Tanzania (mini grids producing less than regulation without a permit or license. 15kW or with fewer than 30 customers) Government does not need or desire to collect data on who supplies electricity. Administrative capacity is low. Mini grids are quite small—on the order of a fewkW in capacity. Registration Requires investors to inform Government prioritizes rapid expansion of Uttar Pradesh (required for all mini grids regulator that mini grid is electricity access through mini grids and does not in 2016 regulation but not enforced as in business and provide want to control who enters the market. of 2017) information on mini grid at Regulator wants to proactively supervise Nigeria (required for isolated mini regular intervals, as required compliance of mini grids with other regulation grids producing less than 100bkW of by the regulator. (tariffs, service standards). distributed power a) Does not require Government wants to collect information on who Tanzania (required for mini grids with approval by the supplies electricity and the level of service provided. an installed capacity of 15kW–1 MW and regulator. with at least 30 customers) There is a low level of administrative capacity. Permitting Administrative procedure Government wants to control entry. Nigeria (required for isolated mini to approve entry of a mini Government wants to use permits to enforce grids that produce at least 100kW of grid business in the market. regulation of tariff and standards. distributed power and have less than 1 It is intended to be less MW of generation capacity). onerous than a license (less Administrative capacity is high. Kenya (required for mini grids with information required and Mini grids are larger. 1–3MW of installed capacity. less scrutiny). Licensing Administrative procedure Government wants to control entry. Cambodia (required for all mini grids to approve entry of a mini Supply of electricity is illegal without a license. since 2001) grid business in the market, Kenya (required for mini grids with typically similar to licenses Administrative capacity is high. installed capacity of more than 3MW) that normal distribution Mini grids are larger. utilities are required to Tanzania (required for mini grids with obtain. installed capacity of 1–10MW) Nigeria (required for mini grids that produce more than 100kW of distributed power and have more than 1 MW of generation capacity). Source: ESMAP analysis. Note: “Suitable where” in this table and the tables that follow does not mean that wherever this condition is fulfilled, the option should be selected. Rather it means that under these conditions, this option may be well suited, while under other conditions, another option may be better suited. kW = kilowatts; MW = megawatts. Annex 5 of the mini grid regulations in Nigeria defines distributed power as “the average active power fed into the distribution network in each 15-min- a.  ute interval of its operation period” (NERC 2016). MINI GRIDS FOR HALF A BILLION PEOPLE    239 registering for entry than when applying for permits and Where governments do not desire to control who enters licenses (table 9.2). the mini grid market, the regulator’s choice will first be guided by whether it needs to know who is operating in The decision tree in figure 9.3 highlights the conditions the market. If it does not need to know, it may opt for no under which the four regulatory options are likely to be regulation of entry. suitable. If the regulator does want to know which mini grid opera- The first question determining the type of entry regula- tors are in the market, its choice may be guided by whether tion is whether electricity supply is legal without a license. it regulates tariffs, service standards, and technical stan- Where it is legal, regulators can choose among the four dards and whether it can do so without a permit or a license. entry regulation options. If it is not legal, they will be limited Registration is best suited if tariffs, service standards, or to the license option. TABLE 9.2 • Information requirements for registration, permitting, and licensing of mini grids in five countries Permitting and Registration Permitting licensing Licensing Uttar Information Pradesh requirement (India)a Nigeria Tanzania Nigeria Kenya Cambodia Nigeria Tanzania Intended ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ location and site (simplified) (simplified) (simplified) (map) (simplified) (detailed) (map) (map) description Financial Not ✔ ✔ ✔ ✔ ✔ ✔ ✔ projections required (simplified) (simplified) (detailed) (five-year (detailed) (10-year (detailed) business business plan) plan) Details on ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ proposed (simplified) (simplified) (detailed) (detailed) (detailed (detailed) (detailed) (if applicable, distribution with map) detailed) system Details on ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ generation (simplified) (simplified) (simplified) (detailed) (detailed) (detailed) (detailed) (detailed) system a Spreadsheet for Not Not required Not required ✔ ✔ ✔ ✔ ✔ tariff calculation required (detailed) Land certificate Not Not required ✔ ✔ ✔ Not ✔ ✔ and building required (proof of (building required (proof of permit land-use permit and land-use right) land-use right) right) Environmental Not ✔ ✔b ✔ ✔ Not ✔ ✔ impact required required assessment Declaration of Not Not required Not required ✔ ✔ Not ✔ Not required compliance required required with health and safety standards Source: ESMAP analysis. Note: “Simplified and detailed” refer to the level of detail the developer must provide as part of the information requirement. Procedure is envisaged in the Draft Implementation Guidelines for the Mini Grid Projects in the State of Uttar Pradesh. The document refers only to a.  “project details” as the information required. Developer may receive provisional registration from the Energy and Water Utilities Regulatory Authority before the commissioning of the project and b.  before submitting the Environmental Impact Assessment certificate, by submitting proof of initiation of the process to obtain the certificate. ✓ = required 240   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 9.3 • Decision tree for regulating entry Entry regulation Electricity supply is illegal Electricity supply is not without a license illegal without a license No desire to control who Desire to control can enter the business who can enter the business Need, desire, or No need, desire, capacity to know or capacity to who is in business know who is in business No regulation or tari , Regulation of tari , service, or service, or technical technical standards standards Not possible to Possible to regulate tari s regulate tari s and standards and standards without permit without permit or license or license License License Permit License Permit Registration Registration No regulation of entry Source: ESMAP analysis. MINI GRIDS FOR HALF A BILLION PEOPLE    241 • Government policy mandates a national uniform tariff. Four options regulate mini grids entry to In this section, we present options for regulating mini grid the market: (1) choosing unregulated mar- tariffs that take into account these factors. ket entry; (2) requiring that mini grids register with How (or if) tariffs are regulated has major implications for the appropriate government entity; (3) requiring the economic viability of mini grids and for developers’ abil- that mini grids obtain a permit to operate, granted ity to attract financing. As a result, without clarity on how by the regulator; and (4) requiring mini grids obtain (or if) mini grid tariffs are regulated (and if they are regu- a license to operate, usually granted by the regula- lated how they will be reviewed), investors will hesitate to tor. Registration, permitting, and licensing usually finance mini grid projects. require more information from the developer and review from the regulator. In some countries, var- Options for regulating retail tariffs include willing buyer, iously sized mini grids have different market entry willing seller, efficient new entrant price cap, individualized requirements (for example, smaller mini grids need cost-based tariff limits, bid tariffs, and uniform national tar- only register, while larger mini grids must get a per- iff. Table 9.3 describes these options and the contexts in mit or license). which they likely to be suitable. The five options presented in table 9.3 address the level of retail tariffs but not their structure, a vital distinction from technical standards are not regulated or if they are regu- the standpoint of business development. Retail tariffs may lated but enforcement mechanisms other than a permit or be based on energy, power, or the number of devices per a license (such as a grievance mechanism) can be used. household. Collection methods also vary, as tariffs can be prepaid or postpaid. Retail tariff structure may differ for Where regulators want to control who enters the mini grid each project depending on technology, developers’ costs, market, they may opt for a permit or a license. The level and customer’s ability to pay. One country may host sev- of administrative capacity may guide the choice between eral tariff structures. permit and license. Generally, entry control that does not demand a great deal of information, is relatively easy to This section focuses on regulating the tariff level rather obtain, and does not grant exclusivity is termed a permit. than the tariff structure, but it is generally recommended to Licenses normally refer to documents that are more for- leave the flexibility to developers to adopt the tariff struc- mal, take more time to complete, and, in some cases, con- ture that best fits their technology and market. In partic- fer exclusivity, particularly in cases where the license is part ular, mini grid developers will need to find the proper tariff of a concession. But there are no strict distinctions among structure and payment model adapted to the local context the terms. to ensure that they will be able to collect payments.2 Not regulating tariff structure, at least initially, may also foster REGULATING RETAIL TARIFFS innovation and redound to the country’s benefit. When Retail tariffs on the main grid tend to be regulated in order designing their tariff regulations, regulators should adopt to protect customers from monopoly power, as the main a light-handed approach so developers can set the tar- grid is usually the sole supplier of electricity. The decision iff structure that best meets the needs of their business to regulate retail tariffs for mini grids is not as clear-cut. On (Tenenbaum and others 2014). the one hand, there are valid reasons not to regulate the Let us consider each of the five options for regulating the retail tariffs of mini grids: level of retail tariff. • They’re competing with kerosene lanterns and backup diesel generators Willing buyer, willing seller Under this option, the level of retail tariff is based on an • Tariffs are already constrained by customers’ scant dis- agreement between the mini grid developer and its cus- posable income tomers. In Nigeria (for systems up to 100 kilowatts [kW] of • Regulating tariffs requires financial and human re- distributed power), Haiti, and Madagascar, an agreement sources, which the regulatory agency may not have. with the community is a regulatory requirement—a form On the other hand, regulating the retail tariffs of mini grids of regulation by contract, as discussed later in this chap- makes sense when ter. In Nigeria and Haiti, the government has developed template agreements between the developer and the • Mini grids have gained pricing power locally community that developers can use. Community agree- • Government subsidies are provided as a way to bring ments are also something mini grid operators consider down the cost of electricity good business sense. 242   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 9.3 • Options for regulating tariffs Option Description Suitable where… Examples Willing buyer, Tariff set at price that the Mini grid operators are unlikely to make Cambodia (all mini grids before willing seller mini grid developer and supernormal profits if given pricing freedom 2001) its customers agree on (where competition with traditional energy Uttar Pradesh (mini grids that do not (that is, the retail tariff is sources sets upper limit on mini grid prices or receive state subsidies [in practice all not regulated). Note that mini grids are small). mini grids as of 2017]) household disposable A policy objective is to expand electrification Nigeria (isolated mini grids with less income sets a de facto limit as rapidly as possible. than 100kW of distributed power) to what the developer can Mini grids compete with a distribution Tanzania (mini grids with less than charge for electricity. company that does not provide reliable 100kW of generation capacity) supply of electricity. Administrative capacity is low. Efficient new Regulator sets single Cost of service is expected to be uniform Bangladesh (all solar PV–diesel mini entrant price benchmark tariff for all across communities. grids) cap mini grids, based on cost Administrative capacity to set the benchmark of service of efficient new is moderate to high. entrant in the market. Individualized Regulator sets limits on There is a risk of monopoly pricing. Cambodia (all mini grids as of 2017) cost-based tariffs for each mini grid The cost of service is not expected to be Nigeria (isolated mini grids with more tariff limits individually, based on full uniform across communities. than 100 kW of distributed power) cost of service. Administrative capacity is high. Kenya (all mini grids as of 2017) Tanzania (mini grids with more than 100kW of generation capacity) Bid tariffs Mini grid investor obtains There is a centrally coordinated approach to Madagascar (65 projects as of 2015) right to serve an area by scaling up electrification through mini grids. Uganda (one tender in 2003; another bidding lowest tariff in Market data are available. in 2017 to electrify 25 villages) competitive tender (either There are enough mini grid developers to where bid tariff is single compete for each area. criterion or in multiattribute bidding processes). Administrative capacity is high. Regulator is willing to accept the winning bid price. Uniform Regulator sets mini grid tariff There is strong political pressure for uniform Cambodia (grid-connected small national tariff at national main grid tariff. electricity prices. power distributors, as of 2017) The subsidy provider is creditworthy. Kenya (plans for mini grids to operate Administrative and financial capacity are high. under contract to national utility and charge national utility’s tariff) Source: ESMAP analysis. kW = kilowatts; PV = photovoltaic. Most private operators clearly recognize that they will the regulator may review their tariffs if 15 percent of cus- be able to construct and operate mini grid systems only tomers complain (EWURA 2017b). if there is “buy-in” or acceptance from the villages that will be served. While the private operators may be legally Efficient new-entrant price cap required to get a license or permit from the national reg- Under this approach, a regulator calculates the tariff that a ulator, the document will be of little or no value unless typical, efficient, newly built mini grid would have to charge the local government and villagers also support the proj- to cover its costs (generation and distribution) and earn a ect (Tenenbaum and others 2014, 78). reasonable return on capital invested (a return equal to the Variants of the willing buyer, willing seller model include (1) opportunity cost of capital) over a specified time period. announcement of a tariff by the mini grid, with interested This tariff is then set as the price cap for all mini grids. The customers signing up for service; (2) individually negoti- tariff is generally set using a financial model of a typical ated tariffs for each customer; and (3) use of a regulatory mini grid.3 The tariff could also be set by observing the tar- safety valve, as in Tanzania, where developers need only iffs bid in tenders, if the country has awarded mini grid con- make their tariff public before installing the mini grid, but cessions competitively using tariffs as the bidding variable. MINI GRIDS FOR HALF A BILLION PEOPLE    243 The efficient new entrant price cap approach is used in variety of common issues, including calculating the asset Bangladesh for solar-diesel mini grids. The Infrastructure base, setting the rate of return, estimating depreciation, Development Company Limited (IDCOL) set the tariff at and taking account of subsidies (box 9.1). the national level of taka (Tk) 30 to Tk 32/kilowatt-hour (kWh) ($0.37–$0.39), plus a monthly line rent.4 The tariff Bid tariff is set based on efficient cost assumptions (competitively • In the bid tariffs option, the government defines areas procured equipment) and allows for a return on equity of (sometimes towns or villages) to be electrified by mini 13–15 percent (World Bank 2017a). grids and competitively tenders the right to supply each area. The winning bidder is the firm that offers the low- The efficient new-entrant price cap approach can be est tariff to supply the area. The tender documents nor- indexed to inflation, the exchange rate, fuel prices, or tax mally define coverage, service levels, subsidy levels, and changes if they are expected to affect the costs of mini other parameters. grids in operation. Such adjustments can be done auto- matically and periodically (for instance, once a year). The • The bid tariff approach aims to reveal the efficient cost tariff can also be adjusted from time to time in response to of service. Bidders would not knowingly bid a tariff that changes in the typical costs and efficiency of capital equip- is below their cost of service, as doing so would fail to ment (for example, decreases in the cost of solar panels make the rate of return investors require. Bidders would and batteries). This type of adjustment should be applied, not bid above their estimated cost of service, for fear however, only to new mini grids, not to mini grids already in that another firm would undercut them and win the ten- operation. It would be unfair to existing operators to require der. Madagascar used a bid approach to award mini grid them to start charging a lower tariff based on lower-cost concessions and set tariffs. equipment not available when they built their mini grids. Nigeria, Sierra Leone, and Uganda are moving toward mul- tiattribute bidding for competitive tenders for mini grids. Individualized cost-based tariffs In 2017 Sierra Leone’s Ministry of Energy initiated pro- These tariffs are set for each mini grid individually at a curement for mini grids in 90 communities. The bids from level calculated to allow the mini grid to recover its rea- five prequalified bidders were judged on more than 20 sonable costs of service and earn a reasonable return on attributes.5 If the tariff is one of the bidding criteria, such capital invested. This approach applies the concepts used a multiattribute bidding process could also be considered for setting the tariffs of large distribution utilities to mini a tariff-setting method if the regulator commits to setting grids. Individualized cost-based limits have been used for allowed tariffs equal to the bid price. Regulators may be mini grids in Cambodia since the early 2000s. They are the reluctant to commit to approving the bid price without first mechanism prescribed for larger mini grids in Nigeria and seeing evidence that there are an adequate number of bid- Tanzania and for government-subsidized mini grids in Uttar ders who compete rather than collude. Pradesh. Regulators use various methods to set the cost-recovery Uniform national tariff tariff. Most are variants of rate of return regulation (also Under this tariff-setting approach, all customers in the called cost of service regulation). Under rate of return reg- same tariff category pay the same tariff for electricity ulation, companies may recover their operating expenses, regardless of where they live (Tenenbaum and others depreciation on their assets, and a rate of return on the 2014). Mini grids are generally required to charge the tariff undepreciated asset base. The rate of return allowed by that applies on the main grid. Uniform national tariffs are the regulators is based on the cost of debt and an allowed most often used for residential customers (although they return on equity. The allowed return on equity is estimated may apply to all customers), because residential tariffs from investors’ opportunity cost of capital and the riskiness have high political visibility. Notably, a uniform tariff regime of the investment. Companies can recover from customers will require significant subsidies to make mini grids a viable only costs that were prudently incurred and expenditures option for electrification. Consider the following two exam- on assets that are used and useful (Brown 2010). ples in Cambodia and Peru. Setting individualized tariffs on a cost-of-service basis is In the most recent stage of Cambodia’s regulatory evolution, a balancing act: allowing mini grids to recover their costs former isolated mini grids that are converted to small power while ensuring they do not charge for costs they did not distributors are required to charge the same tariff as Elec- incur in providing the service, while not charging unreason- tricité du Cambodge (EDC), the utility that supplies most ably high costs because of the mini grid’s own inefficiencies. of the country. These tariffs are made possible by a subsidy To get this balance right while being clear and predictable, fund that covers the difference between each mini grid’s most mini grid tariff regulations spell out the approach to a own cost-recovery tariff and the uniform national tariff. 244   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 9.1 SETTING INDIVIDUALIZED TARIFF CONTROLS Individualized tariff control has many variants. Juris- • treatment of capital subsidies, or permitting opera- dictions may take different approaches to each of the tors “to take depreciation” but not “to earn a profit following components: or return on the equity provided by the grant” (Tenenbaum and others 2014), because capital • regulatory asset base, or the value of the regulated subsidies increase the operator’s regulatory asset assets on which the mini grid operator may earn a base at no cost for the operator, but the operator return. will need to replace the subsidized assets once they • rate of return, or the profit the operator may earn on are fully used. its capital investment. (The value may be set at the Table B9.1.1 provides examples of choices countries weighted average cost of capital, which is then mul- have made in setting individualized tariffs. tiplied by the regulatory asset base.) • depreciation, or the cost of using assets the opera- tor may recover through its cost of service. TABLE B9.1.1 • Individualized tariff features in four countries Feature Cambodia Kenya Nigeria Tanzania Regulatory asset Value of assets is cost Not defined Not defined All assets that are used base at time of acquisition; and useful in provision costs should be of regulated services prudently incurred Rate of return Return on asset base of Internal rate of “Usual non-recourse Reasonable return on 10 percent return of 18 percent commercial debt interest capital, calculated by rate in local currency […] the regulator +plus 6 percent” Depreciation Straight line based on Not defined Straight line based on Straight line over useful standard individual individual asset lifetime economic life of the asset lifetime asset Treatment of Excluded from asset Not defined Subtracted from the Excluded from asset capital subsidies base but included for capital asset base base but included for depreciation depreciation Source: Cambodia, EAC 2007; Kenya, GIZ 2015; Nigeria, NERC 2016, 2017; Tanzania, EWURA 2017a, 2017b. In Peru, the government’s Electric Social Compensation As a rule, regulators use seven criteria to evaluate their Fund provides subsidies to main grid and mini grid con- options (table 9.4). sumers in rural areas who consume less than 100kWh a • Tariff accuracy: Does the tariff reflect costs? Individual- month. The subsidies allow these rural customers to pay ized cost-based tariffs ensure that the regulated tariff is tariffs in line with those paid by similar customers in urban cost-reflective and therefore has high accuracy; a uni- areas. Funding for the consumption subsidies comes from form national tariff is not based on mini grids’ costs and a 2.5 percent surcharge on the bills of residential, commer- therefore has low accuracy. cial, and industrial customers with a consumption of more than 100kWh a month (Tenenbaum and others 2014). • Risk of monopoly pricing: Under the willing buyer, will- ing seller mechanism, the risk that the operators charge Decision tree for regulating retail tariffs monopoly pricing is higher than in other tariff regimes. Which approach to tariff regulation may be appropriate An individualized cost-based tariff presents a low risk, will depend on a country’s objectives for mini grids, the given that each mini grid’s tariff is set based on its costs. government’s administrative capacity, the availability of • Suitability for rapid expansion of access: A bid tariff subsidies, and the regulator’s legal and policy constraints. approach aims to rapidly scale up access in specific MINI GRIDS FOR HALF A BILLION PEOPLE    245 areas. A uniform national tariff may not be appropri- Figure 9.4 presents a decision tree to guide policy makers ate in this context, because it aims primarily to ensure in regulating tariffs. equity among customers in the country and its imple- Some countries want uniformity in tariffs across geo- mentation is likely to take time (it requires setting up a graphic regions, a politically appealing policy. But for subsidy program to finance the gap between main grid regulators to implement such a policy, the only option tariffs and cost-recovery tariffs). The worst outcome available is uniform national tariffs. Regarding this option, would be to impose a uniform national tariff without any therefore, governments should consider prerequisites backup subsidy mechanism. for success. Fiscal capacity ensures the subsidies are • Time to implement: An efficient new-entrant price cap available to bridge the gap between the mini grids’ cost may require time for the regulator to set the benchmark. of service and the uniform national tariff. Administrative An individualized cost-based tariff takes significant time capacity allows governments to calculate the required (to design the tariff model and then calculate the tariff subsidy and disburse it. for each mini grid). If a uniform national tariff is not a policy requirement, the • Regulatory capacity needed: The willing buyer, willing next question is whether subsidies are available. Where seller approach requires no regulatory capacity. Design subsidies are not available (the topmost branch of the deci- of a competitive tender in the case of a bid tariff requires sion tree), any option for tariff regulation must allow mini high regulatory (and government) capacity. grids to recover the full cost of service through the tariff. • Compliance cost for the developer: An efficient new-en- The following four regulatory approaches can do this: trant price cap adds little cost (other than knowing the • Willing buyer, willing seller regulation). Submitting a bid in a competitive tender adds significant costs to the developer (with the risk of • Efficient new-entrant price cap not being selected). • Individualized, cost-based tariff limits • Tariff flexibility: The willing buyer, willing seller approach • Bid tariff offers the most flexibility, as the operator can opt for any Among these four options, the choice depends both on the tariff structure (for instance, tariffs varying depending capacity of the regulatory agency and the importance gov- on the time of day or the level of demand). At the other ernment places on reducing the risk of monopoly pricing end of the spectrum, a uniform national tariff provides compared with its emphasis on rapid expansion by cutting the least flexibility. red tape and allowing attractive returns on investment. • Some mini grid regulations also explicitly consider the If a uniform national tariff is not a policy requirement (and structure of retail tariffs, whether—for example, cus- subsidies are available), all four options above are also tomers are charged per kWh, a flat fee, or per kW; tariffs possible. Slight preferences emerge for one approach increase as consumption rises; or there is a lifeline tariff. over another depending on the objective of the subsidy. In general, though, a light-handed approach to tariff reg- If the overriding objective is to expand access to as much ulation leaves these tariff structure choices to the dis- of the population as possible, then subsidies should be cretion of the developer. TABLE 9.4 • Assessment of tariff options Tariff Risk of Suitability to Time to Regulatory Compliance accuracy monopoly rapidly expand implement capacity cost (for Tariff Option (cost-reflection) pricing access (for regulator) needed developer) flexibility Willing buyer–willing Low Medium High Low/ medium Low Low High seller Efficient new-entrant Medium Medium High Medium Medium Low Medium price cap Individualized cost- High Low Medium High High Medium Medium based tariff limits Bid tariff High Low High Medium High High Medium Uniform national Low Low Low Higha High Low Low tariffa Source: ESMAP analysis. A uniform national tariff can be imposed quickly, given that the tariffs are already set by the electricity utility. But designing a subsidy mechanism to a.  make this option viable is likely to take time. 246   MINI GRIDS FOR HALF A BILLION PEOPLE Figure 9.4 • Decision tree for regulating tariffs Retail tari regulation Uniform national No uniform national tari requirement tari requirement Subsidies No subsidy Subsidies available No subsidy available available available Subsidies provided to lower Subsidies provided cost of service for to close the customers viability gap No good option, Uniform Bid Individualized Individualized E cient Willing Bid Individualized E cient Willing as uniform national tari cost-based cost-based new- buyer— tari cost-based new- buyer— national tari tari tari limits tari limits entrant willing tari limits entrant willing requires price cap seller price cap seller subsidies Source: ESMAP analysis. devoted to closing the viability gap in serving certain Tariff review and automatic adjustment clauses areas. A viability gap arises when it costs more to sup- Not all mini grid tariffs need to be regulated, even in ply an area or household than the revenue that will be countries that regulate mini grid tariffs. It is increasingly generated from that area or household. If the overrid- common that mini grids below a certain capacity thresh- ing objective is to bring the cost of electricity down for old can set their tariffs without regulatory review and as many people as possible around the country, then approval. When tariffs are regulated, however, they usu- subsidies would be provided to developers or their cus- ally undergo periodic review by the regulatory agency. tomers with the primary objective of reducing costs. The These reviews should occur at regular, predefined inter- tariff options better suited to one subsidy scenario or the vals (usually not more than every two years and not less other are shown in the decision tree. But any of the four than every five years). approaches in the bullet list above would be possible in either subsidy scenario. MINI GRIDS FOR HALF A BILLION PEOPLE    247 a tariff calculation template in Microsoft Excel that is used by the developer to submit its tariffs and by the regulator to There are five options for regulating mini review them. Using a tariff calculation template that shows grid tariffs: (1) willing buyer, willing seller, (2) the costs (capital, operating, and finance), revenues, and efficient new-entrant price cap, (3) bid tariff, (4) indi- other key inputs can also mitigate the risk that the regula- vidualized cost-based tariff limits, and (5) national tor assesses mini grid tariffs against what the tariffs are on uniform tariff. The uniform national tariff option is the main grid. only possible for mini grids if significant and sustain- able subsidies are available. The choice among the In general, the only reason a regulator should be allowed remaining four options, each of which can in theory to reject a developer’s proposed tariffs is if they do not enable mini grid developers to recover their costs, conform to the tariff guidelines or rules as stated in the depends on both the capacity of the regulatory regulation. This highlights the point made at the begin- agency and the importance the government places ning of this section: developers need clarity on how their on reducing the risk of monopoly pricing, compared tariffs will be reviewed, and what rules govern how they with the emphasis it places on driving rapid expan- can set their tariffs. sion by minimizing red tape and allowing attractive returns on investment. An additional consideration for regulating retail tariffs: regulatory treatment of subsidies A final consideration for regulators when making deci- sions about retail tariffs is how to treat subsidies. In most Developers are usually not allowed to change their tariffs populous countries lacking sufficient, equitable access to once the tariff schedule has been approved after a peri- electricity, there will be a “viability gap” between the tariff odic review, but a common practice is to allow developers that the developer would need to charge its customers to to adjust their tariffs without a full review by the regulator be economically viable and the tariff that customers can for the following reasons: an increase in the fuel price; an afford to pay given their disposable household income increase in taxes; inflation; and depreciation of the local (or the tariff that developers would be required to charge currency with respect to the currency or currencies in under a national uniform tariff regime). In these cases, which the developer pays its creditors, investors, and sup- subsidies will be required. (Chapter 6 provides a detailed pliers. In these instances, developers are often required discussion on the types and sources of subsidies.) When to show documentary evidence of the impact that the a government mandates, authorizes, allows, or provides change in fuel, tax, inflation, or exchange rate has on the a subsidy, it usually does so as part of a specific policy developer’s costs. designed to close this viability gap to achieve a social objective, such as increasing access to electricity for Furthermore, because mini grid developers are competing low-income households. against self-generation by larger customers, good regula- tory practice suggests that developers should be able to Regulators should adhere to the general principle, below in enter into tariff agreements with larger customers without italics, when accounting for subsidies in their regulation of having to receive regulatory approval for these tariffs. retail tariffs: There may be instances in which the regulator is com- If the government authorized, mandated, provided, or allowed a subsidy, the regulator should refrain from pelled to conduct a review of the tariff outside of the actions that would nullify or reduce the effect of the periodic review process. Because one of their roles is to subsidy. Instead, the regulator should take regulatory protect consumers, regulators need to be responsive to actions that help to ensure that the subsidy is deliv- customer complaints. However, because it is common ered to its intended target as efficiently as possible. The for customers to complain about their tariffs (particularly regulator, however, should periodically inform the gov- if the developer must increase them for one of the four ernment of the costs and benefits of the subsidy (Tenen- reasons stated above), it is helpful to set a predefined baum and others 2014, 122). threshold above which a review of tariffs is triggered. Reg- This principle can be applied to two types of subsidies: ulations in Haiti, Rwanda, and Tanzania, for example, set connection-cost subsidies and cross-subsidies in tariffs. this threshold in terms of a percentage of a mini grid’s High connection costs are a major barrier to accelerating customers. the pace of connecting new customers to mini grids, par- To facilitate the tariff review process, regulators and devel- ticularly in rural and low-income areas, so governments opers should use the same methodology to calculate and and donors sometimes provide grants to mini grid devel- review tariffs. This can be achieved by having, for example, opers to reduce the capital costs involved in connecting 24 8   MINI GRIDS FOR HALF A BILLION PEOPLE new customers (Tenenbaum and others 2014). Meanwhile, REGULATING SERVICE STANDARDS cross-subsidizing different groups of customers enables Service standards are usually regulated on the main grid mini grid developers to provide electricity to the lowest-in- to protect customers from monopoly power and ensure come customers within their service area. good service. They are commonly divided into quality of For connection cost subsidies, if a mini grid developer power (such as voltage and frequency stability), quality of receives a grant from an outside entity to reduce its cap- supply (such as hours or service and reliability), and quality ital costs, the developer should be allowed to take depre- of commercial services (such as connection time and cus- ciation on the equity provided by the grant, but should tomer service). not be allowed to earn a profit or return on this equity. Regulating service standards may be warranted where For cross-subsidies across groups of customers, regula- mini grids have gained market power, since such mini grids tors should explicitly allow developers to charge different may have little incentive to improve their service quality tariffs to different customer classes, with the objective of and may not provide the level of service the market wants. expanding access to a larger number of customers in the The regulation of service standards, however, risks hinder- mini grid service area. National utilities routinely cross-sub- ing the development of mini grids, or connection of certain sidize their residential customers by charging commercial customers, by hiking the cost of energy or connection. It and industrial customers higher tariffs, so this same option may limit the flexibility of developers to offer service levels should be available to mini grid developers. adapted to customers’ needs and willingness to pay. Estab- As a final point, some developers lower the underlying lishing service standards also creates additional compli- capital costs of connecting new customers in innovative ance costs for developers. ways (for example, the use of ready boards) and allow Options for regulating service standards include (1) no ser- customers to pay their connection charges over time in vice standards, (2) reporting standards, (3) differentiated smaller monthly installments. In these cases, the devel- standards, (4) uniform mini grid–specific standards, and oper may incur additional financing and administra- (5) main grid–level standards (table 9.5). tive costs associated with these activities. The general recommendation here is if a mini grid developer offers Reporting its customers the ability to repay the costs of: (1) their In the reporting option, mini grids are not bound to comply connection charge, (2) internal wiring, (3) house-im- with any service standards, but they need to report their provements to meet minimum electricity-connection quality of service to customers and to a central body (reg- standards, and (4) purchasing electricity-powered appli- ulator or other). Information provided to the central body ances and machinery through monthly on-bill install- should be available to the public upon request. In develop- ments, then the regulator should allow the provider to ing the reporting requirements, regulators should keep in recover both the financing and administrative costs it mind that mini grids are much smaller than national grid incurs to provide these loans. This can also help incen- utilities and should adjust their reporting expectations tivize developers to encourage productive uses of their accordingly. Regulators need to determine how much electricity, as described in chapter 3. information is needed to effectively manage the sec- tor while avoiding overburdening mini grids. One variant is to gradually phase in reporting requirements as mini Regulators should adhere to the following grid operators increase their capacity or specific data general principle when accounting for sub- become necessary (USAID 2017). The “Quality Assurance sidies in their regulation of retail tariffs: Framework for Mini-Grids,” developed by the U.S. National Renewable Energy Laboratory, provides additional infor- If a subsidy is authorized, mandated, provided, or mation on how developers can measure and report their allowed by the government, the regulator should quality of service (Baring-Gould and others 2016). not take actions that would nullify or reduce the effect of the subsidy. Instead, the regulator should Differentiated standards take regulatory actions that help to ensure that In this option, service standards are regulated and mini the subsidy is delivered to its intended target as grid–specific, but they are not uniform across areas, mini efficiently as possible. The regulator should, how- grids, or customer types. Service standards may differ ever, periodically inform the government of the across mini grids depending on their size, whether they are costs and benefits of the subsidy (Tenenbaum subsidized, or the type of authorization they are required and others 2014). to obtain. In Nigeria, for example, registered mini grids are subject to recommended service standards that are lower MINI GRIDS FOR HALF A BILLION PEOPLE    249 TABLE 9.5 • Options for regulating service standards Option Description Suitable where… Examples No service Mini grids are not bound to Mini grids are in active competition Cambodia (no service standard standards any service standards or with other energy providers. regulation before 2001) reporting requirements. The mini grid market is incipient. Uttar Pradesh (basic service standards Administrative capacity is low. for all mini grids were created in 2016, but as of 2017 they were not enforced) Reporting Mini grids are not bound Mini grids are in active competition None identified to comply with service with other providers of energy. standards, but they must The mini grid market is incipient. report their quality of service on specified indicators to a Administrative capacity is low. central body and customers. Differentiated Service standards are Mini grids have market power, tariffs Nigeria (differentiated standards based standards regulated but differentiated are regulated, and/or subsidies are on type of authorization; registered across mini grids, areas, or provided. mini grids may follow recommended customers. Service needs, willingness to pay, cost service standards that are less Regulator may add reporting of service, and ability of operators to stringent than standards for mini grids requirements as a variant. manage regulatory compliance differ with permits) across areas. Peru (differentiation of service Administrative capacity is high. standards between urban and rural areas and based on population density) Uniform mini All mini grids must provide There is an established market where Uttar Pradesh (since 2016 basic grid–specific a uniform, regulated level of mini grids have market power. standard regulation has been in place standards service that is different from Tariffs are regulated and/or subsidies for all mini grids; more stringent service the required service on the are provided, and the combination of standards are required of state- main grid. tariffs and subsidies can cover the cost subsidized mini grids, but as of 2017, no Regulator may add reporting of service at regulated standards. operator had applied for the subsidy) requirements as a variant. Administrative capacity is high. Main grid–level Mini grids must comply Providing the same quality of service Bangladesh (hours of service, reliability, standards with main grid-level service across the country is a core policy billing, and metering) standards. objective. Cambodia (since 2001, hours of service, power reliability, power quality, billing, metering, and customer service) Kenya (power quality and reliability) Source: ESMAP analysis. than the mandatory service standards for mini grids with a dards that are different from main grid standards. The permit. The regulator may consider reporting requirements Multi-Tier Framework used by the World Bank to track in combination with service standards. In Nigeria mini grids energy access (Bhatia and Angelou 2015) presents dif- with permits must report their incidents and accidents to ferent levels of access that regulators could use to define the regulator every two years (NERC 2016). the service level. In this framework, Tier 1 corresponds to supply of at least four hours during the day and one hour Service standards may also differ across areas, based on in the evening. Tier 4 corresponds to supply of at least 16 the cost of service and customers’ willingness to pay. “The hours of electricity during the day and 4 hours in the eve- standards should be based on customers’ preferences ning, with no more than 14 disruptions a week and voltage and their willingness to pay for the costs of providing the variations that do not affect the use of appliances. The specified level of quality. The standards need not be uni- regulator may opt for reporting requirements in combina- form across all customer categories or geographic areas” tion with service standards. (Reiche, Tenenbaum, and Torres de Mästle 2006). In Peru, service standards differ for urban and rural areas based on Traditionally, the regulator enforces service standards; in the cost of service and willingness to pay. the case of isolated mini grids, the community could do so. The community may directly negotiate the service Uniform mini grid–specific standards agreement contract with the developer and then moni- In this option, all mini grids must provide a uniform level tor compliance (Tenenbaum and others 2014). The local of service. The regulator defines mini grid–specific stan- government or a rural electrification agency could also 250   MINI GRIDS FOR HALF A BILLION PEOPLE monitor service standards. In Cambodia the regulator Main grid–level standards can be set in various ways. In delegated the monitoring of compliance of technical stan- Kenya service standards are lower in rural areas than in dards to the Provincial Department of Mines and Energy, urban areas (ERC 2016). Cambodia illustrates the contin- which has staff in the field (Tenenbaum, Greacen, and uum between uniform service–standard regulation and Vaghela 2018). Similar synergies could be exploited for differentiated service–standard regulation: service stan- service standards. dards are uniform, with some exceptions. Mini grids must provide power 24 hours a day, except where this is not via- Main grid–level standards ble (ERC 2016). In this option, mini grids must comply with standards set A variant of main grid–level standards could be to impose for the main grid. Bangladesh, Cambodia, and Kenya have main grid–level standards when a mini grid connects to the adopted this option. The regulator may consider reporting main grid (and starts operating as small power distribu- requirements in combination with service standards. tor or small power distributor plus small power producer). FIGURE 9.5 • Decision tree for regulating service standards Service standards regulation Monopoly Nascent mini grid market conditions or good competition High administrative Low administrative No subsidies Subsidies capacity capacity available available Moderate to high Low Policy requirement No policy requirement administrative administrative for grid-level quality for grid-level quality capacity capacity No subsidies available Non-uniform cost of, and Uniform cost of, willingness and willingness to pay for, Subsidies to pay for, service available service across areas across areas Grid No good Uniform Di erentiated No good Reporting Reporting No service level solution mini grid– standards solution standards standards speci c standards Source: ESMAP analysis. Note: “Good competition” refers to the situation where mini grid developers are still competing with alternative technologies in communities they serve, such as backup diesel generators. MINI GRIDS FOR HALF A BILLION PEOPLE    251 Such a mechanism should be defined from the start, to be no good solution if there is a policy requirement for grid- provide certainty to mini grid developers. level standards and subsidies are unavailable. Decision tree for regulating service standards REGULATING TECHNICAL STANDARDS Figure 9.5 summarizes the conditions under which differ- Regulation of technical standards can help ensure safety, ent types of standards may be suitable. the quality of equipment and construction, compliance with environmental standards, and future connection of The first question that will determine what type of service mini grids to the main grid. Technical standards regulate standards regulation is appropriate is whether the mini mini grids’ inputs. They differ from service standards, which grids are competitive entrants or in a monopoly situation. regulate outputs (such as the quality of power and supply). When mini grids are competitive entrants, the choice of If mini grids are already built to grid-compatible standards, service standards could be guided by whether the govern- connection to the main grid will be straightforward. Equip- ment provides subsidies. Given the fragile economics of ment that meets these standards may be more expensive many mini grids, imposing service standards regulation than needed to serve the area in the short term, however, without subsidies may harm the development of the mar- and requirements to use grid-compatible equipment may ket. Thus, if the market is competitive and there are no limit innovation. government subsidies, the regulator may opt either for no Options for regulating technical standards include safety service standards regulation if administrative capacity is standards only, mini grid–specific standards, optional low, or for reporting if administrative capacity is moder- grid-compatible standards, mandatory grid-compatible ate to high. If subsidies are provided, the regulator may require reporting of service standards to monitor the standards, and main grid standards (table 9.6). quality of service. Safety standards only In a monopoly situation, the regulator may consider regulat- In this option, only safety is regulated. Safety standards ing service standards to protect customers from abuse of can be designed specifically for mini grids, or safety stan- monopoly power. If administrative capacity is low, the regu- dards developed for the main grid can be applied. In Uttar lator would need to build up capacity to design and enforce Pradesh, our in-country research indicated that mini grids service standards. Alternatively, the regulator may design are required to follow standards defined in both the state’s the service standards and delegate their enforcement to electricity regulations and its mini grid regulations. local governments, communities, or a rural electrification/ energy agency. If administrative capacity is adequate to Some jurisdictions may choose to recommend but not develop and enforce service standards, decisions about how mandate safety standards, as in Nigeria’s mini grid reg- to regulate the standards will depend on the developer’s ulations. But given the risks to personal safety posed by cost of supply and customer willingness to pay. Grid-level electrical systems, mini grid technical standards minimum standards tend to require not just administrative capacity generally spell out safety requirements. but also fiscal capacity for subsidies to cover the high costs that grid-level standards may impose. Therefore, there will Mini grid–specific standards In this option, technical standards are designed specifically for mini grids. These standards do not guarantee future connection to the main grid. Mini grid–specific standards There are four options for regulating the are not necessarily lower than main grid standards. For service standards of mini grids: (1) requir- example, Nigeria’s regulations for the mini grid include ing only that developers report their service levels, health and safety rules that apply specifically to mini grids. (2) developing and enforcing standards specific to a particular mini grid or portfolio of mini grids, (3) Optional grid-compatible standards developing and enforcing standards that are uni- Mini grids have the option to comply with grid-compatible form for all mini grids, and (4) requiring mini grids standards if they want their assets to be protected when the to adhere to main grid–level standards. Requiring main grid arrives. These standards are generally lower than main grid–level standards is possible only for mini those of the main grid. Under Tanzania’s regulations for grids if subsidies are available. The choice among small power producers (SPPs), the mini grid’s distribution the other options depends on whether mini grids assets are not eligible to be bought out by the main grid if are operating in a monopoly situation, and the level they are not built to the (grid-compatible) standards of the of the regulatory agency’s administrative capacity. Tanzania Bureau of Standards. Meters are considered part of the distribution assets only if they are grid compatible. 252   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 9.6 • Options for regulating technical standards Option Description Suitable where… Examples Safety Only safety standards are Communities are not expected to be connected Uttar Pradesh (mini grids that generate standards regulated. to the main grid in the near future. less than 50kW must comply with only Entrepreneurs have adequate information and safety standards only) the incentive to select appropriate equipment. Nigeria (recommended safety Government subsidies are minimal. standards for registered mini grids) There is enough capacity to regulate dangerous equipment. Mini grid– Mini grids must comply Willingness to pay or desired service levels Nigeria (specific health and safety specific with standards designed is not consistent with use of main grid– standards apply to all mini grids) standards specifically for mini grids. compatible technology. Sri Lanka (government-subsidized Compatibility with main Communities are not expected to be connected mini grid must comply with technical grid is not required. to the main grid in the near future. standards that are lower than the Subsidies are provided (to ensure that projects standards for the main grid) receiving subsidies meet minimum standards appropriate for mini grids). Administrative capacity is adequate. Optional Mini grids have the option There is a reasonable likelihood that at least Tanzania (before 2017, the regulation grid- to comply with grid- some mini grids will connect to the main grid in provided that distribution assets had to compatible compatible standards the future. be at least grid compatible to be bought standards (which are lower than The administrative capacity to design the out by the utility when the main grid main grid standards). standards exists. arrives) If they do, they are eligible to choose an economically attractive option when the main grid arrives. Mandatory Mini grids must comply There is a reasonable likelihood that many mini Tanzania (2017 regulation requires mini grid- with grid-compatible grids will connect to the main grid in the future grids with more than 30 customers compatible standards. or integration of mini grids with the main grid is or at least 15kW of installed capacity standards a policy objective. to follow standards defined by the Subsidies are provided (to ensure that projects Tanzania Bureau of Standards) receiving subsidies meet minimum standards Sri Lanka (mandatory standards for appropriate for mini grids). interconnection of hydro mini grids to Administrative capacity is high. main grid) Main grid Mini grids must comply Mini grids are expected to be taken over by the Bangladesh (mini grids must follow standards with main grid technical main grid within the asset life of the mini grid. main grid standards for rural areas) standards. Standards can be shown to be justified in terms Nigeria (mini grids with permits must of economic costs and benefits in the area. comply with standards that apply to the Combination of tariffs and subsidies (tariffs main grid) only or tariffs plus subsidies) can cover the mini Cambodia (all licensees must comply grid’s cost of providing electricity. with main grid codes and standards) Administrative capacity is sufficient to enforce Kenya (all mini grids must comply with the standards. the national grid code for connection and distribution) Source: ESMAP analysis. kW = kilowatts. Mandatory grid-compatible standards grid arrives. The latter option may be better adapted for In this option, mini grids must comply with safety standards small systems that would not be viable if stringent technical and technical standards that allow them to connect to the standards were applied from the beginning or in countries main grid in the future. Equipment or construction stan- where implementation of the main grid extension plan is dards may also be regulated if subsidies are provided. Reg- uncertain (USAID 2017). Sri Lanka is one of the few coun- ulators may require mini grids to be grid compatible when tries that have developed specific technical standards for they are built or require technical upgrades when the main the interconnection of mini grids to the main grid. The coun- MINI GRIDS FOR HALF A BILLION PEOPLE    253 try has successfully developed interconnection standards Decision tree for regulating technical standards and procedures for hydropower mini grids, in which the mini Figure 9.6 presents the decision tree for regulating techni- grid must carry out a load-flow study and install safety relay cal standards. equipment before interconnecting with the main grid. One of the primary considerations for deciding how to regulate mini grid technical standards is what will happen Main grid standards when the main grid arrives in the service area of a mini grid. In this option, mini grids must comply with main grid tech- nical standards. In Cambodia, mini grids must follow the Where future integration of mini grids into the main grid is 2004 Electric Power Technical Standards and the 2007 not essential, regulators may opt for safety standards regu- Specific Requirements of Electric Power Technical Stan- lation only, or mini grid–specific standards with equipment dards, which also apply to the main grid. The regulator and construction requirements if subsidies are provided or advises on how to build the mini grid system so that it can if the government wants to facilitate future interconnection integrate with the main grid in the future (Tenenbaum, of the mini grids. Greacen, and Vaghela 2018). In Bangladesh mini grids Where future integration of mini grids into the main grid must comply with main grid standards in rural areas, which is expected at least for one or more mini grids, regulators are lower than grid standards in urban areas. may consider optional grid-compatible standards. FIGURE 9.6 • Decision tree for regulating technical standards Technical standards regulation Integration of all Integration with grid Future integration mini grids with the expected for one or more with grid is not a main grid is essential mini grids and desirable policy issue Subsidies No subsidies Subsidies No subsidies Subsidies No subsidies available available available available available available Main grid Main grid standards are standards are economically not economically cost-bene t cost-bene t justi ed justi ed Main grid Mandatory grid No good Optional grid Optional grid Mini grid– Safety standards compatible solution compatible compatible speci c standards standards (+ equipment (+ equipment only and construction and construction quality) quality) Source: ESMAP analysis. 254   MINI GRIDS FOR HALF A BILLION PEOPLE There are five options for regulating the Mini grid regulations should explicitly state technical standards of mini grids: (1) min- the options for what happens when the main imum safety standards only, (2) mini grid–spe- grid arrives in the service area of a mini grid and cific standards, (3) optional main grid–compatible should provide as much clarity as possible about standards, (4) mandatory main grid–compatible how each option would be implemented. standards, and (5) main grid standards. The choice among these options depends on the availability of subsidies (higher standards typically require sub- Second, it is desirable that assets continue to be useful sidies) and what will happen when the main grid when the main grid arrives. If the distribution system is arrives in the service area of a mini grid (main grid– built to grid-compatible standards, it can be intercon- compatible standards are the minimum for enabling nected when the main grid arrives and continue to provide mini grids to interconnect with the main grid). service. The mini grid’s generator can provide power to the main grid or operate as a backup option. Where the integration into the main grid is an essential Mini grids can integrate with the main grid or remain dis- policy objective, and subsidies are available, regulators tinct from it. Options that provide for integration include may consider two options: mandatory grid-compatible becoming a small power distributor (SPD), a small power standards and main grid standards. The two options are producer (SPP), both SPD and SPP, and asset buyout. likely to require subsidies to enable mini grids to cover the Options where the mini grid remains distinct from the main costs at these higher standards. Thus, if integration into grid include coexistence of the main grid and mini grid and the main grid is essential and no subsidies are available, asset abandonment. Table 9.7 describes these options and there is no good option. suggests the contexts in which they are suitable. While it may seem clear in some cases that a mini grid will The following legal and regulatory requirements can help never be connected to the main grid—and hence requiring ensure that mini grid investors are protected when the it to achieve main grid standards would add unnecessary main grid arrives.6 costs—it is worth highlighting that grid-extension plans • In the event that the mini grid developer’s exclusivity can and do change, often for political reasons. As a result, is protected by a legal contract with the government, a an area that was not planned for main grid extension in 10 national utility, or another state-owned entity, the main years may suddenly find itself on the list of areas to receive grid arrival regulations should state clearly how the a main grid connection in 5 years. Ultimately, because mini developer will be compensated for its loss of exclusivity. grids have a useful life of 15–20 years, it may be reasonable to assume that nearly every mini grid built today should be • If mini grids are allowed to become SPDs, the retail tariff prepared for the arrival of the main grid. The next section must include a distribution margin that enables the SPD identifies the options that should be made available to mini to be commercially viable. grid developers upon the arrival of the main grid. • If mini grids are allowed to become SPPs, SPPs should be given the legal right to sell power at a predetermined REGULATING RELATIONSHIPS BETWEEN THE feed-in tariff, through a power purchase agreement. MAIN GRID AND MINI GRIDS Regulating relationships between the main grid and mini • If mini grids have the option to sell all or part of their grids can mitigate risks faced by investors while preserving assets to the main grid operator, this asset buyout the benefits of progressively expanding the main grid. Reg- option should require an objective, fair, and predictable ulation for this purpose is warranted for two reasons. asset valuation method. First, it can reassure developers who worry they may lose Small power distributor their investment when the main grid extends to a commu- A mini grid that converts to an SPD retains its distribution nity they serve because its risks the stranding of assets and assets and connects them to the main grid. The gener- expropriation. To provide certainty to investors, regulators ation assets are transferred to the utility, sold to a third should ensure (1) that options for the mini grid to inte- party, moved to another location, abandoned, or decom- grate with the main grid (or, as an alternative, to exit from missioned.7 In theory, the generation assets may also be the business) are clear; (2) that the rules governing these kept for backup power. The mini grid operator buys elec- options are not subject to regulatory or policy discretion; tricity wholesale from the utility and sells it to its retail and (3) that formulas for determining compensation pay- customers. ments are unambiguous. MINI GRIDS FOR HALF A BILLION PEOPLE    255 TABLE 9.7 • Options for preserving value when the main grid arrives Option Description Suitable where… Examples Small power Mini grid retains its distribution The distribution network is in good Cambodia (since 2001 mini grids distributor (SPD) assets (and becomes an SPD). condition and grid compatible, and are automatically converted to Generation assets are transferred the generator is not competitive SPDs, and diesel generators must be to the utility, retained as backup with generation on the main grid. decommissioned) generation, sold to a third party, There are enough retail customers Tanzania (SPD is one option for mini moved to another location, or for the SPD to operate profitably. grids when the grid arrives) decommissioned and abandoned. The SPD purchases electricity wholesale from the utility and sells it to its customers. Small power The mini grid’s distribution The generator is competitive with Uttar Pradesh (isolated mini grids can producer (SPP) network is transferred to the utility generation on the main grid or can convert to SPPs and sell power at a or abandoned. generate at a cost below the feed-in feed-in tariff determined at the state The SPP produces and sells power tariff. level; none had done so as of 2017) to the utility. The utility is creditworthy. Tanzania (one option available to mini Battery storage may be used to grids is to convert to an SPP) provide services. Bangladesh (mini grids must convert to SPPs) Small power Mini grid retains its distribution Mini grid distribution network is in Nigeria (mini grids have the option to distributor plus and generation assets. good condition and grid compatible, convert to SPD plus SPP) small power Mini grid meets its customers’ and generators are competitive Tanzania (mini grids have the option producer needs by buying from the main with generation on main grid or can to convert to SPP plus SPD) grid, running its generator, or both. increase reliability of service. Mini grid can sell its excess power The utility is creditworthy. to the main grid. Mini grid usually has the option of islanding (disconnecting from the main grid for technical or safety reasons). Asset buyout The utility purchases all or some Integration of mini grids by the Uttar Pradesh (mini grids have the of the mini grid’s assets and utility is a core objective of the option to sell their distribution assets operates, decommissions, or country’s electrification strategy. to the utility, at a cost determined abandons them. Mini grid assets that would be by the distribution company and the sold (that is, the distribution developer) infrastructure) are in good condition Nigeria (mini grids have the right and grid compatible. to sell some or all of their assets The national or regional distribution to the distribution company, at the company has the incentives and the depreciated value of the assets plus money to purchase the mini grid’s 12 months of revenue) assets. a Tanzania (mini grids have the option to sell their distribution assets to the connecting distribution utility at a price based on the distribution utility’s avoided cost) Coexistence of Mini grid’s distribution and Main grid provides low-quality Uttar Pradesh (mini grids are built main grid and generation assets are maintained. service or households cannot afford alongside the main grid in some mini grid Main grid and mini grid are connection to the main grid. areas) physically separate but operate in Distribution network is in good Indonesia (mini grids coexist with the the same service area. condition but not grid compatible main grid in 50 communities) Customers have the choice to (low-voltage DC networks). remain with the mini grid or be Mini grid offers a service/cost connected to the main grid. combination that suits some customers better than the main grid option. 256   MINI GRIDS FOR HALF A BILLION PEOPLE TABLE 9.7, continued Option Description Suitable where… Examples Asset Mini grid’s distribution assets, The government wants to protect Nigeria (mandatory for registered abandonment generation assets, or both are utility or distribution companies mini grids) abandoned, decommissioned, or from competition. Indonesia (150 of 200 mini grids removed and used elsewhere. The distribution network is not grid that faced main grid arrival were compatible. abandoned) Mini grid quality is below main grid Sri Lanka (100 of 250 mini grids mini standard, and it is not cost-effective grids built were abandoned) to upgrade it. Generation assets cost more to run than the marginal cost of main grid generation and are not useful for backup. Source: ESMAP analysis. DC = direct current. In Tanzania, donors pay for 100 percent of the capital costs if TANESCO extends the main grid into previously unconnected villages. At present, donors a.  do not provide TANESCO with any grant money to acquire the existing distribution assets of a mini grid in a village that has been served by a mini grid. Small power producer Coexistence In the SPP model, the mini grid distribution network is In this option, the mini grid operator retains its distribu- transferred to the utility, abandoned, or decommissioned. tion and generation assets. Customers have the option of The operator keeps the generator, along with any batteries, remaining connected to the mini grids or switching to the and sells the power generated to the utility. Depending on main grid. The two grids remain physically separate. This the technical design of the mini grid, the batteries can be option, while rare, exists in parts of India and Indonesia used to provide services, such as shifting power from times where the main grid provides unreliable power. it is less valuable to times it is more valuable.8 Asset abandonment Small power distributor plus small power producer Where the distribution network and the generator are In this model, the mini grid remains a mini grid, with the abandoned or decommissioned, the national or local utility additional options of selling and buying power from the treats the area like a greenfield site, building a new distri- main grid. The mini grid thus becomes both an SPD and a bution system. The regulator can also require the mini grid SPP (also called a grid-connected mini grid). Both the dis- operator to remove its distribution assets. The generation tribution and generation assets are connected to the main assets may be removed and used on another site if the grid. The mini grid operator retains the distribution grid and technology allows it. continues selling power to its customers. It also retains its generator and can meet its customers’ needs by buying Decision tree for integration and exit options when from the main grid, running its generator, or both. It can also the main grid arrives sell excess power to the main grid. The mini grid also has Figure 9.7 presents the decision tree for integration and the option of “islanding,” where the configuration in which exist options for mini grids when the main grid arrives. a portion of the grid becomes temporarily isolated from the If mini grids are built to grid-compatible standards (and if main grid but remains powered by its own generator. it is essential or desirable for the country’s electrification strategy to integrate mini grids into the main grid), the gov- Asset buyout ernment may consider the SPP, SPD, SPP + SPD, and asset In the asset buyout option, the utility purchases and oper- buyout options. If mini grids are not built at grid-compatible ates the mini grid’s assets. A variant includes a partial standards, the government may wish to allow mini grids to asset buyout, in which the utility buys out only the distri- coexist with the main grid if it does not have a policy objec- bution assets. In this case, the generating assets may be tive to protect distribution companies from competition; abandoned, decommissioned, or used at another site if otherwise, it should require the mini grids to decommission the technology allows it. The utility may be required to pur- their assets. chase the assets but not necessarily to operate them (in which case the utility abandons or decommissions them). If mini grid generators are not competitive with generation In Nigeria, for example, regulation does not require the util- on the main grid, the government may require mini grids ity to operate the assets (NERC 2016). MINI GRIDS FOR HALF A BILLION PEOPLE    257 FIGURE 9.7 • Decision tree for integration and exit options When the main grid arrives Mini grids are not built at Mini grids are built at grid-compatible standards grid-compatible standards Integration with Integration with the main grid the main grid is desirable is essential No policy objective Policy objective to Dominant Dominant to protect the protect the generation generation distribution distribution technology is technology not companies from companies from competitive competitive competition competition with generation with generation from the main grid from the main grid Coexistence Asset Option to Option to Distribution Option to Option to Asset Mini grids with the abandonment convert to convert to network convert to convert to buyout operate under main grid (possibly with SPP and SPP and buyout SPD small power option that contract (possibly with compensation SPD distribution option that distributor can be with the main compensation for loss of network can be (SPD) plus exercised grid from for loss of value) buyout exercised small power by main the start value) by the producer grid mini grid (SPP) Source: ESMAP analysis. to remove their generation assets and either allow them ity can buy their assets and that the valuation will be fair. convert to SPD or give them the rights to sell their distri- There are different ways to determine the appropriate bution assets to the main grid. If a mini grid’s generation is compensation due the developer under the asset buy- eligible for feed-in tariffs or is competitive with generation out option, but they generally fall into one of the following from the main grid, the government may offer the possibil- three categories: ity to the mini grid to convert to an SPP, with the option to Compensation based on assets: Under this approach, continue operating as an SPD (SPP + SPD option). the value of the mini grid would be based on the value For options that include asset buyout, the regulations of its assets after accounting for depreciation and any should offer a credible compensation mechanism to the grants the developer received for the assets. Apart from mini grid operators, who will want certainty that the util- the usual physical assets, a mini grid may also have soft 258   MINI GRIDS FOR HALF A BILLION PEOPLE assets such as goodwill, which in the mini grid context might result from the community-engagement efforts of Six options present themselves when the the mini grid developer. These tend to be difficult to valu- main grid arrives in a mini grid’s service area: ate on the one hand, but on the other are important assets (1) allow the mini grid to operate as a small power to the main grid operator: Entry into a service area with distributor (SPD), (2) allow the mini grid to operate existing demand for high-quality electricity services and as a small power producer (SPP), (3) allow the mini familiarity with for grid-based electricity. The compen- grid to operate as an SPD and SPP, (4) allow the mini sation mechanism in the current Nigerian mini grid reg- grid developer to sell its eligible assets to the utility, ulations is based on the “remaining depreciated value of (5) allow the mini grid to coexist with the main grid, assets” but also includes an additional payment equal to and (6) require the mini grid to decommission and the revenue generated by the mini grid in the 12 months remove its assets. The SPP, SPD, and SPP + SPD before the buyout. options require the mini grid to have been built at Compensation based on cash flow: This method is for- least to main grid–compatible standards. The asset ward looking and uses the discounted cash flow to calcu- buyout option requires a credible compensation late the present value of a business’s future cash flows, mechanism. including its terminal value. This method estimates future cash flows in terms of what a new business owner could achieve. It also recognizes the buyer’s cost of capital. It is often difficult to produce reliable estimates of future In the first and third approaches, assets would only be cash flows because markets in which mini grids naturally eligible for compensation if they have been maintained to operate are uncertain and unpredictable. The concession good working order and built to the appropriate technical contract for mini grids in Haiti includes a compensation standards. mechanism based on the greater of the book value of The second option would be particularly applicable if the assets less any grants received toward their purchase, developer’s exclusivity was granted contractually for a set or the net present value of the expected profit (revenues period of time—for example, through a concession, license, minus costs) for the remainder of the concession, using a or permit—and the main grid arrives before the expiration predefined discount rate. of the contract. Compensation based on utility’s avoided cost: Under this approach, the developer is compensated an amount equal COMBINING REGULATORY ELEMENTS to the main grid operator’s avoided cost of having to build grid infrastructure in the area served by the mini grid, after Unpredictability and inconsistency are the two primary accounting for depreciation and any grants that the devel- regulatory risks facing developers. To mitigate against oper received for the assets. Consequently, a requirement these risks, this section proposes ways, first, to combine for this approach is accurate data on the true, unsubsi- regulatory elements into effective packages so the regu- dized costs—typically per kilometer or per customer—of lations themselves are more consistent, and, second, to extending the main grid to rural areas. The mini grid reg- implement regulatory packages in a phased approach that ulations in Tanzania take this approach, stating that com- increases predictability for the sector. pensation is “based on the Rural Energy Agency’s average capital cost for installing distribution equipment in rural COMBINING REGULATORY ELEMENTS INTO areas measured on a cost per kilometer basis over a recent EFFECTIVE PACKAGES calendar or fiscal year minus depreciation” (Government Regulations for mini grids are interrelated. For example, of Tanzania 2019). service standards can be regulated when tariffs are regu- Under the first option, an independent third party can lated, and mini grid assets can be protected when the main conduct a valuation of the mini grid’s assets (splitting the grid arrives, conditioning connection on compliance with valuation costs between both parties). But a major point technical standards, which in turn can be regulated when of contention is likely to be which assets are to be sold to subsidies are provided. the main grid operator. In general, the main grid operator The options for each element can be combined in many should be required to purchase the immovable distribution ways, only some of which yield effective packages. Table infrastructure assets if they are determined to be in good 9.8 summarizes packages that regulators can customize working order and built to the correct standards. to fit their own conditions while ensuring that the results are consistent with the three core principles of regulatory systems: credibility, legitimacy, and transparency.9 MINI GRIDS FOR HALF A BILLION PEOPLE    259 TABLE 9.8 • Effective regulatory packages Package Description Suitable where . . . Examples Laissez-faire No regulation of entry, tariffs, service The government wants to minimize barriers to Cambodia (de facto standards, or technical standards beyond investment to promote rapid electrification through laissez-faire before those for safety. private mini grids. 2001) No provisions for when the main grid arrives. Mini grids compete with other energy providers Uttar Pradesh No subsidies provided. and traditional sources of power, like backup diesel (voluntary laissez- generators. faire) Government administrative and financial capacity are low. Government administrative and financial capacity are high or medium, but the scope for mini grids is not large or the government wants to allocate few regulatory resources to mini grids. Mini grids are community owned, making self- regulation sensible (Reiche, Tenenbaum, and Torres de Mästle 2006); the regulator could step in in case of political takeover. Light regulation Low-cost regulatory interventions that The government wants to minimize barriers to Nigeria (for mini offer clear value while minimizing the risk investment while ensuring safety, knowing which grids with less than of preventing mini grids on the edge of mini grids are in business, and providing certainty 100kW of generation commercial viability from developing. to investors as to what happens when the grid capacity) Requires mini grids to register, agree with arrives. Tanzania (for mini customers on tariffs, report service quality Mini grid developers are in competition with other grids with less than on specified service indicators, and follow energy providers. 100kW of generation safety standards. Government administrative and financial capacities capacity) Provides optional grid-compatible standards, are low. combined with provisions to convert to Government administrative and financial capacities small power distributors (SPD), small power are medium or high, but the scope for mini grids is producer ( SPP), or SPD + SPP; may provide not large or the government wants to allocate few model contracts for tariff setting. regulatory resources to mini grids. Central Comprehensive approach to regulation and Central coordination is appropriate where there Nigeria (centrally coordination subsidies, based on a centrally coordinated is a desire to rapidly expand service in the defined coordinated approach and approach where the government identifies areas. Individualized tariff control is suitable where with least-subsidy individualized areas suitable for electrification by mini grid, mini grids are no longer in competition with other tender planned) tariff defines coverage targets, and proactively sources of energy and have monopoly power. Madagascar (call for attracts operators to serve those areas. tenders) Requires mini grids to obtain a permit or Bangladesh license and to follow regulated service (information on where standards and grid-compatible standards mini grid development Regulates tariffs. is suitable and subsidy Provides options for when the main grid package) arrives. The government may provide information on where it is suitable for mini grids to operate and develop a package of regulation and subsidies that applies in such cases. Uniform tariff Pursues a vision of geographic equity in There is a uniform tariff policy. Cambodia (grid- electricity service. Expanding access rapidly is not a policy priority. connected SPDs, as Requires that all mini grids obtain a permit of 2017) There is a creditworthy subsidy provider, and/or or license, charge a uniform national tariff, cross-subsidies across a wide customer base are Kenya (plans for mini and follow regulated service standards and possible. grids to operate under grid-compatible standards. contract to national The government can make a credible commitment utility and charge the Provides energy subsidies and provisions for that the subsidy will be provided over the long when the main grid arrives. national tariff) term. Administrative capacity is high. Source: ESMAP analysis. kW = kilowatt. 260   MINI GRIDS FOR HALF A BILLION PEOPLE COMBINING REGULATORY ELEMENTS IN A promotes good service at the lowest cost-recovery tar- PHASED APPROACH iffs throughout the stages of development of mini grid sectors. Such a regulatory framework needs to be pre- The timing and extent of regulation are key to the develop- dictable and flexible enough to evolve. ment of mini grid sectors. Mini grid markets evolve as mini grids gain power in local energy markets as they mature. Defining different regulator phases for the short, medium, This maturing requires regulation to plan on adapting to and long terms is one way to provide investors with reg- new circumstances. ulatory predictability and certainty while allowing regula- tion to evolve. An example of a phased approach is shown • In a nascent mini grid market, light regulation may be in table 9.9. enough to achieve good service at the lowest possi- ble tariffs. Mini grids face competition from traditional Over two decades, Cambodia’s regulation evolved from sources of energy when they start operating in regions laissez-faire to a uniform national tariff for grid-connected where the main grid is unlikely to expand soon. This com- SPDs (box 9.2). Regulation was introduced in the early petition incentivizes mini grids to offer better service 2000s and was applied progressively, evolving along- at lower tariffs, which reduces the need for regulation. side the market. The proposed evolutionary framework When mini grids gain market power, heavier regulation approach would be different from Cambodia’s experience, may be required to ensure quality of service at the low- as the different phases of regulation would be defined from est possible tariffs. Without regulation, mini grids’ posi- the start to provide certainty and predictability to inves- tion might allow them to excessively raise their tariffs. tors. Cambodia’s path from unregulated tariffs to uniform The challenge is to design a regulatory framework that national tariffs is not the only path (for example, a country could go from unregulated to a price cap or an individual- ized cost-based tariff). The goal of a good regulatory framework for Thresholds to move from one phase to another should mini grids should be to promote good ser- be established. The transition from the starting phase to vice at the lowest cost-recovery tariffs throughout the medium-term phase could be triggered by any of sev- the stages of development of a country’s mini grid eral criteria, including the number of years since the first sector, taking into account subsidies and broader mini grid was registered and initial market penetration. national electrification strategy. As a result, a regula- The transition from medium-term to long-term regulation tory framework for mini grids needs to be predictable could be triggered by the same criteria, but the levels would yet flexible enough to evolve as the market evolves. be higher (for example, 15 rather than 5 years since the first mini grid registration). TABLE 9.9 • Three phases of evolutionary regulation Area Start Threshold 1 Medium term Threshold 2 Long term Entry Registration Five years after Registration Fifteen years after Permit/license first mini grid is first mini grid is Tariff Willing buyer–willing seller registered Efficient new-entrant registered Individualized cost- Number of price cap Number of customers based tariff customers served served reaches a Service Reporting reaches a certain Differentiated certain level (for Main grid-level standards level (for example, regulated standards example, 80 percent standards 40 percent of of total identified in Technical Safety standards total identified in a Safety standards a national geospatial Safety standards standards Optional grid-compatible national geospatial Optional grid- analysis) Optional grid- standards analysis) compatible standards Average power compatible standards Average power consumption per Subsidies a Implicit subsidies (for consumption per Capital and/or customer reaches Ideally not needed in the example, market or site customer reaches connection subsidies a certain level (for long term, but where information provided for a certain level (for example, 80 percent needed, connection free; land provided for free; example, 40 percent of average main grid subsidies or, in the case favorable tax treatment) of average main grid consumption) of mandatory uniform Capital subsidies consumption) tariffs, energy subsidies Source: ESMAP analysis. Note: The ability to transition from one regulatory package to another also depends on the administrative capacity of the regulator. Moving from left to right in the table requires an increase in the regulator’s capacity. a. See chapter 6 for a detailed discussion on subsidies. MINI GRIDS FOR HALF A BILLION PEOPLE    261 Evolutionary regulation requires capacity at the outset to REGULATORY INNOVATIONS TO develop the credible, legitimate, and transparent system that will apply in the future. To attract investment under FURTHER INCENTIVIZE PRIVATE- such a system, a government needs to credibly com- SECTOR INVESTMENT IN MINI GRIDS mit that it will abstain from intervening until the trigger Even when regulatory frameworks for mini grids are well is reached and will regulate as planned from then on. To designed and implemented correctly and consistently, provide such credibility, governments may consider using many investors will hesitate to commit millions of dollars partial risk guarantees—a form of insurance that provides of capital under a regulatory regime that rests solely on the the private party with compensation for the failure of the discretion of independent regulators. What additional tools, regulator (or other government agency) to implement the then, are available to provide confidence to investors, while regulation as written. BOX 9.2 FROM LAISSEZ-FAIRE TO COMPREHENSIVE REGULATION: CAMBODIA’S SUCCESSFUL ELECTRIFICATION WITH MINI GRIDS In the 1990s, mini grids emerged in Cambodia under lion loan from the World Bank. The REF implemented de facto laissez-faire conditions following the civil war several programs, including a $45 per connection and the Vietnamese occupation. By 2001 an estimated grant that ran from 2005 to 2012 and, starting in 150–300 diesel-fired mini grids provided power to 2013, a zero-interest loan to pay for households’ 72,000–182,000 customers. These mini grids provided connection fees along with viability gap finance and valuable services on a fully privately financed basis, funding for medium- and low-voltage lines. with no regulation of entry, tariffs, service, or techni- Under these circumstances, mini grids expanded elec- cal standards. They provided electricity to towns and tricity access, lowered costs, and improved the quality villages that would not otherwise have been supplied. of service. By 2016 at least 341 mini grids (including Although the charges of about $0.55/kWh were high grid-connected mini grids [SPDs] that have up to by the standards of other countries, customers were 16,000 customers) were serving 1.3 million custom- pleased to get service even at these prices. ers. Most mini grids interconnected with the main grid, In 2001 the Cambodian government started to regulate which allowed them to decommission their expensive the sector, taking a gradual approach. The Electricity diesel generators and buy bulk power from the grid. In Law passed in 2001 required all mini grids to obtain effect, they became SPDs. As a result, average tariffs a license, charge tariffs approved by the Electricity fell to $0.24/kWh in 2016. Authority of Cambodia (EAC), and meet service and Regulation progressively tightened. Since March technical standards. The government used the follow- 2016, the government has been implementing a uni- ing carrot-and-stick strategy to ensure compliance: form national tariff program for grid-connected SPDs. • Most mini grids came to the EAC to seek a license Licensees buying bulk power from the national grid spontaneously, possibly because the license must charge a tariff converging toward the uniform granted exclusive right over the service area. national tariff (except in Phnom Penh and Takhmao, where Electricité du Cambodge [EDC] has a different • EAC advised mini grid operators on compliance with tariff). This uniform tariff is below SPDs’ cost-recov- the standards and granted longer licenses to oper- ery tariff; the REF covers the difference with funds ators who showed improvements. It allowed tariffs provided by EDC. This ongoing subsidy amounted to to cover the investments required to upgrade licens- about $30 million in 2016. The EAC continues to cal- ee’s distribution systems. culate cost-recovery tariffs for individual SPDs, but the • In parallel, the Cambodian government distributed methodology has become more stringent, with SPDs’ subsidies through the Rural Electrification Fund distribution margins decreasing from $0.16/kWh in (REF), established in 2005 and financed by a $40 mil- 2011 to $0.05/kWh in 2016. Source: World Bank 2017b. 262   MINI GRIDS FOR HALF A BILLION PEOPLE also responding to unexpected events and developments? REGULATION BY CONTRACT And how can this be achieved while assuring customers Regulation by contract refers to a situation where the rules that their point of view is being considered and avoiding that govern tariffs (including mechanisms to adjust tariffs), the risk of public backlash after being seen as offering too services standards, and other conditions are embodied in much protection to investors? a contract.10 The rules and processes for adjusting terms Two innovative regulatory mechanisms can address these and conditions are also embodied in the contract. Regula- questions: tion by contract is often accomplished through rural elec- trification concessions, such as those in Madagascar, Mali, • Regulation by contract, in which regulatory decisions Senegal, and Uganda (World Bank 2015). Where contract are made subject to rules agreed to by both the devel- law is respected, but regulators have not yet developed the oper and the government; and track record needed to be credible to investors (or where • Arbitration-style appeal mechanisms, in which appeals formal regulatory agencies do not exist), regulation by con- against regulatory decisions are decided by a special tri- tract may be a good option to boost investor confidence. bunal constructed to balance the interests of the inves- Contracts cannot be unilaterally changed, which can make tor and the consumers. regulation by contract more credible than independent dis- These two mechanisms coexist alongside the different cretionary regulation. Independent regulators often have options discussed so far in this chapter. Regulation by con- broad discretion to change tariffs, or to change the rules tract usually includes provisions for market entry, tariffs, and formulas by which tariffs are set. Under regulation by and service and technical standards, as well as an agree- contract, these changes cannot be made unless both par- ment as to the level of subsidy to be provided if provided ties agree.11 “From a legal perspective the concession is a by one of the contract signatories. Arbitration-style appeal contract between the government and the regulated com- mechanisms can coexist in any regulatory environment. pany, the terms and conditions of which can only be altered Table 9.10 considers these mechanisms in turn. by mutual consent” (Brown 2010). The essential point that differentiates regulation by contract from independent dis- TABLE 9.10 • Two innovations in regulation that can further incentivize private-sector investment in mini grids Option Description Suitable where . . . Examples Regulation Rules that govern tariffs, services Contract law is respected, and contracts Nigeria (tripartite contract between by contract standards, and other conditions can be enforced quickly and fairly. developers, the local distribution are embodied in a contract Regulators do not exist or do not have company, and communities served, between a public party (such as the track record needed to be credible to for grid-connected mini grids) an REA, a municipality, or a local investors. Mali (rural electrification community) and the developer. concessions) There is a legal framework that enables The public party to the contract contracts between government and Madagascar (rural electrification may be the central government, an private companies for the supply of concessions) REA or fund, a local government, or public services. Senegal (rural electrification a customer collective. There is a centrally coordinated concessions) approach to mini grid development with Haiti (tripartite contract between competitive tenders. developer, local government, and There is administrative capacity on grant-giving government entity) both sides to negotiate the terms of the contract. Arbitration- Appeal: mechanism that enables Investors are concerned about the Jamaica (Appeal Tribunal provided style appeal investors to appeal a regulator’s quality of regulatory decisions. for in the license of the privately mechanism discretionary decisions to a neutral The cost of arbitration is not owned electricity utility) body. disproportionate compared with the Arbitration: appeal model based value of sums at stake. on a balanced composition of There is enough capacity on both sides the appellate body (each party to appoint arbitrators and manage the chooses a member and the third is appeal process. agreed upon by both parties). Judicial or administrative appellate processes are inadequate. Source: ESMAP analysis. REA = rural electrification agency. MINI GRIDS FOR HALF A BILLION PEOPLE    263 cretionary regulation is that under the first, public authori- the regulatory terms of the contract. A PRG is a form of ties cannot change the rules unilaterally. insurance that the government entity signing the contract will adhere to the terms in the contract. If the government A variant of regulation by contract occurs when one of the agency fails to do so, the PRG could compensate the pri- signatories is providing a subsidy. In Bangladesh, IDCOL, a vate party. Within the broader electricity sector, PRGs government-owned financial institution, provides loans and have been used to support regulatory contracts that grants to the mini grids in the country’s mini grid program. accompanied the successful privatizations of large elec- The financing agreement with IDCOL controls the tariffs tricity distribution enterprises in Uganda and Romania. and sets the service and technical standards of the mini The PRGs provided for financial compensation when the grids, providing a practical and flexible way to ensure that regulator failed to implement provisions of the contract, financing and regulatory parameters are aligned. It also such as tariff-setting provisions. The PRG contributed to avoids duplication by having a single agency responsible the success of these regulatory contracts. for both subsidy and regulation. This model does not yet, however, protect mini grids from independent discretion- One lesson learned from this experienced is that the trans- ary regulation because, legally, they still require licenses action costs of establishing a PRG to back up a regulatory from the Bangladesh Energy Regulatory Commission contract can be considerable. Therefore, PRGs are prob- (BERC) and may be subject to BERC’s regulatory control. ably feasible only for larger groups of mini grids selected Government policy is to remove mini grids from the licens- through a centralized competitive procurement, rather ing requirement, but this policy has not yet been made than for individual mini grids that come into existence legally effective. through spontaneous initiatives. Tripartite contracts are another variant of regulation by ARBITRATION-STYLE APPEAL MECHANISM TO contract. In Nigeria, grid-connected mini grids must sign COMPLEMENT A LIGHT-HANDED REGULATORY a tripartite contract with the distribution company whose APPROACH lines the mini grid is connected to, and the community it An arbitration-style appeal mechanism enables investors serves. The tripartite contract is designed to complement to appeal a regulator’s discretionary decisions to a neutral the formal mini grid regulations, and allows for the mini body that can overturn those decisions on grounds such grid to act as if it were a delegate of the licensed distribu- as bias, failure to follow due process, or noncompliance tion company, charged with improving service in an under- with stated regulatory principles and rules.12 “The appel- served area. The Nigerian Electric Regulatory Commission provided a template of a tripartite contract in Annex 11 of late process, at a minimum, is largely designed to accom- the 2016 mini grid regulations (NERC 2016). In Myanmar, plish three things: assure that the regulator adheres to and a model tripartite contract between the developer, the does not exceed its legal authority and powers, to protect government (Department of Rural Development), and against arbitrary exercise of its powers, and to assure that the village electrification committee has also been devel- all required legal processes were followed” (Brown 2010). oped. Since there is no national mini regulatory system in Of the three main types of appeal mechanisms—judicial or Myanmar, the tripartite contract is the de facto regulatory quasi-judicial, governmental, and arbitration—the arbitra- system for new mini grids (DRD 2017). Haiti developed a tion model can be used where there is a desire to insulate similar tripartite agreement between the developer, the regulatory decisions from politics and where there are con- local government, and the Ministry of Public Works, Trans- cerns that the judiciary is ill-equipped (Brown 2010). The portation, and Communications, which provides a subsidy essence of the arbitration model is the balanced compo- to the developer. In the absence of a formal mini grid regu- sition of the appellate body. Following standard principles latory framework, the template contract can both kickstart for commercial arbitration, each party choses one member the nascent mini grid sector and lay the groundwork for a of an appeals panel. Those two members then agree on a more formal framework based on how the sector evolves. third panel member, who becomes the chair of the appeals The template contract was initially drafted with support panel. There is usually a provision for a neutral party— from ESMAP in consultation with mini grid developers, presidents of the engineers’ association or law society—to municipal governments, and the grant-giving ministry appoint the chair of the appeals panel, if the first two mem- through a series of workshops. bers cannot agree. Contracts for mini grids could be bolstered by partial risk Once constituted, the appeals panel follows a process that guarantees (PRGs) issued by the World Bank or other is a similar to a court, but with more flexibility allowed—for international finance organization (Tenenbaum, Greacen, example, in consideration of evidence and use of experts. and Vaghela 2018). The PRG would create an added The panel may be given a broad authority, including setting incentive for the government or regulator to comply with aside a regulatory decision and requiring the regulator to 264   MINI GRIDS FOR HALF A BILLION PEOPLE retake it after correcting flaws identified by the appeals tration model. Although Jamaica’s electricity utility is much panel, or substituting the appeals panel’s own judgement larger than a mini grid—with a peak demand of 600 mega- for that of the regulator. watts—it provides a model that could be adapted to mini grids.13 In Jamaica, this appeals mechanism was important Jamaica has a well-developed system of independent reg- in creating the investor confidence needed to attract mil- ulation where decisions of the Office of Utilities Regulation lions of dollars of private finance for investment in electric- with respect to regulation of the privately owned electricity ity distribution and generation. The independent appeal utility may be appealed to a panel constituted on the arbi- tribunal in Jamaica is described in box 9.3. Two regulatory innovations have emerged INVESTORS’ PERSPECTIVE ON MINI to further incentivize private-sector invest- ment in mini grids: (1) regulation by contract, in GRID REGULATIONS which mini grid regulations are embodied in a legal Through conversations with private-sector investors and contract; and (2) an arbitration-style appeal mecha- building on the International Finance Corporation’s exten- nism, in which an independent entity or tribunal can sive experience investing in infrastructure projects around be asked to review a regulator’s decisions, with the the world, we have identified four characteristics of mini authority to overturn the regulator’s decision. grid regulations that investors perceive as barriers to investing in mini grids. BOX 9.3 INDEPENDENT APPEAL TRIBUNAL IN JAMAICA Jamaica has a well-developed system of independent one member shall be a former Judge of the (a)  regulation, where the regulator’s decision with respect Supreme Court or the Court of Appeal and shall to regulation of the privately owned electricity utility be the Chairman of the Tribunal: may be appealed to a panel constituted on the arbitra- one member shall be appointed on the recom- (b)  tion model. mendation of the Licensee; and The regulator (Office of Utilities Regulation, OUR) has one member shall be appointed on the recom- (c)  regulatory discretion with respect to regulation of the mendation of the Office. Jamaica Public Service Company (JPS), the privately Although the license mentions that the members of owned national electricity utility. JPS’s license, however, the Appeal Tribunal are “appointed by the Minister,” in contains detailed rules that spell out how tariffs are to effect one is nominated by the utility, one by the reg- be set in a way that limits discretion and resembles a ulator, and the deciding vote in case of disagreement regulatory contract. The licensee, JPS, has the right to rests with the third, who must be a retired high-rank- appeal a decision by OUR through a tribunal. The deci- ing judge. In Jamaica, the judiciary has a good repu- sion of the tribunal is binding for both parties (regulator tation for independence, which is partly why investors and licensee). accepted this formulation. This tribunal is similar to arbitration in the way that the This appeal mechanism was instrumental to the pri- members are appointed. Condition 32 of JPS’s electric- vatization of JPS, which needed guarantees that it ity supply license states as follows: would be able to charge reasonable tariffs. Ultimately, If the Licensee is aggrieved by a decision of, or failure it helped create the investor confidence needed to to act, by the Office, under this License, the licensee attract millions of dollars of private finance for invest- is allowed to appeal to the Appeal Tribunal (hereinaf- ment in electricity distribution and generation; accord- ter called “the Tribunal”) ing to JPS’s annual reports, the utility has invested The Tribunal shall consist of three (3) members more than $500 million over the past 10 years. appointed by the Minister as follows: Source: Jamaica Gazette 2016. MINI GRIDS FOR HALF A BILLION PEOPLE    265 • Unclear or overly burdensome market entry pro- for all relevant permits and approvals for all bureaucratic cesses: Unclear rules on licensing procedures and processes related to starting and operating a mini grid registration of assets is a major barrier to investment. business as a “nice to have” (though not a “must have”), In some countries, there is a lack of clarity on which which makes the mini grid sector more attractive as an licenses or permits are required, which entities issue investment opportunity. Additional streamlining can be them, and what the requirements and timing are for achieved by combining licenses and permits for genera- renewing them. This poses a significant risk from an tion, distribution, and electricity supply. investor’s perspective. Furthermore, a government • Clear procedures for tariff review and approval: Simpli- track record of complicated market-entry procedures fied and transparent tariff review, approval, and revision or lack of capability that has led to multiyear delays processes, as well as clear benchmarks or calculations in mini grid developers obtaining licenses or permits against which mini grid tariffs are reviewed, are “must would send a signal to many investors that the market have” elements of the tariff review and approval process is not fit for private-sector investment. for investors. An additional “nice to have” that investors • Politically motivated or opaque tariff regulations: If the have suggested is the availability of external assistance tariff-setting process is politically motivated or opaque, for the developer when it applies to the regulator for its it is difficult for the regulator to maintain support for tar- tariff schedule approval. iffs that are economically viable for mini grid developers, • Clear main grid expansion plans, combined with viable and it is difficult for developers to design business plans options for the developer when the main grid arrives: if they do not know how their tariff will be evaluated. This The existence of grid expansion plans, ideally with clear bias or lack of clarity introduces uncertainty around tar- explanations of how the plans were developed (for exam- iffs that would dissuade investors from financing mini ple, least-cost approach to electrification), combined grids. with economically viable compensation or a continuity • Unclear rules or lack of protection for the developer plan for when the main grid arrives, are “must haves” if the main grid arrives: If the mini grid regulations do for investors. In addition, investors will be looking at not give economically viable options to mini grid devel- whether the grid-arrival options, including a compensa- opers for what happens when the main grid arrives, tion mechanism, cover the developer over a long period. investors are not likely to make long-term investments • Compensation mechanism for government breach of in mini grids. This risk is compounded if there is a lack of contract: Any contract the developer enters into with transparency on where and when the national grid will the government or the utility should include a compen- be extended, which makes it challenging for developers sation mechanism if the government or utility defaults to identify potential project sites, and for investors to on its contractual obligations. commit to investing in the sector. • Lack of protection from government breach of con- tract: Legal contracts play an important role in reg- CONCLUSION AND RESOURCES ulations, not only in countries that opt for a regula- tion-by-contract approach, but also for regulatory In this chapter, we have presented the different ways that regimes that have license agreements or other contrac- regulators in many countries have dealt with five key reg- tual arrangements between the government and the ulatory decisions. We suggested ways to combine different developer. If there is a need for the developer to enter approaches into coherent regulatory packages, taking into into a contract with the government or the utility, a lack account how a mini grid sector might evolve over time. We of protection for the developer if the government or util- also presented two innovative regulatory mechanisms that ity defaults on its contractual obligations would be a red can further incentivize private-sector investment in mini flag for potential investors. grids. Finally, we presented an investor’s perspective that highlights the main barriers for mini grid regulations and The above four barriers are balanced by the following four also the “must haves” to attract private-sector investment. “must haves” in a regulatory regime necessary to attract external investment into the mini grid sector: No single approach to regulating mini grids works best in all settings. But when developing regulations for the first time, • Adequate capacity by the regulator to implement its changing regulations, and implementing them, the overar- regulations: Investors require at least a minimum thresh- ching approach should be light-handed. This would entail old (though this threshold is not well defined) of regulator the following four principles (Tenenbaum and others 2014): capacity to be able to carry out its duties, such as review and approval of tariffs; issue licenses; and respond to • Minimize the amount of information required by the reg- complaints. Investors have also cited a “one-stop shop” ulator 266   MINI GRIDS FOR HALF A BILLION PEOPLE • From the Bottom Up: How Small Power Producers and Mini-Grids Can Deliver Electrification and Renewable The overarching approach to regulating mini Energy in Africa (Tenenbaum and others 2014) is a grids should be light-handed, which entails guide focusing on regulatory and policy decisions that (1) minimizing the amount of information required African electricity regulators and policy makers must by the regulator, (2) limiting the number of separate make to create a sustainable decentralized track to regulatory processes and decisions, (3) standardiz- electrification. ing documents and forms, and (4) acknowledging • “Electrification and Regulation: Principles and a Model and using related decisions made by other govern- Law” (Reiche, Tenenbaum, and Torres de Mästle 2006) ment or community bodies. defines regulation, sets out general principles to apply, and offers a typology of electricity supply models and a model electrification law. • Limit the number of separate regulatory processes and decisions • Standardize documents and forms ACKNOWLEDGMENTS FOR THE • Acknowledge and use related decisions made by other SIX-COUNTRY CASE STUDIES THAT government or community bodies. INFORMED THIS CHAPTER Working together, regulators and policy makers must The authors would like to thank the peer reviewers, Emily select the regulatory options that are appropriate for their Chessin and Jakob Schmidt-Reindhal, for their useful rec- specific needs and that can attract private-sector invest- ommendations, and all of the stakeholders interviewed to ment in mini grids to ensure timely and low-cost access to prepare the six case studies on which this chapter draws. electricity. The following resources provide helpful guidance for devel- BANGLADESH oping mini grid regulations: ESMAP wishes to thank Monowar Islam, of the Bangladesh Energy Regulatory Commission, and Md. Enamul Karim • Mini Grids and Arrival of the Main Grid-Lessons from Pavel and Ms. Farazana Rahman, of the Infrastructure Cambodia, Sri Lanka, and Indonesia (Tenenbaum, Development Company Limited (IDCOL), who shared their Greacen, and Vaghela 2018) is a comparative analysis of knowledge and contributed insights into Bangladesh’s mini regulatory, commercial, and technical characteristics of grid framework and experience. ESMAP is also grateful to mini grids before and after the main grid arrived in vil- Chris Purcell of Catalyst Off Grid Advisors for his important lages previously served by isolated community-owned contributions; Mahfuzur Rahman, of IDCOL, for organizing and privately owned mini grids. This report also high- the team’s visit to mini grid sites; Fuad Sakib, of IDCOL, lights recent technical, commercial, and regulatory for showing the team the sites; and the SuperStar Group, developments affecting mini grids in different countries. SolarGao, and Solar Electro Bangladesh Ltd., for providing • Practical Guide to the Regulatory Treatment of Mini- insights on their experience in developing mini grid projects grids (USAID 2017) provides guidance for creating an in Bangladesh. enabling regulatory framework for mini grid develop- ment on the issues of mini grid policy and planning, CAMBODIA retail service regulation, and technical standards. While ESMAP thanks H.E. Dr. Ty Norin, Badri Rekhani, Teng Sok- this report mostly focuses on regulatory decisions once homal, and H.E. Yim Viseth of the Electricity Authority of policy decisions have been made, it offers guidance on Cambodia, who shared their knowledge and insights on policy decisions and how regulators can play a role in Cambodia’s mini grid framework and experience. ESMAP these decisions. It also recommended steps for policy also benefitted greatly from substantive inputs from senior makers and regulators once decisions have been made. officials and managers in government agencies and organi- • Policies and Regulations for Private Sector Renewable zations whom they met during their field trip, including Dr. Energy Mini-grids (IRENA 2018) gives practical advice Loeung Keosela, of the Rural Electrification Fund Depart- on policy and regulation design, covering licensing, tar- ment of Electricité du Cambodge, and H.E. Victor Zona, iffs, arrival of the main grid, and access to finance for of the Ministry of Mines and Energy. The case study was different technologies and levels of service, based on enriched by reviews and contributions from experts in the a wealth of examples from Africa, South America, and power sector in Cambodia, including Rin Seyha, Un Roeurn, Southeast Asia. and Sophal. ESMAP is also grateful to mini grid operators MINI GRIDS FOR HALF A BILLION PEOPLE    267 Keo Nguon and Chaek Sean for organizing the team’s visit the Rural Electrification Fund. The case study was enriched to mini grid sites and providing insights on their experience by reviews and contributions from developers and experts in developing projects in Cambodia. in the power sector in Nigeria, including Bolade Soremekun, of Rubitec Solar; and Dolapo Kukoyi, of Details Commercial UTTAR PRADESH (INDIA) Solicitors. ESMAP is grateful to mini grid developers Ifeanyi ESMAP thanks Vikas Chandra Agarwal, S.K. Agarwal, and Orajaka, of GVE Projects, and Okenwa Anayo Nas, of Nayo Sanjay Srivastava of the Uttar Pradesh Electricity Regula- Tropical Technology, for organizing the team’s visit to mini tory Commission, and Shruti Deorah, of the Central Elec- grid sites and providing insights on their experience in tricity Regulatory Commission, who shared their knowledge developing mini grids projects in Nigeria. and insights on Uttar Pradesh’s mini grid framework and TANZANIA experience. ESMAP benefitted greatly from substantive inputs from senior officials and managers in government ESMAP thanks Eng. Godwin Samwel and Eng. Massawe, agencies and organizations met during the field trip, includ- of the Energy and Water Utilities Regulatory Authority ing Disha Banerjee, of the Smart Power for Rural Develop- (EWURA); Advera F. Mwijage, of the Rural Energy Agency; ment/Rockefeller Foundation; Deepak Gupta and Saloni and Patrice Tshaklara, of the Tanzania Electric Supply Sachdeva, of the Shakti Foundation; and Sangeeta Singh, Company Limited (TANESCO), who shared their knowl- of the Uttar Pradesh New & Renewable Energy Develop- edge and insights on Tanzania’s mini grid framework and ment Agency. The case study was enriched by reviews experience. The case study was enriched by reviews and from experts in the power sector in Uttar Pradesh, includ- contributions from experts in the power sector in Tanzania, ing Debajit Palit, of the Energy and Resources Institute; and including Anastas Mbwala. ESMAP is grateful to mini grid contributions from mini grid developers, including Ketan developers Joanis Holzigel, of Rafiki Power, and Prosper Bhatt, Rohit Chandra, Sarraju Narasinga Rao, and Anil Raj Magali, of Ensol, for organizing the team’s visit to mini grid of OMC Power; Nikhil Jaisinghani and Sandeep Pandey, of sites, and to Khalfan Hussein, of Rafiki Power, and Alicia Mera Gao Power; and Pramod Singh, of Tata Power. ESMAP Rutajumbukilwa, of Ensol, for showing the team the mini is grateful to OMC Power for organizing the team’s visit to grids in Changombe and Mpale, respectively, and providing mini grid sites and providing insights on its experience in insights on their experience in developing mini grids proj- developing mini grids projects in Uttar Pradesh. ects in Tanzania. KENYA ESMAP thanks Silas Chebo, Leah Jara, Caroline Kimathi, REFERENCES Patrick Kungu, and Eng. Joseph Oketch, of the Energy Regulatory Commission; Eng. Isaac Kiva, of the Ministry of Baring-Gould, I., K. Burman, M. Singh, S. Esterly, R. Mutiso, and C. McGregor. 2016. Quality Assurance Framework for Mini-Grids. Energy; and Eng. Ephantus Kanweru and Francis Mutua, of Technical Report No. NREL/TP-5000-67374. Golden, CO: National the Rural Electrification Agency, who shared their knowl- Renewable Energy Laboratory. https:/ /www.nrel.gov/docs/fy17o- edge and insights on Kenya’s mini grid framework and sti/67374.pdf. experience. The case study was enriched by reviews and Bhatia, Mikul, and Niki Angelou. 2015. Beyond Connections: Energy contributions from experts in the power sector in Kenya, Access Redefined. ESMAP Technical Report 008/15. Washing- including David Mwangi and Benard Muok. ESMAP is grate- ton, DC: World Bank. https:/ /openknowledge.worldbank.org/han- ful to mini grid developer Stephen Nakholi, of RVE.Sol, for dle/10986/24368. organizing the team’s visit and showing its members mini Brown, A. 2010. “Infrastructure: The Regulatory and Institutional Dimension.” In Services, Development, and Trade: The Regulatory grids in Sidonge, and to David Hirsh, of Powerhive, for pro- and Institutional Dimension of Infrastructure Services, Vol. 1, edited viding insights on his experience in developing mini grid by Mina Mashayekhi, 71–90. New York: United Nations Conference projects in Kenya. on Trade and Development. Brown, Ashley C., Jon Stern, and Bernard Tenenbaum. 2006. Handbook NIGERIA for Evaluating Infrastructure Regulatory Systems. Washington, DC: ESMAP thanks Dr. Musiliu O. Oseni and Dr. Abdussalam World Bank. https:/ /openknowledge.worldbank.org/handle/10986/ 7030. Yusuf, of the Nigerian Electricity Regulatory Commission, DRD (Department of Rural Development). 2017. “Form NEP-9: who shared their knowledge and insights on Nigeria’s mini Developer Application for Mini-Grid Screening.” https://www.ger- grid framework and experience. ESMAP benefitted greatly man-energy-solutions.de/GES/Redaktion/DE/PDF-Anlagen/ from substantive inputs from senior officials and managers Marktnachrichten/20170922-myanmar-1.pdf?__blob=publication- in government agencies and organizations met during the File&v=3. field trip, particularly Damilola Ogunbiyi and Lolade Abiola, EAC (Electricity Authority of Cambodia). 2007. “Regulations on General of the Rural Electrification Agency; and Dr. Sanusi Ohiare, of Principles for Regulating Electricity Tariffs in the Kingdom of Cambo- 268   MINI GRIDS FOR HALF A BILLION PEOPLE dia, ss 5(10) and 5(12).” Phnom Penh. https://policy.asiapacificen- Tenenbaum, Bernard, Chris Greacen, Tilak Siyambalapitiya, and James ergy.org/sites/default/files/Regulation%20on%20General%20 Knuckles. 2014. From the Bottom Up: How Small Power Producers Principles%20For%20Regulating%20Electricity%20Tariff%20 and Mini-Grids Can Deliver Electrification and Renewable Energy in in%20the%20Kingdom%20of%20Cambodia_2007.pdf. Africa. Washington, DC: World Bank. https:/ /openknowledge.world- ERC (Energy Regulatory Commission). 2016. “Kenya National Distribu- bank.org/handle/10986/16571. tion Code.” https://www.kengen.co.ke/sites/default/files/Kenya% USAID (United States Agency for International Development). 2017. 20National%20Distribution%20Code.pdf. Practical Guide to the Regulatory Treatment of Mini-Grids. Washing- ESMAP. 2022. “Regulatory Indicators for Sustainable Energy (RISE) ton, DC: World Bank. https://pubs.naruc.org/pub/E1A6363A-A51D- Scores.” https://rise.esmap.org/ 0046-C341-DADE9EBAA6E3. EWURA (Energy and Water Utilities Regulatory Agency). 2017a. “The World Bank. 2015. Evaluation of Rural Electrification Concessions in Electricity (Development of Small Power Projects) Rules 2017, s 32(4) Sub-Saharan Africa Detailed Case Study: The South African Solar and s.41.” Dodoma, Tanzania. http:/ /www.ewura.go.tz/wp-content/ Home. Washington, DC: World Bank. http:/ /documents.worldbank. uploads/2015/03/The-Electricity-Development-of-Small-Pow- org/curated/en/708211498162775424/pdf/116659-WP-PUBLIC- er-Projects-Rules-2017.pdf. P150241-37p-Detailed-Case-Study-South-Africa.pdf. EWURA. 2017b. “Tariff Application and Rate Setting Rules 2017, s.18 World Bank.2017a.“Mini Grids in Bangladesh:ACase Study of an Incipient and Schedule 1.” Dodoma, Tanzania. https:/ /www.ewura.go.tz/ Market.” ESMAP Technical Paper, World Bank, Washington, DC. http:/ / wp-content/uploads/2017/12/EWURA-Tariff-Application-and-Rate- documents.worldbank.org/curated/en/669331512390210193/ Setting-Rules-2017-GN-452.pdf. Mini-grids-in-Bangladesh-a-case-study-of-an-incipient-market. GIZ (Deutsche Gesellschaft für Internationale Zusammenarbeit). 2015. World Bank. 2017b. “Mini Grids in Cambodia: A Case Study of a Suc- How Do We License It? A Guide to Licensing a Mini-Grid Energy Ser- cess Story.” ESMAP Technical Paper, World Bank, Washington, DC. vice Company in Kenya. Eschborn and Bonn: GIZ. https:/ /www.giz. https://openknowledge.worldbank.org/handle/10986/29019?- de/en/downloads/GIZ2015-ProSolar-Licensing-Guidebook.pdf. show=full. Government of Tanzania. 2019. “The Electricity (Development of Small Power Projects) Rules, 2019.” Government Notice No. 280. http:/ / www.ewura.go.tz/wp-content/uploads/2019/05/The-Electrici- NOTES ty-Development-of-Small-Power-Projects-Rules-2019-GN-280.pdf. IDCOL (Infrastructure Development Company Limited). 2015. “Financ- 1. The most frequently used statistical definition of an outlier is ing Agreement between Super Star Renewable Energy and IDCOL.” any data point that lies outside a range defined by 1.5 times the IDCOL, Dhaka, Bangladesh. “inter-quartile range” (IQR), or the difference between the 1st and 3rd quartiles. IRENA (International Renewable Energy Agency). 2018. Policies and Regulations for Private Sector Renewable Energy Mini-grids. Abu 2. Collecting payment on a regular basis is often a challenge in the Dhabi: IRENA. https:/ /www.irena.org/publications/2018/Oct/Poli- early stages of mini grid operations, after the initial enthusiasm cies-and-regulations-for-renewable-energy-mini-grids. has settled in the community. This is mainly due to two reasons: (1) consumers are not used to regular expenses of this kind, and (2) Jamaica Gazette. 2016. “Jamaica Public Service Company Limited, Elec- the cost per kWh is a parameter that is not easily compared with tricity License 2016, License No. 19A, Condition 32.” CXXXIX (19a). the cost of traditional energy services, but it may be easily com- https://www.our.org.jm/ourweb/sectors/jps-electricity-licence- pared with the kWh tariff of the main grid. Information from email 2016-published-ja-gazette-wed-2016-jan-27-0. exchange with mini grid expert from INENSUS (mini grid devel- Lazard. 2018. “Levelized Cost of Storage Analysis.” https://www.lazard. oper), June 2018. com/perspective/levelized-cost-of-energy-and-levelized-cost-of- 3. The methodology would be to (1) create a stylized efficient mini storage-2018/. grid by estimating the financial and technical parameters, includ- NERC (Nigerian Electric Regulatory Commission). 2016. “Regulation ing generation capacity, planned outage rate, capacity factor, heat for Mini Grids 2016.” NERC, Abuja, Nigeria. http://www.nercng.org/ rate (for thermal plants), plant life, construction period, total capital index.php/library/documents/Regulations/NERC-Mini-Grid-Regu- cost (generation and distribution), fixed and variable operating and lation/. maintenance cost, and the weighted average cost of capital; and (2) NERC. 2017. “MYTO Mini Grid Model.” NERC, Abuja, Nigeria. set the tariff following the approach for an individualized cost-based tariff limit. Reiche, Kilian, Bernard Tenenbaum, and Clemencia Torres de Mästle. 2006. “Electrification and Regulation: Principles and a Model Law.” 4. The line rent is set at Tk 120–300 ($1.46–$3.60) a month, depend- Energy and Mining Sector Board Discussion Paper 18, World Bank, ing on the type of consumer (IDCOL 2015). Washington, DC. http:/ /siteresources.worldbank.org/INTENERGY/ 5. In Sierra Leone the tariff is one of the criteria in the multiattribute Resources/EnergyPaper18.pdf. process, but it remains to be seen whether the tariff approved by Republic of Kenya and World Bank. May 2016. “Current Activities the regulator will be set at the winning bid tariff. The bid product and Challenges to Scaling Up Mini Grids in Kenya.” World Bank, will depend on the business model in the public-private partnership Washington, DC. https:/ /www.esmap.org/sites/esmap.org/files/ arrangement (it may be an engineering, procurement, and con- DocumentLibrary/ESMAP_Kenya%20Roundtable_May%202016_ struction contract, as in Kenya, or a build-own-operate contract (as formatted-v4.pdf. in Nigeria) (Tenenbaum, Greacen, and Vaghela 2018). Tenenbaum, Bernard, Chris Greacen, and Dipti Vaghela. 2018. Mini- 6. A recent ESMAP technical report presents case studies for three Grids and the Arrival of the Main Grid: Lessons from Cambodia, Sri Asian countries (Cambodia, Indonesia, and Sri Lanka) of what Lanka, and Indonesia. ESMAP Technical Report 013/18. Washing- happened to existing community-owned and privately owned mini ton, DC: World Bank. https:/ /openknowledge.worldbank.org/han- grids when the main grid arrived at their villages (Tenenbaum, dle/10986/29018. Greacen, and Vaghela 2018). MINI GRIDS FOR HALF A BILLION PEOPLE    269 7. Abandonment means that the assets are left in place as is. Decom- the company is subject, definition of territory to be served, length missioning means the assets are entirely or partly dismantled and of concession, capital investment expectations, potential liabilities no longer functional. and other risk exposures, and a host of other requirements (Brown 8. Battery technology at mini grid scale costs about $0.25 to $0.40/ 2010). kWh, which may not be cheaper for the utility than running peaking 11. Rules in a contract might specify how the tariff could adjust over time, plants (Lazard 2018). but the base tariff or the formulas by which the tariff is set would not 9. According to Brown, Stern, and Tenenbaum (2006), regulatory be changed unilaterally. systems are credible when investors “have confidence that [it] will 12. This section draws on regulatory theory and examples in the electric- honor its commitments,” legitimate when consumers are “con- ity sector. No examples in the mini grid sector have been found, given vinced that [it] will protect them from the exercise of monopoly the relatively nascent state of mini grid regulation. power,” and transparent when “investors and consumers ‘know the No examples of a fully developed arbitration-style appeal mecha- 13. terms of the deal.’” nism were found in the mini grid sector, although arbitration clauses Concessions usually include such terms and conditions as the price 10. are relatively common in power purchase agreements (for example, paid for the concession, pricing methodology, provisions on when in Tanzania’s standardized power purchase agreement) and other and how tariffs might be changed, monopoly or nonmonopoly sta- contracts between a developer and the government (for example, in tus, service expectations, degree of regulatory discretion to which Haiti’s concession agreement). 270   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 10 CUTTING RED TAPE FOR A DYNAMIC BUSINESS ENVIRONMENT CHAPTER OVERVIEW This chapter presents four mechanisms that can help develop an enabling business environment for mini grids: (1) reducing red tape through standardized agreements and processes; (2) using technology platforms to connect developers with investors and suppliers in large-scale mini grid tenders; (3) eliminating duplication of government oversight by delegating authority to a single entity (or a formal regulator); and (4) setting up e-government ser- vices to reduce the overhead cost of business registration, land and building permits, and environmental approv- als. All four mechanisms can be deployed in parallel, and each has the potential to make it significantly easier, and less risky, to do business as a mini grid. WHY AN ENABLING ENVIRONMENT MATTERS Mini grid developers often must navigate a complex and murky path of permits, regula- Private-sector players—both investors and developers— tions, and contracts as they build and operate their operate within business-enabling environments that give mini grids. Sometimes more than 20 clearances them confidence to invest. They will be putting their money are needed before a company can start operating into assets with 10- to 20-year lives and will earn a return a business. From an investment standpoint, each on their investment only if the business of selling electricity additional approval, review, permit, license, or other is profitable. However, in our conversations with investors bureaucratic process constitutes a risk—and there- and developers, a common theme emerges: “We’re not fore a cost—to the business. willing to commit millions of dollars of capital to mini grids in countries where it’s just too difficult to set up and run a mini grid business.” Mini grid developers must navigate multiple layers of ticularly important in countries with high electricity-access bureaucracy as they build, own, and operate their mini deficits. grids. From an investment standpoint, each additional This chapter offers four ways to increase the ease with approval, review, permit, license, or other bureaucratic pro- which mini grid developers can set up and operate their cess constitutes a risk to the business. Sometimes more businesses. than 20 clearances are needed before a company can start operating a business. Simplified and standardized • Reduce red tape through standardized agreements and approaches to overseeing mini grid businesses would help processes. reduce risk—and overhead costs—for developers. The end • Use technology platforms to connect developers with result would be more mini grids providing more access to investors and suppliers and to stage large-scale tenders affordable electricity for large groups of people who would to reduce inefficiencies in the market. otherwise wait years for a main grid connection. This is par- MINI GRIDS FOR HALF A BILLION PEOPLE    27 1 • Eliminate overlaps in government oversight by delegat- tors require that the rules governing their business be sta- ing authority to a single entity, whether or not that entity ble and predictable. is a formal regulator. UNPREDICTABILITY AND THE NEED FOR • Set up e-government services to reduce overhead asso- FLEXIBILITY ciated with business registration, land and building per- mits, and environmental approvals. Unpredictable changes in demand, technological innova- tion, and other factors will have significant effects on the All four mechanisms can be deployed in parallel, and each profitability of mini grids over their lifetime. For example, has the potential to make it significantly easier, and less demand for electricity grows as consumers gain access to risky, to do business as a mini grid. it. When existing customers start to consume more energy, The four strategies presented in this chapter do not the average generation cost per kilowatt-hour (kWh) will go address all of the bureaucratic processes that mini grid down. If the tariff stays at the same level, the mini grid will developers must navigate, many of which are described in earn returns above its cost of capital. earlier chapters. However, the options presented here can The costs of many mini grid components are also decreas- be implemented in the near term using existing resources ing over time (see chapter 1). New battery technologies and institutional structures. They would complement most will bring down the cost of supplying power at night. Other other initiatives described earlier, including those aimed at factors—such as exchange rate depreciations, unfavorable reducing the risks of corruption and expropriation. changes to the tax code, or extreme weather events that Each of the four strategies presented in this chapter damage systems and require expensive rebuilds—push requires a certain amount of government capacity to costs in the opposite direction. implement. We address this issue throughout the chapter If policies, taxes, and rules and regulations for doing busi- and refer readers to chapter 7, which covers training and ness kept pace automatically with exogenous factors like skills building. those mentioned above, investors could have complete certainty about how their investment would be affected. In practice, an unknowable number of factors may change CHARACTERISTICS OF DOING over time, and political and regulatory decisions can nei- BUSINESS AS A MINI GRID ther keep pace with nor account for all of them. Thus, at least some of the current rules and policies that define the Mini grids operating in low-income countries have several business environment will lag behind the actual conditions unique characteristics that affect how they interact with, experienced by businesses. In other words, no contract is and are affected by, the business environment. ever complete. This phenomenon has been called the con- LONG-LIVED SUNK ASSETS tractual incompleteness problem. Parties can write a solid contract now, but it cannot possibly account for every As in other infrastructure sectors, most of the investment eventuality (Brown and others 2006, 114). As a result, in a mini grid business—particularly when powered by developers need a business environment that is conducive renewable sources of energy—goes into assets with high to operating flexibly, so that they can adapt—taking advan- upfront capital costs, long lifetimes, and typically no alter- tage of changes that drive down their cost of electricity native uses. That is, most mini grid businesses have high and absorbing shocks in a way that does not put them out sunk costs. of business. The importance of sunk costs becomes clear when com- THE POLITICAL NATURE OF ELECTRICITY paring mini grids’ capital assets with assets that are mobile, such as those of a bus operator. Buses can be sold at full In many countries where large portions of the population market value to bus companies in another jurisdiction if the do not have access to electricity, electricity is a political business environment suddenly became unfavorable. This tool. Those seeking to win favor from the public will prom- is not the case for a mini grid operator. While some assets— ise to provide or reduce the price of electricity (Bakovic, such as containerized battery systems or solar photovol- Tenenbaum, and Woolf 2002, 12–14). If reducing the price taic (PV) modules—can be moved and sold, others, such of electricity led to shortages of electricity, the problem as the distribution grid or a mini hydropower plant, can- would self-correct: no one votes for power cuts. However, not. The only way an investor can recover its investment in because most of the costs of a mini grid are sunk, mini sunk assets, and earn a return, is by earning revenue over grids can continue to operate (for a while), even at tariffs the asset’s life (except in special cases when the main grid far below their average cost of service. Cutting tariffs is a arrives, as discussed in chapter 9). For this reason, inves- tempting target for politicians. Investors in the mini grids 27 2   MINI GRIDS FOR HALF A BILLION PEOPLE would lose their investment, but the political calculus is that investors’ votes are few compared with the many votes of electricity consumers. Mini grids operating in low-income countries have several unique characteristics that This path is particularly tempting in communities that will- underpin how they interact with, and are affected ingly agreed to a tariff before they had electricity and move by, the business environment. To start with, most of quickly from being delighted just to have electricity to a the investment in a mini grid business—particularly phase of being pleased to have electricity but complain- when it is powered by renewable sources of energy— ing about its cost. In the village of Bisanti, in Nigeria, after goes into assets with high upfront capital costs and receiving electricity for the first time, inhabitants proudly long lifespans. Changes in demand, technological told visitors about how they were the only place in Nigeria innovation, policies, taxes, and rules and regulations to have had uninterrupted power supply for two years and are often unpredictable. Finally, electricity is a politi- how pleased they were to have electricity. All customers cal tool, so political expediency risks compromising interviewed quickly moved on to explain that the cost of economic viability. These factors make it even more power was too high and that a way had to be found to bring important that mini grid developers operate in an the costs down.1 environment that is conducive to starting and oper- There are enough episodes of political pressure creating ating a business. unfavorable business environments for mini grid compa- nies in low-income countries that investors are wary of this risk. • In Cambodia the regulator has been tightening the regulator to remove the existing stipulation that feed-in calculation of full-cost-recovery tariffs, reducing distri- tariffs would not be changed retroactively. This proposal bution margins for grid-connected small power distrib- was not implemented, but the example illustrates that utors (Tenenbaum, Greacen, and Vaghela 2018). investors are reluctant to invest wherever they fear that the rules could be changed after they have sunk their • In Nigeria all mini grids built and planned prior to 2017 investment.4 were “registered” (which allows for unregulated tariffs), and no investor opted for a “permit” (which comes with Some elements of the business environment—notably a regulation of tariffs and service standards, as well as strong legal system that includes competent, independent protection of assets when the grid arrives). This pattern judicial oversight or review—can mitigate these political suggests that investors were waiting to see whether the risks (Brown 2012). In countries that lack effective rule of business environment—and in particular, the additional law and judicial protections for investors, other mecha- regulation that comes with receiving a permit—would nisms are needed. These need to be credible and to give allow them to earn the returns they need.2 One Nigerian investors confidence that commitments will be honored mini grid operator commented that political risks were and that they will be dealt with fairly. the biggest threat. He has invested in containerized gen- eration and battery storage units in part so that he can remove these units if political factors affect the viability FOUR COMPLEMENTARY OPTIONS of his business.3 TO MAKE IT EASIER FOR MINI GRID • In Tanzania several investors interviewed expressed reluctance about investing because of concerns about DEVELOPERS TO DO BUSINESS how the regulation would be enforced or fears that This section presents four complementary options that rules might change after they invested. One developer can make it more attractive for investors and developers to expressed concerns about the lack of clarity on how the do business. They emerged from in-depth case studies of Energy and Water Utilities Regulatory Authority would mini grids in Bangladesh, Cambodia, Kenya, Nigeria, Tanza- enforce tariff regulation. Others expressed concerns nia, and Uttar Pradesh, as well as from World Bank experi- about how the authority might use its discretionary ence in other countries, including Haiti, Mali, Myanmar, and power to approve the conversion to a small power pro- Nigeria. The options are not mutually exclusive and may be ducer (SPP) or small power distributor, or how it would complementary. compensate assets once the main grid arrived. One mini grid developer indicated that it was holding off from The four options are summarized in table 10.1 before being investing in a project because of a proposal made by the described in separate sections of the text. MINI GRIDS FOR HALF A BILLION PEOPLE    27 3 TABLE 10.1 • Options for making it easier for mini grid developers to do business Option Description Advantages Disadvantages Suitable where . . . a Examples Reduce red Standardized power Standardized PPAs Standardized Standardized PPAs Nigeria: A tape through purchase agreements and standardized asset documents and and asset transfer standardized standardized, (PPAs) govern what transfer templates processes are not agreements can be environmental and preapproved happens when the reduce risk for tailored to the enforced as contracts. social management templates and mini grid connects to developers by increasing unique needs and The utility is able process is being contracts. the main grid and sells their negotiating power characteristics of to pay on time for developed. electricity to the main with the main utility or individual projects the electricity that Tanzania: A utility company. other purchasing entity. or deals. This risk it purchases from standardized PPA A standardized asset Standardized social can be mitigated by the mini grid under cap is in place for transfer template is and environmental allowing the parties standardized PPAs. mini grids and small used when the mini grid management systems to the agreement—or power producers. The purchasing entity sells its assets to the reduce costs for the developer for its is able to pay on ESMAP developed a main utility company or developers compared environmental and time for the assets it standardized asset a distribution network with creating their own social management agrees to buy. transfer agreement operator. environmental and system—to make Formal requirements for the regulator in social management changes to the A standardized social to monitor the Tanzania. system. They also standardized and environmental environmental and reduce costs for the template, though management system social aspects of mini regulatory or oversight these changes would makes it easier for grid projects are in authority, which would require review and developers to track and place or are being otherwise have to approval by the manage their social and developed. monitor compliance. appropriate authority. environmental impacts. Use technology A technology platform Dramatically Relies on a single A large cohort of Nigeria: The platforms receives data from streamlines the process platform, which may developers exists Odyssey platform is to connect developers about their for connecting investors create a bottleneck in possessing the being used for the developers with mini grid projects, with developers. the market if there are capacity to provide tender of 250 mini investors and standardizes the data, Reduces the time and technical problems data on their projects grids. suppliers and and makes it available resources required to with the platform and as required by the Economic to run large- to investors and run a large-scale tender if the platform is the platform. Community of scale mini grid suppliers. for mini grids. only way developers Large-scale tenders West African States tenders. The goal is to help can legally bid for mini are being planned for (ECOWAS): The Levels the playing field investors finance mini grid sites. dozens or hundreds of Odyssey platform by enabling all mini grid grid portfolios and to projects to be evaluated May pose privacy mini grids. was used for a help developers secure by potential investors concerns for some Financial investors renewable energy financing and attractive across the same developers and are seeking to invest mini grid tender. deals from component metrics. investors. in portfolios of mini suppliers. May crowd out small- grids. Serves as a useful The platform can also tool to share market scale developers be used for large-scale intelligence in a or community-led tenders of mini grid controlled way. mini grid projects projects and portfolios by catering more Standardizes where the standardized to larger-scale applications for mini presentation of developers with grid tenders, facilitating developers’ projects greater technical transparent comparison facilitates bid review. and administrative of applications capacity. (reducing opportunities for corruption). Makes it easier for the government to manage large amounts of data. 274   MINI GRIDS FOR HALF A BILLION PEOPLE Option Description Advantages Disadvantages Suitable where . . . a Examples Eliminate In the absence of a Gives investors Some local No formal regulatory Nigeria, Haiti, duplication of formal regulator, the confidence that only one governments may agency is tasked with and Myanmar: government national government government entity will lack financial and overseeing mini grids. Community oversight by delegates oversight of have primary oversight human resources to A legal framework agreements. delegating mini grids to a single authority over their design and enforce explicitly grants Bangladesh: Grant- authority to a government entity business. a regulation, and authority to local making government single entity that knows the context Enforcement of different jurisdictions governments or to a entity has authority. when no formal well (typically a rural regulation may be easier may have different specific government Mali: Rural regulator for electrification agency, in the case of delegation rules, making it more entity. electrification mini grids local government, or to local governments. difficult for developers agency has exists. government agency to build in multiple Complex interfaces authority. providing financial jurisdictions. between agencies support for rural Delegating authority can be avoided if the development to a grant-giving subsidizing agency also activities). agency may result in acts as the regulator. conflicts of interest. Set up These initiatives The main advantage E-government Governments have Nigeria: “National e-government provide businesses for developers is that initiatives require adequate data e-Government services (and citizens) with a it speeds up many security and data protection and online Strategies” to reduce way to interact online processes that would protection on both security available to Kenya: “e-Citizen” overhead costs. with government normally take days or ends (government protect sensitive data. and citizen or India: agencies online to weeks to accomplish in business), and “e-Governance perform bureaucratic person. these may not yet Infrastructure” tasks such as business It can also increase registration, land be possible in some Ghana: Recently transparency, as and building permits, countries. launched “Forum on electronic records can and environmental E-government e-Governance” be accessed by the approvals. Ideally this relevant stakeholders as initiatives can also would be done through needed. “disenfranchise” a one-stop shop small businesses specific to mini grids. and households that do not have reliable access to the internet. Source: ESMAP analysis. “Suitable where…” does not mean that wherever this condition is fulfilled, the option should be picked. Rather, it means that under these conditions, a.  this option may be well suited, while under other conditions, another option may be better suited. ESMAP = Energy Sector Management Assistance Program. REDUCING RED TAPE THROUGH STANDARDIZED, PREAPPROVED TEMPLATES The Energy Sector Management Assistance Mini grid developers often must navigate a complex and Program (ESMAP) has identified four com- murky path of permits, regulations, and contracts as they plementary options to make it easier for mini grid build and operate their mini grids. Chapter 9 discusses the developers to do business: (1) reducing red tape mini grid–specific permits and approvals that developers through standardized pre-approved templates; (2) must typically obtain. In addition, developers usually other using technology platforms to connect developers must wade through other types of red tape (table 10.2). with investors and suppliers and to run large-scale mini grid tenders; (3) eliminating duplication of These approvals may be costly and take time to obtain, government oversight by delegating authority to making the business environment less conducive to pri- a single entity if no formal regulator for mini grids vate-sector investment. Box 10.1 provides some brief exists; and (4) setting up e-government services to examples.5 reduce overhead associated with business registra- Three standardized templates could significantly reduce tion, land and building permits, and environmental the cost—in time and money—as well as the risk that devel- approvals. opers incur as they build and operate their mini grids. MINI GRIDS FOR HALF A BILLION PEOPLE    275 TABLE 10.2 • General types of bureaucratic processes that mini grid developers navigate Type of bureaucratic Examples process Right to operate a Registering as a business business Obtaining construction or building permits Registering as a tax-paying entity Land and natural resource Registering the property rights Proving ownership or usage rights to land Securing approval of the right to use a specified amount of water or other natural resource Environmental approvals Undergoing environmental review at the level specified by the national environmental agency Undergoing review and approval by the river or irrigation authority Obtaining statements from the relevant government agency that the project is not in a protected area. Source: Tenenbaum and others 2014, 84. BOX 10.1 THE EFFECT ON INVESTMENT OF PERMITS OUTSIDE THE ELECTRICITY SECTOR Many countries require mini grid operators to obtain In Tanzania, Rift Valley Energy was the first operator various permits. (other than TANESCO, the national utility) to obtain a license. In addition to the actual electricity license, the In Nigeria, all mini grids must comply with environmen- operator had to obtain 27 permits, licenses, or agree- tal legislation. Developers must obtain an Environmen- ments from various government bodies. tal and Social Impact Assessment Certificate from the Federal Ministry of Environment. According to Nayo In Zimbabwe, mini grids must get clearance from the Tropical Technology, which was building a mini grid in Environmental Management Agency and the Zim- September 2017, obtaining this clearance could take babwe National Water Authority. The two agencies two years. The ministry and the Rural Electrification required fees to evaluate the impact of the project. Agency are currently developing a more streamlined Combined, those fees could represent 6.5 percent of process. total project costs. One of the agencies recently low- ered its fees. Sources: IRENA 2016; World Bank 2017a, 2017b. • Standardized power purchase agreements (PPAs) that very high levels of uncertainty and risk if they want to con- detail what happens when the mini grid connects to the nect to the main grid and sell electricity to the main utility, main grid and the mini grid sells electricity to the main because mini grid developers hold very little bargaining company. This would correspond to the “SPP” and “SPP power against the main utility when negotiating on the price plus SPD” options discussed in chapter 9. of electricity and the conditions of its sale. Similarly, a lack of bargaining power makes selling assets to the main utility • A standardized asset transfer template that describes or a distribution network operator risky for the developer what happens when the mini grid sells its assets to because it will have difficulty negotiating a fair price. In the the main utility or a distribution network operator. This absence of a standardized environmental and social man- would correspond to the “buyout” option discussed in agement framework, developers are often left with cumber- chapter 9. some, and sometimes conflicting, reporting requirements • A standardized social and environmental management and operating rules that they must follow. framework to make it easier for developers to track and manage their social and environmental impacts. The three templates help create an enabling business envi- ronment that makes it easier not only to start a business The templates touch on three areas developers frequently but also to operate one. The first two templates simplify mention as being difficult, time consuming, and risky. In developers’ interaction with the main grid if and when it the absence of standardized PPAs, developers are left with 276   MINI GRIDS FOR HALF A BILLION PEOPLE arrives in the mini grid’s service area, a topic discussed A well-drafted standard PPA is not enough in itself to pro- further in chapter 9. Developers can also present these tect investors. There must be a requirement in the law for templates to their investors to show that their investments the main grid to sign the PPA. The standard PPA contains will not be lost if the main grid arrives. The third template protections on top of the existing legal framework (against makes it easier to navigate the environmental and social late payment, for example) that are critical to investors. clearances that developers need when starting and oper- However, the contract is binding only once it is signed by ating their mini grids. both parties. Regulation or law should specify that the util- ity must enter the standard PPA when it connects to an iso- Standardized power purchase agreements lated mini grid that is converting to an SPP. In addition to making it easier and less risky for mini grid In Tanzania the Electricity Rules require the main grid to developers to sell their electricity to the main grid, stan- enter the standard PPA in cases where the isolated mini grid dardized PPAs can reduce the administrative capacity is allowed to become an SPP (EWURA 2017). In Cambodia required by the regulator and eliminate the need for a regu- there is no standard PPA and no obligation in the law for the latory review of each negotiated PPA (Tenenbaum and oth- main grid to connect to the SPP and enter into a PPA. ers 2014). Standardized PPAs typically contain clauses on a variety of terms and conditions related to the price and sale Standardized asset transfer agreements of electricity, the obligations of the utility and the developer, We have not yet encountered a standardized asset transfer and safety and technical requirements, among others. agreement in the mini grid sector. ESMAP has created such Table 10.3 contrasts key clauses from two good examples an agreement for the regulator in Tanzania, but it has not of PPAs, a PPA with a 1.5-megawatt rice husk gasification yet been used in practice. However, asset sale agreements facility in Cambodia with the standardized PPA in Tanzania. TABLE 10.3 • Key provisions of power purchase agreements in Cambodia and Tanzania Provision Cambodia Tanzania Obligation to Under the power purchase Under the PPA, the buyer must make the connection to the SPP, and the SPP connect agreement (PPA), the buyer bears the interconnection costs. must make the connection to the small power producer (SPP), with the latter bearing the interconnection costs. Obligation to A “must-take” clause covers The PPA states the contract to be a take or pay contract, and the seller’s facility to buy the SPP’s 40 percent of the generation be a must-take facility.a However, because risk of the main grid not being able to output capacity, with an option to buy receive the SPP’s power is largely allocated to the SPP, this PPA is not a take-or- up to 80 percent of the capacity. pay contract in accordance with the usual terminology. The SPP must produce no less Grid risk is largely allocated to the SPP through the exclusions to the buyer’s than the must-take amount. obligation to purchase the buyer’s entitlement. PPA Article 2(h) states that the buyer may “interrupt, reduce or cease to purchase and accept delivery of all or a portion of the buyer’s Entitlement to the extent necessary under Good Utility Practice in order to install equipment, make repairs, replacements, investigations or inspections of the buyer’s electric system.” PPA Article 2(h) provides that the buyer may “curtail or interrupt” taking power from the SPP “whenever the buyer’s system or the systems with which it is directly interconnected experience an Emergency, or whenever it is necessary to aid in the restoration of service on the buyer’s system or on the systems with which it is directly or indirectly interconnected.” This clause omits the language above related to ceasing to purchase; accordingly it is unclear whether the buyer is still required to pay for buyer’s entitlement not taken in such circumstances. As financiers and potential SPPs are likely to perceive grid risk as substantial, the allocation of risk stated above may make it difficult or impossible for SPPs to obtain finance. Pricing The regulator preapproved this The price of the energy sold is defined by regulation. tariff. The tariff is denominated PPAs signed after August 2015 are denominated in dollars and indexed on the U.S. in dollars and is constant in Producer Price Index. This arrangement places the currency exchange risk on the nominal terms. offtaker, which is often a requirement to enable project finance in foreign currency. (In PPAs signed before August 2015, prices were denominated in local currency and indexed to local inflation.) MINI GRIDS FOR HALF A BILLION PEOPLE    27 7 TABLE 10.3, continued Provision Cambodia Tanzania Enforcement Late payment attracts default Late payment attracts default interest at the prime rate (announced by the Bank of payment interest at a rate of 20 percent a of Tanzania), compounded monthly. year. If payment is two months Eventual nonpayment gives rights to the SPP to terminate the contract as an late, a right to withhold supply event of default (which may be triggered if the buyer does not meet its obligations arises. This contract feature under the PPA). gives the supplier leverage and protection in case of nonpayment; without this provision, the seller would still have to supply following the other provisions in the PPA, even if it is not being paid. Dispute Disputes are to be resolved After a 60-day informal dispute resolution period, either party can appeal to the resolution through mutual discussion regulator to resolve the dispute. The SPP could also seek international arbitration mechanism during a 30-day period. If the to resolve it, in which case then has 20 days to do so. The SPP must reimburse the discussion fails, domestic buyer’s travel expenses in case of arbitration. No arbitral rule is specified, and the courts resolve the dispute. parties must agree on the location of the arbitration. Risk sharing The SPP bears transmission The SPP bears transmission and distribution risk (the buyer can claim force and distribution risks, majeure on its system). The buyer bears some government risk (in case of failure commercial and market risks, to achieve the commercial operation date for the seller, caused by failure of the and security-of-supply risks. government to grant necessary permits). Changes in The PPA does not contain a The PPA does not contain a provision covering changes in law. law b provision covering changes in law. Force majeure Force majeure is defined as Force majeure is defined as events outside the reasonable control of the parties events that cannot be managed and not resulting from any of the parties’ failure or negligence, including acts by the buyer or seller that of God, fires, floods, epidemics, earthquakes, civil disturbances, insurrections, prevent them from meeting strikes, and war. their obligations, including war, riots, demonstration, flooding, and earthquakes. Lender step-in The PPA does not mention The PPA contemplates lender step-in rights. If the buyer is to claim an event of rights lender step-in rights. default against the seller, it needs to notify lenders and give them “reasonable time, access, and opportunity” to resolve the default, and cooperate with them to this end. Article 3(c)2 appears to provide both parties with mutual rights to step in to resolve events of default attributable to the other party. This would provide the SPP with the ability to step in and cure events of default on the buyer’s grid, so long as the SPP satisfies the requirement that it “has the skills and means to carry out the work necessary” (PPA, Article 3(c)2), which is very unusual. Termination A party can terminate the The PPA defines conditions of default. There is an immediate right to terminate agreement if the other party the contract upon the occurrence of an event of default, although several events breaches its obligations and of default also have default and rectification periods specified within them (PPA does not remedy the breach Article 3(c)1). within 30 days following written Conditions for default include failure to complete the project on time, bankruptcy, notification. In case of serious failure to meet the obligations in the agreement and remedy within 60 days breach, the delay is 15 days. after written notice, failure to make undisputed payment within 90 days, and reorganization of the buyer preventing it from performing its obligations. Sources: EWURA 2014, 2015. A take-or-pay contract typically specifies an amount of electricity that the offtaker must purchase from the generator. If the contracted amount of kilo- a.  watt-hours is not used by the offtaker, the offtaker is still responsible for paying for it. Must-take typically applies to “nonfirm generators” (intermittent renewables) and means that the offtaker must buy whatever amount of electricity is generated by the generator. Changes in law would mean negotiation of the SPP, which may be very burdensome for the regulator and would defeat the purpose of having a stan- b.  dardized PPA (Tenenbaum and others 2014). 278   MINI GRIDS FOR HALF A BILLION PEOPLE are common in other sectors. A standardized asset trans- the use of a third party to assist with the valuation of fer template for mini grids would be intended primarily for the assets. It is too early to evaluate the effectiveness of agreements between privately owned mini grids and the these different approaches. main utility or distribution network operator. Before apply- • Notification to relevant entities. This section identifies ing a standardized template to a specific deal between the entities that each party to the agreement must two parties or incorporating it into a broader regulatory notify. These typically include the regulator, the develop- framework, care should be taken to ensure that the tem- er’s investors and/or board, and the utility’s or distribu- plate adheres to existing laws, rules, and regulations and is tion company’s investors and/or board. appropriate for the local context. • Signatures. This section contains the signatures of the In general, a standardized asset template would cover the parties to the agreement, which are typically accompa- following topics: nied by the signature of one or two witnesses. • Eligibility. This section establishes that the mini grid’s • Annexes are often used to attach additional informa- assets are eligible to be transferred to the utility/distri- tion, such as detailed inventories of the assets, account- bution network operator and that the mini grid owner ing information on grants received, or the developer’s is eligible for compensation under the regulations expenditures on the assets up for sale. It can be partic- governing what happens when the main grid arrives. ularly useful to provide a worksheet describing how to Assets are typically eligible for sale if they were built to calculate the compensation due to a mini grid owner for grid-compatible standards and have been maintained its assets. Another important annex is the certificate of well. Typically, only distribution infrastructure—lines, deposit in an escrow account or a bank guarantee letter poles, transformers, and meters (if compatible with from the purchasing entity’s bank, stating that funds are the utility’s billing system)—would be considered for available and have been set aside exclusively for the pur- the sale. Movable assets like generators, batteries, chase of the assets. and solar panels are typically not sold to the utility but can be repurposed by the developer. Developers are A standardized asset transfer template that was developed typically eligible if they have registered with, or have for Tanzania is available on the companion website to this received a license or permit from, the regulator and handbook: www.esmap.org/mini_grids_for_half_a_billion_ are operating as a legal business within the laws of the people. country. Standardized social and environmental management • Bill of assets for sale. This section identifies and lists frameworks the assets that are to be transferred. It should include One of the most intensive clearance processes that devel- as much information as possible about the assets, opers can face is a review of the mini grid’s environmen- including when they were purchased, when they were tal and social impacts. With these clearances often comes installed, and whether they have had any major repairs. ongoing monitoring of environmental and social safe- The purchasing entity typically has the right to a third- guards. Streamlining and clarifying the environmental and party evaluation of the quality and quantity of the assets social management process can reduce risk and overhead and mutually agrees with the developer which assets costs, making it easier for mini grids to do business. are to be sold. An example of how to streamline this typically resource- • Compensation mechanism. This section states how and time-intensive process for both the developer and the developer is to be compensated for its assets. Sev- the government oversight entity comes from Nigeria. The eral approaches are possible, such as those outlined in Federal Ministry of Environment is responsible for monitor- chapter 9. The most common types are (1) compensat- ing mini grid projects, including final reviews of all environ- ing the developer for the fair market value of its assets mental and social impact assessment reports. The Rural after accounting for depreciation and any grants that the Electrification Agency (REA) developed a standardized developer received, plus some amount of recent annual Environmental and Social Management System (ESMS) revenues (this is the approach in Nigeria); and (2) com- that has been approved by the Ministry of Environment. pensating the developer based on the utility’s avoided The system is designed to manage potential environmental cost of having to build comparable infrastructure to and social risks while simplifying the clearances that devel- connect the mini grid’s customers, after accounting opers must obtain. for depreciation and any grants that the developer has received (this is the approach in Tanzania). The devel- The ESMS is effectively a standardized process for both oper and the purchasing entity can mutually agree on developers and the REA. Before the standardized ESMS, MINI GRIDS FOR HALF A BILLION PEOPLE    27 9 FIGURE 10.1 • Nigeria’s Environmental and Social Management System for minimum-subsidy tenders for mini grid development Process steps E&S Tasks and responsibilities SITE SELECTION Create list of potential sites • Include E&S information questions in site surveys • Analyze and verify survey information Conduct site surveys • Apply E&S Exclusion Criteria for Mini Grid and Power Generation Sites • Estimate land requirement for each site based on expected generation capacity Select project sites based • Prepare tender documents, including E&S requirements on site survey results TENDERING PROCESS Prepare for competitive tendering process • Include E&S information from demand surveys in lot package – Lot package • Include E&S criteria for developers as part of overall criteria for appraisal and – Advertising selection of bids (ESMS + clean track record with no environmental or labor nes) – Developer information and capacity • E&S aspect in bid workshop for developers building – Online platform • Grievance Redress Mechanism • Prepare and submit required documents (ESMS as part of business plan) to Submit bidding package REA for review and veri cation • Con rm clean E&S track record • Review bids to ensure all E&S criteria are met Review and select winning • Liaise with developers if improvement to ESMS is needed (until REA is satis ed bidder developers that a robust system exists) Sign agreement Grant agreement b/w REA and developer includes E&S clause CONSTRUCTION AND OPERATION • Conduct E&S screening • Classify sites into E&S risk category (I, II) • Inform REA of outcomes of screening (consolidated report), including if sites fall under E&S exclusion criteria for mini grid and power generation sites Prepare for construction • Prepare ESIA, ESMP, RAP/LRP, as applicable • Obtain E&S permits required by law • Conduct stakeholder engagement • Submit relevant documents to REA • Verify developers’ information (sample checks for adequacy of the ESIAs, ESMPs, Review and verify stakeholder engagement process, RAPs/LRPs) Testing and commissioning can preparation documents provide a good platform for veri cation • Conduct construction & maintain good compliance • Maintain a grievance mechanism to address community concerns Construction and operation • Inform REA of incidents or accidents • E&S reporting as part of progress report to REA • Monitor E&S performance throughout project Monitor construction • Maintain an REA-level grievance mechanism to address any project-related and operation feedback Source: Rural Electrification Agency of Nigeria. Note: Gold shading indicates steps for the REA; blue shading indicates steps for mini grid developers. E&S = environmental and social; ESIA = Environmental and Social Impact Assessment; ESMP = Environmental and Social Management Plan; ESMS = Environmental and Social Management System; RAP/LRP = Resettlement Action Plan / Livelihood Restoration Plan; REA = Rural Electrification Agency. 280   MINI GRIDS FOR HALF A BILLION PEOPLE FIGURE 10.2 • Nigeria’s Environmental and Social Management System for performance-based grants for mini grid development Process steps E&S Tasks and responsibilities PROPOSAL ACCEPTANCE INTO THE PROGRAM Program announcement • Publish site’s selection / eligibility criteria, including E&S exclusion criteria for mini grid and power generation sites • Publish E&S quali cation criteria for developers (ESMS + clean track record with no environmental nes or labor nes) • Maintain REA-level grievance mechanism • Prepare and submit required documents (ESMS as part of business plan) to REA for Prepare and submit proposal review and veri cation package • Con rm clean E&S track record • Review proposal package to ensure all E&S criteria are met Review and approve developers • Liaise with developers if improvements to ESMS are needed (until REA is satis ed that a robust system exists) Sign agreement Agreement between REA and developer to include E&S clause DESIGN VERIFICATION FOR SITE • Conduct E&S screening • Classify sites into E&S risk category (I, II) • Inform REA of outcomes of screening (consolidated report), including if sites fall under Prepare sites E&S documents E&S exclusion criteria for Mini Grid and Power Generation Sites • Prepare ESIA, ESMP, RAP/LRP, as applicable • Obtain E&S permits required by law • Conduct stakeholder engagement • Submit relevant documents to REA Review and verify preparation • Verify developer’s information (desk review of ESIAs, RAPs, sample site visits) during documents design veri cation with aim to minimize E&S impacts CONSTRUCTION & OPERATION • Conduct construction & maintain good compliance • Maintain a grievance mechanism to address community concerns Construction and operation • Inform REA of incidents or accidents • E&S reporting as part of progress report to REA Monitor construction and • Monitor E&S performance throughout project operation • Maintain an REA-level grievance mechanism to address any project-related feedback Source: Rural Electrification Agency of Nigeria. Note: Gold shading indicates steps for the REA; blue shading indicates steps for mini grid developers. E&S = environmental and social; ESIA = Environmental and Social Impact Assessment; ESMP = Environmental and Social Management Plan; ESMS = Environmental and Social Management System; RAP/LRP = Resettlement Action Plan / Livelihood Restoration Plan; REA = Rural Electrification Agency. MINI GRIDS FOR HALF A BILLION PEOPLE    281 each developer had to devise its own environmental and Advantages. Standardized PPAs and asset transfer agree- social management framework and have it approved by ments reduce costs—in time and money—for both parties the Ministry of Environment. Ensuring that it was adher- to the agreement. Standardized environmental and social ing to the policies and rules related to environmental and management systems reduce the costs to the developer social management was often time consuming and oner- compared with having to create its own systems; they also ous for the mini grid business. The purpose of the ESMS reduce the costs to the regulatory or oversight authority is to set out a ministry-approved standardized process that would monitor compliance, because all developers for how developers and the REA can jointly ensure that would be using the same template. the mini grid meets its environmental and social require- Standardized PPAs and asset transfer agreements reduce ments. Because it is preapproved by the Ministry of Envi- the risk to developers by increasing their negotiating power ronment, the ESMS significantly reduces the number of with the main utility or other purchasing entity. Standard- additional clearances that the developer must obtain by ized environmental and social management systems increasing the role of self-evaluation and implementing reduce developers’ risk by ensuring that they know exactly light-handed administration by the REA. Figures 10.1 and how their performance across specific environmental and 10.2 depict the process. Because the standardized ESMS social indicators will be monitored. is a recent innovation, we do not yet have results for how well it works in practice. Disadvantages. Standardized agreements and systems are not tailored to the unique needs and characteristics of In the minimum-subsidy tender approach to mini grid individual projects or deals. This risk can be mitigated by development (figure 10.1), developers are expected to allowing the parties to the agreement, or the developer submit their environmental and social safeguard strategy, in the case of an environmental and social management including a verifiable clean bill of environmental and social system, to make changes to the standardized template, compliance over a period of at least three years. These which would require review and approval by the appropri- requirements coincide with the developer’s responsibilities ate authority. to conduct requisite site screening and risk categorization, secure required environmental permits from the Ministry Where suitable. The three standardized templates are well of Environment, and prepare applicable site-specific safe- suited in markets with a large number of mini grids with rel- guards for clearance with the REA. The responsibilities of atively similar characteristics and where large tenders for the developer go beyond the tendering process to include dozens or hundreds of mini grids are being prepared. the construction and operation phase, keeping up-to-date Requirements for success. The most important require- records of environmental and social incidents, and submit- ment for successfully rolling out standardized PPAs and ting progress reports on environmental and social compli- asset transfer agreements is the enforceability of con- ance to the REA. tracts. These documents are intended as legally enforce- In the performance-based grant approach to mini grid able contracts; their ability to reduce the risk for developers development (figure 10.2), in which developers receive per- is contingent on the developer being able to count on the formance-based grants based on the number of mini grid legal system should any adverse circumstance invoke one connections at sites that they identify and develop them- or more clauses of the agreement. Contract enforcement selves, developers are expected to conduct all necessary requires substantial government capacity and could be an environmental and social screening of their proposed site. area where targeted capacity building is required. While an independent third party is expected to monitor A specific requirement for the success of standardized environmental and social compliance, all developers must PPAs is the ability of the utility to consistently pay on time prepare a strategy for battery disposal as part of their envi- for the electricity that it purchases from the mini grid. ronmental and social responsibilities in line with the Minis- Where the buyer is commercially insolvent, long payment try of Environment’s guidelines. delays are likely. Summary of advantages, disadvantages, suitability, A specific requirement for the success of a standardized and requirements for success of standardized asset transfer template is the ability of the purchasing templates entity to pay on time for the assets it agrees to purchase. Standardized templates have several advantages com- For standardized environmental and social management pared with project- or deal-specific agreements, although systems, the most important requirement ensuring that they have some important disadvantages as well. They are the system is preapproved by the relevant government not suitable everywhere and have several critical require- agencies that would oversee the environmental and social ments for success. aspects of mini grids. 282   MINI GRIDS FOR HALF A BILLION PEOPLE BOX 10.2 Three standardized templates have the potential to significantly reduce the costs WORLD BANK EXPERIENCE WITH of bureaucracy: (1) standardized power purchase ODYSSEY ENERGY SOLUTIONS IN agreements governing sales of mini grid electricity NIGERIA to the main grid; (2) standardized asset transfer agreements for governing sales of mini grids’ eligi- The World Bank selected Odyssey as its technology ble assets to the main utility; and (3) a standardized platform to run a large-scale tender for mini grids environmental and social management system that for the Nigeria electrification project. The project is applied when mini grid developers must obtain involves a $150 million investment in mini grids, of environmental approvals. The first two templates which $70 million was allocated to mini grid pro- help mini grid developers negotiate with the national curement using a top-down approach. utility on more even and transparent terms. The third Nigeria’s Rural Electrification Agency conducted simplifies the process for obtaining approval of the extensive geospatial and socioeconomic analysis developer’s environmental and social management on an initial list of 8,000 potential mini grid sites system. The first two templates should be used only to identify 250 that have good potential to sup- in countries where contracts are enforceable. The port commercially viable mini grids. In effect, the third should be used only after it has been approved REA performed the due diligence that has typically by the government agencies responsible for the been performed by developers in other countries. environmental and social aspects of mini grids. Data for these sites is made available on the Odys- sey platform, and developers are preparing bids by entering their project data into the platform. USING TECHNOLOGY PLATFORMS TO The process generates standardized bids for the CONNECT DEVELOPERS WITH INVESTORS REA and the World Bank to review, with success- AND SUPPLIERS AND TO CONDUCT LARGE- ful bids eligible to receive an upfront capital cost SCALE MINI GRID TENDERS subsidy. It is expected that the tender will extend access to 110,000 new mini grid customers. More Creating an enabling business environment for mini grids information on how the Odyssey platform is used in goes beyond reducing red tape. It also entails removing Nigeria appears in chapter 2 of this handbook. market inefficiencies and increasing market intelligence. One way to do this is to use a technology platform to con- Source: ESMAP analysis. nect developers with financial investors and component suppliers and to run large-scale tenders for mini grids at the regional or national scale. platform provides governments and donors with an effi- The leading example of such a platform is Odyssey Energy cient way to review mini grid projects, because each project Solutions,6 which has developed a web-based platform that is presented in a standardized way. In this way, it increases streamlines the process of building investable portfolios of market intelligence and makes government procurement mini grids. The platform helps developers analyze site data, more transparent, thus reducing the risks of corruption forecast demand, design their systems, model tariffs, and and expropriation. Third, a technology platform can also produce pro forma financial statements and analytics. make procurement of mini grid components easier for Developers can share these data on the website with finan- developers by providing a hub to connect developers and ciers and suppliers to secure investment and seek attrac- technology suppliers. tive procurement deals. The World Bank’s experience with Technology platforms like Odyssey that can present mini Odyssey in Nigeria is described in box 10.2. grid data in a standardized format also facilitate large-scale One important reason to consider using a technology tenders for mini grids at the regional or national scale. The platform like Odyssey in the rollout of national mini grid fact that all projects are presented the same way stream- programs is that it significantly increases the market effi- lines and expedites the review of developers’ bids. In addi- ciencies of several key processes. First, it creates a single tion, built-in analytics in the platform help developers hub connecting developers and financiers, thus making check their financial and technical models for errors and deal sourcing and deal making easier and less costly for inconsistencies, increasing the likelihood that the technical both financiers and developers—in essence, reducing the and financial specifications of bids selected for the tender transaction costs of deals. Second, use of a technology will be met during project implementation. MINI GRIDS FOR HALF A BILLION PEOPLE    283 Advantages. Using a technology platform like Odyssey both buyers and suppliers in the markets the platform can streamline the process for connecting developers with intends to facilitate—namely, the procurement market for investors and component suppliers. In this way, it helps cre- mini grid components, the investment market for portfo- ate an enabling business environment not only for develop- lios of mini grids, and the market for mini grids created by ers but also for technology vendors and financial investors. government and donor tenders. In addition, while govern- Technology platforms can also significantly reduce the ments may be familiar with conducting tenders, they may time and resources required to run a large-scale tender by not be familiar with using a digital platform like Odyssey for creating efficiencies for both developers and bid reviewers. that purpose. As a result, capacity building targeted at the Platforms can level the playing field by enabling investors use of these types of platforms will likely be required. to evaluate mini grid projects across the same metrics, through a standardized presentation of project data. ELIMINATING DUPLICATION OF GOVERNMENT OVERSIGHT BY DELEGATING AUTHORITY TO A Disadvantages. Reliance on a single platform can create SINGLE ENTITY a bottleneck in the market if technical problems with the platform cause major disruptions to the market or tender In many countries, in the absence of a formal regulator process. Use of an online platform may also pose privacy tasked with overseeing mini grids, developers face double or concerns for some developers, who may not feel comfort- triple layers of government oversight—from rural electrifica- able giving their project-level data to a third-party website. tion authorities to local governments to agencies providing Using a technology platform like Odyssey as the go-to hub subsidies, among others. Each additional layer of oversight for investment and procurement deals may also crowd out compounds the risk and cost of the mini grid business, small-scale developers and community-led mini grid proj- raises the likelihood that rules and regulations for develop- ects by catering to larger-scale developers with greater ers will change, and increases the chances that those rules technical and administrative capacities. and regulations will be inconsistent or conflicting. Suitability. A technology platform like Odyssey is suitable One way to reduce multiple layers of oversight (again, when when there is a large cohort of developers with the capac- there is no formal regulator) is to delegate primary over- ity to provide data for their projects covering all metrics sight authority for mini grid–specific activities to a single required by the platform. It is also applicable in countries government entity. This does not mean that that entity will that are planning large-scale tenders for dozens or hun- wield the exclusive authority over mini grids. For example, dreds of mini grids, particularly when geospatial data are a government agency with a legal mandate to protect the available for the sites on which developers are bidding. This environment could not legally cede its authority to another technology platform–enabled strategy is especially suit- government entity. What it does means is that only one able in markets where financial investors are seeking com- authority will be responsible for decisions that would nor- mercial investments in portfolios of mini grids. mally fall to a formal regulator: decisions related to market entry (permits, registration, licensing), tariffs, service stan- Requirements for success. To be successful, a technology dards, technical standards, and what to do when the main platform like Odyssey requires active participation from grid arrives. These decision areas mirror those discussed at length in chapter 9, but they are discussed again here for two rea- sons. First, a lack of clarity on who holds authority for these The use of technology platforms to con- decisions when there is no formal regulator represents nect developers with investors and suppli- a significant risk to mini grid businesses. The underlying ers and to conduct large-scale mini grid tenders assumption of this chapter is that a regulator or other has the potential to improve market efficiencies in government entity has already been tasked with making three ways: (1) by reducing the transaction costs regulatory decisions about mini grids. Second, multiple financiers face when investing in mini grids; (2) by government entities having authority over these decisions increasing market intelligence and making govern- represents a constraint on the ease with which mini grid ment procurement more transparent, thus reduc- developers can do business. ing risks of corruption and expropriation; and (3) by simplifying procurement of mini grid components The two most common candidates for overseeing mini grid easier for developers by connecting developers with activities in the absence of a formal regulator are the local potential suppliers. To be successful, the platform government and the agency that provides grants or subsi- needs active participation from developers, inves- dies to mini grid developers (Tenenbaum and others 2014). tors, suppliers, and the government. Local government. If a village hosting a new mini grid has an effective local government, it can regulate the mini grid, 284   MINI GRIDS FOR HALF A BILLION PEOPLE typically through a contract signed with the developer. This as an independent party balancing the interests of devel- is a variant of a “community agreement” that takes advan- opers and customers, as a traditional regulator would.8 tage of an existing local government body to represent the In addition, grant-making agencies have competencies in community. Such a local governance framework should be financial analysis but may not be trained for cost-of-service consistent with the local government’s legal authority.7 One calculations or other technical decisions that factor into important drawback of delegating authority to local gov- mini grid regulations. ernments is that developers that want to build mini grids in An important task for the government entity assigned to several different jurisdictions would need to adhere to each oversee mini grid activities is to anticipate what will happen local government’s set of rules and guidelines. For this rea- when the main grid arrives in the mini grid’s service area. son, delegating oversight of mini grids to local governments Discharging this task requires the grant-making agency or is not likely to be compatible with a portfolio approach to local government to ensure that its technical specifications development. for mini grids are compatible with eventual interconnection Grant-making agency. Where mini grids are contracted or with the main grid, and to stipulate clear terms for trans- subsidized by rural electrification agencies (REAs) or other ferring oversight authority to the entity that regulates the grant-making agencies, complex interfaces between agen- main grid (Tenenbaum and others 2014). cies can be avoided if the subsidizing agency also acts as Table 10.4 provides a summary of the local government and the regulator. REAs that provide subsidies often perform REA options, with advantages, disadvantages, and exam- reviews of mini grid business plans, which is similar to a ples. traditional cost-of-service review that a regulator would undertake (Tenenbaum and others 2014). Suitability. Delegating oversight of mini grids to local gov- ernment or a public grant-making agency is well suited for In Bangladesh, for example, the Infrastructure Develop- countries where no regulatory agency is tasked with over- ment Company Limited (IDCOL), the grant-making agency, seeing mini grids or where no formal regulations specific is the de facto regulator. In Mali, AMADER (Agency for the to mini grids are in place. This strategy may also be suit- Development of Domestic Energy and Rural Electrifica- able when used in conjunction with a centrally coordinated tion), the grant-making agency, is legally responsible for approach to mini grid development, where the grant-mak- traditional regulatory responsibilities (Tenenbaum and oth- ing entity uses competitive tenders to select operators for ers 2014). This technique—delegating regulatory functions concession areas (USAID 2017). to the agency that distributes subsidies—can also be found in industrialized countries. In the United States, the New Requirements for success. Two conditions must be met York State Development Authority has provided subsidies if governments wish to successfully delegate oversight to private providers of community solar projects in return of mini grids to a single entity in the absence of a formal for control over the prices that the developers will charge to regulator. The most important requirement is that the low- and moderate-income customers. legal delegation of authority be clear and specific. In Haiti, for example, a 2006 decree legally delegates to municipal However, delegating regulatory power to financing agen- governments the authority to oversee the production, dis- cies may create conflicts of interest and may not be suitable tribution, and sale of electricity within their jurisdictions where those agencies lack adequate capacities or skills. In (Government of Haiti 2006). Second, the entity to which Bangladesh, some developers view IDCOL’s double role as authority has been delegated should specify which laws a conflict of interest. As a financing institution, it has its own and regulations will take precedence over its own authority. financial interests, and some developers do not perceive it TABLE 10.4 • Advantages and disadvantages of two options for mini grid oversight Option Advantages Disadvantages Examples Local government Local government is more accessible Some local governments may lack financial Community to developers and customers than a and human resources to design and enforce a agreements used national entity would be. regulation. in Haiti, Nigeria, Enforcement of regulation may be Different jurisdictions might have different rules, and Myanmar facilitated by the physical presence which makes it significantly more difficult for of the regulator in the community. developers to build in multiple jurisdictions. Rural electrification Complex interfaces between agencies May lead to conflict of interests. Bangladesh agency or grant- can be avoided if the subsidizing May not be appropriate if the grant-making Mali making body agency also acts as regulator. agency does not possess adequate skills. Source: USAID 2017, 46–48. MINI GRIDS FOR HALF A BILLION PEOPLE    285 tion of the e-government website (for example, using the “https” protocol instead of the “http” protocol) and robust In countries where no formal regulator is data management practices for government employees tasked with overseeing mini grids develop- who interact with information submitted by businesses ers often face double or triple layers of government and citizens. oversight. To reduce the layers of bureaucracy, gov- ernments can formally delegate oversight authority As more and more commerce is conducted online, demand to a single entity. The two most common options from citizens and businesses for e-government services will are local governments or the public agency that pro- almost certainly increase. Meeting demand will require col- vides grants or subsidies to mini grid developers (for laboration between the government and companies, and example, a rural electrification agency). For either between the government and its citizens, to identify the of these options to be successful, there must be a types of services that can be offered online. One example of clear legal delegation of authority to the entity, and the collaborative process of establishing an e-government the entity must explicitly state which laws and reg- initiative is the “Forum on e-Governance” that the Ghanaian ulations will take precedence over its own authority. government recently launched. Actually building out the e-government capabilities typi- cally requires a partnership between the government and In parallel, the oversight entity must have sufficient capac- the private sector. In Nigeria, a public–private partnership ity to oversee the mini grid sector effectively and credibly. was established to develop the online infrastructure—both In most cases, this entity will require significant support the back end of the website that government employees during the early stages of its role. would use, and the front end that citizens and businesses would use. SETTING UP E-GOVERNMENT SERVICES TO Advantages. The main advantage of an e-government ini- REDUCE OVERHEAD tiative for businesses is that it speeds up many processes As access to the internet expands rapidly in developing that would normally take days or weeks to conduct in per- countries—including through mobile data networks— son. It can also increase transparency, as electronic records countries are beginning to establish “e-government” ini- can be accessed by stakeholders as needed. tiatives. These initiatives provide businesses (and citizens) Disadvantages. The main disadvantage is that e-govern- with a way to interact with government agencies online to ment initiatives require security and data protection on accomplish tasks such as business registration, land and both ends (government and citizen or business), and these building permits, and environmental approvals. may not yet be possible in some countries. E-government Several recent examples of e-government initiatives initiatives can also disenfranchise small businesses and include those in India, Kenya, and Nigeria. Across these and households that do not have reliable access to the internet. other initiatives, the objective is to increase the use of inter- Suitability. E-government initiatives are suitable where net-based services to streamline and make more transpar- governments have adequate data protection and online ent both internal government operations and the ways that security available to protect sensitive data, and where there citizens and businesses interact with the government. is demand from citizens and businesses to be able to inter- In some cases, e-government initiatives will have separate act with the government via the internet. sets of services for individuals and businesses. The e-gov- Requirements for success. To be successful, an e-govern- ernment web portal in Kenya offers one such example. ment initiative requires a participatory approach to iden- Under the “e-Business” section of the website,9 companies tifying the types of services that citizens and businesses can apply for, pay for, and receive PDF copies of their busi- would like to have available via the internet. It also typi- ness licenses. Under the “e-Citizen” section,10 individuals cally requires a public–private approach to building out can apply and pay for a variety of government services. the back- and front-end capabilities of the e-government Most e-government websites require users to create an website and related suite of services, since most govern- account to access online services. This typically includes ment agencies will not have the technical capabilities to providing certain personal information. Because this will develop the platform in-house. Finally, a successful initia- be stored online—usually on servers that are owned by tive requires robust data protection—not just within the the government or by a company contracted by the gov- website, but also with respect to the government employ- ernment—the long-term success and reputation of e-gov- ees who interact with the sensitive information that citizens ernment initiatives depend significantly on the quality of and businesses submit online. data protection. Protection includes high-level encryp- 286   MINI GRIDS FOR HALF A BILLION PEOPLE On setting up e-government services to reduce overhead costs E-government initiatives can significantly • Barrier: The underlying institutional framework behind reduce the overhead costs of interacting with the e-government service is complicated, ineffective, the government to apply for permits, licenses, and and opaque. Online technologies will not solve structural approvals. To be successful, the initiative requires a institutional issues. participatory approach to identifying which services should be offered online, a public–private partner- • Must have: Sufficient institutional capacity—and coordi- ship approach to building out the e-government nation—are necessary to run the back end of e-govern- website and associated services, and robust data ment services. protection measures both online and for govern- Beyond the four options described in this chapter, investors ment employees handling sensitive information. have also pointed to three factors in a country’s enabling environment that they assess when considering invest- ment opportunities. These factors, along with the char- acteristics that investors perceive as barriers and “nice to INVESTORS’ PERSPECTIVES ON THE haves” (as opposed to “must haves”) follow. FOUR OPTIONS PRESENTED ABOVE Qualified local personnel Through conversations with private-sector investors, and • Barrier: The supply of qualified local personnel (for building on the International Finance Corporation’s exten- example, project managers and engineers) is insuf- sive experience investing in infrastructure projects around ficient to manage and operate mini grids and to close the world, we have identified characteristics of each of the tenders and complete the technical design phase of four options discussed in this chapter that investors per- mini grid projects. ceive as either barriers or “must haves” when considering a mini grid investment. • Nice to have: Labor legislation carve-outs should allow the importation of essential workers to build and main- On reducing red tape through standardized tain mini grid projects in the face of shortages of skilled preapproved templates local labor. Regional licensing standards for operations and management should ensure adequate labor quality • Barrier: The template documents are not adaptable to and minimize developers’ labor search costs. Training different mini grid business models and are not updated programs should exist for operators, with associated as new regulations or policies come into effect. labor exchanges and a centralized labor pool. • Must have: The preapproved templates must protect Chapter 7 provides an in-depth discussion on training and the investment from government delay, expropriation, skills development for the mini grid sector. or changes to the law. Transaction costs On using technology platforms to connect • Barriers: Costs to import and deliver materials to the developers with investors and suppliers and run mini grid site are high because of tariffs and customs large-scale mini grid tenders duties. Poorly managed procurement processes are • Barrier: The government lacks sufficient capacity to use overly burdensome for developers. the technology platform effectively. • Nice to have: Tariffs and custom duties are reduced for • Must have: The technology platform must integrate major components needed to construct mini grids. Bid- both technical and financial aspects of mini grid projects ding processes have clear criteria that value operational at the level of each mini grid and at the portfolio level. experience and quality. Development partners help to finance feasibility studies, develop portfolio-based bid On eliminating duplication of government oversight packages, and provide transaction advisors. by delegating authority to a single entity • Barrier: The governmental body to whom oversight Financial infrastructure authority has been delegated does not have the ade- • Barrier: Telecom coverage and mobile finance sectors quate capacity to design and enforce rules and policies have limited access to remote areas of the country, lim- that govern the mini grid sector. iting the viability of data transfer and mobile payment • Must have: The authority to which oversight is delegated platforms. Microfinance options for potential produc- must have a good record of fair, transparent, and effec- tive-use customers are lacking. tive governance. MINI GRIDS FOR HALF A BILLION PEOPLE    287 • Nice to have: Mobile banking services reach a large EWURA. 2017. “The Electricity (Development of Small Power Proj- portion of the population, including in nonurban areas. ects) Rules, 2017, s 32(4).” http://www.ewura.go.tz/wp-content/ uploads/2015/03/The-Electricity-Development-of-Small-Power- Microfinance institutions are willing to lend—or have Projects-Rules-2017.pdf. experience lending—to micro- and small enterprises Government of Haiti. 2006. “Livre III: Des Relations des Collectives Ter- for the purchase of income-generating machines and ritoriales.” In Le Moniteur. IRENA (International Renewable Energy appliances. Agency). 2016. Policies and Regulations for Private Sector Renew- able Energy Mini-grids. Abu Dhabi: IRENA. https:/ /www.irena.org/ publications/2016/Sep/Policies-and-regulations-for-private-sec- tor-renewable-energy-mini-grids. CONCLUSION Tenenbaum, B., C. Greacen, and D. Vaghela. 2018. Mini-Grids and the Arrival of the Main Grid: Lessons from Cambodia, Sri Lanka, and Indo- Mini grids are businesses and therefore navigate the com- nesia. ESMAP Technical Report 013/18. Washington, DC: World Bank. plex world of permits, licenses, approvals, and clearances. https:/ /openknowledge.worldbank.org/bitstream/handle/10986/ Bureaucratic processes generally include permits and 29018/134326.pdf?sequence=6&isAllowed=y. processes related to setting up and operating a business, Tenenbaum, B., C. Greacen, T. Siyambalapitiya, and J. Knuckles. 2014. acquiring land rights, and receiving approval for the envi- From the Bottom Up: How Small Power Producers and Mini Grids ronmental and social impacts that the project may have. Can Deliver Electrification and Renewable Energy in Africa. Washing- ton, DC: World Bank. https:/ /openknowledge.worldbank.org/han- From an investment standpoint, these bureaucratic pro- dle/10986/16571. cesses constitute risks to the business; from the devel- USAID (United States Agency for International Development). 2017. oper’s perspective, they also increase the cost of doing Practical Guide to Regulatory Treatment of Mini-Grids. Washington, business. Therefore, simplifying and streamlining the envi- DC: World Bank. https:/ /pubs.naruc.org/pub/E1A6363A-A51D- 0046-C341-DADE9EBAA6E3. ronment in which developers do business can make the World Bank. 2017a. “Mini Grids in Nigeria: A Case Study of a Promis- mini grid market more attractive to the private sector. A ing Market.” World Bank, Washington, DC. https:/ /openknowledge. more appealing market means more mini grids providing worldbank.org/handle/10986/29016. more access to affordable electricity to more people. World Bank. 2017b. “Mini Grids in Tanzania: A Case Study on Regula- tions, Subsidies, and Exit Strategies.” World Bank, Washington, DC. World Bank. 2019. “Doing Business: Measuring Business Regulations.” REFERENCES World Bank, Washington, DC. http:/ /www.doingbusiness.org/en/ data/doing-business-score. Bakovic, T., B. Tenenbaum, and F. Woolf. 2002. “Regulation by Contract: A New Way to Privatize Electricity Distribution?” Working Paper 14, World Bank, Washington, DC. https:/ /openknowledge.worldbank. org/handle/10986/15078. NOTES Brown, A. 2012. “Infrastructure: The Regulatory and Institutional 1. Field interviews conducted by Castalia in Bisanti, Nigeria, August Dimension.” In Services, Development, and Trade: The Regulatory 2017. and Institutional Dimension of Infrastructure Services, Vol. 1, edited 2. However, the framework is very recent (adopted in 2017), and mar- by Mina Mashayekhi. Geneva: United Nations Conference on Trade ket reaction may change in time. and Development. 3. Field interviews conducted by Castalia with developers (2017) Brown, A., J. Stern, B. Tenenbaum, and D. Gencer. 2006. Handbook for Evaluating Infrastructure Regulatory Systems. Washington, 4. Field interviews conudcted by Castalia with developers (2017) and DC: World Bank. https:/ /openknowledge.worldbank.org/han- email exchange with developers (2018). dle/10986/7030. 5. Tenenbaum and others 2014 (87, 104-14) discuss regulatory pro- EWURA (Energy and Water Utilities Regulatory Authority). 2014. “Order cesses and nonelectricity approvals and provide recommendations 014-025: The Electricity (Standardized Small Power Purchase Tar- on designing a good regulatory review and approval system in a way iff for Year 2014) Order.” https://www.ewura.go.tz/2015/08/26/ that minimizes barriers to investment. USAID 2017 (73-–76, 103–6) the-electricity-standardized-small-power-purchase-tariff-for-year- suggests different options for designing the regulatory process and 2014-order-2014/. environmental review for mini grids, including adapting environ- mental reviews to size. EWURA. 2015. “Standardized Power Purchase Agreement for Purchase of Electric Energy from a Generation Facility Connected to the Main 6. See https://www.odysseyenergysolutions.com/. Grid; Power Purchase Agreement between SOMA Energy Co, LTD 7. Formal community agreements are used in Nigeria, Myanmar, and and BVC Power Development Co, Ltd.” Haiti. These agreements are a form of regulation by contract. Chap- ter 9 provides a detailed discussion on regulation by contract. 8. Email exchange with a mini grid developer, 2017. 288   MINI GRIDS FOR HALF A BILLION PEOPLE CHAPTER 11 CALL TO ACTION CHAPTER OVERVIEW In this final chapter, we present a call to action for key mini grid stakeholders, from policy makers and regulators to developers to investors and suppliers. The goal is to help build 210,000 mini grids to connect half a billion people by 2030. We also highlight some topics for future research. Concrete actions that various stakeholder groups can or community bodies. In addition, regulators can collab- undertake in the near term to help connect half a billion orate with regulatory agencies from other countries to people to mini grids by 2030 are described below. develop harmonized technical specifications. Doing so will expand the total market for developers and their suppliers by making it easier to do business in multiple countries. POLICY MAKERS Policy makers can leverage the latest geospatial analysis DEVELOPMENT PARTNERS technology to develop high-quality national electrification plans to guide investment in mini grids, main grid extension, Development partners can coordinate the design and and solar home systems. They can develop legislation that funding of strategic interventions to crowd in public and supports the electrification plan with a clear institutional private investment. For the mini grid sector to scale up framework. They can ensure that existing legislation—for rapidly over the next decade, investment from all sources example on renewable energy, rural development, and for- will need to increase dramatically. Coordinated funding eign trade and import duties—is conducive to large-scale from donors will be essential to de-risk—and therefore deployment of mini grids powered by renewable energy. catalyze—early large-scale investment in mini grids in countries marked by large deficits in access to energy. In parallel, development partners will need to support the REGULATORS development, diffusion, and uptake of actionable knowl- edge throughout the sector. Regulators can adopt a light-handed approach and provide clear guidance in the five key areas discussed in chapter 9: market entry, retail tariffs, service standards, technical INDUSTRY ASSOCIATIONS standards, and arrival of the main grid. Regulators can ensure that the process of developing or changing regula- Industry associations should hold members and stake- tions is transparent and involves extensive public consulta- holders to account by developing and tracking progress tion. Regulators can adopt a light-handed approach that (1) toward key performance indicators (KPIs) that are linked minimizes the amount of information required of develop- directly to enlarging the mini grid sector. In the overview to ers, (2) limits the number of separate regulatory processes this handbook, we presented a set of KPIs that are relevant and decisions, (3) standardizes documents and forms, and for mini grid developers; similar KPIs can be developed for (4) integrates related decisions made by other government other associations of stakeholders, such as solar photovol- MINI GRIDS FOR HALF A BILLION PEOPLE    289 taic manufacturers and appliance manufacturers. To track RESEARCHERS progress toward their KPIs, industry associations should collect data from their members on a regular basis and Both qualitative and quantitative research is needed to make appropriate data available to the public in a clear and address the knowledge gaps identified earlier in this chap- transparent way. ter and to identify statistically significant causal relation- ships that can inform investment decisions. When carrying out research, it is important to ensure that people in low-in- come countries are actively engaged in the knowledge MINI GRID DEVELOPERS development process, with a particular focus on engaging Developers should work individually and collectively female scholars, students, and research subjects. toward two KPIs that will help the industry grow at scale: (1) increasing the pace of deployment through a portfo- lio approach to mini grid development, and (2) providing TOPICS FOR FUTURE RESEARCH service of superior quality. With support from develop- ment partners and other stakeholders to crowd in finance In the course of assembling this handbook, we have identi- and establish enabling business environments in key fied the following avenues for future research. access-deficit countries, the cost of mini grid electric- ity could fall to $0.25/kilowatt-hour (kWh) by 2025 and COLLECTING DATA ON INSTALLED AND PLANNED $0.20/kWh by 2030. To facilitate collective action, mini MINI GRIDS grid developers can join an industry association and sup- In the overview and chapter 1, we present data and analy- port other initiatives that promote continued professional- sis from extensive surveys of mini grids around the world. ization of the industry. However, the data sets that underpin these chapters can be improved. The global database of installed and planned mini grids discussed in the overview does not include data INVESTORS from several countries that are likely to see large numbers of mini grids—for example, Brazil and several countries Investors should strive to develop financing vehicles to in North Africa and the Middle East. In addition, there are channel investment—debt, equity, risk-sharing instru- likely to be a large number of small diesel-only mini grids ments, and convertible notes—from large and small inves- scattered around the world that do not appear in our data tors into portfolios of mini grids. The only way to reach the set. The data set for chapter 1 contains detailed informa- scale of investment needed to connect 490 million peo- tion for more than 400 mini grids, but statistical analyses ple to mini grids by 2030 is to enable current investors to will be even more robust and generalizable if they are con- invest more and to attract new investors. Existing financial ducted on a larger number of mini grids from a wider set flows are typically the result of one-off deals. What the sec- of countries. tor needs are new types of financing vehicles, developed for Future data collection efforts should focus on countries not investors by investors, that remain actively managed and well covered in the databases discussed in the overview sustainably funded. and in chapter 1, and on diesel-only and renewable-only mini grids. Establishing teams of researchers focused on particular countries or regions would be a good way to SUPPLIERS improve data collection. In addition, regular surveys—for example every two years—will help improve trend analyses Suppliers are urged to take a longer view of the mini grid on the global mini grid market and mini grid costs. Demand industry than what quarterly reporting might suggest. That for mini grid services from different types of customers, means preparing now—when the market is nascent—for ideally over time, is another key data gap to fill. 2025, when the market will be expanding rapidly at scale. To demonstrate their commitment to a long-term view, COMBINING MINI GRIDS, SOLAR HOME suppliers can consider two types of discounts to mini SYSTEMS, AND MAIN GRID EXTENSIONS INTO grid developers. First, suppliers could allow developers to AN INTEGRATED ELECTRIFICATION STRATEGY AT receive bulk-order discounts for all orders made within a THE LOCAL LEVEL certain timeframe—say, two years. Second, suppliers could In chapter 2, we focused on geospatial analysis and other offer significant discounts for prepaid orders of compo- tools to help countries and developers prepare portfolios nents to be shipped at a date in the future—say, two years of mini grids. As we highlight in that chapter, as well as in from the order date. the overview, mini grids are one of the three main strategies 290   MINI GRIDS FOR HALF A BILLION PEOPLE for increasing electricity access, alongside solar home sys- As we note in chapter 6, if mini grid developers are going tems and extending the main grid. Electrification planning to attract investment, they will need to develop a viable that incorporates all three strategies tends to delineate geo- business plan. What makes a good business plan? In par- graphic areas where each strategy is the least-cost solution. allel, developers may need to design a “pitch deck”—a However, a topic that has received less attention is how best short presentation to show investors as a way to attract to combine two or all three of these technologies in a single their investment. What does that entail, and how should area. For example, while a mini grid may be the best solution developers pitch their businesses to investors? Answers for a town that is densely populated but far from the main to these and other questions about how to own and oper- grid, it may not be economically viable for the mini grid to ate a mini grid business in key energy access-deficit coun- connect every customer in that town. In situations like this, tries could be the focus of future research, in partnership combining mini grids and solar home systems in the same with business scholars who have studied these topics in geographic area is a solution worth considering. other contexts. A 4,000-person town in Togo serves as a cutting-edge In addition, while much has been written about mini grid example of how mini grids can be combined with solar business models, we do not address this topic directly any- home systems. A partnership between Bboxx, General where in the report. The term business model is often used Electric, and Togo’s Ministry of Energy has deployed a loosely and interchangeably with ownership structure (as solar-diesel–hybrid mini grid and solar home systems to in ACP-EU Energy Facility 2012) or profit-making status (as serve every household, school, and shop in the town—and in Schnitzer and others 2014). The framework in table 11.1 to provide streetlights (Ross 2019). Notably, Bboxx, as from academic research on business models (Knuckles the developer, uses its proprietary digital technology to 2016; Baden-Fuller and Haefliger 2013), can provide the remotely monitor not only the mini grid but also the solar basis for a more structured discussion of mini grid busi- home systems and streetlights, thus bringing the manage- ness models in future research. ment control of different electrification solutions under one A more structured discussion of mini grid business mod- metaphorical roof. els is useful for at least three reasons. First, it can offer Some examples of places where combined electrification insight into how mini grid sectors are evolving over time. strategies may make sense are described below. The first-, second-, and third-generation mini grids that we describe in this handbook display significantly differ- • Mini grids and main grid extensions. Urban and peri-ur- ent business models. Second, it can identify areas where ban areas underserved by the main grid, or where the developers and suppliers can innovate. For example, main grid is unreliable. recent innovations in monetization include pay-as-you-go • Mini grids and solar home systems. Towns far from the metering paid using mobile money. Third, it can help reg- main grid with dense population centers and house- ulators and policy makers develop enabling environments holds far from the town center. that accommodate a variety of business models and • Solar home systems and main grid extension. Very permit innovation in the business model over time. Spec- low-income urban and peri-urban areas where some ifying options for what business models are available to households cannot afford a connection to the main grid mini grid developers when the main grid arrives is just one or lack a permanent housing structure. example of this. • Mini grids, solar home systems, and main grid exten- POLICIES AND BUSINESS ENVIRONMENT sion: Urban or peri-urban areas that are underserved FACTORS THAT AFFECT MINI GRIDS by the main grid and where some households cannot Chapters 9 and 10 discuss the regulatory and general busi- afford a connection to the main grid or mini grid or lack ness environments within which mini grids must operate. a permanent housing structure. By focusing on regulations, chapter 9 does not enter into BUSINESS TACTICS AND STRATEGIES FOR MINI a detailed discussion of policies that affect mini grids. GRID DEVELOPERS However, various policies have a direct impact on mini grid businesses, from national electrification and rural develop- In chapter 5, we zoom in on the private-sector entities that ment plans to policies that restrict or encourage trade and participate in the mini grid value chain; in chapter 6 we dis- others that protect the environment, workers’ rights, and cuss how to channel financing to private-sector mini grids. gender equality. While it is beyond the scope of this book, A related topic not covered in those chapters is business an assessment of how different policies affect mini grids techniques and strategies that private-sector developers would constitute an important contribution to the sector. can use to raise money for their business and to operate In chapter 10, we did not address the full set of bureau- the mini grid. MINI GRIDS FOR HALF A BILLION PEOPLE    291 TABLE 11.1 • Analytical framework to guide future research on mini grid business models Business model dimension Example topics Customer Grid-connected versus stand-alone: When is the utility a viable customer? Anchor-Business-Community model: What makes a good “B” or “A” customer? Productive users of electricity: What are effective ways to target local businesses? Value proposition Affordability and willingness to pay: How do developers find the right price point? Tier of service and quality of service: What levels of service do customers want? Appliances and productive use: How much value do these add, and for whom? Value chain Who designs, builds, owns, operates, and maintains the mini grid? Is the industry moving toward specialization or vertical integration? How can the supply chain for components be made more efficient? Monetization Tariff level and structure: What works best, when, and why? Connection charges: How can these be made affordable for customers? Cross-subsidization: When do cross subsidies work and why? Financing: Should developers finance connection charges and appliances in-house or go through third parties? Source: Adapted from Knuckles (2016). cratic processes that affect the ease with which mini grid REFERENCES developers do business. Land rights, business permitting, ACP-EU (Africa,Caribbean,and Pacific Group of States–European Union) taxes and import duties, and corruption are just some of Energy Facility. 2012. “Sustainability—Business Models for Rural the important business environment topics that future Electrification.” Thematic Fiche no. 7, European Commission, Brus- research on mini grids should consider. The World Bank’s sels. https:/ /europa.eu/capacity4dev/public-energy/documents/ Doing Business initiative could serve as an excellent sustainability-business-models-rural-electrification. resource and partner for this work. Baden-Fuller, C., and S. Haefliger. 2013. “Business Models and Techno- logical Innovation.”Long Range Planning 46 (6): 419–26. https://www. Connecting 490 million people to mini grids by 2030 is a sciencedirect.com/science/article/pii/S0024630113000691. monumental task that will require unprecedented levels of Knuckles, J. 2016. “Business Models for Mini Grid Electricity in Base investment, innovation, and commitment from developers, of the Pyramid Markets.” Energy for Sustainable Development 31 investors, development partners, governments, and other (April): 67–82. https:/ /www.sciencedirect.com/science/article/ stakeholders. We hope that the ideas presented in this pii/S0973082615001349. book can serve as motivation and guidance for decision Ross, K. 2019. “Minigrid Combines Solar and Digital to Electrify Togo makers all along the mini grid value chain. Village.” Renewable Energy World, April 25, 2019. https:/ /www. renewableenergyworld.com/articles/2019/04/minigrid-com- bines-solar-and-digital-to-electrify-togo-village.html. Schnitzer, D., D. S. Lounsbury, J. P. Carvallo, R. Deshmukh, J. Apt, and D. M. Kammen. 2014. Microgrids for Rural Electrification: A Critical Review of Best Practices Based on Seven Case Studies. Washington, DC: United Nations Foundation. https:/ /www.researchgate.net/ publication/304969460_Microgrids_for_Rural_Electrification_A_ critical_review_of_best_practices_based_on_seven_case_studies. 292   MINI GRIDS FOR HALF A BILLION PEOPLE Energy Sector Management Assistance Program The World Bank 1818 H Street, N.W. Washington, DC 20433 USA esmap.org | esmap@worldbank.org