MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Reigniting Growth for All October 2021 © 2021 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. 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 as long as full attribution to this work is given. Attribution—Please cite the work as follows: “World Bank. 2020. Mozambique Country Economic Memorandum © World Bank.” All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover Design: Florencia Micheltorena. Photo by Isono / IStocks, Photo by Gustavo Mahoque / World Bank. Page 4: Photo by Nathaniel Tetteh / Unsplash. Page 10: Photo by John Hogg / World Bank . Page 30: Photo by bfk92 / IStock. Page 48: Photo by ilyankou / Shutterstock. Page 70: Photo by RossHelen / IStock. Page 94: Photo by Nathaniel Tetteh / Unsplash Acknowledgements This report was prepared by a team led by Shireen Mahdi (Lead Country Economist, ELCDR), and Fiseha Haile (Senior Economist, EAWM1). The team included Albert Pijuan (Senior Economist, EAEM2), Anna Carlotta Allen Massingue (Economist, DFCII), Elwyn Davies (Economist, ETIFE), Fernanda Ailina Pedro Massarongo Chivulele (Research Analyst, EAEM2), Martin Molinuevo (Senior Private Sectror Specialist, ETIRI), Jakob Engel (Economist, ETIRI), Patrick McSharry (Consultant, EAEM2), Hannah Nielson (Consultant, EAEM2), Halfdan Mangueira (Consultant, EAEM2), Dirk van Seventer (Consultant), Lillyana Daza Jaller (Consultant, ETIRI), Adriana Conconi (Consultant, EAEF2), Taciana Lopes (Consultant), Gabriela Schmidt (Economist, EAEM2), Brigida Tchamo (Program Assistant, AECS2), Adelina Mucavele (Program Assistant, AECS2), and Nani A. Makonnen (Senior Program Assistant, EAEM2). Fiona Hinchcliffe provided editorial support. Florencia Micheltorena designed the report. The report was prepared under the overall guidance of Idah Pswarayi-Riddihough (Country Director, AECS2), Mathew A. Verghis (Practice Manager, EA1M2), Carolin Geginat (Country Manager, ECCAR), and Paulo Guilherme Correa (Program Leader and Lead Economist, EACS2). The team would like to also acknowledge the crucial comments received from Andre Herzog (Senior Urban Specialist, SAEU2) and all the participant from the seminar session that preceeded the preparation of the report. The report was supported by the generous financial support from the Foreign Commonwealth and Development Office (FCDO former UK Government Department for International Development -DFID). It also benefited from the Umbrella Facility for Trade Trust Fund. The team also gratefully acknowledges the collaboration with the Government of Mozambique. Regional Vice President: Hafez Ghanem Country Director: Idah Pswarayi-Riddihough Global Practice Director: Marcelo Estevao Practice Manager: Mathew Verghis Task Team Leaders: Shireen Mahdi and Fiseha Haile Table of Contents Abbreviations 3 Executive Summary 5 Chapter 1. Rethinking Mozambique’s Current Growth Model 11 1.1 Economic growth has lost momentum in recent years 11 1.2 What’s driving growth in Mozambique 14 1.3 Why the current model needs rethinking 23 References 28 Chapter 2. Measuring and Macro-Managing the Resource Boom 31 2.1 Introduction 31 2.2 LNG investments and impacts 31 2.3 There are substantial macroeconomic risks linked to LNG development 37 References 46 Chapter 3. Mozambique’s Growth Story from a Spatial Perspective 49 3.1 Introduction 49 3.2 Satellite data helped to build a district-level GDP database 51 3.3 The results reveal the growth dynamics behind spatial poverty trends 53 3.4 Poverty will persist without boosting agricultural productivity 57 3.5 Policy recommendations 66 References 68 Chapter 4. Can Services Drive Economic Growth 71 4.1. Introduction 71 4.2. Can Services Become a Driver of Economic Growth 76 4.3. Strengthening the Regulatory Environment for Services 84 4.4. Recommendations 91 Chapter 5. How Does Corruption Affect Firm Performance 95 5.1. Introduction 95 5.2. What are the broader governance challenges in Mozambique 96 5.3. How do firms in Mozambique experience corruption? 98 5.4. Recommendations 104 References 106 Annexes 108 Abbreviations and Acronyms AMPETIC Mozambican Association of ICT companies CEM Country Economic Memorandum CEMPRE Mozambican Enterprise Census CFMP Current medium-term fiscal framework (Cenário Fiscal de Médio Prazo) CIP Centro de Integridade Pública CPSD Country Private Sector Diagnostic FDI Foreign direct investment FID Final investment decision GDP Gross domestic product GDPpc Gross domestic product per capita GoM Government of Mozambique GVC Global value chain IMF International Monetary fund INE National Institute of Statistics (Instituto Nacional de Estatística) IOF Household Budget Survey (Inquérito sobre Orcamento Familiar) ITES IT-enabled services LIC Low-income country LNG Liquefied natural gas MTPA Million tons per annum MZN Mozambique metical (currency) PFM Public financial management PIH Permanent income hypothesis PIM Public investment management PPP Public-private partnerships PPP Purchasing power parity SAM Social accounting matrix SME Small and medium-sized enterprise SOE State-owned enterprise SPV Special purpose vehicle SSA Sub-Saharan Africa TCF Trillion cubic feet USD United States dollars WB World Bank WDI World Development Indicators 4 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Executive Summary Mozambique is endowed with rich natural resources, but Growth in recent decades has been driven by investment it continues to face substantial development challenges. and a rapid expansion of services and industry. Since the The discovery of huge gas reserves is particulary expected to mid-2000s, the advent of extractives-focused megaprojects provide Mozambique with a transformative opportunity for has boosted investment growth, supported by a surge in FDI. sustained and inclusive growth. Mozambique has generally FDI inflows into Mozambique accounted for 15 percent of been considered a development success story given its rapid total inflows into SSA in recent years. However, investment and sustained growth economic over the last two decades. has declined in recent years, partly reflecting the fiscal However, poverty remains high and income inequality has consolidation efforts that followed the hidden debt crisis. The been rising. Almost half of the population is classified poor, rapid growth in the services sector has offered a wider path most of whom live in rural areas, concentrated in the Northern to jobs outside agriculture, but services remain dominated and Central regions. The deteriorating security situation and by low-productivity commerce and informal activities. A armed insurgency in Cabo Delgado also presents another large share of the population is still engaged in informal rural pressing challenge. In addition, the country’s weak human activities, predominantly smallholder agriculture, which are capital development undermines the prospects for long- associated with high rates of poverty. While sectors which term growth. Mozambique is now at a crossroads with the exhibit the most dynamism generate few jobs, those which many opportunities and risks associated with the discovery employ the most people lack dynamism. of some of the largest natural gas (LNG) fields in the world. These opportunities and risks need to be managed in the Making the most of Mozambique’s natural gas resources context of the challenges posed by the COVID-19 crisis. and bringing growth closer to the poor will require a new ambitious growth model that goes beyond extractives. The Mozambique’s decades-long remarkable growth existing growth strategy is limited in its capacity to support performance lost momentum since 2016. It has been one of accelerated poverty reduction and drive equitable growth. the fastest-growing economies in sub-Saharan Africa (SSA), This CEM explores Mozambique’s current growth model in supported by the return of peace in the early 1990s. Political great analytical detail. From this analysis it makes concrete stability provided a foundation for sound macroeconomic recommendations for what a new growth model could look policies. Improved economic management, in turn, helped like. attract foreign direct investment (FDI) and large donor support. A rapidly expanding resource sector has also The report consists of five chapters. Chapter 1 provides an played a major role. Today, Mozambique is no longer a star overview of Mozambique’s current growth model, asking growth performer. Growth decelerated sharply following the what’s driving growth and outlining why this model needs hidden debt scandal in 2016 and will enter negative territory rethinking. The remaining chapters explore selected issues in 2020 for the first time in almost three decades. The in further detail. Chapter 2 provides analysis of the potential economic fallout from the debt crisis has been exacerbated impact of Mozambique’s resource boom on GDP, exports, by the impact of the tropical cyclones in 2019, highlighting revenue and employment, and discusses how to make Mozambique’s high level of vulnerability to climatic and good use of the opportunities and manage the associated policy-induced shocks, and the COVID-19 pandemic in risks. Chapter 3 tells Mozambique’s growth story from a 2020 (see Box). These shocks have exposed the country’s spatial perspective. It constructs a unique district-by-district existing vulnerabilities, stemming from overreliance on sectoral GDP database to identify the main growth nodes exports of primary commodities and capital-intensive mega- in Mozambique and understand why there is a weak link investments with limited local linkages. between growth and poverty reduction. The services sector MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 5 is the subject of Chapter 4, exploring how to overcome Growth has become progressively less inclusive over the bottlenecks to deliver on its potential to drive growth in last two decades as Mozambique increasingly depended Mozambique. Chapter 5 continues this theme, examining on capital-intensive megaprojects, with limited linkages with the challenges posed to private sector growth by weak the rest of the economy. Inter-sectoral linkages have been governance and rising corruption. All five chapters make undermined by poor infrastructure and large geographical policy recommendations for the way forward. distances between areas of highest agricultural potential and population mass (rural centre and north) and areas of The identification of thematic areas followed the greatest effective demand (urban south). In addition, rural Bank’s new Country Economic Memorandum (CEM) areas—where poverty is deeply entrenched—suffer from 2.0 analytical framework.1 The first phase of the CEM 2.0 periodic cyclones, droughts, floods and epidemics. approach consisted of a Country Scan of key constraints to growth in Mozambique.2 Based on the data scan and Weak human capital undermines the prospects for long- consultations3, three pillars emerged as prominent issues: term growth in Mozambique. Although the country has one (i) raising productivity and competitiveness (of the services of the highest social spending rates in SSA, human capital sector); (ii) broadening growth beyond the major cities; formation remains weak. Mozambique ranks 148 out of 157 in and (iii) making the most of the resource boom. Together, the World Bank’s 2020 Human Capital Index, below most of these pillars have the objective to steer Mozambique its peers in SSA. A child born in Mozambique today will be 36 towards a more diversified and inclusive growth path percent as productive when she grows up as she could be if whilst leveraging off the highly anticipated exploration of she enjoyed complete education and full health. The country its natural resources. The chapters in this CEM address key also has one of the lowest low level of learning compared to issues in the aforementioned three areas. The report also those of peers, which is a critical obstacle to faster and more discusses governance challenges to private sector growth inclusive growth in Mozambique. Enabling Mozambicans, in Mozambique, with a deep dive assessment on corruption. particularly the poor and vulnerable, to contribute to and The CEM 2.0 framework is not expected to be exhaustive benefit from economic progress requires investing in their and Mozambique may face challenges that are outside the human capital. Human capital formation will be essential to scope of the report, some of which have been addressed ensure that jobs in more productive sectors can be created elsewhere. For instance, constraints and opportunities for and filled. Starting in infancy, many Mozambicans—especially private sector development are discussed in the Country those in rural areas and in the Central and Northern regions— Private Sector Diagnostic (2020) while the role of agriculture lack access to the services and investments that are essential in fostering growth and poverty reduction are addressed in for productive lives by adulthood. Households in these the Rural Income Diagnostic (2020). The thematic areas in provinces may be unable to benefit from growth due to a this CEM are largely in line with the Mozambique’s Five-Year lack of productive assets (including human capital), which Development Plan which seeks to “adopt a more diversified limits their access to economic opportunities. and competitive economy, promote the productive sectors to increase income generation and create more jobs, Recent macroeconomic conditions have not been especially for young people.“4 favourable to growth. Despite the most recent crisis of economic governance, the tail end of which has coincided The current growth model needs with COVID-19, macroeconomic management has been largely sound. Nonetheless, Mozambique is currently facing rethinking high levels of public debt that may impede investment through several channels. Productive investments (for Mozambique’s current growth model has been associated example, electricity and infrastructure) are needed to with a rise in inequality and has had a limited impact on crowd in private investment and to ensure sustainable poverty reduction. Although the poverty rate fell from 60.3 growth. However, there is limited fiscal space to undertake percent in 2002/03 to 48.4 percent in 2014/15, people in public investment given the implications for fiscal and debt the bottom 40 percent of the income distribution have sustainability, much more so with the ongoing crisis. In been largely left behind. It is now amongst the most addition, policy uncertainty and the fear of higher future unequal countries in SSA, having seen its Gini coefficient tax rates to address a tight fiscal situation may hurt investor increase from 0.47 to 0.56 between 2008 and 2014. confidence and limit private investment. Tax incentives may 1 For more details, see the Concept Note and Country Scan for the Mozambique Country Economic Memorandum. 2 The country scan was based on a review of a wide range of data, an update of the growth diagnostics indicators and relevant analytical work, including the recent jobs diagnostic, poverty assessment and urbanization reports for Mozambique. 3 Consultations were conducted at concept stage with government officials, including the Ministry of Economy and Finance, the Bank of Mozambique, and the National Institute of Statistics, as well as representatives from the banking sector and the Confederação das Associações Económicas (CTA) private sector association. 4 To achieve the main objective of the Five-Year Development Plan, the Government identified the following strategic areas: (i) transforming and modernizing agriculture; (ii) investing in economic and social infrastructure; (iii) promote the potential of extractive industry; (iv) establish links between tourism and other economic sectors; (v) strengthen the development of artisanal fishing. 6 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM help attract private investment, but the government would role. Tackling these weaknesses may help boost the future lose much-needed tax revenues that could help reduce debt contribution of this important sector to growth. and finance development in the short to medium term Competitiveness and productivity in Mozambique are The LNG boom raises both opportunities and risks that further constrained by weak governance and corruption. need to be carefully managed. The major challenges that The private sector could become an engine of economic Mozambique faces during LNG development differ between transformation and job creation, but it is being held back the investment and production phases. During the investment by several constraints. A weak governance framework is phase (the first ten years), the main challenge is to balance hampering regulatory reforms needed for services sector the tight fiscal space and need for fiscal consolidation with growth, while corruption is cited by Mozambican firms as one continued borrowing for the LNG projects in a context of of the top three obstacles to their operations. A regulatory already high levels of debt. Revenue will start to increase from assessment with a focus on the backbone services of ICT, around 2033 when sites are operational and cost recovery telecoms, transport and logistics identifies the key regulatory on LNG projects has been achieved. During the production bottlenecks which partly stem from poor implementation phase, Mozambique will need to address revenue volatility of existing rules rather than restrictive regulations per se. and the loss of competitiveness of non-tradable sectors. Analysis in this CEM also finds that corruption is increasing and that it has a negative effect on firms’ sales and Agriculture’s stagnant performance is keeping the rural productivity growth. It particularly reveals that corruption is poor in a poverty trap. Low agricultural productivity is most damaging to high-performance firms (those with high trapping the rural poor in poverty, while growth in industry sales and productivity growth rates) that pay relatively small and service sectors is dynamic, with growth nodes emerging bribes. The results strongly suggest corruption is a serious amongst poor rural districts—especially those with good constraint on private sector growth in Mozambique. connections to urban areas. Mozambique’s crop yields are not only low by regional standards but are also far below Mozambique also faces considerable challenges in the country’s agricultural potential. Increasing agricultural ensuring gender equality in access to opportunities and productivity could have a major impact on poverty reduction resources. With regards to legislation, Mozambique has in the short-to-medium term. Relative isolation and high made progress in advancing equality of rights between transport costs are important major barriers for farming men and women. Under the Constitution, men and women households in accessing input and output markets and are ensured equal rights under the law. However, major raising productivity. challenges remain with regards to the implementation of the various laws and rights. However, women still have lower Innovative analysis in this CEM finds that growth was access to land compared to men due to discriminatory fastest in rural areas, albeit from a low base, supported by customary practices. The enforcement of policies remains non-agricultural sectors, while agriculture, where most of weak partly due to deficient institutional development. the poor are, stagnated, contributing to the perpetuation of Women in Mozambique achieved, on average, only 1.4 the poverty trap. Growth in secondary towns and cities has years of schooling, two years below the average schooling been shown to contribute to more inclusive growth and to among men of 3.4 years which is also very low. Women’s support agricultural markets and structural transformation lower education and skills make it more difficult for them in rural areas. Given the proximity of small towns and to find work in the formal sector and negatively impact secondary cities to rural areas, their growth tends to improve their potential as entrepreneurs. Overall learning levels are consumption linkages, create demand for non-farm labor, significantly lower in the North and Center of the country. and improve market linkages between actors in agricultural The impact of COVID-19 on women’s income is expected value chains. to be large due to the containment measures imposed by the pandemic which limited their ability to acquire and sell The services sector has been the largest driver of products. Not only are Mozambican women less likely than employment and growth in the past two decades, but its men to work in the small formal sector, they are also paid less. productivity is declining. Services has provided employment opportunities to workers shed by agriculture. However, the What would a new model for growth sector remains dominated by low-productivity commerce and informal activities mostly clustered in urban areas. look like? Skills and marginal productivity of new workers entering Mozambique will have to go beyond extractives. It needs to the services sector are below average. As more and more move away from the dual focus on megaprojects and low- labor shifts from agriculture to services, in the absence productivity agriculture towards a more interconnected and of significant capital investments in services, the sector’s competitive economy that shares growth more equitably. productivity growth is expected to continue to decline. The It is essential to address the excessive reliance on external large share of small informal firms also plays an important savings and develop a clear (vision for a) domestic engine MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 7 of growth. This will involve: (i) developing and upgrading the Seeking Growth Elsewhere, services economy; (ii) modernizing and raising agricultural productivity; and (iii) maximizing comparative advantages in and for Everyone small-scale manufacturing, as well as strengthening linkages Orient policies towards strengthening linkages between the with the extractives sectors. extractives sector and the rest of the economy. As extractive industries continue to play an increasingly important role To mitigate the fiscal risks associated with the LNG in the economy, stronger inter-sectoral linkages would projects, several measures are necessary. The Government help to share the benefits of resource-driven growth more of Mozambique (GoM) needs to consider implementing widely. Public investment, budget allocation and revenue different strategies at different phases of the LNG project mobilization will need to take spatial disparities into account. cycle to achieve fiscal consolidation on the one hand and For example, public investment and access to infrastructure manage volatile revenues on the other: needs to be balanced across regions. It is essential to establish specific fiscal targets to reach underserved areas in • Develop a strong and credible fiscal framework with the five-year plan (Plano Quinquenal do Governo). Budget a clear, simple fiscal rule. While revenue from LNG allocation formulas could be updated to take access gaps projects will significantly increase only in about a into account. The continued development of strategic decade, clear fiscal rules and targets are needed now growth corridors can both encourage private investment in to have a well-functioning framework for managing upstream and downstream sectors and spur diversification in revenue volatility in the future. the non-resource economy. • Contain the public sector wage bill and review the role of SOEs in the economy. Mozambique’s public sector Increase agricultural productivity, focusing on areas wage bill remains amongst the highest compared to with greatest agricultural potential. For growth to make peers. It increased from 8 percent of GDP in 2008 to a significant dent in poverty, more needs to be done to 11.6 percent (nearly 60 percent of total tax revenues) raise agricultural productivity and increase rural mobility in 2019. Mozambique’s underperforming SOEs have to connect farmers to markets and the growing non-farm placed added pressure on the budget. There is a need opportunities in their districts. There is significant unrealized to conduct a thorough mapping exercise of the role of potential to achieve efficiency gains in crop production in SOEs in the economy. This should consider revisions of the north and center of the country. Improving access to the participation terms of the state-owned hydrocarbon major corridors and markets can increase productivity, company. as seen in eastern Mozambique. Given the GoM’s limited • Improve public investment management. In the short- fiscal space, investments should focus on selected types of term, it is important to avoid significant frontloading roads and areas with greatest agricultural potential. Policies of public investment and delink public investment should support the adoption of improved seeds and fertilizer from LNG revenues. Mozambique needs to continue by liberalizing the agricultural input market and expanding improving projection appraisal and selection, including smallholders’ access to finance. Further, it is crucial to by SOEs, through multi-year planning and consolidating build resilience to climate risks by expanding the coverage the existing digital platform for investment management. of safety nets, increasing the availability and use of local • Create a well-designed sovereign wealth fund with weather information, and fostering risk transfer mechanisms. a dual mandate of achieving short-term stabilization given revenue volatility and long-term savings for future Transition to a more complex type of service sector growth generations. to bolster its contribution to growth and strengthen its role • Transparency and accountability will be critically as a backbone of the economy. Mozambique could grow its important for managing the natural resource boom services sector in size and sophistication, while also pursuing successfully. Among other measures, Mozambique industrial sectors where the country has a comparative should continue to work towards full compliance with advantage. Growth in the services sector—in particular in the Extractive Industries Transparency Initiative (EITI). backbone services, such as telecoms and transport and logistics—has the potential to support the expansion of Gradually scale up investment once significant progress is the economy and to spur growth in other export-oriented made on fiscal consolidation and structural reforms. The activities, including agribusiness and mining. Development gradual increase of investment should be accompanied of these sectors, which are still severely underdeveloped, by well-formulated sectoral policies. While prudent should remain Mozambique’s top policy priority. macroeconomic management is necessary for sustained growth, policies should go beyond that to boost the Support an open trade and investment regime with competitiveness and productivity of non-LNG sectors. a transparent and effective regulatory framework. Measures are also needed to ensure more equitable Mozambique’s generally open and non-discriminatory distribution of resources. Some of these reform options are regimes are hampered by weak governance and poor addressed below. 8 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM implementation. Strengthening these will be key to developing sectors of the economy. For these IT-enabled services to the services sector. For example, while the GoM has tried grow, they must rely on a solid telecommunications market to improve regulation of the telecoms sector, weaknesses overseen by a sound and transparent regulatory framework. in governance and regulation add hurdles to the expansion See Table below. of telecom infrastructure. A strong telecommunications infrastructure, including broadband connectivity based on a Policy recommendations are presented in each chapter of fiber optic network, is crucial to the information technology the report. Selected ones are summarized in the table below. (IT) sector and to IT-enabled services that support other Main Policy Recommendations Measuring and Macro-Managing the Resource Boom Responsibility Timeline for Strengthen medium-term fiscal framework and adopt fiscal rules Establish medium-term fiscal targets and rules DEEF, DNPO Medium term Develop medium-term budget framework with sectoral three-year budget ceilings DNPO ST to MT Rationalize the SOE sector Consider revising the legal instruments that set the terms of the state-owned MEF, ENH, Medium term company forthe participation in hydrocarbon projects and setting a maximum MMRE investment amount Create a well-designed stabilization fund Establish a sovereign wealth fund with a dual mandate of achieving both short-term MEF, BoM ST, MT stabilization and long-term savings for future generations Mozambique’s Growth Story from Spatial Perspective Scale up access to public services, notably infrastructure, in rural districts Establish targets to reach underserved areas in the Plano Quinquenal do Governo DNO, MEF Short term Update budget allocation formulas to take access gaps into account. Improving agriculture productivity Increase public investments in small-scale irrigation infrastructure MAP, MEF Short term Build resilience to climate risks Expand coverage of formal safety nets, including adaptive schemes MITADER, MAP ST to MT Increase the availability and use of local weather information to manage risks MITADER, MAP Short term Can Services Drive Growth in Mozambique? Transfers of funds Ensure all foreign firms, notably SMEs, obtain access to foreign transfers, even in the BoM Short term absence of Investment Authorization by APIEX, in compliance with the Investment Law. Provide commercial banks with clear guidelines and trainings to allow international BoM Short term transfers by already registered foreign firm -Decree 49/2017 + Notice 20/GBM/2017 Labor regime Consider abolishing quota system and labor market tests (LMT). Replace by AdR Medium term knowledge transfer schemes to increase domestic capacity Telecom Implement “special regime” for access to land by telecom MLE Medium term Update infrastructure sharing regulations and enforce compliance ARECOM Short term How Does Corruption Affect Firm Performance? Improving the regulatory framework and business environment in which firms operate Consolidate licensing for all economic activities under a single entity and mandating MRA. MADER, Medium- term centralized payment of licensing fees. MISAU, MIC Digitize the company registry and make it available to the public, which would Ministry of enhance transparency and allow for easy access by ministries and civil society. Justice Medium- term Enforcing laws and regulations involving firms Enhance institutional capacity to implementing existing laws and regulations MRA, MIC, MEF Short- to with a focus on reducing the cost of contract enforcement and observing public medium- term procurement rules. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 9 1. 10 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Rethinking Mozambique’s Current Growth Model5 Resource inflows from natural gas exports are expected to provide Mozambique with a transformative opportunity for growth. But making the most of these opportunities and bringing growth closer to the poor will require shifting the growth model to broaden the drivers of growth and to raise productivity in sectors with the highest employment potential. Extractives are not enough. The existing growth strategy is limited in its capacity to support accelerated poverty reduction and drive equitable growth. A new growth model with an ambitious focus on achieving sustained and more inclusive, diversified growth is essential. 1.1. Economic Growth has Lost Poverty has steadily declined, but at an uneven pace. The poverty rate fell from 60.3 percent in 2002/03 to 48.4 Momentum in Recent Years 5 percent in 2014/15. Although rapid growth helped reduce poverty, people in the bottom 40 percent of the income Mozambique has experienced sustained high economic distribution have been largely left behind—their share in growth for over two decades. Real gross domestic product private consumption declined over the same period (World (GDP) growth accelerated remarkably following the end of the Bank, 2020). The weak link between growth and poverty protracted civil war in 1992 (Figure 1.1a). Growth averaged 7.9 reduction can be partly explained by deep-seated structural percent over 1993–2015, compared to 0.5 percent in 1980- challenges as well as transitory economic factors (World 1990. The growth recovery during the 1990s largely reflects Bank, 2020). The pattern of growth has become progressively aid-financed post-war reconstruction. The return to political less inclusive over the last two decades as Mozambique stability provided a foundation for sound macroeconomic has increased its dependence on export-oriented, capital- policies. Improved economic management, in turn, helped intensive megaprojects, with limited linkages with the rest attract foreign direct investment (FDI) and large donor of the economy. The non-inclusive nature of recent growth support, which financed infrastructure investment and raises concerns about its future sustainability. In addition, expansion in public services. This, combined with a rapidly rural areas—where poverty is deeply entrenched—suffer from expanding resource sector, made Mozambique one of the periodic cyclones, droughts, floods and epidemics. Further, fastest-growing economies in sub-Saharan Africa (SSA) infrastructure investment, notably road transport, has been (Figure 1.1b). Over 1997–2006, growth was underpinned skewed towards urban areas and export corridors (World by the first wave of capital-intensive megaprojects. During Bank, 2019). 2007-2016, economic activity was mainly driven by the second wave of megaprojects and FDI inflows for coal, and Mozambque’s existing growth strategy has been limited liquified natural gas (LNG) exploration.6 Expansionary macro in its capacity to generate productive jobs and support policies also supported growth. accelerated poverty reduction (Lachler and Walker, 2018).7 Although Mozambique saw a considerable growth in 5 Prepared by Fiseha Haile (Senior Economist, EAEM2) building on the Country Scan and Concept Note. 6 Well-known examples include the Mozal aluminum smelter, the Vale coal mine, and the Sasol gas project. 7 More analysis and discussion on poverty and employment can be found in: World Bank (2018)–Mozambique Job Diagnostic; World Bank (2020) Mozambique Rural Income Diagnostic; World Bank (2020) Fiscal Incidence Analysis. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 11 > > > Figure 1.1. Historically Mozambique’s Growth has Outpaced those of its Peers a. Mozambique and SSA: real GDP growth (1993-2009) 15 Post-conflit First phase of megaprojects Second phase of reconstruction megaprojects 10 5 0 2000 2008 2006 2009 2004 2002 2003 2005 2007 2001 2010 1998 2018 1996 1999 1994 2016 2019 2014 1993 1995 2012 2013 2015 1997 2017 2011 -5 SSA Mozambique Source: World Development Indicators. b. Growth in recent decades places Mozambique among SSA’s top performers 12 Rising stars ETH Top performers GDP % growth, 2005 - 2019 10 8 RWA GHA MOZ 6 ZMB TZA UGA BFA NER KEN AGO GIN SLE TGO CIV MWI SYC MRT NGA STP SEN BWA MLI BEN SDN 4 CMR TCD CPV GNB COG NAM MUS LSO MDG GMB SWZ BDI GABCOM ZAF 2 CAF Weak performers Backsliders 0 -2 0 2 4 6 8 10 GDP % growth, 1995-2005 Source: World Development Indicators. Note: Dotted lines represent averages for SSA countries. labor productivity in recent decades, the earnings of some notably in agriculture8, and promoting private investment in household’s rose faster than those of others, leading to a labor-intensive firms, which created wage employment in more unequal income distribution. Good jobs did not expand urban areas. Mozambique needs to emulate such growth fast enough to absorb the growing, better educated labor pattern. This is all the more important as the working age force. Today, most jobs do not provide a sufficient stream of population is growing rapidly. An estimated 500,000 people income to lift workers and their families above the poverty are projected to enter the labor force each year over the line. The majority of Mozambicans are self-employed or next decade—almost twice as many as over the last decade. work in unpaid family jobs, with wage jobs remaining Going forward, the challenge is to help the workforce scare, even in urban areas. To achieve inclusive growth, (particularly the youth) increase their earnings by creating Mozambique needs to create higher quality jobs. Countries opportunities for more productive employment. that managed to achieve strong and relatively inclusive growth, including Uganda, Rwanda, and Bangladesh, did so The country’s growth performance has outpaced those by investing in the sectors where the poor earn their living, of its regional and structural peers.9 Mozambique has 8 Bolstering rural income growth is crucial to accelerate poverty reduction. Agriculture is the sector with the largest potential to raise the incomes of the rural population and make a significant dent on poverty. World Bank (2020) shows that growth in agriculture will decrease poverty and inequality over three times faster than growth in any of the other sectors. 9 In this report, Mozambique is benchmarked against a set of regional, structural, and aspirational peers to gain insights into areas where reforms could help promote sustained and shared growth. Regional peers are average SSA and LIC countries. Structural peers are countries with economic and structural characteristics similar to those of Mozambique. These include Côte d’Ivoire, Ghana, Guinea, Tanzania, and Uganda. Aspirational peers are countries that set a good development precedent and that Mozambique often refers to for policy inspiration. These include Botswana, South Africa, Malaysia, Mongolia, and Mauritius. 12 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM witnessed one of the most robust post-conflict economic to the suspension of aid in the wake of the debt crisis, and recoveries in SSA. At an average of 4.8 percent in 2000– given the prospects of substantial revenue from gas (Figure 2015, Mozambique’s per capita growth far exceeded the 3d). The economic fallout from the debt scandal has been SSA average (2.2 percent). This allowed the country to exacerbated by the impact of the tropical cyclones in gradually catch up with its SSA peers (Figure 1.1a). Its per 2019, and the COVID-19 pandemic in 2020. Growth will capita income was only 16 percent of the average for SSA enter negative territory in 2020 for the first time in nearly countries at the end of the civil war in 1992, but grew to three decades. The aforementioned shocks exposed the 35 percent in 2019—significant progress, albeit from a very vulnerability of the existing growth model stemming from low base. The country’s growth per capita also surpassed heavy overreliance on the exports of primary commodities its structural peers: Côte d’Ivoire (0.3 percent), Ghana (2.9 and capital-intensive mega-investments with limited local percent), Tanzania (2.5 percent), and Uganda (3.5 percent). linkages. However, Mozambique’s income per capita remains below the averages for SSA and low-income economies, barely The COVID-19 pandemic has had a severe impact on rising from the 3rd lowest in the world in 1993 to the 13th economic activity. COVID-19 is affecting the country lowest in 2019. through external and domestic channels. Externally, the economy suffered from the disruptions in global value However, economic activity decelerated sharply following chains and trade, falling international commodity prices the hidden debt crisis in 2016. In the first quarter of 2016, (such as of coal, gas and aluminum) and lower FDI. the revelation of previously undeclared loans plunged Domestically, the containment measures to curb the spread Mozambique into an economic crisis and macroeconomic of the virus reduced demand for goods and services, which instability, derailing its track record for high growth.10 adversely impacted business volumes, employment and Growth halved from 7.7 percent over 2000-2016 to 3.3 household income, further amplifying the impact of external percent over 2016–2019. The hidden debt crisis led to a shocks (Box 1).11 The economy has contracted by 2 percent severe crisis of economic governance and worsened the between April and September 2020 compared to the same economic slowdown that was already well underway due period in 2019 and is expected to contract for the first time in to the commodity price slump that started in 2014. FDI three decades. The global pandemic will likely undo years of inflows dropped from 36.4 percent of GDP in 2013 to 18.4 hard-won progress in curbing poverty and improving living percent in 2018, as Valle (major coal miner) finished its standards. Moreover, disruption in the delivery of public construction phase and global commodity prices dropped services will worsen the significant infrastructure deficit and around mid-2010s (Figure 1.2b). Official aid also fell from low productivity, which may inflict long-term damage on 17.5 to 12.4 percent of GDP over the same period, due partly potential output. > > > > > > Figure 1.2a. Mozambique and Peers: Figure 1.2b. Global Trends in Key Income per capita Commodity Prices 2000 150 Constant 2010 US$ 1500 100 1000 50 500 0 0 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 2018 2010M01 2010M08 2011M03 2011M10 2012M05 2012M12 2013M07 2014M02 2014M09 2015M04 2015M11 2016M06 2017M01 2017M08 2018M03 2018M10 2019M05 2019M12 2020M07 Mozambique SSA LIC Coal Natural gas Aluminium Source: World Development Indicators. Source: World Bank Commodity Price Data (The Pink Sheet). 10 The metical depreciated drastically from about MZN 31/USD 1 to MZN 80/USD 1 while inflation surged, reaching 17.4 percent in 2016. In 2018/19, the metical stabilized at around MZN/USD 60, while inflation came down to about 3 percent. 11 The services sector output declined by 2 percent in the first 9 months of 2020, the hospitality and restaurant sector was particularly hit contracting by almost 25 percent as a result of the reduced people flow. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 13 > > > Figure 1.3. Investment Has Been a Key Driver of Growth a. Mozambique: Supply-side decomposition b. Mozambique: Demand-side decomposition 15 10.0 7.9% 8.0 7.0% 10 Percentage points 6.4% Percentage points 5.9% 6.0 5 4.0 0 2.4% 2.0 -5 0.0 -10 1992- 1997- 2007- 2012- 2017- 1992- 1997- 2002- 2007- 2012- 1996 2006 2011 2016 2019 1996 2001 2006 2011 2016 Agriculture Industry Services Real GDP Growth Consumption (C+G) Investment (I) Net Exports (X-M) Source: World Bank Sta estimates. Source: World Bank Sta estimates. c. Sectoral contributions to annual growth (2000-2019) 15% 10% 5% 0% -5% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Agriculture Extractives Manufacturing Private services Other and Public services Source: World Bank (2020). 1.2. What’s Driving Growth in investment growth. Megaproject investments accounted for 49 percent of gross investment over 2010–2016 (Figure 3c). Mozambique? FDI played a key role in the development of megaprojects, particularly from 2010 onwards, and averaged 25 percent Investment Has Underpinned Growth on the of GDP over 2010-2019—representing 15 percent of total Demand Side FDI inflows into SSA (Annex A1, Figure 4.a). Accommodative macroeconomic policies also supported growth. Public On the demand side, the growth acceleration since the expenditure rose from 30 to 40 percent of GDP between mid-2000s has been mainly driven by investment (Figure 2008 and 2019. In addition to expansive fiscal policy, the 1.2a). In the 1990s, growth was mostly underpinned by country had several years of expansionary monetary policy, consumption and, to a lesser extent, investment, reflecting with domestic credit to the private sector rising from 12.2 a post-conflict recovery in domestic demand. In the first half percent of GDP in 2000-2010 to 26 percent in 2011-2019 of the 2000s, trade began contributing to growth as exports (Figure 3e). from one of the first megaprojects (aluminum) accelerated. Over 2007–2016, the second wave of megaprojects boosted 14 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Box 1.1. The Impact of COVID-19 on the Mozambique Economy COVID-19 has triggered the deepest global recession in decades. The crises, combined with measures to slow the spread of the virus, has caused the deepest postwar recession. Global GDP is expected to contract by 4.4 percent, from a pre-covid growth estimate of 2.5 percent. The SSA economy is expected to contract by 3.6 percent. Despite the promising prospects of COVID-19 vaccines and the expected economic recovery, global economic activity and income will likely below pre-pandemic levels in the medium-term. The crisis hit the Mozambique economy as it was attempting to recover from the economic slowdown triggered by the hidden debt crisis and tropical cyclones in 2019. Real GDP is projected to contract for the first time in three decades by 0.8 percent in 2020, compared to a pre-Covid estimate of 4.3 percent. The pandemic affected the economy through both external and domestic channels. Externally, lower global demand, combined with the commodity price slump, led to a substantial drop in the country’s exports and reduced investor appetite, especially in the LNG sector. Domestically, containment measures reduced domestic demand, with particularly significant impact on the services sector. Despite the substantial uncertainty, the economy is set to recover from next year, but growth will remain below pre-pandemic levels. The impact of COVID-19 on firms has been severe (see Chapter 4). A survey conducted by the National Statistics Institute (INE), indicates that about 90 percent of the firms in Mozambique have been negatively affected by the crisis. The commerce and retail sectors, which represent more than 70 percent of the small firms in the country, were the most affected. Firms responded to the drastic reduction in revenues by cutting labor costs. A common action by firms was reduction of working hours, as well as the suspension and termination of labor contracts, the latter of which affected about 2.3 percent of the workforce. Households are feeling the impacts of COVID-19 through loss of earnings and employment. The crisis particularly affected urban low-income households engaged in informal services. According to the results of a high frequency survey (HFS) conducted by the INE, 67 percent of the urban population were not working in June 2020, whereas over one-third were working before COVID-19 in March. Over 70 percent of workers indicated that they were unable to work normally – of these, only 9 percent received their wages in full. About 41 percent of the urban households interviewed reported a reduction in their wage income. The loss of earning and jobs have undermined food security, with more than 50 percent of urban households reportedly running out of food. This is likely to hold back urban poverty reduction, and combined with the closure of schools to contain the spread of COVID-19, may dent progress in building human capital. COVID-19 has exposed the pre-existing fragilities of the economy, highlighting the need for structural reforms to avoid lasting scars on potential output. Looking ahead, it is important to reignite the agenda of structural reforms once the crisis subsides, while maintaining the immediate focus on reinforcing targeted support to viable firms to minimize the loss of productivity capacity. Further, social protection programs should be scaled up to extend support to informal entrepreneurs, and incentivize children to return to schools as they reopen. In the medium-term, the Government should continue to strengthen the macro-fiscal framework to enhance fiscal and debt sustainability, and make the most of the resource boom. It is also critical to promote economic diversification and use gains from the LNG sector to support inclusive growth and job creation by enhancing linkages with the domestic private sector. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 15 > > > Figure 1.4. Mozambique’s Private Investment Levels Could Be Stronger a. Private capital stock (in percent of GDP) is well below the SSA average (2015-2017) 3.5 3.0 GHA CPV Private capital stock 2.5 MRT SYC LBR GAB CHE NER AGO 2.0 ZMB ZAF TZA BWA BEN CMR MUS LSO TGO 1.5 SDN COD ZWE UGASEN NAM MDG KEN TCD SOM NGA ETH 1.0 CIV MOZ RWA DJI 0.5 0.0 0.0 0.5 1.0 1.5 2.0 2.5 Public capital stock Source: IMF Investment and Capital Stock Dataset (2019). Note: Dotted lines represent averages for SSA economies. b. Gross investment (Percent of GDP) c. Gross fixed capital formation (2010-2016) 50 Mozambique 40 SSA 30 Construction Infrastructure 20 Manufacturing Megaprojects 10 0 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Source: World Development Indicators. Source: INE. d. Aid and FDI inflows (% GDP) e. Domestic credit to the private sector (% GDP) 20 40 40 15 30 10 20 20 5 10 0 0 0 2003 2005 2007 2009 2011 2013 2015 2017 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Total o cial aid inflows FDI inflows Mozambique Ghana Cote d'Ivoire Guinea Tanzania Source: World Bank and OECD. Source: World Development Indicators. Note: Total aid is on the right-hand side. 16 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Box 1.2. Climate Shocks have mad significant negative impacts on the economy Mozambique is prone to recurrent natural hazards, namely, droughts, earthquakes, floods, and cyclones. According to the National Institute for Disaster Management (INGC), the occurrence of natural hazards has increased in number and magnitude since 1960. Existing databases, which include Emergency Events Database (EM-DAT), DesInventar, and registries from the Ministry of Economy and Finance recorded 71 events (droughts, earthquakes, floods, landslides, or storms) in Mozambique between 1956 and 2016. Additionally, 741 droughts, 437 floods, and 137 cyclones were recorded between 2000 and 2012, and five earthquakes were registered since 2002. Natural disasters have caused the economy massive and permanent damages and losses. The 2019 IDAI and Kenneth cyclones costed the country more than US$ 3 billion (about 20 percent of 2019 GDP) in damages of infrastructure, and loss of production and human lives in 6 provinces (Sofala, Manica, Zambezia, Inhambane, Tete and Cabo Delgado). World Bank Estimates indicate that natural disasters led to an annual average of losses and damages of MZN 4 billion, between 1984 and 2014. However, it is likely that the calculations were underestimated due to the information deficit. Between 2000 and 2014, a period when more information was available, the average of losses and damages was about MZN 7.5 billion. Economic growth has also been impacted by natural disasters, with lower rates being registered in years with climate shock compared to the previous years (figure 6). The drop in growth, mainly reflects losses of production in the agricultural sector, human and physical capital. Additionally, limitations in post disaster recovery processes have led to permanent loss of the productive capacity. For example, in 2013 and 2014 natural disaster events costs were estimated at MZN 11.6 billion, but recovery allocation totaled only MZN 3.5 billion and disbursement was only MZN 1.4 billion. > > > Figura 1.10. Climate shocks have constrained GDP growth Real GDP growth, 1997 – 2019 14% 12% 11.7% 10% Cyclone Chedza – 8% 7.0% 7.7% largest impact in Zambézia/Nampula 6% Cyclone Jokwe – 4% largest impact in Tropical cyclone Nampula/Zambézia Cyclone Idai – largest 2% Elaine – largest impact impact in Sofala and in Gaza/Inhambane Manica 0% 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 20 20 20 20 20 20 20 20 20 20 19 20 20 20 20 19 19 20 20 20 20 20 Pre-2019 Idai Projection Actual Source: National Statistics Institute and World Bank projections MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 17 Box 1.2. continued The negative impacts of climate shocks have constrained long-term growth prospects and reversed development gains. More than half of the Mozambique’s population live along the coastline, making them vulnerable to tropical storms. According to the World Bank Poverty team analysis, more than 50 percent of the population in areas with significant exposure are already poor, vulnerable, and tend to get worse-off with climate shocks. Evidence shows that after the Cyclone Jokwe, which occurred in the provinces of Nampula, Zambezia and Sofala in 2008, affected households saw consumption levels and asset ownership reduced by 25 percent, compared to non-effected households. In general, the provinces more affected by cyclical occurrence of natural disasters are also the ones with the highest poverty levels and inequality, including in terms of access to public services and infrastructures. This is partially explained by the above-mentioned limitations of the post recovery allocations and execution, which contribute to amplify the deficits. Mozambique has taken some steps towards climate change adaptation and mitigation. A disaster management fund was established in 2017 with an initial allocation of 0.1 percent of GDP. Also, government has taken a holistic approach to disaster impact mitigation by undertaking initiatives to develop catastrophe insurance, both sovereign and microinsurance. The legal framework has also been strengthened with a recent approval of the disaster risk management law and a public investment decree. These legal instruments include provisions to force infrastructure investment projects to consider elements of resilience to natural disasters. More, fiscal risk assessments have taken into consideration climate shock risks, however the de facto utilization of the recommendation of the analyses seems to be limited by the availability of fiscal resources to set a mitigation budget. Additional steps have been taken to improve the collection of natural disaster data and to improve capacity in assessing risks and estimating mitigation measures. It also important that these efforts are combined with a strategy that promotes green growth and contributes to reduce the probability of ocurrence of disasters. Source: World Bank et all (April 2018). Financial Protection against Disasters in Mozambique. After increasing steeply for almost a decade, investment Decomposition analysis shows that capital accumulation has declined in recent years, partly reflecting the fiscal made the largest contribution to growth, while the country consolidation efforts that followed the debt crisis – performs poorly on total factor productivity (TFP) and primarily capital spending cuts. However, Mozambique’s human capital (Figure 1.7). Growth can either be driven by an investment rate remains above the SSA average. Gross increase in the factors of production (e.g. labor and capital) investment rose from 14.3 percent of GDP in 2000–2010 to or by a more efficient combination of resources to produce 29.2 percent in 2011–2019 (Figure 3b). This is higher than more output for a given level of input (total factor productivity its structural peers: Côte d’Ivoire (18.9 percent), Ghana (22.5 or TFP). While growth in the 1990s was largely due to percent), and Uganda (25 percent), though slightly lower TFP and labor accumulation, the contribution of TFP has than Tanzania (33.6 percent). It also exceeds the threshold declined in recent decades (Figure 1.7). Capital accumulation suggested by the experiences of the high-growth cases has accounted for three-quarters of GDP growth since identified by the Commission on Growth and Development 1997, which is partly explained by natural capital. The small (2008) — “Overall investment rates of 25 percent of GDP contribution of TFP in recent years partly reflects the fact or above are needed for strong, enduring growth.” These that land expansion, rather than productivity growth, has high-growth economies also invested another 7-8 percent driven agricultural growth, while the use of inputs has been of GDP in education and health, compared to X percent in modest (Figure 1.7d). In contrast, the contribution of human Mozambique (Annex A.1). Despite rising capital investment capital remained low and stagnant. This is reflected in the rates, Mozambique still suffers from a large infrastructure fact that Mozambique ranks below many SSA countries in deficit, notably a poor-quality road network. Moreover, the the World Bank’s 2020 Human Capital Index.12 concentration of basic infrastructure in and around the capital offers limited development prospects for the rest of Mozambique stands out among SSA countries for its low the country (see Chapter 3). level of private capital stock. Investment—by both private 12 A child born in Mozambique today will be only 36 percent as productive when she grows up as she could have been had she enjoyed complete education and full health. The country also has one of the lowest levels of learning of its peers. 18 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM and public entities—is a key driver of growth. Sustaining rapid helped spur growth in services sectors. Driven by over 30 growth without high investment rates is rare. Over time, as megaprojects, services have expanded substantially, with economies become more developed, private investment its share of total employment rising from 9 percent in 1996 takes on a greater role and contributes to growth and job to 24 percent in 2014 (Figure 5b), partly reflecting non-farm creation by directly supporting productive capacity. In self-employment in the informal sector in urban and rural Mozambique, while public capital stock is marginally higher areas. than the SSA average, private capital stock is far below average. Its private investment rate averaged 18 percent of Mozambique, like other LICs, faces a changing global GDP during 2010–2016, lower than those of its structural context for industrial competitiveness. The manufacturing- peers and the SSA average (Figure 1.4a). The private sector led pathway to growth, particularly through export- in Mozambique is underdeveloped, characterized by a led production that was behind East Asia’s economic high degree of informality and a large number of micro transformation, is narrowing for today’s least developed enterprises.13 The informal sector is estimated to account economies. While low-skilled and cheap labor attracted for 89 percent of enterprises but only for 30.9 percent of manufacturing FDI and pushed these countries up the GDP—compared to 22, 39, and 33 percent in South Africa, income ranks, advances in technology are changing the Tanzania, and Zambia, respectively (World Bank, 2020). context. Global value chains are increasingly mechanized as labor-substituting technologies become cheaper. There Over Time, There Has Been a Limited Shift in the is growing concern that—in the face of shifting global Structure of the Economy Away from Agriculture technology and trade policy changes—the traditional path of catch-up via manufacturing-led industrialization may On the supply side, Mozambique saw a shift in its growth no longer be accessible for developing countries. The patterns and drivers since the late 1990s. Post-conflict manufacturing sector has been in stagnation or decline reconstruction in 1993–1997 saw workers return to farms, in low-income countries, including Mozambique, while which helped boost agricultural growth. Agriculture services has been emerging (Annex A.1, Figure 5). accounted for nearly 40 percent of the real GDP growth during this period. However, as noted above, the country Dynamic Sectors Generated Few Jobs, while Those increasingly shifted into an export-oriented growth that Employ the Most People Lacked Dynamism model in the late 1990s on the back of capital-intensive megaprojects. Although investments in the resource Over time, Mozambique has seen a limited shift in the sector have dominated growth since 1998, the sector has structure of the economy away from agriculture. Over generated only limited formal employment opportunities. the last three decades, three main phases are apparent. Megaprojects in extractives, notably coal production and On the return of peace in the early 1990s, Mozambique LNG exploration, took off in 2004 and 2011, respectively. was predominantly an agricultural economy. Starting in The second wave of megaproject investments was the late 1990s, the country experienced strong growth in characterized by rapid growth in the services sector while manufacturing, dominated by capital-intensive investments, the contributions of agriculture and manufacturing declined particularly in the aluminum sector. The late 2000s steadily (Figures 1.3b and 1.3.c). witnessed saw a shift to a third phase, characterized by an emphasis on extractives, driven by large increases in FDI Non-megaproject manufacturing did not take off with the into coal and LNG. Agriculture’s share of total employment investment boom. In fact, the output share of manufacturing fell from 86.6 percent in 1997 to 71 percent in 2015 (Figure declined from 12 percent in 1997 to 8.8 percent in 2018, 1.4b). In contrast, its share of total output barely declined, below the SSA average of 10 percent. During the first wave from 28.7 percent in 1997 to 24 percent in 2019 (Figure of megaprojects, manufacturing made a higher contribution 1.4a). This clearly indicates that agriculture remained a to growth and job creation (Figure 1.2c). However, low productivity activity (Figure 1.6c). The output share manufacturing remains the sector with the lowest labor share of industry rose from 15.3 to 23.6 percent over the same of total employment, at 4-5 percent since 1997. Its weak period, mostly due to mining. However, the relatively performance indicates that the benefits of megaprojects fast growth of industrial output did not translate into have not spilled over into other sectors of the economy. employment growth. The share of jobs in industry has only Much of the megaproject-driven demand for goods was increased from 4.4 to 4.9 percent. The services sector now fulfilled through imports. Manufacturing’s contribution to represents the largest share of GDP, reaching 46.5 percent export growth is low as only 11 percent of manufacturing in 2019. The employment share of services increased from firms are exporters.14 However, the investment boom has 9 to 24 percent between 1997 and 2015. 13 See Lachler and Walker (2018) and World Bank (2020) Country Private Sector Diagnostic for more discussions on informality and inclusive growth. 14 World Bank (2018b). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 19 > > > Figure 1.5. Most Jobs are Still in Low-Productivity Agriculture a. Sectoral shares in gross value added b. Sectoral shares of employment 60% 100% 9.0% 16.1% 15.0% 24.0% 50% 80% 40% 60% 30% 86.6% 40% 80.5% 80.4% 20% 71.0% 20% 10% 0% 0% 1997 2003 2009 2015 91 94 97 00 03 06 09 12 15 18 20 20 19 20 19 20 19 20 20 20 Agriculture Industry Services Agriculture Extractive Manufacturing Services Source: World Development Indicators. Source: World Bank Sta estimates based on data from INE. c. Correlation between sector productivity and change in employment shares (1997-2015) 4.0 Finance Log (sectoral productivity 3.0 /total productivity) Construction Extractives 2.0 Transport 1.0 Manufacturing Other services Commerce 0.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0 10.0 15.0 Agriculture -1.0 -2.0 Change in employment share (percentage points) Source: World Bank Sta estimates based on data from INE. Note: The size of a circle represents the relative size of each sector (as measured by its employment share) in 2015. The horizontal axis captures the change in a sector’s employment share between 1997 and 2015 while the vertical axis measures the logarithm of the share of sectoral productivity in total productivity. In Mozambique, as in many other SSA countries, the bulk services remain dominated by relatively low-productivity of labor movements is still taking place within the informal commerce and informal activities mostly clustered in urban sector. Figure 1.4c plots the changes in sectors’ employment areas. Therefore, while the sectors which exhibit the most shares against relative productivity (as measured by the log of dynamism generate few jobs, those which employ the most the ratio between sectoral productivity and total productivity people lack dynamism (Lachler and Walker, 2018). in 2015). A positive correlation between the direction of labor flows and productivity in individual sectors suggests that Labor productivity contributed to growth as structural growth-promoting structural transformation is taking place. change progressed. Figures 1.7c and 1.7d show the Reflecting the classic pattern of structural change, agriculture decomposition of output growth, which can increase is in the bottom-left quadrant (representing low productivity for various reasons: rising productivity within each sector and a declining labor share) and the relatively more dynamic (if each worker produces more), structural change (if sectors in the top-right quadrant (higher labor productivity workers move from low- to higher-productivity activities), and a rising labor share). The rapid growth in services has demography (if the share of the working age population offered a wider path to jobs outside agriculture. However, rises) and employment (if a larger share of the working age 20 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 1.6. Productivity is Declining and Overly Reliant on Services a. Average labor productivity by sector b. Mozambique: Output per worker (% change) 6000 7% 1996 2003 Câmbio Constante USD 2010 5000 2008 2014 6% 4000 5% 4% 3000 3% 2000 2% 1000 1% 0 0% Agriculture Industry Services Total 1996- 2003- 2008- 2014- 2003 2008 2014 2016 Source: World Bank Sta estimates. Source: World Bank Sta estimates. c. Growth in output per worker d. Formal: sales per employee in constant 2011 international $ (adjusted by 2011 PPP) 3.5 1997-2001 2002-2006 70000 2007 3.0 2007-2011 2012-2016 60000 2018 2.5 50000 Percentage points 2.0 40000 1.5 30000 1.0 20000 0.5 10000 0.0 0 y od g ion l ism es tai er rin Agriculture Industry Services etc. Intersectoral vic Fo Re hin -0.5 ct ur tu Reallocation er tru To ac ac rS ns uf E ect M he Co an , als Ot rM et he M Ot Source: World Bank sta estimates. Source: World Bank (2018b). Note: Permanent and temporary employees winsorized at 95th percentile. population is employed). Productivity improvements were has the highest average productivity, while agriculture has the main drivers of growth, partly reflecting the shift of labor the lowest. However, productivity growth has been steadily to more productive sectors (Figure 1.7a). Both within-sector falling across all sectors since 2003, notably in agriculture and between-sector productivity changes have contributed and services. Transport and wholesale trade had the highest significantly to productivity growth. However, there are sales per employee, on a par with some capital-intensive notable differences across sectors. Lower employment manufacturing sub-sectors (Annex A.1, Figure 1). Labor rates have slightly reduced growth, which indicates that productivity has also picked up in non-retail services whilst job creation has not kept pace with the country’s rapid productivity in manufacturing has fallen. Sales per employee economic growth. (returns) dropped in retail (Figure 1.8d). In sum, the services sector has played a prominent role in sustaining overall However, productivity growth has been declining in productivity growth, but the average productivity level of recent years, especially in services. Since the launch new jobs in the sector is low and declining. Nonetheless, of the first megaprojects in 1997, labor productivity has the shift from agriculture into services still generates increased by an annual average of 5.3 percent. The level of productivity gains because agricultural productivity is even labor productivity increased in almost all sectors. Industry lower. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 21 > > > Figure 1.7. Physical Capital, Rather than Human Capital, Has Driven Recent Growth a. Mozambique: Growth decomposition (1992-2019) b. Benchmarking: Growth decomposition (2000-19) 10 8 8 6 6 4 4 2 2 0 0 ire a a ue ia da an ine an -2 an vo biq 1992- 1997- 2002- 2007- 2012- 2017- Gh nz Gu d’I Ug am 1996 2001 2006 2011 2016 2019 -2 Ta te oz -4 Co M Capital Labor TFP Real GDP Growth Capital Labor TFP GDP Growth Source: World Development Indicators. Source: World Development Indicators. c. Mozambique: Contributions to real GDP growth d. Mozambique: decomposition of agricultural output growth Annual Change (percentage points) 6.5 6% 5.5 4.5 4% 3.5 2.5 2% 1.5 0.5 0% -0.5 1980- 1990- 2000- 2010- -1.5 -2% 1990 2000 2010 2016 1992-2016 1992-1996 1996-20002000-20052005-2010 2010-2016 y=4.57% y=5.2% y=4.94% y=5.54% y=4.44% y=3.21% Demographic Change Participation Rate TFP Input Intensification Employment Rate Productivity Land Output Growth Source: Based on data from USDA. Source: Based on data from USDA. The pattern of structural change observed in Mozambique released by agriculture has mostly been absorbed by low- is characterized by the absence of labor movements to productivity services activities. Despite this, Mozambique’s high-productivity sectors (Box 1.2). A process of structural experience is broadly similar to the process of structural transformation that yields sustained productivity growth and change observed elsewhere in SSA. job creation is marked by: (a) declining employment (or hours worked) and output shares in agriculture; (b) rising The Economy Has Become Less Diversified employment and output share in services; and (c) inverted over Time U-shaped evolution of employment and output shares in manufacturing.15 Although Mozambique has largely Given its heavy reliance on megaprojects, Mozambique’s followed this general pattern, it is lagging in the process. export basket remains concentrated on primary products, The employment and output shares of agriculture and although export volume has more than doubled in the last services have changed at a much slower pace than those decade. Export volume increased from USD 2.4 billion in 2007 of other SSA countries. Manufacturing’s employment share to USD 5.2 billion in 2018. Growth was accompanied by trade remained stagnant while its output share continues to fall. expansion, but Mozambique has a low level of per capita In addition, despite the decline in agriculture’s employment income for its volumes of trade. A narrow range of natural share, the country still has a very large share of agricultural resources and low-value-added agricultural commodities employment compared to countries with similar levels dominate its export basket. In 1996, Mozambique’s export of income per capita. Further, the bulk of the labor force basket was dominated by marine and agricultural exports 15 World Bank (2020e); Duarte and Restuccia (2010, 2018); Herrendorf, Rogerson, and Valentinyi (2014) 22 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 1.8. Mozambique’s Export Basket Lacks Sophistication Compared to its Peers a. Economic complexity and per capita GDP (2018) 12 Qatar Singapore 11 UAE Ireland Switzerland Japan Malaysia 10 Mauritius Libya China Algeria EgyptGeorgia 9 Nigeria Angola India PNG Ghana 8 Guinea Ethiopia Mali Uganda 7 Mozambique 6 -3 -2 -1 0 1 2 3 Economic Complexity Index Source: World Bank Sta estimates. b. Export composition in 1996 (% of total goods exports) c. Export composition in 2018 (% of total goods exports) Source: World Bank Sta estimates. Source: The Observatory of Economic Complexity. (Figure 1.8c), while by 2018, coal, gas, and minerals had 1.3. Why the Current Model Needs emerged as major additional products (Figure 1.8d). Despite the changing composition of the export basket, all new Rethinking and old major exports share the feature of being primary Mozambique needs to move away from dependence on products. Its high concentration on certain products has megaprojects and low-productivity subsistence farming made the country prone to global commodity prices towards a more interconnected and competitive economy shocks. It also reflects the low productivity of other sectors. that shares growth more equitably. A comparison of Mozambique’s performance on the Economic Complexity Mozambique’s growth strategy with those of 13 economies Index,16 which measures the sophistication of an economy’s that have grown at more than 7 percent for 25 years or more export basket, has been falling since the late 1990s, with pinpoints areas where Mozambique can adjust its model.17,18 the country lagging behind its peers with similar per capita While it shares some features with these historically fast- income (Figure 1.8a and b). 16 Harvard University Growth Lab (undated). 17 The 13 economies include Botswana; Brazil; China; Hong Kong, China; Indonesia; Japan; the Republic of Korea; Malaysia; Malta; Oman; Singapore; Taiwan, China; and Thailand (Commission on Growth and Development 2008). 18 Commission on Growth and Development (2008). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 23 growing economies, it diverges in some important aspects, economic opportunities between rural and urban areas. especially conception and outcomes. Table 1 in Annex A.1 Household consumption is nearly three times lower in compares the common characteristics of these economies rural areas than in urban areas and it has been growing with Mozambique. It lists five major characteristics which at a slower pace among households at the bottom of the have promoted fast growth: (i) they fully exploited the world income distribution. Moreover, in terms of non-monetary economy; (ii) they maintained macroeconomic stability; (iii) dimensions, there are also large differences in human and they mustered high rates of saving and investment; (iv) they physical capital between urban and rural areas. Although let markets allocate resources; and (v) they had committed, the poor have gained from growth, wealthier segments credible, and capable governments. Mozambique shares, of society, primarily located in urban areas, have been the to some extent, its exploitation of the world economy. It is main beneficiaries. This reflects the extent to which growth also similar in terms of ensuring relative macroeconomic has been concentrated in urban centers, agglomerating stability – despite the deterioration related to the hidden economic activity and jobs in these areas, including higher- debt crisis – and government leadership in the development skilled jobs in the service sector. process. Disparities in access to basic infrastructure have However, it has made limited progress in mustering particularly been growing between rural and urban areas, high savings rates and letting markets allocate resources especially in the rural parts of Mozambique’s central and efficiently. Limited competition in key sectors, including northern provinces. This is particularly evident in the roads transport, continues to create an uneven playing field, and sector and is linked to a pattern of public investment that hampers private sector growth (World Bank, 2020). The underemphasizes funding to rural areas in northern and country has undergone limited structural transformation central zones.20 Capital investment in roads has been skewed and uneven regional development, with the impact of towards urban areas, contributing to the declining rates of growth felt unevenly across the population (Chapter rural connectivity, whereas the non-road capital budget was 3). Mozambique also features a number of what the more balanced, potentially reflecting progress in rural water, Commission identifies as suboptimal policies, such as electricity and health access. open-ended protection of some sectors and SOEs from competition (World Bank, 2020). Despite the setbacks Without new measures, natural resource exploitation will associated with the hidden debt crisis, Mozambique has continue to be detached from the rest of the economy. traditionally fared well in terms of government leadership in To avoid this and share the benefits of resource-driven the development process. growth more widely, policies should be orientated towards strengthening linkages between the sector and the rest of Growth Needs to be More Inclusive the economy. As extractive industries continue to play an increasingly important role in the economy, stronger inter- Mozambique’s current growth model has contributed to a sectoral linkages would help expand the distribution of the rise in inequality and has had a limited impact on poverty returns to growth. The continued development of strategic reduction. It has been among the fastest-growing economies growth corridors can both encourage private investment in in SSA for decades, but it has struggled to translate its rapid upstream and downstream sectors and spur diversification growth into widespread poverty reduction. Growth has in the non-resource economy. These issues are the topic of been increasingly driven by extractives and expansion of Chapter 2. services, especially in rural areas (Chapter 3), with limited linkages to the broader economy. Mozambique’s impressive Productivity and Competitiveness Need growth rates have disproportionally benefited those at the Strengthening top of the income distribution in urban areas (Figure 1.10a). Poverty rates in Mozambique have been considerably less Mozambique needs a viable strategy to transition its sensitive to economic growth than in other SSA countries.19 output and employment towards better capitalized, Although the country has the potential to use its resource more productive activities that can generate increased wealth to promote inclusive growth, continued focus on earnings in a changing world. Mozambique will have to capital-intensive megaprojects could further accentuate the look for growth everywhere. This will involve: (i) developing weaknesses of the current model. and upgrading the services economy; (ii) modernizing and raising agricultural productivity; and (iii) maximizing Growth has disproportionately benefited large urban comparative advantages in small-scale manufacturing, as centers, widening the gap in living standards and well as strengthening linkages with the extractives sectors. 19 World Bank (2020a). 20 World Bank (2019). 24 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 1.9. Recent Growth Gains Have Mainly Benefitted the Better O a. Consumption growth incidence curves (GIC), b. Basic infrastructure access index by district (2015) 2008/09-2014/15 Q5 12 10 Q4 Annual growth rate, % 8 Q3 6 4 Q2 2 Q1 0 0 10 20 30 40 50 60 70 80 90 -2 Q5 = highest access Q1 = lowest access Urban GIC Rural GIC Note: dotted lines in each graph show 95% confidence intervals. Consumption growth incidence curves with 95% confidence intervals nationwide, urban and rural, 2008/09-2014/15. Source: World Bank using IOF-2008/09 and IOF-2014/15 c. Spatial concentration of economic activity d. Access to basic social services: rural versus urban areas20 90 80 70 60 50 40 30 20 10 0 Owns phone/ mobile education Electricity Sanitation Owns a fridge Owns a TV Close to market Close to road Years of Safe water Urban Rural Source: World Bank Urbanization Review. Source: World Bank Poverty Assessment using IOF-2014/15 e. Households experiencing three or more monetary and/or non-monetary deprivations Urban Rural 2002/03 78.6% 21.4% 2002/03 99.6% 0.4% 2008/09 57.2% 42.8% 2008/09 97.9% 2.1% 2014/15 32.0% 68.0% 2014/15 89.4% 10.6% Poor Non-poor Poor Non-poor Note: Multidimensional poverty is defined by deprivation along a set indicators: education, access to services (no access to electricity, improved water and improved sanitation); housing conditions (poor quality dwelling), asset ownership (no ownership of at least two of the following assets: fridge, TV, phone, bicycle, car or motorcycle) and the prevalence of monetary poverty (consumption per capita is below the poverty line). Source: World Bank using IOF-2002/03, IOF-2008/09 and IOF-2014/15 (Household Budget Survey or Inquérito sobre Orcamento Familiar). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 25 Box 1.3. The non-inclusive nature of the growth model has helped fueled the fragility Although there is no clarity about the nature and causes of the Cabo Delgado insurgency, there is a widespread consensus that among the driving forces behind the insurgency is the systematic sense of exclusion and grievances that were capitalized by extremist groups. Four main areas of tension have been identified: (i) ethnic—the province coexists with tensions between its two main ethnic groups, where the Mwanis that occupies the costal area have had less access to opportunities compared with the Makondes, located in the hinterland; (ii) Disenfranchised youth— the youth remain in a status of “waithood”. With limited access to education, political voice, technical knowledge, and resources they are pushed away from the development benefits. Most of the time options gravitate between taking low productivity informal jobs in urban centers, opting for illicit activities or, in this case, being recruited by extremist groups; (iii) Geographical sense of divide—Mozambique inherited a colonial South-North divide, in which the distribution of public resources and State presence skewed to the South, consolidated a feeling of marginalization and exclusion in the Central and Northern regions, were Cabo Delgado is located. The sense of neglect was exacerbated by the recent increased presence of the State in the province with a heavy-handed approach to formalization of extractives activities. This pushed local population away from extractives activities without alternative forms of employment or access to services; and (iv) political participation – systematic exclusion from political participation and violence against alternative group have perpetuated the use of violence as a form of social participation. Besides these four factors, possible external drivers have also been cited in the analyses of the conflicts. These includes the expansion of violent extremist groups, illicit traffic, international networks, and interests in controlling natural gas revenue. Nevertheless, regardless of the exact motivation and objectives of the insurgents, they were capitalized on existing local grievances and opposition to the State. These conflicts pose a risk to the economic prospects of the economy. The country’s economic growth averaged at 3 percent between 2016 and 2019. Projections point to recovery and acceleration to levels above 8 percent in the next decade with the materialization of the LNG projects. The attacks perpetrated by the insurgents in areas near the site of the projects have already led to operations distress, with the Total led project interrupting on site activities. This adds another layer of challenges to projects still underway on search of financial disclosure, notably the Exxon led project estimated at US$ 15 billion. Simulations from the World Bank show that a full cancelation of the Exxon project may compromise the economic sustainability of the LNG portfolio and constrain growth prospects. The conflict in the center perpetrated by the RENAMO subgroup, has limited the country North-South circulation, besides leading to a displacement of several families. Both conflicts have contributed to double military spending to almost 2 percent of GDP in 2020, adding pressured to the already tight fiscal space. Socio-economic vulnerabilities have been further exacerbated. With poverty rate at 50 percent, the northernmost province of Cabo Delgado is one of Mozambique poorest according to the 2014/15 household survey data. Evidence shows that the province districts are poorly served in terms of infrastructure and public services, with deficit in transport, access to markets, school, health facilities and electricity World Bank (2019). These deficits are further aggravated as the insurgents’ attacks have culminated with loss of life, displacement, and destruction of social infrastructure. Since 2017, the conflict costed the country more than 2 thousand lives and led to more than 500 thousand displaced. About 76 schools and 130 health units have been shut down or destroyed. The quality of public service has deteriorated, as in the areas under attacks public servants fled and military and/or policy personnel have overseen most government institutions. The humanitarian pressures from the conflict add to the impacts of the Cyclone Kenneth that in 2019 fustigated some of the localities attacked. Addressing the key sources of fragility is fundamental not only to curb current conflicts, but also to avoid replication and expansion. Thus far, government has mainly resorted to security interventions, with increased reliance on mercenaries. But, the insurgency, which started with attacks from a group poorly armed continues to increase in range and sophistication, and in cases supported by local population. Locals cooperation signals the importance of addressing the socio-economic dimensions of conflict, including ethnical, social, economic, and geographical layers. The constitution of an Integrated Development Agency of the North (ADIN) with focus on creating economic opportunities is a step on that direction. However, efforts need to be expanded to allow political participation, internal non-violent dialogue, improved access to public services, and improved intergenerational connection. This can be done with the support of existing stakeholders, including the donor community, the civil society and religious organization. The role of women and youth should be considered. It is important that existing risks are considered and addressed, including expectations about upcoming LNG benefits, increasing economic inequality, youth unemployment, natural disasters and growing political tensions. Source: World Bank (2020) Risk and Resilience Assessment. 26 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Productivity needs to be improved in the services sector. As more and more labor shifts from agriculture to services, in the absence of significant capital investments in services, the sector’s productivity growth is expected to continue to decline. Firm level data confirm the growth in low-skill segments of the service sector (e.g. retail) and the large share of small informal firms. Tackling this may help boost the future contribution of this important sector to growth. This is discussed further in Chapter 4. Agricultural productivity can also be boosted considerably. Mozambique’s crop yields are not only low by regional standards but are also far below the country’s agricultural potential. Increasing agricultural productivity could have a major impact on poverty reduction in the short-to-medium term. Linked to this are the low adoption rates of productivity- enhancing technologies (e.g. improved seed and fertilizer) amongst Mozambican farmers because of lack of financing and credit, keeping yields low. Mozambican farmers are also highly exposed to price and climate-related shocks, and lack risk mitigation instruments. Relative isolation and high transport costs are other major barriers for farming households in accessing input and output markets and raising productivity. This is discussed further in Chapter 3. Gaps in infrastructure access, skills and digital development are holding back competitiveness. Mozambique’s level of competitiveness, measured by the Global Competitive Index (GCI), has been on a declining path since the early 2010s. Ranking 133 out of 140 countries in 2018, Mozambique has fallen below other low-income countries (LICs) and the average for SSA.21 Whilst the country has fared relatively well in developing trade infrastructure (including ports, customs and primary roads linked to export corridors), a thin internal road network limits market connectivity. Skills development is progressing but at a slow pace, even by regional standards. Deficits in access to electricity, ICT adoption, and regulatory constraints further depress firms’ competitiveness in Mozambique. Competitiveness and productivity are constrained by a weak governance setting and corruption. Corruption is noted by Mozambican firms as one of the top three obstacles to their operations.22 Unlike in 2007, corruption ranked at the top of the list of obstacles to firms with more than five employees in the 2018 Enterprise Survey,23 indicating a rise in rent-seeking practices between the two periods, especially for smaller firms in the services sector. This is the topic of Chapter 5. 21 World Economic Forum (2019). 22 World Bank (2020b). 23 World Bank (2018b). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 27 References Christiaensen, L., Weerdt, J. D., and Kanbur, R. (2014). Vandercasteelen, J. Beyene, S., Minten, B. and J. Swinnen “Urbanization and Poverty Reduction: The Role of (2018). “Big cities, small towns, and poor farmers: Secondary towns in Tanzania,” Analysis and Policy Brief Evidence from Ethiopia”. World Development, 106(2018); No. 18, Institute of Development Policy and Management, 393-406. University of Antwerp. World Bank (2020a). Mozambique Rural Income Diagnostic: Christiaensen, L., Weerdt, J. D., and Kanbur, R. (2017). Cultivating opportunities for faster rural income growth “Migration Equilibrium and Income Distribution with and poverty reduction. Washington, DC: The World Bank. Multiple Destinations: Cities, Towns, and Poverty.” In World Bank (2020b). Mozambique Country Private Sector process. Diagnostic. World Bank: Washington, .C. Duarte, M. and D. Restuccia (2018). Structural Transformation World Bank (2020c). Wealth of Nations database. Washington, and Productivity in Sub-Saharan Africa. University of DC: The World Bank. Toronto, Canada. World Bank (2020d). Africa’s Pulse. October 2020. Duarte, M. and D. Restuccia (2010). “The role of the structural Washington, DC: The World Bank. transformation in aggregate productivity.” Quarterly World Bank (2019). Mozambique Economic Update: Closing Journal of Economics 125 (1): 129–73. rural infrastructure gap key to achieving inclusive growth. Harvard University Growth Lab (undated), The Atlas of Washington, DC: The World Bank. Economic Complexity. www.atlas.cid.harvard.edu. World Bank (2018a). The Growth Report: Strategies Herrendorf, B., R. Rogerson, and A. Valentinyi (2014). “Growth for sustained growth and inclusive development. and structural transformation.” In Handbook of Economic Commission on Growth and Development. Washington, Growth, volume 2, ed. P. Aghion and S. Durlauf, 855–941. DC: The World Bank. Elsevier. World Bank. (2018b). Mozambique Enterprise Survey. Lachler, U. and I. Walker (2018). Mozambique Jobs Diagnostic. Washington, DC: The World Bank. World Bank: Washington, D.C. https://openknowledge. World Economic Forum (2019). The Global Competitiveness worldbank.org/handle/10986/30200. Report. World Economic Forum: Geneva. McMillan, M. and K. Harttgen (2014). “What is driving the ‘African growth miracle’?”. NBER Working Paper No. w20077. Available at SSRN:  https://ssrn.com/ abstract=2436706. 28 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Chapter 2. Measuring and Macro-Managing the Resource Boom: Summary of recommendations Recommendations Responsibility Timeline for Strengthen medium-term fiscal framework and adopt fiscal rules Establish medium-term fiscal targets and rules DEEF, DNPO Medium term Develop medium-term budget framework with sectoral three-year budget ceilings DNPO ST to MT Fiscal consolidation Conduct a systematic review of employee compensations to simplify remuneration MEF Medium term and eliminating nonessential allowances (e.g. the 13th-month bonus). Carry out a comprehensive functional review to identify overlapping functions and MEF Medium term redundancies, and to clarify organizational goals and responsibilities. Improve public investment management and SOE management Improve projection appraisal and section through multi-year planning and continue MEF, IGEPE Short term establishing a digital platform for investment management incl. by SOEs Establish a sovereign wealth fund with a dual mandate of achieving both short-term MEF, BoM ST, MT stabilization and long-term savings for future generations Create a well-designed stabilization fund Establish a sovereign wealth fund with a dual mandate of achieving both short-term MEF, BoM ST, MT stabilization and long-term savings for future generations MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 29 2. 30 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Measuring and Macro- Managing the Resource Boom The discovery of natural gas is expected to provide Mozambique with a transformative opportunity for inclusive growth. One of the most anticipated gains from LNG is an increase in fiscal revenues when exports take off. But production and price volatilities could affect the potential impact. In addition, large inflows of foreign currency could have a negative effect on the economy by causing a loss of competitiveness. This chapter explores how LNG production and export could affect the main economic variables, such as GDP, exports, revenue and employment. It highlights the macroeconomic risks associated with increased LNG production during the development and investment phase, as well as in the next decade when increased revenues start coming in from production. It outlines strategies to mitigate the risks identified, emphasizing the need for fiscal consolidation now, while strengthening institutions and systems and setting up a well- designed stabilization fund to manage revenue volatility in the future. 2.1 Introduction macroeconomic risks during the investment and production phases (Section 2.3). This highlights the potential impact The production and export of LNG could have a profoundly on public debt and emphasizes the need for prudent fiscal positive impact on Mozambique. The expected large- management. The analyses set the stage for a range of scale production and export of LNG beginning in 2024 will mitigation strategies, discussed in Section 2.4. significantly boost GDP, exports and fiscal revenue. It will also have an impact, though more limited, on employment, the 2.2 LNG Investments and Impacts local community and infrastructure development. However, delays to the start of production, lower production volumes or Substantial Investments Are Needed to lower prices could all affect the potential impact. In addition, Realize LNG’s Potential for Mozambique Dutch disease24 effects could negatively affect the economy. To mitigate these factors and maximize the benefits for the The discovery of large natural gas reserves in the Rovuma Mozambican economy, sound policies and institutions are Basin could be transformative for Mozambique. Proven gas needed in order to address the changing economic context reserves in the entire country currently stand at about 190 and achieve stable and equitable economic growth. trillion cubic feet (TCF), of which 180 TCF were confirmed in the Rovuma Basin.25 Mozambique, therefore, holds the 13th This chapter begins by exploring the impacts of different largest natural gas reserves globally and the 2nd largest in LNG investment scenarios (Section 2.2). This includes Africa, behind Nigeria and just ahead of Algeria. Within the next assessing the direct and indirect impacts on fiscal revenue, decade, the country could become Africa’s largest producer as well as on jobs and growth. It then looks at the and a leading exporter of liquefied natural gas (LNG).26 24 Dutch disease refers to the negative impact that large inflows of foreign currency can have on an economy by shifting the structure of production towards non-tradable sectors and through a loss in competitiveness. Whilst associated with natural resource discoveries, Dutch disease can occur from any events that result in large inflows of foreign currency, including a surge in commodity prices, foreign assistance or foreign direct investment 25 An additional 10 TCF of proven reserves have been discovered in the Moçambique Basin. 26 Poulsen (2016). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 31 The investments needed to realize this potential are travel restrictions. It also assumes an 18-month delay to substantial, but subject to vast uncertainties. Since the the Exxon-led Mamba project, with a final investment discovery of its gas reserves in the Rovuma basin in 2009- decision (FID) expected in the second half of 2021. 12, Mozambique has secured two investments for LNG • Under the second scenario (Mamba delay scenario), production in the area.27 The country aspires to realize a the Exxon-led Mamba project is placed on hold for number of additional projects to exploit the rest of its vast four years and exports begin in 2030, whilst all other reserves. But the likelihood and the potential pace of these assumptions remain the same. investments is highly unpredictable and complicated by the swings in global commodity market. As this report is written, The analysis in this chapter does not account for the delays oil markets are in turmoil having seen an unprecedented in the two major LNG projects in 2021 due to the escalation drop in prices as a result of COVID-19. These swings spill of insurgency in Northern Mozambique and the COVID-19 over into market prices for gas and contributed to setbacks pandemic.30 in the investment potential for Mozambique.28 Adding to this, an intensifying terrorist insurgency in Cabo Delgado, The Impact on Fiscal Space is Subject to the province were the gas developments are located, has Uncertainties further soured the setting for the industry. One of the most anticipated rewards from LNG is For these reasons, two investment profiles are considered an increase in fiscal revenues when exports take off. in this chapter. The combined liquefication capacity under Mozambique follows a production sharing contract both investment profiles is the same, at 29.8 million tons under its agreements with consortium operators. As a per annum (MTPA), and the required total investment is result, fiscal receipts from the sale of gas will be received estimated at around USD 60 billion, but the profiles differ in through state profit petroleum once cost recovery has terms of project implementation timings:29 been achieved, as well as through income tax receipts and non-tax revenues.31 These are set to noticeably widen • The first scenario assumes a six-month delay to the Coral the fiscal space.32,33 Under the first scenario, by 2040 project and a one-year delay to the Golfinho project when all LNG sites are expected to be operational, total due to the impact of COVID-19 on supply chains and LNG revenue is expected to be around MZN 1,444bn > > > Figure 2.1 Assumed Investment Needs and Production Levels under the Two Scenarios a. Estimated investment by site (USD billion) b. Estimated production (exports) by site (billion cubic feet per annum) 15 2000 1500 10 USD billion 1000 5 500 0 0 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 Mamba Golfinho Coral Mamba delay Scenario Mamba Golfinho Coral Mamba delay Scenario Source: World Bank sta estimates. Source: World Bank sta estimates. 27 Coral South Floating LNG – approximately USD 7bn investment started in 2017; and Golfinho/Atum – approximately USD 23bn in investment started in late 2019. 28 For instance, Exxon announced a delay to its Area 4 Mozambican LNG Mamba project in early 2020 as oil prices dropped sharply. 29 See Annex A.1 for a description of the LNG projects. 30 The intensification of conflict in the Cabo Delgado province led to the suspension of the Total-led project. In April 2021, TotalEnergies declared Force Majeure, following a suspension of operations in March the same year. The second onshore project led by the multinational Exxon has been substantially delayed due to the compounded effects of COVID-19 and local insurgencies. The off-shore Coral South LNG project remains on track and is expected to start production in 2022 and achieves full capacity in 2023. 31 Profit petroleum means the amount of production, after deducting cost oil production allocated to costs and expenses, that would be divided between the participating parties and the host government under a production sharing contract. 32 Annex A.2 gives an illustration of the LNG revenue channels for one of the projects (Coral South). 33 In addition, Mozambique has already received significant tax revenue through its capital gains tax provisions (fixed rate of 32 percent) from the sale of assets between LNG investors. For example, the recent takeover of Anadarko investments in Mozambique by Total yielded USD 880mn (approx. 6 percent of GDP) in 2019. 32 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM (US$19.6bn) or about 1.43 times higher than without LNG The volatility of the LNG price poses a significant risk for production. A four-year delay to the Mamba project would government revenues, underscoring their sensitivity to such result in revenue in 2040 being 1.37 times higher than shocks. A separate price shock scenario demonstrates this without LNG production. In both cases, receipts from LNG risk (Figure 2.). A 20 percent drop in the export price relative to will help widen the fiscal space to fund the government’s the baseline in 2025, similar to the one experienced in 2020 expenditure program. as a result of the COVID impact on oil prices, causes overall revenues to fall by USD 4 billion by 2040, thereby reaching However, the impact will only be significant after 2033 only about 80 percent of the revenue expected without the – more than ten years from the start of the investment. price shock. In this scenario it is assumed that despite the lower The scenarios highlight that, during the development phase price, LNG production remains unchanged as LNG producers of the projects and the first years of production, additional seek to maintain contracted volume sales. Thus, the impact revenue will be limited. Revenue will start to significantly on revenues would be even more severe if production levels increase from around 2033 when sites are operational and were to drop in a lower price context. The outcome from cost recovery on LNG projects has been achieved (Figure this scenario illustrates how sudden changes in the global 2.). Until then, even though LNG exports are expected to market, leading to lower LNG prices, could dramatically affect already have increased drastically, the contribution to fiscal the government’s public finances. Given the expected size revenue remains relatively small. This is because most of the LNG sector in the economy, such shocks could be of these revenue streams will only be positive once full destabilizing and highlight the need to develop mechanisms production levels have been reached, considering the debt to cushion government resources against them.35 repayment profiles on the projects. The profiles also assume that free cashflow available to the state-owned oil and gas The Government also expects that dividends from the company Empresa Nacional de Hidrocarbonetos (ENH; see involvement of ENH will add to fiscal receipts (Box 2.1). Box 2.1) from one project will be used to cover expenses However, this depends on how much profit ENH can on other projects. Thereafter, corporate income tax will generate and by when. A major constraint in this regard increase significantly as will profit gas, which is the share of is the level of debt that ENH will be carrying to finance its produced gas the government receives after the contractor participation in the LNG projects. The higher and costlier has recovered its costs (exploration, development, operating the debt, the more distant are its prospects of contributing and financing costs).34 through dividends, and the higher the fiscal risks and vulnerabilities. > > > Figure 2.2. Revenue Will Start to Increase from 2033 a. LNG contribution to fiscal revenue (% of non-LNG GDP) 10% 8% 6% 4% 2% 0% 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 Upstream Training Fee Production Bonuses Corporate Income Tax State Profit Petroleum Stamp Duties ENH Cash Flow After Carry Repayment Source: World Bank sta estimates. 34 See also Annex A.2. 35 Unlike for crude oil, there is no harmonized global price for LNG. Instead, the LNG price is segmented into regional markets. For Mozambique, the main markets will be Asia and Europe. In Asia, the price is linked to the crude price index, while in Europe, LNG pricing is tied to the National Balancing Point, a virtual trading location for the sale, purchase and exchange of UK natural gas (US Department of Energy 2017). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 33 > > > Figure 2.3. A Price Shock Scenario Could Reduce Tax Revenue a. LNG price index, historical and projected (2018=100) b. Impact of price shocks on revenue 160 1 500 140 120 1 000 100 80 60 500 40 20 0 0 1992 1995 1998 2001 2004 2007 2010 2013 2016 2019 2022 2025 2028 2031 2034 2037 2040 2019 2021 2023 2025 2027 2029 2031 2033 2035 2037 2039 LNG price index (historical) LNG price index (projection) Revenue no price shock LNG price index (projection with 6% price shock) Revenue with 20% price shock LNG price index (projection with 20% price shock) Revenue with 6% price shock Source: World Bank sta estimates. The Impact on Growth and Jobs Could be implementing the Mamba project would put real GDP 60 Significant percent higher than without LNG production. In the first scenario, GDP per capita would reach almost USD 1,000 by A growth acceleration is another anticipated reward from 2033 (from USD 485 estimated for 2020). Exports are also the development of the LNG sector, both through direct and expected to increase substantially, by almost three times the indirect effects. Growth impacts are expected through two level that would be realized without LNG production by 2033 main channels. The first channel is through project spending (Figure 2.3). As imports are only likely to increase much less, on equipment, services and personnel during the investment there would be a significant positive impact on the overall and operations phases, with most of the spending occurring external position. This is likely to see an expected decline in during the investment period. The second channel is through the current account deficit from 73.9 percent of GDP at peak the production and export of gas. For both channels, the investment in 2021 to a surplus current account of almost 12 impact on growth can occur through direct and indirect percent of GDP by 2033. Furthermore, the composition of effects. The direct impact under the first channel is limited Mozambique’s GDP will change markedly with the growth of to the project’s demand for domestically produced goods the LNG sector, leading to a significant increase in the share and services and the cost of personnel (the local content), of the extractives/mining sector between 2019 and 2033 for estimated at just 10 percent of the project spending. But both production profiles (Figure 2.58). indirect impacts are wider in that they encompass linkages from the project’s demand to other sectors of the economy. The indirect effects on the economy are equally important. With the second channel, production contributes directly The multiplier model estimates the indirect impacts of LNG to national output, but its indirect effects are a function of on growth through the spending channel. With the estimated the share of export receipts spent domestically, in addition multipliers, the direct and indirect impact of LNG production to how they are spent. Particularly important is the question on output and other variables can be evaluated.36 When of how future fiscal revenues will be used, and what the considering only the indirect impact, which is not captured associated fiscal multipliers will be. These effects have been by the MAPIO model, GDP will increase by an additional USD modelled, as described in Box 2.2 and Annex A.3. 3.1bn between 2019 and 2033, which is an increase of 20 percent over baseline 2019 GDP. This confirms the importance The direct impact on the Mozambican economy will be of the positive economic impact of the LNG projects through significant once all four sites become operational. The indirect channels. The indirect GDP impact peaks coincide MAPIO model results show that real GDP is expected to with the peaks of the investment phases (Figure 2.6a). The be 163 percent higher than without LNG production by sectors with the highest multipliers are agriculture, financial 2033 under the first scenario (Figure 2.3), whilst a delay in and business services, transport and industry (Figure 2.6b). 36 The estimated output multiplier of 1.39, for example, implies that an increase in total expenditure by USD 1mn would lead to an increase in output of USD 1.39mn. A more comprehensive table with the main results is given in Annex A.3. 34 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Box 2.1. The role of ENH Through Empresa Nacional de Hidrocarbonetos de Moçambique or ENH, the state-owned hydrocarbons company, Mozambique has maintained a 10-15 percent sharea as an investor in the LNG exploration and development consortia. The decision to maintain a share in the projects dates back to the early 2000s. Participation is both a political decision to maintain state involvement in the development of these significant resources, as is the case for numerous resource exporting economies, and an investment to generate returns through direct participation in the projects. ENH is a state-owned enterprise (SOE), designated in the Hydrocarbons Law 21/2014 to be the exclusive representative of the state to participate in oil and gas projects, with a minimum stake of 10 percent. Given its lack of financial capacity, ENH is borrowing to fund its equity share for participation in LNG projects. The projects are financed through a combination of equity and debt from export credit agencies and commercial banks.b The debt share is contracted under project financing arrangements, whereby a special purpose vehicle (SPV) is established as the borrower, and with lender recourse limited to the revenue generated by the project.c Hence, this borrowing is not defined as public sector debt. In contrast, ENH borrowing to finance its equity share sits on its balance sheet and is defined as public sector debt.d ENH’s planned non-concessional borrowing for the Coral and Golfinho equity share in 2020 is estimated at USD 227 million (1.5 percent of GDP). In addition to ENH’s direct borrowing, the GoM issued a sovereign guarantee during the construction period for the Golfinho project, adding a further 2.8 percent of GDP to public debt stock in 2020 (and 15 percent of 2020 GDP by 2025). This, along with ENH’s exclusive representation of the state, means there is little room for flexibility in the selection of projects for ENH’s participation and investment decisions based on cost-benefit analyses are minimal. As a result, ENH finds itself in a position where borrowing to finance its participation in LNG projects remains on an upward trajectory – a trend likely to continue as more projects come along. Whilst ENH’s participation ensures additional revenues, albeit only at a later stage in production once all financial obligations have been met, it also means that both ENH and Mozambique remain overleveraged, which in turn contributes to keeping the cost of borrowing up. a. ENH equity stake to date comprises a 10 percent stake of Area 4 projects and a 15 percent stake Area 1. b. The debt-to-equity ratio is around 60:40 on average. c. ENH owns only 10-15 percent of the SPVs. d. This includes amounts accumulated to cover its share in the exploration phase, which was 100 percent equity financed. > > > Figure 2.4. Di erent Investment Scenarios A ect Key Macroeconomic Variables a. Exports (USD billion) b. Real GDP (USD billion) 40 40 30 30 20 20 10 10 0 0 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 Exports without LNG GDP without LNG Exports with LNG GDP with LNG Exports with LNG - Mamba delay GDP with LNG - Mamba delay Source: World Bank sta estimates (MAPIO model). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 35 Box 2.2. The two models used to assess the impact of LNG development and production in Mozambiquea The potential benefits associated with LNG development and production in Mozambique have been evaluated using two different models. The first – based on the Macroeconomic Assessment of Public Investment Options model (MAPIO) – is a static macroeconomic model that estimates the direct impact of LNG on growth through the project spending and export channels.b To estimate indirect effects, a second model was developed using multiplier analysis. The MAPIO model assesses the direct impact of the LNG project spending and exports on growth. It evaluates their impact in terms of (i) the increase in demand due to the implementation of the different projects; and (ii) the additional impact of the completed projects on the economy’s output through exports. Given the significant uncertainties linked to the global environment, the domestic business environment and volatile prices due to demand and supply shocks, the model estimates a baseline and alternative scenarios. Since the MAPIO model does not consider indirect effects or externalities, a multiplier model has been developed using Mozambique’s most recent social accounting matrix. This model makes it possible to assess the direct as well as the indirect impact of LNG investment on the economy. The results show the impact on gross value of production, the supply of goods and services, the earnings by the factors of production and households as well as employment for a range of labor categories. The indirect effects measured by the multiplier model are limited to the project spending channel (investment and production). It does not consider indirect effects through the production/export channel because of insufficient information on how fiscal receipts from exports will be used. The results of both models should be interpreted cautiously as best-case scenarios. This is because the estimations assume no delays in project implementation or major price fluctuations in the baseline. The baseline results should be interpreted as an upper bound of the potential impact of LNG projects in the absence of shocks or disruptions. To address this, the MAPIO model shows the impact of a price shock and also has two investment scenarios. a. For more details about the models and their underlying assumptions, please see Annex A.3. b. It is based on the Macroeconomic Assessment of Public Investment Options model (MAPIO) developed by Beguy et al. (2015), which assesses the impact of large-scale investments on key macroeconomic indicators. > > > Figure 2.5. The Extractives Sector Will Account for the Largest Share of GDP, 2025-2033 a. % contribution of sectors to the increase in GDP over 2025-2033 with and without LNG development (MAPIO model), baseline and low production 100% 80% 60% 40% 20% 0% without LNG with LNG 2025-2033 with LNG - Mamba delay Agriculture Exractives Manufacturing / Secondary sector Private services Public services Taxes Source: World Bank sta estimates. 36 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 2.6. Indirect Impacts on the Economy Will be Significant a. Indirect impact on GDP, 2019-2033 b. Multipliers for di erent sectors, 2019-2033 (Multiplier Model) (Multiplier Model) 400 0.12 350 0.10 300 0.08 mil milhões USD 250 0.06 0.04 200 0.02 150 0.00 100 Agriculture Financial and Business Transport Industry Trade Services Construction Goods Services Services Communication Accommodation and 50 0 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 Source: World Bank sta estimates. Significant employment opportunities are expected to be 2.3 There are Substantial created, mostly for low-skilled labor. Over the whole period, the multiplier model estimates that about 1.6 million worker Macroeconomic Risks Linked to LNG year opportunities would be created directly and indirectly. Development This increase in employment is spread over 15 years with considerable annual variation depending on the timing of The macroeconomic risks are substantial and need to be the capital and operational expenditure (Figure 2.10a).37 The mitigated by strategies adapted to each phase of the LNG majority of the employment opportunities, however, are for projects. The major challenges that Mozambique faces low-skilled labor, i.e. workers with primary education. Direct during LNG development differ between the investment and indirect impacts on employment by education level are and production phases. During the investment phase (the shown in Figure 2.10b. Similar to GDP, the indirect impact first ten years), the main challenge is to balance the tight of LNG investment by far exceeds the direct impact. Direct fiscal space and need for fiscal consolidation with continued employment opportunities represent just 8 percent of the borrowing for the LNG projects in a context of already high total. Household income is estimated to increase by an levels of debt. During the production phase, Mozambique additional USD 3.3bn over 2019-2033 when combining the will need to address the loss of competitiveness of the non- direct and indirect impact. tradable sectors. The following section outlines these risks to set the stage for possible mitigation strategies, discussed This section shows that the LNG projects will have in Section 2.4. a significant beneficial impact on the Mozambican economy, although with a substantial lag and associated Limited Fiscal Space and Debt Accumulation risks that should not be underestimated. Two different are Key Early Risks models illustrate that LNG production and exports affect the economy both directly and indirectly, leading to significantly Although there has been progress in fiscal adjustment higher GDP, exports, government revenue and more in recent years, fiscal space will be limited during the employment opportunities, thereby potentially providing LNG investment phase as anticipated LNG revenue will immense benefits to Mozambique. However, most of these take time to materialize. Mozambique’s budget has faced positive effects will only materialize about a decade after reduced fiscal space since 2016: external funding levels production begins. Moreover, different production and price dropped by 3 percent of GDP between 2015 and 2018 and trajectories can lead to substantially different outcomes, debt servicing costs increased by 2.4 percent of GDP. Whilst highlighting the vulnerabilities of the economy to the challenging, the budget has managed to adjust to these volatilities of the LNG market. The following sections discuss conditions through measures to reduce spending, primarily the associated risks and mitigating strategies further. through downsizing the capital budget and eliminating 37 The employment elasticity used is based on estimates for Sub-Saharan Africa, which are typically around 0.75, see Kapsos (2005). The elasticity is applied indiscriminately to all types of labor. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 37 > > > Figure 2.7. Significant Employment Opportunities Will Be Created a. Direct and indirect impact on employment, b. Impact on employment by education level, worker years, 2019-2033 (Multiplier Model) 2019-2033 (Multiplier Model) 250000 2.5 200000 worker years, millions 2.0 worker years 150000 1.5 100000 1.0 50000 0.5 0 0.0 2019 2021 2023 2025 2027 2029 2031 2033 Direct Impact Direct and indirect Not completed primary Completed primary Completed secondary Completed Tertiary Direct impact Indirect impact Source: World Bank sta estimates. subsidies to fuel, transport and bread prices. Despite the a significant shock since the revelation of previously economic downturn, Mozambique has also managed undisclosed borrowing in 2016. Less than two decades after to enjoy a relatively stable collection of tax revenues the HIPC debt relief process, public sector debt reached 126 at 26 percent of GDP by strengthening administration percent of GDP in 2016 with four-fifths of it being accounted mechanisms. Together, these measures have contributed to for by external debt and, thus, being affected by the sharp an estimated reduction of 4 percent of GDP in the primary currency depreciation that took place in 2016.40 Since then, fiscal deficit, from 5.9 to 1.5 percent of GDP between 2015 currency appreciation and limited external borrowing have and 2019. Even so, the public sector’s gross financing helped to lower public sector and external debt to 109 and needs remain high and should be reduced further to lower 90 percent of GDP respectively, but Mozambique remains the debt burden. highly indebted. Limited access to external financing sources has increased the public sector’s reliance on costlier Mozambique’s vulnerable fiscal position underscores the domestic borrowing and progress on debt restructuring has need for further consolidation in the short term, as well been slow.41 as prudent management of fiscal revenues once they materialize. In the scenario where fiscal consolidation Whilst public debt levels are expected to decline over the reforms result in a primary fiscal surplus by 2023, debt medium term, Mozambique will remain at high risk of debt indicators would remain on a downward trajectory and distress over the coming decade. At 109 percent of GDP, would only reach sustainable levels by 2033 (Figure 2.11a).38 public sector debt is amongst the highest in Africa, despite Conversely, a scenario where spending remains high, the 15 percent of GDP narrowing experienced since 2016. consistent with the five-year average of the primary deficit Going forward, the pace of the fiscal adjustment and choice (ranging between 2.8 and 3.5 percent of GDP), would drive of financing instruments will continue to play a key role in up public debt levels and keep them above the prudent easing debt levels, although SOE operations will cause an threshold for countries with weak debt carrying capacity additional source of pressure. Even with added reforms to (35 percent of GDP) – as is the case for Mozambique – bring the primary fiscal deficit to a surplus by 2022-23 and throughout the forecast period.39 the acceleration of GDP growth due to LNG production and exports, overall public sector debt levels are expected to In the meantime, high public debt burden casts a remain above the DSA’s prudent threshold for total public long shadow. Mozambique’s economy has undergone debt (35 percent of GDP) until 2032. 38 The fiscal reforms scenario does not assume spending of additional revenues brought about my LNG operations. 39 Under this spending scenario the borrowing strategy mirrors the recent developments in the Mozambican economy and assumes the following sources for additional financing needs: issuance of treasury bills (20 percent), 1-3 years treasury bonds (40 percent), 3-7 years treasury bonds (30 percent) and treasury bonds with more than 7 years (10 percent). Interest rates associated with these are expected to range between 8 and 15 percent over the forecast period. 40 Under the Heavily Indebted Poor Countries (HIPC) Initiative, external debt narrowed from 160 percent of gross national income in 1998 to a far more manageable 33 percent 10 years later. 41 The authorities succeeded in reaching an agreement with holders of the notes on Mozambique’s sovereign bond (formerly EMATUM bond) in September 2019, three years after the hidden debt scandal. Negotiations on the other two loans, MAM and Proindicus, are at a standstill following legal action taken by the authorities to nullify the sovereign guarantees on the loans. 38 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 2.8. Debt Scenarios Underscore the Need for Prudent Fiscal Management a. Public sector debt under di erent spending b. Public sector debt servicing scenarios % of GDP, 2018-2038 % of revenue and grants, 2018-2038 150 50 40 100 30 20 50 10 0 0 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 Public sector debt Debt service Public sector debt with additional borrowing Debt service with additional borrowing Total public debt threshold Source: 2019 IMF-World Bank Debt Sustainability Analysis; World Bank sta estimates. Borrowing for LNG project participation will have an to maintain discipline in a fiscal consolidation context. To additional impact on external position in the coming five do this successfully, and to make progress in improving years. Mozambique is expecting LNG investments of close investment and service delivery simultaneously, will entail to USD 55bn in the next decade, with the bulk of this being a medium-term fiscal program anchored by an adjustment frontloaded in the next five years in the development phase target and a comprehensive package of spending efficiency for the Coral, Golfinho and Mamba projects. Government reforms. involvement in these projects through the 10-15 percent stake held by ENH implies that an estimated USD 7.5bn will Mozambique Will Also Face the Risk of Loss of be needed to finance its participation (through debt and Competitiveness equity),42 thus adding to the already high debt levels. External debt levels are expected to narrow significantly once the The risk of loss of competitiveness exists, including construction phase for each of the projects is concluded in through Dutch disease effects, but may be relatively 2024-25, but the additional borrowing will keep external debt contained during the LNG investment phase. Dutch above the debt sustainability analysis (DSA) prudent threshold disease effects (Box 2.3) are likely to have been contained level of 35 percent of GDP until 2034 – two years later than in the LNG exploration phase. External balance assessments if the government had had no direct participation in the ENH indicate the real exchange rate was mostly in line with projects. Similarly, debt service as a share of revenues and fundamentals during the investment boom over 2011– grants will pick up significantly once LNG production begins, 2015.43 During this time, the local content of the FDI boom, drawing on much of the potential returns from investing in which was mostly LNG-exploration linked, was low. Hence, the projects (Figure 2.7b). FDI inflows were almost entirely offset by imports of goods and services. As a result, pressure on the real exchange Mozambique will remain on a consolidation path during rate to appreciate through this channel was limited. At the the LNG development phase. LNG revenues will increase same time, demand and investment in non-tradable sectors fiscal space and lower the debt burden, but not for another increased, contributing to the expansion of construction and eight or nine years. Production is not expected to begin the service economy. Much of this growth occurred in the until 2024-25. The staggered investment profile for the LNG informal lower-skill segments of the sector where surplus sites also means significant revenues are only expected labor was available. once peak production is achieved in the late 2020s or early 2030s. In the meantime, debt is projected to be high and However, real exchange rate pressure could increasingly fiscal space limited. In this context, Mozambique will need become a constraint to competitiveness in the coming 42 This figure represents both debt and equity financing for ENH’s participation in the three projects. Additional financing for a fourth project estimated at USD 15-20bn, where ENH has a 15 percent stake, would add to financing needs. 43 See IMF Article IV reports 2011-2019 for these and the following assessments. During the LNG production phase, other methods should be used as complements to the standard approaches for external balance assessments, see IMF (2019a). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 39 years. As the LNG projects progress from the investment to the economy. During the production phase, net inflows to the development and production phase, the risk of from export receipts could be even larger. This risk is acute Dutch disease effects will become more pronounced. As of it is not mitigated by an adequate fiscal framework that mentioned above, during the investment stage the volume of moderates spending in line with absorptive capacity for FDI inflows is largely offset by imports of goods and services. productive investment (see Box 2.4 for Sudan’s experience). Yet, the demand effects from the local content share, even This underscores the need for strong fiscal planning tools; in the range if 5 to 10 percent, could be significant in the processes will need to be even stronger as public investment Mozambican context given the size of the inflows relative is scaled up. Box 2.3. The experience of Sudan – lack of fiscal discipline and exchange rate appreciation In Sudan significant oil exports began in 1999 and the economy quickly became heavily dependent on oil. In 2010, which was the year before the secession of South Sudan, the oil sector accounted for 16 percent of GDP and 45 percent of domestic revenues and grants. The government spent nearly all oil revenues and undertook major investments in infrastructure. For example, both the road network and electricity generation roughly doubled from 2000 to 2008. At the same time, there were declines in the efficiency of public investment and the wage bill increased substantially. During the oil investment phase and in the early 2000s, there was initially no clear evidence of Dutch disease despite oil-related currency inflows of about 10 percent of GDP during 1996-2000. The real exchange rate appreciated to a limited extent in the context of a fixed nominal exchange rate regime. Traditional exports declined in the early 2000s but only in line with previous trends and partially due to temporary trade restrictions. In the mid-2000s, a combination of pro-cyclical fiscal policy, oil export revenues, large remittances and (mainly non-oil) FDI inflows caused the real exchange rate to appreciate more substantially. The agriculture sector suffered (not only from the exchange rate appreciation but also from neglect by policymakers) and the share of consumption goods in total imports increased. To help overcome these challenges, the Sudanese Government attempted to establish a fiscal framework for managing natural resource revenue in 2002. The Oil Revenue Stabilization Account (ORSA) was a locked account for the government at the central bank and it was created to help smooth expenditure. Initially funds were accumulated in the account between 2002 and 2005, reaching 2 percent of GDP at the end of 2005, but when revenue shortfalls arose as a result of delays in new oil fields leading to fiscal deficits, substantial draw-downs in 2006 left only USD 9.3mn in the account. The ORSA was built up again over 2007 and early 2008, but was depleted again in late 2008 and 2009 in the wake of the global crisis and collapsing oil prices. Following the secession of South Sudan, the ORSA has disappeared. The case of Sudan in the 2000s shows that even though the country had the foresight to establish a stabilization account, the attempt to smooth expenditure during the oil economy failed within a few years due to poor management. Source: World Bank (2015); Wohlmuth et al. (2009). 2.4 How to Macro-Manage the LNG Strengthen the Macro-Fiscal Framework and Boom? Adopt Fiscal Rules The current medium-term fiscal framework needs To mitigate the risks associated with the LNG projects, substantial reform. An evaluation of the current medium- several measures are necessary. The Government needs term fiscal framework (Cenário Fiscal de Médio Prazo - to implement different strategies at different phases of the CFMP) has highlighted its weaknesses and challenges.44 LNG project cycle to achieve fiscal consolidation on the While it has objectives that are consistent with international one hand and address revenue volatilities on the other. This good practice – such as predictability, clear links to policy section outlines the areas where reforms are necessary if and comprehensive coverage – weak implementation the country is to be prepared to face the major challenges reflects the institutional challenges within its formulation and ahead. management process. The main weaknesses are missing 44 World Bank (2018b). 40 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 2.9. A Conceptual Fiscal Policy Framework for a Resource-Rich Developing Country Non-resource income Resource income Total government income Factors to consider • Size of the resource envelop • Public expenditure needs Fiscal-policy framework • Public expenditure capacity • Fiscal volatility • Revenue-reserve horizon • Institutional strength Fiscal targets for achieving policy objectives (e.g. ensure public spending is in line with expenditure capacity or • Expectations (politician, public) address volatility of resource revenues) Fiscal rule or guideline to support the target (e.g. expenditure-growth rule or the structural budget-balance rule) Improving the government’s capacity to Improving the government’s capacity to manage public spending manage public savings Source: World Bank (2014b). inputs, such as a lack of a medium-term debt strategy and the CFMP recommended separating the current framework accurate debt forecasts; missing fiscal anchors, such as fiscal into: (i) a medium-term fiscal framework for macroeconomic or debt targets; no guidelines on spending growth; as well forecasts and aggregate revenue and spending targets; and as limited co-ordination regarding forecasting and review. (ii) a medium-term budget framework for sectoral three-year Weak forecasting has led to credibility concerns as inflation, budget ceilings while improving co-ordination between the primary fiscal deficit and expenditure have deviated by a framework and the budget process. Given Mozambique’s large margin from their projections. The current framework, vulnerability to natural disasters, a fiscal framework needs therefore, is not an effective instrument or foundation for appropriate flexibility, for example, through escape clauses. fiscal and budgetary planning. However, implementation and enforceability are key for credibility. A strong and credible fiscal framework will be important. Figure 2.9 outlines the various components of such a Establishing a clear, simple fiscal rule now as part of the framework. A well-structured medium-term fiscal framework framework will be essential for managing revenue volatility that is anchored in a fiscal target or rule (see Box 2.5) will in the future. Currently, no fiscal rules or targets have been be essential to maintain fiscal discipline and achieve debt officially adopted by the government. While revenue from reduction. This framework would provide policymakers with the LNG projects will only increase in about a decade, a basis for managing risks and vulnerabilities over the medium establishing a clear fiscal rule now is needed to have a well- term. This is especially important for resource-rich countries functioning framework, including a sovereign wealth fund, in order to mitigate the volatility and procyclical spending in place when surplus LNG revenues start coming in. The bias associated with resource revenues.45 The evaluation of fiscal rule will be needed to manage the volatility of resource 45 World Bank (2014b). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 41 Box 2.4. Options for fiscal rules in Mozambique The choice of the fiscal rule depends on the fiscal target in the fiscal policy framework. Fiscal rules have been adopted by many countries in order to reach their fiscal targets. The fiscal rule needs to be tailored to the country’s circumstances and adjusted through an iterative process to adapt to the country’s objective and institutional structure. Mozambique could first pilot a fiscal rule and then adjust it based on the lessons learned before its formal adoption. The objective of the target that the fiscal rule supports could be the ability to increase public spending while maintaining macroeconomic, particularly debt, sustainability while taking into account the limited capacity of government agencies. For Mozambique, the rule would also need to address the volatility of increasing resource revenue. To manage the resource revenue volatility, a price-based rule might be a good option. Through the price-based rule, expenditure is delinked from revenues that vary as commodity prices fluctuate. Chile, for example, adopted a price-based rule in the 1980s to limit the impact of fluctuations in copper prices on expenditures. As prices increased above a threshold, excess revenue was saved in a fund. As prices fell below another, lower threshold, withdrawals were made to compensate for the shortfall. To address capacity constraints and to mitigate pressures to increase spending, it may also be desirable to limit the pace of expenditure growth. Expenditure rules could be designed in different ways depending on preferences and objectives. Targets for spending increases could help manage the expectations of the public as well as of government agencies regarding the spending of increased resource revenues. Targets might be adjusted as capacity increases; however, frequent adjustments might lead to a loss of credibility. World Bank (2014b) provides some illustrations of two possible such rules. One rule ties the pace of expenditure growth to the nominal growth rate of the non-resource economy. The second rule limits expenditure growth to the decade-average for a group of peer countries. To improve credibility and further address revenue volatility, an expenditure-growth or price-based rule could for example be combined with a non-resource primary balance or structural primary balance target. Aiming for a positive non-resource or structural primary balance could further support the decoupling of fiscal policy from the volatility stemming from resource revenues. Segura-Ubiergo et al. (2013) provide some illustrative analysis of different price-based structural balance rules for Mozambique based on example rules used in Ghana, Trinidad and Tobago and Mongolia. The analysis shows that such rules can protect the budget against substantial volatility. revenue with a medium to long-term perspective to avoid government wage bill is higher than those of most of the pro-cyclical spending and to ensure inter-generational country’s peers (Figure 2.9b). The wage bill increased from equity. The fiscal rule itself should be simple, based on easily 8 percent of GDP in 2008 to 11.3 percent in 2016 and is understandable fiscal variables with numerical targets; and projected to reach 12 percent of GDP in 2021 without further flexible yet enforceable to help anchor fiscal policy and reforms. The wage bill growth was driven by ad-hoc rises ensure debt sustainability. Moreover, it should be possible in basic salaries and other compensation elements, an to monitor and enforce the fiscal rule by putting in place increase in employment and public financial management an adequate legal and institutional framework. Political weaknesses.47 Going forward, wage bill consolidation will support is essential, as is the integration into the country’s require structural reforms that target sectors with excessive public financial management framework.46 Some possible employment and wage levels. Aligning job-specific considerations for a fiscal rule for Mozambique are outlined requirements with compensation, a systemic review of pay in Box 2.5. levels and much-needed institutional-level restructuring of the public sector are amongst key reforms needed. Wage Bill Control Should be a Key Feature of Fiscal Consolidation Avoid any Significant Frontloading of Public Investment in the Near Term Despite some measures to control wage bill growth since 2016, the public sector wage bill remains high. Since the Given the current fiscal position and the need for fiscal hidden debt crisis, Mozambique has taken measures to limit consolidation, it would not be prudent to pursue any subsidies and staff promotions in the public sector. While significant frontloading of public investment in the near these have helped contain wage bill growth, the central term. Considering the current limited fiscal space, it would 46 The extent to which fiscal rules are implemented partly depends on political commitments to broader fiscal discipline and the presence of institutional constraints on the executive. 47 Xiao et al. (2017). 42 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 2.10. Government Expenditure is Higher than the SSA and LIC Average a. General government expenditure, 2018 (% of GDP) b. Central government wage bill (% of GDP) 35 18.5 16.8 30 15.0 25 13.3 11.5 20 9.8 Median across countries = 6.6 percent of GDP 15 8.0 10 6.3 4.5 5 2.8 0 1.0 Nigeria Madagascar Tanzania Uganda Zimbabwe Angola Ghana Kenya Zambia Mozambique South Africa Bangladesh Uganda Sudan Rwanda Nepal Kenya Mali Guinea Guinea-Bissau Gambia, The Cameroon Central African Republic Tanzania Madagascar Chad Niger Malawi South Sudan Haiti Senegal Nicaragua Benin Burkina Faso Cambodia Togo Sierra Leone Tajikistan Ethiopia Burundi Mauritania Eritrea Moldova Zambia Ghana Comoros Kyrgyz Republic Honduras Mozambique Liberia Zimbabwe Lesotho SSA average Source: IMF, Fiscal Monitor. Source: Xiao et al. (2017). be advisable to establish constraints on fiscal policy well in is funded by LNG revenues, as well as non-LNG revenues advance of the start of LNG production and exports and any and debt issuance, while additional savings are accumulated increase in revenue. While the present weak fiscal position in a stabilization fund. Compared to the spend-as-you-go may be a sufficient deterrent for increased spending in rule, this results in more efficient investment, a more stable the near term, the prospects of high resource revenues in path of private investment, consumption and non-LNG the long term might be an incentive to increase spending output, the mitigation of the Dutch disease effect through in anticipation of those revenues. However, given the the resources accumulated in the stabilization fund, and a differences in revenue streams between the LNG investment more sustainable debt path. and production phases, the constraints for fiscal policy in the near term would need to be stricter than the longer- A gradual increase of investment to take limited absorptive term fiscal rules. Therefore, the fiscal consolidation process capacity into account needs to be accompanied by well- should continue over the coming years to avoid the risks formulated sectoral policies and reforms.49 Measures related to borrowing against future LNG revenues to to mitigate the possible negative effects of Dutch disease frontload expenditure (see also Box 2.7 on the experience would need to go beyond prudent macroeconomic of Ghana). management. While adequate macroeconomic policies are essential to avoid Dutch disease, other measures are needed If fiscal consolidation continues and significant progress to attract investment and boost firms’ competitiveness in the can be made in reforming fiscal processes in the next few non-LNG sectors. Policies would also have to ensure that years, gradually scaling up investment is recommended resources are distributed more evenly. These aspects of the as LNG revenues start to come in. Once Public Financial necessary reforms are addressed in other chapters of the Management (PFM) and Public Investment Management CEM, especially Chapter 4. (PIM) reforms have been implemented, the fiscal position has improved as a result of fiscal consolidation, and LNG Improve Public Investment Management in revenues have started to materialize, the government will be the Short to Medium Term in a better position to increase public investment. According to the IMF, delinking public investment from LNG revenues The focus needs to be on the quality of public investments. and gradually scaling up investment would yield superior PIM capacity is currently weak in Mozambique. Poor- results than a scenario in which LNG fiscal revenues are used quality investment projects are likely to only create to fund public investment as they become available (spend- additional financial obligations and challenge debt reduction as-you-go-rule) without the accumulation of resources in a efforts. Reforms are currently underway to improve PIM, stabilization fund.48 Under the first approach, the investment including developing methodologies for project appraisal, 48 IMF (2019b). 49 Limited absorptive capacity might also constrain the implementation of public investment projects. Absorptive capacity constraints could cause the rate of return on investment to decline by limiting the ability to rapidly scale up public investment. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 43 establishing a digital platform for investment management In summary, Mozambique should establish a sovereign and strengthening the supporting legal framework (see wealth fund with a dual mandate of achieving short-term below). Other areas where improvements can be made are stabilization given revenue volatility and long-term savings multi-year planning and project selection and considering for future generations. The rules governing deposits in the recurrent budget implications of capital projects. and withdrawals from the fund should be clearly stated in Improved PIM will enable the authorities to select the the country’s fiscal framework. However, no institutional investment projects with the highest return which would structure can guarantee successful management of a lead to a higher fiscal multiplier. This will help increase sovereign wealth fund unless there is broad commitment to productive capacity, especially if the investment strategy is fiscal discipline in general and to the country’s fiscal rules well designed to benefit agriculture and other key non-LNG in particular. Box 2.7 illustrates Ghana’s experience of what sectors. Chapter 3 provides some pointers with regards to happens when there is no broad-based commitment to the spatial impact on growth of future public investment fiscal and monetary discipline. programs. Review the Role of the SOE Sector in the While Mozambique has established a regulatory Economy framework for public-private partnerships (PPPs), further efforts are needed to strengthen the framework. The World The state’s participation in the business sector should be Bank benchmarks the regulatory PPP frameworks of 135 reviewed thoroughly. Mozambique’s underperforming SOEs economies (including Mozambique) against internationally have placed added pressure on the budget, with issuance recognized good practices in preparation, procurement, of domestic debt to cover SOE operations, including debt contract management and treatment of unsolicited restructuring, accounting for over half of the net issuances proposals.50 Comparing Mozambique to the average, the between 2016 and 2018. A thorough mapping exercise of the areas with the most scope for development are preparation state’s enterprise sector, including subsidiaries, autonomous (score of 46 compared to an average score of 50) and the agencies, and other potential sources of fiscal risks, is urgent. treatment of unsolicited proposals (score of 33 compared Similarly, a rationalization of the role of SOEs in the economy to an average score of 56). Compared to the average SSA would be useful to have clear criteria for state engagement country, Mozambique scores above average in all areas in the business sector. except the treatment of unsolicited proposals. Legal reform is needed to give ENH more flexibility Create a Well-designed Stabilization Fund regarding investment decisions. The current law requires that the state-owned oil and gas company ENH has a stake When resource revenues are volatile, some of the revenues of at least 10 percent in all LNG investments. Through this, should be saved for precautionary reasons so that they ENH has been automatically forced to borrow to finance its can be used in periods when revenues are lower than equity participation in the LNG projects. In addition, sovereign expected. Both commodity prices and production volumes guarantees had to be issued to cover the share of ENH in the can fluctuate over time, causing revenues to vary in an financing of the LNG projects. These borrowing decisions unpredictable way. The magnitude of the appropriate savings have been taken irrespective of the cost and benefit of such for precautionary reasons depends, among other factors, on an involvement. Therefore, the GoM should reconsider the revenue volatility and policymakers’ risk aversion.51 law to grant ENH more flexibility regarding its decision when to invest and when not. One option could be to remove Inter-generational equity considerations would favor the mandated minimum stake for ENH and potentially set a saving revenues for future generations to use after the maximum investment amount. non-renewable resources have been depleted. There are different ways of achieving inter-generational equity. Enhance Transparency and Accountability A natural benchmark is the permanent income hypothesis (PIH) under which the level of net wealth is stabilized, and Transparency and accountability will be critically important the profile of spending is smooth, implying a constant for managing the natural resource boom successfully. fiscal space (Annex Box 1). Norway is one example, having A strong and sustained commitment to transparency and saved all its resource revenue in a fund and only spending accountability will encourage public support and help ensure the interest income. However, it is also possible to let the that the government pursues appropriate policies. A lack of fiscal space increase gradually over time or, conversely, to transparency, on the other hand, makes it hard to hold the create higher fiscal space in the short term by frontloading government accountable, in turn making it less likely that investment spending. resource revenues will be used appropriately. To ensure 50 World Bank (2018b). 51 For examples of quantitative studies, see van der Ploeg (2010) and Cherif and Hasanov (2012). 44 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Box 2.5. The experience of Ghana – deterioration of fiscal and monetary discipline despite nominal fiscal rules and central bank independence The case of Ghana demonstrates that reforms of the institutional frameworks for fiscal and monetary policy are not enough to guarantee appropriate policies in the absence of a broad-based commitment to fiscal and monetary discipline. Following an extended period of macroeconomic instability, a new central bank law was passed in 2002 to insulate the central bank from political interference and fiscal dominance. The primary objective of the Bank of Ghana would be to maintain price stability independent of instructions from government or any other authority. The law also imposed quantitative limits on government borrowing from the central bank (10 percent of the government’s estimate of the current year’s revenue). After the discovery of oil in 2007, the government studied best practices among oil-producing countries and pursued public consultations on the management of Ghana’s oil resources. The Petroleum Revenue Management Act (PRAM) was passed in 2011, establishing limits on spending in terms of the government’s forecast of oil revenue for the current year. The excess revenues were to be paid into the Petroleum Stabilization Fund (intended to sustain public expenditure capacity during periods of unanticipated oil revenue shortfalls) and the Heritage Fund (intended to sustain public expenditure capacity during the post-oil period). To enhance transparency, Ghana became compliant with the Extractive Industries Transparency Initiative (EITI) and set up a Public Interest Accountability Committee to monitor the utilization of oil revenues and ensure government accountability. However, when oil was discovered in 2007, the economy was weakening, and the government was facing an election in 2008. There was a sense that the future oil revenues provided fiscal space and the government pursued an expansionary fiscal policy. The budget deficit increased from below 5 percent of GDP to 6.5 percent in 2008 and subsequently reached double digits in 2012 and 2013 as expenditure increased further in the period leading up to the 2012 elections. The government wage bill accounted for much of the increase, reaching 72 percent of total revenue in 2012 up from 46 percent in 2008. The higher wage bill, combined with a higher interest burden from the government’s large-scale borrowing, crowded out other priority spending. As a result, capital expenditure fell from a pre-oil average of 12 percent of GDP in 2004-2008 to 5 percent in 2014. Despite the nominal central bank independence, there was a dramatic increase in the central bank’s financing of the government. From 2008 to 2014, central bank financing increased by a factor of 7. In the absence of strict limits on monetary financing, the central bank was not able to withstand pressure from the government. success, the GoM should make a clear commitment to improved budget transparency include a greater emphasis implementing transparent and accountable processes and on audit processes, the timely publication of pre-budget policies for all aspects of resource management. Chapter 5 statements and mid-year reviews, as well as the inclusion of discusses this in more detail. additional information in the budget proposal and year-end report. Various governance indicators confirm progress on transparency, but challenges remain. Mozambique is Mozambique should continue to work towards full currently classified as “weak” by the Resource Governance compliance with the Extractive Industries Transparency Index, with a score of 50 out of 100 (RGI 2017).52 This is an Initiative (EITI). Mozambique joined the EITI in 2009 and improvement over 2013, when it was classified as “failing” its latest validation by the EITI Board was in October 2019. and scored 37. Mozambique is currently ranked 41st out of The country was found to have achieved meaningful 89 countries worldwide and 12th out of 31 in SSA. It scores progress overall in implementing the 2016 EITI standard, especially poorly on open data (26 of 100), government with considerable improvements in specific requirements.53 effectiveness (34 of 100) and national budgeting (35 of 100). The EITI board encouraged Mozambique to ensure Improvements are reflected in the Open Budget Index score the sustainability of transparency and multi-stakeholder for transparency, which was 42 (out of 100) in 2019, up governance in the extractive sector by completing the from 35 in 2015, but still lower than in 2012 (47). Areas for institutionalization of the EITI Secretariat.  52 https://resourcegovernanceindex.org/country-profiles/MOZ/oil-gas. 53 See www.eiti.org/mozambique for more detailed information. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 45 In the current context of high debt levels and extremely future LNG revenues, and to clarify that increased spending limited fiscal space, the public should be aware that before these revenues materialize is not desirable. Decisions increased government spending is not an option. Public on the management of the current and future resources expectations need to be managed to contain unrealistic and on creating a sovereign wealth fund should be carefully hopes concerning the fiscal space raised by the government’s communicated to the public. References Beguy, O., S. Dessus, A. Garba, J. Hayman and J. Herderschee Segura-Ubiergo, A., M. Poplawski-Riberiro and C. Richmond (2015). “Modeling the impact of large infrastructure (2013). “Fiscal challenges of the natural resource boom”. projects: a case study from Niger. Macroeconomic In D. Ross (ed.) Mozambique rising – building a new assessment of public investment (MAPIO)”, MFM Global tomorrow, Chapter 10, pp. 122-140. Washington DC: Practice, Discussion Paper No. 7, November 2015. International Monetary Fund. Washington DC: The World Bank. US Department of Energy (2017). Understanding Natural Gas Corden, W. M. and J. P. Neary (1982) ‘Booming Sector and De- and LNG Options. October 2017. industrialisation in a Small Open Economy’,  Economic Van der Ploeg, F (2010). “Aggressive oil extraction and Journal, vol. 92, pp. 825–48. precautionary saving: coping with volatility,” Journal of Cherif. R. and F. Hasanov (2012). “Oil exporters’ dilemma: Public Economics, Vol. 94, pp. 421-33. how much to save and how much to invest?” IMF Wohlmuth, K., R. Alabi, P. Burger, A. Gutowski, A. Jerome, T. Working Paper, WP/12/4. Washington DC: International Knedlik, M. Meyn and T. Urban (eds.) (2009). New Growth Monetary Fund. and Poverty Alleviation Strategies for Africa – Institutional Dabla-Norris, E., J. Brumby, A. Kyobe, Z. Mills and C. and Local Perspectives. Münster: LIT Verlag. Papageorgiou (2012). “Investing in public investment: World Bank (2019). Education Service Delivery in an index of public investment efficiency,” Journal of Mozambique – A Second Round of the Service Delivery Economic Growth, Vol. 17, Issue 3, pp. 235-266. Indicators Survey. Washington, DC: The World Bank. Hubert, D. (2018). Government Resources from Coral FLNG, World Bank (2018a). Procuring Infrastructure PPPs. Resources for Development Consulting, December Washington, DC: The World Bank. 2018. World Bank (2018b). Review of the Mozambique Medium- IMF (2019a). The Revised EBA-lite Methodology, May 2019. Term Fiscal Framework: turning the CFMP into a more Washington DC: International Monetary Fund. effective tool for fiscal and budgetary planning. June. IMF (2019b). “Mozambique’s natural gas resources: tradeoffs Washington, DC: The World Bank. and opportunities”, IMF Country Report No. 19/167, June World Bank (2015). Sudan: Realizing the potential 2019. Washington DC: International Monetary Fund. for diversified development. Country Economic Kapsos, S. (2005). “The employment intensity of growth: Memorandum. Washington, DC: The World Bank. Trends and macroeconomic determinants”, Employment World Bank (2014a). Mozambique Public Expenditure Strategy Papers, 2005/12. Geneva: International Labour Review: Addressing the challenges of today and seizing Office. the opportunities of tomorrow. Washington, DC: The Kojo, N. C. (2014). Demystifying Dutch Disease. Policy World Bank. Research Working Paper 6981. Washington DC: World World Bank (2014b). “Generating sustainable wealth from Bank. Mozambique’s natural resource boom”, World Bank Moolman, E. (2003). “An econometric analysis of labour Mozambique, Policy Note, January 2014. Washington, demand at an industry level in South Africa”, Trade and DC: The World Bank. Industry Policy Strategies, Working Paper 5-2003. Xiao, Y., C. Abdallah, C. Bender and R. Martins (2017), Poulsen, H. (2016). “Mozambican petroleum outlooks: future “Republic of Mozambique – wage bill and civil service lng markets, benchmarking with peer group countries”. reform”. Fiscal Affairs Department, Technical Assistance Presentation by Henrik M. Poulsen, Rystad Energy, Report, September 2017. Washington DC: International Maputo, February 2016. Monetary Fund. 46 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Chapter 3. Mozambique’s Growth Story from Spatial Perspective: Summary of recommendations Recommendations Responsibility Timeline for Scale up access to public services, notably infrastructure, in rural districts Establish specific targets to reach underserved areas in the five-year plan (Plano Short to DNO, MEF Quinquenal do Governo) Medium term Update budget allocation formulas to take access gaps into account. Consider DNO, MEF Short term establishing a fiscal target for investment in basic infrastructure. Invest in road infrastructure focusing on selected types of roads (secondary roads, reconstruction or rehabilitation of bridges and culverts) and areas with greatest MTC, MEF Medium term agricultural potential, especially in the northern and central regions Improving agriculture productivity Increase public investments in small-scale irrigation infrastructure MAP, MEF Short term Support the adoption of improved seeds and fertilizer by liberalizing the agricultural MAP Short term input market and giving smallholders access to finance Promote soil and water conservation measures (such as climate smart agriculture) MAP Medium term Promote agricultural markets Create enabling environment for private sector operators (enforce certification and AIPEX, MAP Short term regulation framework, address distortions by free seed distribution) Address bottlenecks to competition in the import and distribution of fertilizers MAP Medium term Build resilience to climate risks Expand coverage of formal safety nets, including adaptive schemes MITADER, MAP Medium term Increase the availability and use of local weather information to manage risks MITADER, MAP Short term Promote risk transfer mechanisms through insurance arrangements MAP, BoM Medium term MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 47 3. 48 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Mozambique’s Growth Story from a Spatial Perspective Despite strong growth in recent decades and improvements in overall living standards, poverty in some regions and among some groups remains persistent in Mozambique. Understanding why there is a weak link between growth and poverty reduction is necessary, particularly as the country enters its resource boom. As the analysis is hindered by the lack of local-level data on economic activity, this chapter presents an innovative approach to constructing such a dataset using satellite imagery and a range of other sources. The main objective of the chapter is to present new data that shows growth trends at the district level, thus plugging a major knowledge gap in Mozambique, where growth data series are available only at an aggregate level. It finds that agriculture’s poor performance is trapping the rural poor in poverty, while growth in industry and service sectors is relatively dynamic, with growth nodes emerging amongst poor rural districts – especially those with good connections to urban areas. This has some important policy implications. 3.1 Introduction 2017). Cities and towns account for a small proportion of total population, but for over half of the national GDP. Widening inequalities, stemming from a concentrated Despite the considerable growth in urban employment (see pattern of growth, have been a defining characteristic of Chapter 1), the urbanization process in Mozambique remains Mozambique’s growth trajectory. Its high-growth episode gradual compared to those of other countries in SSA. Natural helped to reduce poverty from 60 to 48 percent between population growth accounts for most of the increase in 2002 and 2015. But inequality increased at the same time, urban population, while migration accounts for a small mainly because growth disproportionately benefited proportion. Urbanization has been limited partly because wealthier groups, and as poverty levels remained stubbornly cities are not offering greater employment opportunities high in some northern and central provinces.54 Since 2016, than rural economies. Urbanization allows for, among the pace of growth has slowed as Mozambique has suffered others, “agglomeration economies”, in which proximity a number of shocks, the most recent being COVID-19. to suppliers and customers produces benefits including GDP growth fell to 3.3 percent between 2016 and 2019 on an increase in productivity. Faster urbanization can be average, barely above the rate of population growth. Poverty instrumental in driving broad-based economic development simulations suggest large potential impacts of slower growth when combined with sound institutions and policies to on the poor.55 reduce spatial disparities in welfare (Box 3.1). Urban development has been an important feature of Understanding why there is a weak link between growth structural transformation in Mozambique, with economic and poverty reduction in Mozambique will be important activity concentrated around the largest cities (World Bank, as the country enters its resource boom. While national 54 World Bank (2018); World Bank (2017a); World Bank (2019). 55 See World Bank (2020, forthcoming). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 49 Box 3.1. Spatial transformation, Urbanization, and Agglomeration As a country develops, the process of structural transformation implies that people move out of rural areas and agricultural employment into urban-based manufacturing and service sectors. The successful European and Asian economies are good examples of urbanization with structural transformation. Labor relocated away from agriculture either due to either ‘push’ or ‘pull’ factors or both. The ‘push’ factor stems from productivity growth in agriculture—a Green Revolution—which releases labor for the modern sector (Parente and Rogerson, 2007; Michaels et al., 2012) while the ‘pull’ factor is associated with a rise in non-agricultural productivity—an Industrial Revolution—which attracts underemployed labor from agriculture into the modern sector. In contrast, urbanization and structural transformation have been limited in Mozambique. As discussed in Chapter 1, while crop yields in Mozambqiue are low by regional standards while the manufacturing and modern service sectors are relatively small and unproductive. Urbanization and agglomeration can be a catalyst for broad-based economic development and job creation. “No country has grown to middle income without industrializing and urbanizing”, as shown in the 2009 World Development Report. Economic development is typically characterized by the the rising concentration of people and production in towns and cities. As economies grow, production tends to become more concentrated spatially. Spatial transformation, that is manifested in urbanization, is closely related to the sectoral transformation of countries from agrarian to industrial and then to services. The rate of urbanization (the share of the population living in towns and cities) increased from 29.1 percent in 2000 to 36.5 percent in 2019, despite Mozambique’s rapid and sustained economic growth during this period. This reflects the fact that Mozambique witnessed a limited shift in the structure of the economy away from agriculture (Chapter 1). The spatial disparity in living standards that results from concentrations of economic production can be addressed through putting in place sound policies and institutions to promote economic integration (WDR, 2009). Agglomeration of economic activity can enhance productivity via three key mechanisms: transport cost savings, developing markets for specialized services, and labor market matching. The concentration of production in leading areas and people in lagging areas is linked with geographical divergence in welfare. With progressive policies and institutions, policy makers can deal with the challenge of striking the right balance between allowing “unbalance” economic growth and ensuring inclusive development. These include getting regulations right (including on land, labor, trade and basic services), investing in infrastructure to connect leading and lagging areas and, where necessary, (spatially) targeted interventions. Unleashing productivity-driven growth that creates more, better, and inclusive jobs— including for lower-income, lower-skilled people—is essential. There is a need to create productive employment opportunities in cities and towns to attract underemployed workers from agriculture. Efficient implementation of these policies would enable the economy to recover faster and thrive in the post-COVID-19 world. accounts provide a breakdown of GDP into sectoral districts56 and three sectors: agriculture, industry and services, components, there exists little data about economic activity and extractives/megaprojects.57 The methodology used to at either the province, district or municipality level. Where generate the dataset innovates on established approaches has growth been concentrated? Is growth dominated by for using satellite-based information to estimate sub-national major cities, or are towns and secondary cities also playing GDP. Second, using this data, spatial growth patterns are a role? And what has been driving the distribution of growth presented along with a discussion on the potential drivers dividends in recent decades? These are important questions of the distribution of growth. The chapter concludes with that remain unanswered in the absence of evidence. policy recommendations. This chapter seeks to present Mozambique’s growth story 3.2 Satellite Data Helped to Build a from a spatial perspective. It begins by estimating GDP at the subnational level by combining official national accounts District-Level GDP Database with satellite imagery and geographic information system The first step in understanding the spatial nature of growth (GIS) data, resulting in the first ever dataset for Mozambique’s is to view growth patterns at a local level. Districts were GDP at the district level. The data cover 2000-2019 for 142 56 Mozambique’s political administrative division continues to evolve as different administrative posts are upgraded to district level. Currently, the country has about 156 districts, up from 128 in 1986. The analysis opted for the 142-district administrative division as it is the set that is most common in various subnational databases. 57 This chapter seeks to understand how to broaden growth outside of the main cities by identifying growth nodes and determinants. Although ideally the analysis would have to go beyond districts to better understand urbanization dynamics, the methodology used allowed disaggregation only until the district level. Using the estimated district-level data, the analysis attempted to account for other dynamics, including urbanization, by creating dummies for districts with municipalities (i.e. a variable that indicates the portion of a district that is urban) and defining a proxy for city classification using total population and 50 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM considered as the unit of analysis, as it tends to be the are available. The National Institute of Statistics has been base of many development policies. The available data and releasing data gradually, so data on some key variables methodologies used allowed for a disaggregation down to such as skills and employment were not available. the district level. As these data are lacking in Mozambique, the analysis drew on data from multiple sources to maximize The result was a panel dataset of GDP for three sectors the district-level country data on the following aspects: (agriculture, industry and servicesand extractives) for 142 districts spanning the period 2000 to 2019 (Figure 3.2, • Sectoral GDP: It used the official annual figures for GDP Annex 5A). at the national level and disaggregated them to the district level, mostly using satellite-based information: The Dataset Provides Information that is night-time lights information was used as a proxy for Consistent with Available Evidence industry and services-related sectors58, while agricultural output was measured using land cover classifications To assess the reliability of the data, the estimates were and vegetation greenness data (Box 3.1).59 compared with consumption levels as measured in the two • Output from extractives and megaprojects (extractives most recent national household surveys, the Household products, aluminum and hydropower): this was Budget Survey (Inquérito sobre Orçamento Familiar – IOF) measured separately (and disaggregated to those of 2008/09 and 2014/15. The data generated are largely districts producing them) using export statistics consistent with the surveys, showing Maputo as the province corresponding to the commodities produced. with the highest output/consumption share and comparable • Agricultural population: Many of the districts are rural; shares for the other provinces. The outcome is similar in these districts their population is treated as being when the data are compared with the National Institute of agricultural for the purpose of the analysis. A different Statistics’ (INE) provincial-level GDP estimates (Figure 3.3). approach (based on night-time lights data) was needed Finally, comparing industry and services GDP with data from to address cities and towns where a limited amount of the national firm census shows further complementarity agricultural activity takes place. It was then possible to between the estimated GDP distribution and the distribution combine the ratio of the urban and agricultural habitable of formal firms around the country. areas with estimates of the urban and rural population densities to estimate the agricultural population. Overall, the dataset provides information that is largely • Socio-economic indicators: a database covering consistent with existing regional growth trends. For various socio-economic indicators was constructed to instance, the share of GDP from agriculture in 2017 is highest support the analysis of spatial growth patterns. Databases in Zambezia and Nampula and in the center and north of on socio-economic indicators (e.g. poverty, health, the country, regions well known to be the highest producing agricultural production, and infrastructure) only exist in regions based on annual agricultural production data. In Mozambique at the provincial level. When district-level contrast, the share of GDP from industry and services is data are available, the coverage period or the depth are highest in Maputo and Gaza and generally in the south (Figure limited. For example, the national official database on 3.414). Similarly, GDP per capita (excluding extractives) in roads only covers classified roads to the provincial level. 2017 is highest for the provinces of Maputo and Gaza and Obtaining data on road coverage at the district level lowest for Zambezia and Niassa. would require contacting 156 district administrations individually. To overcome this type of limitation, the analysis resorted to a variety of sources. For example, the Malaria Atlas and the Coca-Cola Last Mile project both contain data on connectivity, but each database was only available for one point in time (2015 and 2018, respectively). Although population census data (2007 and 2017) provides information on some of the variables, including demographics and socio-economic 3.3 The Results Reveal the Growth indicators, not all data variables from the 2017 census Dynamics Behind Spatial Poverty population density. Other variables considered in the analysis include distance to cities, and access to infrastructures. However, although the analysis makes important contributions in terms of understanding the distribution of growth, were there is almost no district data, more needs to be done in the future narrow down the unit of analysis. 58 Nighttime light (NTL) data are certainly far from perfect and while they need to be used with caution, they offer a means of understanding activity that is otherwise invisible.  The two sources of NTL data used are NASA Defense Meteorological Satellite Program (1992-2013) and visible infrared imaging radiometer suite (VIIRS) (2012-2018). The Chapter acknowledges the instability of the time series and used regression to smooth out the annual values in order to extract the trends from the NTL. Unfortunately, neither district nor provincial-level GDP estimates are available and hence the desire of being able to use a ground-truth dataset for validation is a serious challenge. 59 Constant GDP (MT) at 2009 prices published by the Instituto Nacional de Estatisticas (INE) [http://www.ine.gov.mz/]. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 51 > > > Figure 3.1. Satellite Imagery and Sectoral GDP are Well Correlated a. Mozambique agriculture GDP and vegetation b. Mozambique industry and services GDP and index (NDVI), 1990-2015 night-time lights, 1990-2015 12.0 Real GDP Growth, % 12.0 11.5 11.8 26.5 11.0 11.6 10.5 11.4 25.0 In (Lights) 11.2 In (GDP) 10.0 1990 1995 2000 2005 2010 2015 2020 11.0 6000 25.5 10.8 10.6 5500 25.0 10.4 NDVI 5000 10.2 4500 10.0 4000 1990 1995 2000 2005 2010 2015 1990 1995 2000 2005 2010 2015 2020 Source: Own calculations based on NASA Defense Meteorological Satellite Program (DMSP), European Space Agency (ESA), INE. > > > Figure 3.2. Maps of District Sectoral GDP Were Prepared from the Dataset a. Agriculture per capita GDP by district b. Industry & services per capita GDP by district Note: the darker the color, the higher the values Note: the darker the color, the higher the values 52 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Box 3.2. Using satellite imagery to estimate sectoral GDP Various studies have used satellite imagery, especially night-time lights, to estimate economic activity, building on an established relationship between these indicators and the levels of output in the literature. Some studies have used night-time lights to measure the global distribution of economic activity (Ghosh et al., 2010a), to map global poverty (Elvidge et al., 2009), or measure the size of the informal economy (Ghosh et al., 2010a). Njuguna and McSharry (2017) measure multidimensional poverty by combining night-time lights with mobile call detail records. Others measure economic growth (Henderson et al. 2012) or focus on subnational income per capita estimates (Ebener et al., 2005; Bundervoet et al., 2015). This analysis makes an additional contribution by using other satellite imagery-based sources to estimate the distribution of agricultural activity. The spatial distribution of agricultural GDP was estimated based on the size of the population engaged in agriculture per district and the estimated level of productivity. First, for each district, the number of people working in agriculture was estimated using the district’s population and rural population density. To do this, European Space Agency satellite imagery was used to distinguish between urban and agricultural habitable areas. With this information, the number of people working in agriculture was estimated. Agricultural productivity was estimated using the Normalized Differentiated Vegetation Index (NDVI), which is an indicator of health and density of vegetation derived from satellite imagery, as a proxy. The NDVI shows a statistically significant correlation with agriculture output in Mozambique (Figurea). To account for variation in productivity across districts, a scaling exponent was estimated such that NDVI was used to modulate the fraction of agricultural productivity attributed to each district. GDP for industry and services was estimated using night-time lights data, available from the NASA Defense Meteorological Satellite Program. Given the nature of the data, it was not possible to disaggregate industrial activity from services or to distinguish between their sub-sectors. Thus, this segment represents the bulk of non-agricultural output. More specifically, it is dominated by the services sector, which represents 79 percent of industry/ services, given the small share of manufacturing in total output. Industry and services GDP are strongly correlated with data derived from night lights (Figureb). Trends Most of the fastest growing districts are in the northern Structural change is being driven by growth in industry and central regions of Mozambique, which, being poorer, and services. This section discusses the observed spatial are starting from a lower base. In comparison, the wealthier growth patterns over 2008–2017.60 Overall GDP per capita southern districts grew at a slower pace, but from a higher (GDPpc) growth (as opposed to average annual growth) over base. Notably, almost all of the fastest-growing districts are this period stood at 31 percent, and at 4 percent excluding rural, indicating growing dynamism in the non-agricultural extractives, with growth occurring in 94 percent of districts. parts of the rural economy, particularly in northern Most of the non-extractive growth came from industry and Mozambique. This dynamic is now the engine for structural services, which grew by 12 percent; and to a much lesser change as rural areas grow their non-farm economies. extent from agriculture, which expanded by just 0.6 percent. However, it is only in the southern provinces of the country With regards to structural change, the share of agriculture in that industry and services’ share has increased to levels GDP declined, on average, by 3.5 percent while the shares above 50 percent. of industry and services rose by 8.6 percent in 86 percent of districts (Figure 1.5 in Chapter 1). As discussed in more Growth in Agriculture is Below Average detail below, these two sectors exhibited markedly different trends over the period, shedding some light on the growth Unlike industry and services, growth is limited in agriculture. dynamics behind spatial poverty trends in recent years. With an average per capita growth of 0.6 percent over 2008- Growth in Industry and Services is Dynamic, 2017 (10 percent for rural districts) and an improvement in Especially in Rural Areas 83 percent of districts, the sector has barely kept up with 60 Growth rates are based on a comparison of average growth for the period 2008-2012 with growth over 2013-2017. These periods and panels are selected as the reference period throughout this chapter to be consistent with the population census and firm census data, which are used in subsequent sections to explore the context for growth. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 53 > > > Figure 3.3. World Bank Estimates Correlate Well with Mozambique Survey Data a. GDP estimates and IOF consumption shares by b. GDP estimates and IOF consumption shares province 2008/09 by province 2014/15 45 60 40 50 Thousands MZN Thousands MZN 35 30 40 25 30 20 15 20 10 10 5 0 - 0 o ut o ne az a la ca te la zia a ss lgad ut o ne az a te la ca la zia sa do ba fa ni Te pu be Nia ba Te fa ni pu be ias elga ap m G So Ma am m e ap m G So Ma m m N M ha D a D In N Za o M ha N Za o ab In ab C C GDP per Capita 2009 Real Consuption 08/09 GDP per Capita 2015 Real Consuption 14/15 Source: World Bank based own GDP estimates, on IOF 2008/9 and 2014/15 c. INE and WB provincial GDP shares (average 2011-2019) 45 40 35 Thousands MZN 30 25 20 GDP per Capita 2009 15 Real Consuption 08/09 10 5 0 o ne a te la a la zia sa do ut ba az Te fa ic pu ias ap G So an be ga M m M am m N el ha N Za o D In ab C Source: World Bank based on own GDP estimates and INE GDP population growth in most places.61 For example, agricultural three exceptions that started with relatively higher levels, GDPpc growth rates have declined since 2006 and stagnated saw a significant increase in the output share of industry between 2013-2017 for Manica, Zambezia, Nampula and and services, suggesting incipient structural transformation. Niassa provinces in central and northern Mozambique. Of the 128 rural districts, the worst-performing 28 districts Overall, Mozambique is performing poorly compared to provide evidence of structural change associated with African peers in this regard, as shown in Figure . Agricultural urbanization.62 These 28 districts have, on average, higher performance is poor partly because urbanization has not population and population density (202,000 and 48 per achieved its potential to create markets for rural products km2) than the average rural district (172,000 and 43 per (World Bank, 2017). The World Development Report (2009) km2). The regional distribution is mixed, with districts from on Reshaping Economic Geography shows that urbanization all three regions (south, center and north) being amongst and agglomeration, if managed well, can make substantial these three groups. contributions to productivity and economic growth. There are also signs of incipient structural transformation. Low agricultural output growth reflects the low level For example, Marrupa, Namaacha and Sussundenga, the of agricultural productivity in crop production, both by 61 91 percent of districts registered growth rates between 2 and -2 percent. 62 These districts experienced per capita agricultural output growth of less than 0.5 percent. 54 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 3.4. Correlations with Regional Growth Estimates Are Strong a. Agricultural GDP shares and agriculture production b. Share of industry and services GDP and number of levels by province (2017) firms (2015) 40% Map 25% Zam Nam Share of 2017 Agriculture GDP 2015 Share of IS GDP 20% 15% Tet 20% Cab Gaz 10% Man Sof Nam Nia Inh Inh Sof Gaz Tet Map Man Zam 5% Nia Cab 0% 0% 0% 10% 20% 30% 40% 0% 10% 20% 30% 40% 50% Share of 2017 Agriculture Production 2015 Share of Total firms Source: World Bank based on IOF 2008/9 and own GDP estimates Source: WB based on IOF 2008/9 and own GDP estimates > > > Figure 3.5. Growth in Industry & Services is Fastest in the Poorest Districts Growth in IS GDPpc versus IS GDPpc level (baseline 2008-2012) by urbanization (left) and region (right) CAGR: compound annual growth rate; GDPpc: gross domestic product per capita; IS: Industry & services. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 55 > > > Figure 3.6. Agricultural Output per Capita Growth Rates are Below Other Countries in the Region Constant USD 2010 10% 5% 0% Mozambique Kenya Rwanda Ethiopia Tanzania Zimbabwe Sub-Saharan Africa -5% -10% 2002/2007 2008/2013 2014/2019 Source: WDI > > > Figure 3.7. Rainfall Patterns Explain the Stable Growth Distribution in Agriculture a. Growth in AG GDPpc versus AG GDPpc (2008-2012) by b. Growth in AG GDPpc versus AG GDPpc urbanization level (left) and region (right). (2008-2012) with rainfall level: high (blue); medium (green) and low (red) CAGR: compound annual growth rate; GDPpc: gross domestic product per capita; AG: agriculture. Source: World Bank (2017). 56 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM historical and comparative standards.63 Mozambique has Rural Mobility one of the lowest  cereal yields per hectare in Southern Africa. There has been little progress in the last two decades What do these spatial growth trends tell us about the in bringing yields in traditional crops closer to the regional link between growth and poverty reduction? To explore average. Compared to SSA peers, Mozambique has the this further, the analysis expands the existing measure of lowest levels of land and labor productivity (Figure 3.8).64 growth elasticity of poverty from the national to the sub- Nonetheless, there is significant unrealized potential to national level by combining data from the two most recent achieve efficiency gains in crop production in the northern household budget surveys and the sub-national GDP data and central regions (Fiugure 3.9). Together with area from the dataset constructed for this chapter.67 Figure 3.10 expansion, modest increases in productivity are observed in shows the results at the provincial level.68 Mozambique after the early 2000s, albeit with great variation from year to year. Relative to other countries in SSA, yield Growth has not helped reduce poverty in the poorest parts gaps (the difference between actual yields and potential of the country. Growth has been associated with poverty yields) in Mozambique are particularly pronounced. Yield reduction in all but two provinces between 2008/9 and gaps in staples and cash crops feature more prominently in 2014/5 (with the strongest link in Maputo city and province). the Central Region and the highlands in the Northern Region, In the two poorest provinces, however (Nampula and Niassa), zones characterized by favorable agro-climatic conditions growth has been accompanied by an increase in poverty; and soil quality.65 Some parts of Mozambique have strong in Cabo Delgado poverty reduction has been minimal. The potential to improve productivity by adopting small-scale province with the highest per capita growth rate had only irrigation and transitioning into higher-value cash crops, with the sixth highest elasticity. Industry and services’ growth agricultural extension support.66 rates in these northern provinces, mainly Cabo Delgado 3.4 Poverty Will Persist Without and Niassa, have been amongst the highest but with limited Boosting Agricultural Productivity and impact on poverty reduction. In sum, the growth elasticities > > > Figure 3.8. Land and Labor Productivity in Mozambique are Among the Lowest in the Region a. Land b. Labor 1000 2500 800 1990 2000 1990 2014 2014 600 1500 400 1000 200 500 0 0 A a i e da ia a a e A a i e da ia a a e aw aw ny nd bi ny nd bi qu bw qu bw SS SS an an an an m m al al Ke Ke ga ga bi bi nz nz ba ba Rw Rw Za Za M M am am U U Ta Ta m m oz oz Zi Zi M M Source: World Bank (2017). 63 World Bank (2020). 64 Land expansion has contributed to agricultural output growth over the last two decades, but this has put pressures on forests. Two-thirds of forest losses and degradation in Mozambique are attributed to small-scale agriculture. 65 Another explanation for the poor performance of the agriculture sector may have to do with the fact that urbanization has not achieved its potential to create the consumers markets for rural products. Rural areas produce just enough to survey while the urban areas have access to high quality imported agriculture products. 66 Irrigation is limited in Mozambique despite being a downstream country with large seasonal flows of big rivers in the region such as the Zambezia and the Limpopo. Less than 4 percent of cultivated area is irrigated, well below the total potential (70 percent or 3.1 million hectares) (World Bank, 2020). Innovation in production support services, notably agricultural extension, also offer opportunities to raise productivity. There has also been little transition, if any, into higher-value cash crops by small farmers. 67 Growth elasticity of poverty refers to the percentage reduction in poverty rates associated with a percentage change in average per capita income, as per the World Bank’s Mozambique Poverty Assessment (World Bank, 2018). 68 The analysis is at the provincial level to fit the structure of the household survey data. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 57 > > > Figure 3.9. Mozambique’s Agricultural Production Value is Far Below Potential a. Agriculture production value, USD million b. Agriculture production potential, USD million Source: World Bank (2017). show that that growth has not helped reduce poverty much to find productive employment in more dynamic non- in the poorest parts of the country because of the weak farm activities. However, available evidence on rural-urban performance of agriculture in these regions, despite the migration suggests this is limited in Mozambique (Figure strong performance of services and industry (albeit from low 3.12). The overall rate of urbanization has been low and base). movement of labor to urban areas has been concentrated in Maputo city and province (Figure 3.13). Moreover, the Consistently weak agricultural growth rates, despite share of households in rural areas earning their income strong growth in other sectors, points to the persistence from agriculture has barely shifted over the past 20 years, of poverty traps linked to agriculture’s performance. dropping from 94 percent in 1997 to 91 percent in 2014, Figure 3.11 shows how agricultural growth remains capped compared with a more robust transition in urban areas at around 2 percent even when industry and services GDP (Figure 3.14). For growth to make a significant dent in poverty, grew at rates well above 20 percent in per capita terms. We especially in northern Mozambique, more needs to be done can also see that most of the districts in this situation (high to raise agricultural productivity and increase rural mobility growth in industry and services but low growth in agriculture) to connect farmers to markets and the growing non-farm are located in the north of the country in the provinces of opportunities in their districts. Cabo Delgado, Nampula and Niassa, where poverty has been least responsive to growth. Weak agricultural performance limits improvements in the livelihoods of farming households, unless they can migrate 58 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 3.10. Growth is Not Reducing Poverty in the Poorest Provinces GDP per capita growth elasticity of poverty at the provincial level and agriculture and industry per capita GDP, 2008/9-2014/15 15.0 100% 10.0 50% 5.0 0% 0.0 a la do a l zia te ne la a ity o na ss az ic ut pu fa Te C ba ga be an io ia ap G So am o N at m el m M M ut N D ha Za N ap -50% -5.0 o In M ab C -100% -10.0 Agriculture GDP 2009/2015 Industry & Servides 2009/2015 Income/Poverty Elasticity 15-09 Note: Growth elasticity of poverty (at the provincial level) refers to the percentage reduction in poverty rates associated with a percentage change in average per capita income. Source: World Bank Calculations using IOF 2008/9 and 2014/5 and GDP estimates > > > Figure 3.11. Agricultural Growth Has Remained Low While Industry & Services GDP Growth Has Driven Overall Growth in Cities and Towns AG GDPpc growth versus IS GDPpc growth showing cities (red), towns (blue) and rural (green) districts Source: World Bank Sta Estimates. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 59 > > > > > > Figure 3.12. Rural-urban migration was lower Figure 3.13. Most rural-urban migration is than in SSA up to the middle of the 2000s concentrated on Maputo Annual urban population growth Annual urban population growth 6 5% 4% 5 Annual urban growth (%) Annual urban growth (%) 3% 3.0% 4 2.9 2% 3 1% 1.7% 1.7 1.5% 2 0.5 0% 2.9 -1% -2.0% 1 1.8 2.0 1.7 -2% 0.4 0 Europe Asia Africa Mozambique -3% (1800 – 1910) (1960 – 2010) (1960 – 2010) (1997 – 2007) Maputo Province Maputo City Migration Natural growht rate Natural growth Migration > > > Figure 3.14. Rural households appear to be trapped in low-productivity agriculture (Structure of employment) a. Rural b. Urban 100% 100% 10 15 23 28 80% 80% 16 11 8 10 10 60% 60% 20 23 94 93 93 91 24 40% 40% 67 48 20% 20% 45 34 0% 0% 1997 2003 2009 2014 1997 2003 2009 2014 Agriculture Private Wage (Non-Farm) Agriculture Private Wage (Non-Farm) Public Wage (Non-Farm) Self-Employed (Non-Farm) Public Wage (Non-Farm) Self-Employed (Non-Farm) Source: World Bank using Integrated Agriculture Survey (2015). Source: World Bank using Integrated Agriculture Survey (2015). Weak commercialization is at the core of low agricultural from the Northern and Central regions is engaged in markets. income growth and high rural poverty (World Bank, 2020).69 About 5 percent of the total production of staples and cash Smallholders have weak market orientation, which impedes crops in Mozambique is traded in local market, much lower productive investments and diversification. The rates of than those of Tanzania (52 percent), Uganda (42 percent) or commercialization are low across all regions, although a Malawi (39 percent). Less than half (45 percent) of farmers relatively higher share of small-scale farmers (13–18 percent) growing cash crops sell any part of their production in 69 For more detailed discussion, see World Bank (2020) Mozambique Rural Income Diagnostic. 60 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > markets. Labor markets in agriculture are also thin, even for Figure 3.15. Rural areas outstrip towns and temporary or seasonal work. Promoting commercialization cities for their rates of growth will require addressing three key barriers (World Bank, 2020): (i) infrastructure gaps, (ii) underdeveloped agricultural input markets, and (iii) poor organization and functioning of output markets and value addition. 1.0 0.8 Rural 3.5 Growth is Closely Linked to 0.6 All AG GDP pc CAGR, % Urbanization and Connectivity 0.4 0.2 Rural Areas and Towns Are Driving Growth 0.0 -0.2 An analysis of growth by level of urbanization shows that rural areas and towns, not cities, have been the drivers of growth -0.4 City and structural change. To explore the relationship between -0.6 growth and urbanization, the analysis categorizes districts -0.8 into three groups according to their level of urbanization: (i) Town -1.0 mainly urban (main cities). i.e. total population and population IS GDP pc CAGR, % density are respectively above 50,000 inhabitants and 1,500 inhabitants per square kilometer; (ii) partly urban, with Note: Growth in AG GDPpc versus growth in IS GDPpc for di erent levels of urbanization, with size of bubble reflecting features of small cities or towns, i.e. total population between GDPpc excluding extractives (2008-2012) 5,000 and 50,000 and population density between 250 and 1,500 inhabitants per square kilometer; and (iii) mostly rural, with total population below 5,000 and population density below 250. The list of districts in each category is presented in Annex 5D, Table 0.1.70 The growth nodes identified echo the regional patterns discussed previously (Figure 3.5). To identify districts that Towns represent 18 percent of GDP and have the highest behave as growth nodes we first ranked growth in agriculture levels of per capita GDP growth, twice as high as cities.71 GDPpc and growth in industry and services GDPpc. The Towns have grown at a faster pace than cities in terms of product of these two ranks was then calculated and divided industry and service output (Figure 18 and Table 3.3), despite into two groups depending on whether the base level of a higher rate of population growth (Figure 21). Their negative GDPpc (excluding extractives) was below or above the agriculture growth rates (like cities) indicate that they are in median. An additional ranking of this product then provided the later stages of structural change. Overall, towns have the top districts in each growth node category (Annex been the growth engine of the urban economy. Their robust 5D, Table 0.2). Growth nodes from a low base (GNL) are growth, starting from a high base, is creating opportunities dominated by rural districts from the north, making up 60 and absorbing labor. percent of the districts identified. In contrast, growth nodes from a high base (GNH) contain no districts from the north Rural areas, however, have the fastest rates of growth and exactly half from the south and center of the country (mostly from a low base) and structural change. As (Figure 20). previously discussed, rural areas have been experiencing robust acceleration in industry and services output, which Three main indicators distinguish the growth nodes from has accelerated their growth performance and makes them other rural districts: poor market connectivity, access to a key engine of growth. To identify the rural growth nodes, public and financial services. Rural growth nodes present the analysis looked for districts that demonstrate strong much higher growth rates, but much lower levels of access GDP per capita growth rates, even after the direct effects to infrastructure, than other rural districts. This is explained by of extractives are excluded. These are typically rural areas the fact that these fastest-growing districts have started from that are making a structural transition and where jobs are a very low base, including very low levels of infrastructure. being created outside of agriculture. Some growth nodes Analysis for 2007–2017 confirms that growth nodes have are mostly rural areas that are growing from a low base, limited connectivity (Table 3.2). The average distance to the and others are further along the structural transition and are nearest road is 20 percent higher than in other rural districts growing from a higher base (Figure 3.16). and three times the average distance for towns and cities. 70 The use of two variables (population and population density) to classify districts into three categories is based on a previous categorization used by INE and recent research by the World Bank on the need to have a data-driven classification of urbanization: https://blogs.worldbank.org/sustainablecities/how-do- we-define-cities-towns-and-rural-areas 71 Average share of total GDP between 2000 and 2019. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 61 Table 3.1. Output growth rates Districts/sector AG 2000 AG 2019 IS 2000 IS 2019 EM 2000 EM 2019 All 73.6% 48.4% 25.1% 48.7% 1.3% 2.9% City 1.0% 1.3% 99.0% 98.7% 0.0% 0.0% Town 9.5% 3.8% 86.9% 94.0% 3.6% 2.2% Rural 82.2% 54.2% 16.7% 42.7% 1.1% 3.1% North 85.5% 60.5% 14.5% 38.3% 0.0% 1.2% Centre 75.9% 53.8% 21.4% 42.2% 2.7% 4.1% South 51.8% 21.5% 47.2% 74.6% 1.0% 3.9% Note: AG: agriculture sector; IS: industry and services sectors; EM: Extractives and mining sectors Source: WB estimates > > > Figure 3.16. Growth Nodes Are Mainly Rural a. GDP per capita growth 12.0% 11.1% 10.0% 8.4% 8.0% Agriculture GDPpc growth 6.0% 5.0% Industry and Services 4.6% GDPpc growth 4.0% 2.1% 1.8% 1.7% 2.0% 1.0% 1.0% 0.4% 0.0% -0.4% -0.7% -2.0% All City Town Growth Growth OtherRural nodes from a nodes from a low base high base This may be because road expenditure is biased to urban as well as higher consumption rates than rural areas, areas (World Bank, 2019)72, contributing to low connectivity generating greater demand for off-farm goods and services. levels in rural areas.73 Mobile phone ownership is also lower The population of small and medium-sized towns is growing than in other groups of districts, but the pace of growth was almost twice as fast as in cities (4.7 percent compared to 40 percent higher than other rural districts. 2.9 percent; Table 3.2). The average per capita consumption in towns is double the rural average. The clustering of rural Rural non-farm income growth and employment off-farm jobs around belts with better connectivity and opportunities in Mozambique are greater in areas higher population density is particularly pronounced in the closer or better connected to small and medium-sized southeastern part of the southern region – where most of towns, cities or along transport corridors.74 Previous the rural non-farm income growth is observed. studies of Mozambique have identified access to credit, technology, inputs, physical infrastructure and markets as Access to public services is also growing quickly in rural critical bottlenecks when starting a household enterprise, growth nodes. In rural growth nodes, less than 1 and 3 particularly in highly isolated rural areas.75 Towns and cities, percent of the population has access to electricity and tap in contrast, tend to have greater access to public services water, respectively. This is below the rural average (4–6 72 Data on the distribution of spending in roads used was from the Administracao Nacional de Estradas. The findings echoed the significant increase in distance to transport reported by rural households in IOF 2015. 73 World Bank (2019). 74 World Bank (2020). 75 Fox et al. (2016). 62 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 3.2. Growth Nodes Share Certain Characteristics Unit/Variable Period All City Town Rural Other Level Growth Rural of Nodes Districts signif Agriculture growth per capita 2007–17 1.00% -0.40% -0.70% 1.10% 1.30% Industry/services growth per capita 2007–17 5.00% -0.40% 2.10% 16.6% 5.30% *** Demographic Population growth 2007–17 3.00% 4.10% 4.70% 3.00% 2.80% Working age pop growth 2007–17 2.80% 3.90% 4.80% 2.60% 2.60% Years of schooling Average 2017 6.3 8.2 7.9 6 6.2 * Connectivity Roads / area Km 2002-16 5.20% 8.30% 12.40% 3.70% 4.60% ** Roads / area growth 2002-16 5.20% 8.30% 12.40% 3.70% 4.60% ** Distance to roads Km 2015 5.2 1.7 1.9 6.3 5.4 ** Mobiles Per capita avg 2017 21.90% 37.70% 37.20% 18.10% 20.40% Growth in mobiles Per capita avg 2017 27.40% 9.90% 13.50% 37.90% 27.50% *** Public services Electricitya Per capita avg 2017 6.80% 44.50% 31.40% 0.80% 3.50% ** Water b Per capita avg 2017 10.20% 48.70% 43.00% 2.80% 6.40% * Growth electricity 2007–17 22.00% 8.50% 12.00% 32.40% 21.50% *** Growth water 2007–17 18.90% 7.00% 7.40% 26.10% 19.20% ** Private sector Number of formal firms Average 2014-15 303 3665 687 51 126 *** Number of mobile money agents Per capita avg 2018 0.06% 0.38% 0.24% 0.02% 0.04% ** Growth Firms 2003-15 2.80% 6.50% 3.60% 3.20% 2.50% Growth Money Agents 2016-18 70.70% 31.00% 58.50% 31.40% 81.40% Source: Population Census 2007 and 2017, CEMPRE and 2002/3 and 2014/5, Bank of Mozambique, PLM project, and GDP estimates. Note: a) Refers to electricity as a source of energy (including cooking and heating) according to the 2017 population census; b) Refers to access to tap water inside the house and in the yard. percent). But growth in access to these public services was Inhambane and Manica having the lowest levels of access over 35 percent higher than other rural districts and more (less than half of the access levels of Maputo). This partly than 3 times higher than in cities and towns. This may be reflects the fact that road expenditure78 tended to be higher explained by the fact that investment in infrastructure (other in urban areas, thus contributing to the observed decline than roads, such as schools, hospitals, water points) is in rural connectivity, whereas non-road spending had the inversely correlated with urbanization rate.76 opposite tendency. More generally, World Bank (2019) finds significant regional disparities in access to infrastructure, with There is an urban bias in access to infrastructure and the central and northern regions being the areas with lowest public services, namely transport and education (Figures levels of coverage (Figure 22 and Figure 23). Households in 3.21 and 3.22).77 With the exception of health and water, Maputo and Gaza have the highest levels of access to basic the gap between rural and urban areas has widened across infrastructure, while the two most lagging provinces, Tete all sectors, with a deterioration in rural connectivity. Rural and Zambezia, consistently remain at the bottom of the areas had lower access to basic infrastructure (transport, scale. electricity, education) than urban, with Zambezia, Tete, 76 World Bank (2019). 77 World Bank (2019). 78 Road expenditure calculation was based on the road funds data. The results were compared with the urbanization levels, which were calculated using the data from the politico administrative division used by INE in the administration of IOF. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 63 > > > > > > Figure 3.17. Growth Nodes are Concentrated Figure 3.18. The Population of Small and Medium- in the North and South of the Country Sized Towns is Growing Faster Than Cities (Annual population growth rates, 1997-2007) 25 Cities Namacurra Towns Nametil 20 Average annual population growth (%) Inhaminga Manhiça 15 10 Matola Nampula 5 Beira Maputo 0 -5 -10 1 10 100 1 000 Population size, 2007 (1000s) Note: Solid dots show average population growth rates Source: Dorosh et al (2017) Connectivity is an Important Channel for with an extractive or megaproject activity; (ii) districts on Growth a development corridor; and (iii) districts near a city or a large town.79 Figure 26 plots these district groups by their The spatial growth narrative for Mozambique, and the agriculture and industry/services growth rates. The data potential to accelerate it, has traditionally emphasized show that all three district groups have higher GDPpc two main channels. The first is linked to development of levels than the average for all districts and the average for the extractive industries and mega projects which, through large urban centers (as shown in the bubble sizes in Figure growth spillovers, could boost demand locally and stimulate 26), confirming the importance of these channels for the output. The other channel is the role of development levels of output. The evidence shows that all three districts corridors which, if sufficiently upgraded, could lead to the groups have similar levels of agricultural growth. However, emergence of growth poles propelled by trade flows along districts that are near cities and large towns experience the corridor (World Development Report, 2009). A third and higher levels of industry and services growth than those on under-discussed channel is the role of large urban centers a development corridor or with an extractive/megaproject in spurring growth in neighboring areas. This has potential industry. With growth having been largely consumption- given that they represent a large share of demand and driven in recent decades, the role of cities as hubs for consumption. demand is an important channel for growth that could be stimulated by increasing demand linkages with neighboring So how does growth performance vary for these different rural areas. channels? The analysis categorizes districts into: (i) districts 79 Defined as districts located along or connected to the three main east-west transport corridors connecting provinces and neighbor hinterland countries to ports, and industrial zones. The categorization considered the three main development corridors: (i) Maputo that connects South African neighbor provinces and Swaziland to Maputo port; (ii) Beira, the transport infrastructure connecting the Beira port to Zimbabwe, Malawi, Zambia and the Moatize coal mining in Tete; and (iii) Nacala that connects the Nacala port and Zambia, Zimbabwe and Malawi, passing through Zambezia, Niassa, Cabo Delgado, Nampula and Tete (in the latter province it serves coal exports). Of the total 142 districts, 36 were identified as being in a trade corridor and 75 were considered connected to a trade corridor. 64 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > > > > Figure 3.19. There are Significant Regional Figure 3.20. Access to Infrastructure and Disparities in Access to Infrastructure Poverty Reduction are Linked Access to basic infrastructure and population growth Access to infrastructure and poverty reduction (2009-2015) (2009-2015) 1 1 Bubble size = population Bubble size = % Maputo Maputo 0.9 0.9 increase 2007 - 2017 below the poverty ine Maputo City City 0.8 in 2014/15 0.8 Maputo 0.7 0.7 Gaza Acess 2015 Gaza 0.6 Acess 2015 0.6 Manica Manica 0.5 Sofala 0.5 Inhambane Sofala Inhambane Cabo Delgado Cabo Delgado 0.4 0.4 Tete Niassa Niassa 0.3 Tete 0.3 Nampula Nampula 0.2 Zambezia 0.2 Zambezia 0.1 0.1 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Acess 2009 Acess 2009 Source: World Bank (2019). Source: World Bank (2019). > > > > > > Figure 3.21. Insu cient Progress Has Been Made Figure 3.22. …With Expenditure Trends in Channeling Resources to Underserved Areas… Mirroring Access Indicators Investimento a nível distrital por quintil - 2009/15 Índice de acesso à ifra-estrutura básica por distrito 2015 Source: World Bank (2019). Source: World Bank (2019). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 65 > > > Figure 3.23. Proximity to a Large Town or City is a Key Driver of Growth Growth in AG GDPpc versus growth in IS GDPpc for di erent spatial groups (2008-2012) 1.0 0.8 Near Development Corridor 0.6 All Near Extractives & Megaprojects 0.4 AG GDP pc CAGR, % Extractives & Megaprojects 0.2 Near Cities & Large Towns 0.0 On Development Corridor -0.2 -0.4 -0.6 -0.8 Cities & Large Towns -1.0 0 2 4 6 8 10 12 14 16 IS GDP pc CAGR, % Note: The size of bubble reflects GDPpc, excluding extractives. 3.6 Policy Recommendations to improving agricultural output. Small cities connect farmers to input and output markets, while medium-sized The analysis finds that the areas with the highest levels cities serve as logistical and transport hubs and host larger of poverty, especially the rural districts of Niassa, Cabo consumer markets. Improved linkages between extractive Delgado and Nampula, experienced the most rapid growth industries and the local economy can also expand access rates in industry and services. So, why are poverty levels in to the benefits of resource-driven growth. A comprehensive these areas persistently high if growth has been respectable? strategy that improves the business climate and supports The analysis finds that agriculture, the sector which employs small and medium-sized enterprises (SMEs) could help the majority of the poor, has seen modest growth rates. maximize the benefits of growth corridors and enhance the This has been the case even in the rural districts that have developmental impact of megaprojects (see Chapter 4). experienced rapid growth in other sectors. This suggests a somewhat different narrative for Mozambique’s rising growth Improve Agriculture Productivity to Grow the disparities. As opposed to non-inclusive growth, dominated Incomes of the Rural Poor by the large cities, there is a pattern of widespread growth in industry and service sectors with growth nodes emerging Increasing agricultural productivity could have a major amongst poor rural districts. It is agriculture’s stagnant impact on poverty reduction as most of the rural poor performance that is keeping the rural poor in a poverty trap. are locked into staple agriculture. Input intensification and technology adoption can increase productivity. Farmers These findings have important policy implications, which are who adopt technologies such as irrigation, fertilizer and listed below. pesticides, are nearly 30 percent more productive than those who do not.80 Yields are also responsive to other investments Strengthen Intersectoral Linkages to Promote (mechanization), production methods (crop rotation and Growth in Rural Areas line sowing) support services (agricultural extension),81 and ICT innovations (mobile banking). The availability Greater rural-urban connectivity could better integrate of improved seeds and fertilizer could be promoted by, the poor into markets. The growth of cities is central among others, liberalizing the agricultural input market and 80 Access to and use of modern agricultural inputs such as improved seeds, fertilizer, and pesticide remain low due to the limited supply of quality inputs at affordable prices (World Bank, 2020a). 81 Only a small fraction of farmers has access to extension in Mozambique. This problem prevents farmers from tapping into the best production inputs and practices. Digital technologies can be leveraged to improve the efficiency of delivery of extension services. 66 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM giving smallholders access to finance. Efforts to improve investment in basic infrastructure could be considered to transportation infrastructure and develop agricultural value help investment policy in reaching its goals. chains could increase the productivity of smallholder farmers and promote more inclusive growth (World Bank, 2020a). Farmers are also highly exposed to price and climate-related shocks and lack risk mitigation instruments. Strengthen Investment Policy to Foster Investment in Non-Extractive Sectors The investment law and investment promotion capacity need to be updated.82 The law defining the framework for investment policy has remained largely unchanged. Mozambique’s investment regime suffers from outdated and fragmented laws, and an absence of implementing regulations. There is a need to modernize the Investment Law, and other related legislation, and align it with international investment policy commitments. This would include implementing a registration mechanism instead of a screening mechanism within the Investment & Export Promotion Agency of Mozambique (APIEX) (Chapter 4, Section 4.3). Moving to a simple registration process in the Investment Law would considerably simplify the process of establishing new investment in the country. Scale Up Access to Public Services, Notably Infrastructure, in Rural Districts Improving transportation networks and logistical capacity will better connect rural areas to urban centers and export points. Connectivity to cities and towns is a channel for growth in rural areas. Expanding transport corridors will be essential to promote productive farming practices, enhance smallholder participation in markets, and increase their incomes. Given fiscal constrraints and absorptive constraints, investments should prioritize selected road types (feeder roads, rehabilitation of bridges and culverts) in areas with strong agricultural productivity potential, especially in the northern and central regions (World Bank, 2020a). Ensure Public Investment and Budget Allocation Take Spatial Disparities into Account Public investment programs need to take spatial disparities into account if they are to succeed in reversing the growing gaps in access. For instance, the Plano Quinquenal do Governo, the GoM’s five-year plan would benefit from explicit targets that identify underserved areas and whether they are catching up or falling behind. Tackling disparities also entails restructuring budget allocation formulas to take gaps in access to public services (notably infrastructure) into account, then ensuring that the formulas are applied in practice (World Bank, 2019). Further, a fiscal target for 82 This section draws on World Bank (2020a). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 67 References Barro, R. J. and Sala-i-Martin, X. (1995). “Economic Growth”. Islam, N. (1995). “Growth empirics: a panel data approach”. McGraw-Hill, New York. The Quarterly Journal of Economics 110(4): 1127-1170. Baumol, W. J. (1986). “Productivity growth, convergence Mendes-Resende, G. (2013). “Spatial Dimensions of Economic and welfare: what the long-run data shows?”. American Growth in Brazil”. ISRN Economics, vol. 2013, Article ID Economic Review, LXXVI: 1072-85. 398021, 19 pages. https://doi.org/10.1155/2013/398021. Bundervoet, T., Maiyo, L. and Apurva, S. (2015). Bright lights, Njuguna, C. & McSharry, P.E. (2017) “Constructing big cities: measuring national and subnational economic spatiotemporal poverty indices from big data”. Journal of growth in Africa from outer space, with an application to Business Research 70: 318-327. https://papers.ssrn.com/ Kenya and Rwanda. Policy Research Working Paper 7461. sol3/papers.cfm?abstract_id=2852722 World Bank Group. Vidyattama, Y. (2010). “A Search for Indonesia’s Regional Ebener, S., Murray, C., Tandon, A. and. Elvidge, C. (2005). Growth Determinants”. ASEAN Economic Bulletin 27(3): “From Wealth to Health: Modeling the Distribution of 281-294. Income per Capita at the Sub-National Level Using Night- World Bank (2016). “Mozambique – Systematic Country Time Light Imagery”. International Journal of Health Diagnostic”. World Bank: Washington, D.C.: Geographics 4(5): 1–17. World Bank (2017a). “Mozambique Urbanization Review”. Elvidge, C., Sutton, P., Ghosh, T., Tuttle, B., Baugh, K., Bhaduri, Africa Urban and Resilience Unit. World Bank: Washington B. and Bright, E. (2009). “A Global Poverty Map Derived DC. from Satellite Data”. Computers and Geosciences World Bank (2017b). “Let’s Work Mozambique Country Pilot 35:1652-1660. – Jobs Diagnostic Report”. Policy Report. World Bank: Fox et al. (2016). “Household Enterprises and Poverty Washington, D.C. Reduction in Sub-Saharan Africa.” Development Policy World Bank (2018). “Strong but not broadly shared growth. Review, 34 (2): 197-221 Mozambique Poverty Assessment”. World Bank: Maputo, Ghosh, T., Powell, R., Elvidge, C., Baugh, K., Sutton, P. and Mozambique. Anderson, S. (2010a). “Shedding light on the global World Bank (2019). “Mozambique Economic Update: distribution of economic activity”. The Open Geography Mind the Rural Investment Gap”. World Bank: Maputo, Journal (3): 148-161. Mozambique. Ghosh, T., Powell, R.L., Anderson, S. Sutton, P.C. Elvdige, C.D. World Bank (2020a). Mozambique Rural Income Diagnostic: (2010b). “Informal economy and remittance estimates of Cultivating Opportunities for Faster Rural Income Growth India using nighttime imagery”. International Journal of and Poverty Reduction. Washington, D.C.: World Bank Ecological Economics & Statistics, 17. Group. Gini, C. (1909). “Concentration and dependency ratios World Bank (2020b). Country Private Sector Diagnostic. (Translated into English)”. Rivista di Politica Economica, Washington, D.C.: World Bank Group. 87 (1997), 769–789. World Bank (2020, forthcoming). Mozambique Economic Henderson, V., Storeygard, A. and Weil. D. (2012). Measuring Update. World Bank: Washington, D.C Economic Growth from Outer Space. American Economic Review 102(2): 994-1028. 68 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 69 4. 70 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Can Services Drive Growth in Mozambique?83 Services have been an important source of growth and job creation. However, the country needs to develop a more competitive service sector to strengthen its role as a backbone of the economy. This chapter presents the evolution of services sector performance in Mozambique and explores the potential of services to drive growth. It finds that an open trade and investment regime, supported by a transparent and effective regulatory framework, is key to developing a services sector. A regulatory assessment with a focus on the backbone services of Information and Communication Tecnology (ICT), telecoms, transport and logistics identifies the regulatory bottlenecks that restrict trade and investment services. The chapter concludes with policy recommendations for a regulatory reform agenda for boosting growth and productivity in these key services. 4.1 Introduction 83 of productivity growth, driven primarily by less complex service sectors such as trade, transport and retail, seems Mozambique’s recent economic trajectory reflects the to be reaching its limits, while more complex services contemporary challenge of finding growth beyond the such as financial and professional services and ICT are not traditional manufacturing-led development paradigm. sufficiently developed to be competitive. This has only been Following an investment boom that started in the late 1990s, exacerbated by the global pandemic, which has caused Mozambique’s productivity increased as labor moved out of a rapid decline in domestic demand for and significant agriculture. Services have now become an important source disruptions in services sectors and a worsening external of growth and job creation (Lachler and Walker 2018). Since environment more generally for export sectors (Box ). the early 2000s, services output has grown considerably faster than manufacturing, though in recent years both This chapter argues that focusing on services still offers have stagnated at less than 5 per cent growth per annum Mozambique significant growth potential. However, the (Figure 4.1a). This reflects findings across Africa, where the economy needs to transition to a more complex type of reallocation of resources to services has been a prominent services sector to bolster its contribution to growth and characteristic of structural change. Mozambique’s services strengthen its role as a backbone of the economy. A central sectors have consistently been the most productive sector, premise of this Country Economic Memorandum is that achieving the fastest growth in output per worker over the Mozambique will have to look to look for growth elsewhere, past two decades. rather than only pursuing the traditional manufacturing-led growth model. Mozambique could grow its services sector Even before the COVID pandemic, there were nascent in size and sophistication, while also pursuing industrial signs that the impact of services as a driver of growth was sectors where the country has a comparative advantage. diminishing. The rate of productivity growth in services The dynamics underlying Mozambique’s declining service has declined, from about 3 percent from 2007-11 to 1.7 sector productivity growth, and the potential of more percent from 2012-16 (Figure 4.1b). As such, the initial wave complex services to drive job creation and economic 83 Based on a background paper prepared by Martin Molinuevo (Senior Private Sector Specialist, ETIRI), Lillyana Daza Jaller (Consultant, ETIRI) and Jakob Engel (Economist, ETIRI). Inputs by Adriana Conconi (Consultant, EAEF2), Taciana Lopes (Consultant), Gabriela Schmidt (Economist, EAEM2) and Dirk van Seventer (Consultant) are gratefully acknowledged. Overall oversight and helpful guidance were provided by Shireen Mahdi (Senior Economist, EMNM1). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 71 transformation, represent a significant knowledge gap. This also contributed the most to growth, making up over one- chapter aims to improve understanding of the composition, third of total growth in each of the last three decades, and role and potential of the services economy in supporting almost half in 2010–2018 (Figure 4.227b). Mozambique’s growth path. There has been considerable growth in the number of Services Have Played a Key Role in services firms, with these making up the largest number Mozambique’s Economy Since Independence of firms country-wide, both in 2002-03 and in 2014-15 (Figure 4.3a). Over time, the number of firms has grown Private sector services have played a leading role in in all sectors except agriculture, but the number of Mozambique’s economy since independence. While private services firms has increased most rapidly, almost the share of manufacturing in GDP has fallen from about doubling between 2002-03 and 2014-15. Even though the 15 percent in 1994 to less than 10 percent, services have extractives sector has the lowest number of firms, it has consistently made up 40-50 percent of output (Figure 4.227a), experienced the fastest growth, with firm population almost approximately in line with the average for sub-Saharan Africa quadrupling. Agricultural establishments, meanwhile, have but significantly more than in low-income countries. Private seen a decline though this does not include smallholder sector services (defined in the note under Figure 4.227) have agricultural firms. Box 4.1. The impact of COVID-19 on the services sector The service sector is the worst hit by the COVID-19 crisis. Firms have been affected by a combination of channels which include (i) measures to curb the spread of the virus that reduced the flow of people, availability of the workforce, and overall demand for domestic services, (iii) disruptions in global value chains reduced international demand for domestic services, (ii) limitations imposed in borders led to constraints to import and export. About 86 percent of the firms negatively affected by the pandemic operate in the service sector. Firms reported a 54 percent reduction in business volumes in the second quarter of 2020 (year-on-year). Economic activity deteriorated across all services activities. Firms in the hospitality industry were particularly affected due to movement restrictions. The commerce and retail sector, representing over 74 percent of the small firms in the country, suffered from import constraints. This hit informal cross-border traders, that were no longer allowed to cross the border with their goods. And, the financial sector volume of transactions reflected the overall decline in economic activity. The authorities took measures to alleviate the impacts of the crisis on services as part of broader support to the private sector. The Government and the Central Bank provided support tht incuded: (i) postponement of tax payments and compensation of private sector tax credits; (ii) reduction of electricity tariffs for businesses; (iii) reduction in monetary policy rates and Central Bank’s reserve ratio; (iv) credit lines at discounted interest rates; and (v) credit lines to finance transactions linked to the imports of essential consumer goods. The resources set aside for private sector support amounted to about 1.5 percent of GDP, with available budget reaching 0.7 percent of GDP as of November 2020. However, a small proportion of this has been disbursed so far. The effectiveness of the measures was limited— either being of insufficient scope or else hindered by procedural bottlenecks. The measures implemented so far did not fully target the services activities most affected by the pandemic, such as hospitality and tourism, transport, informal micro firms or household enterprises. Incentives, cash injections and tax relief were not tailored to the characteristics of firms in these industries. Support to the private sector needed to be more rapid, transparent and time bound to address immediate liquidity challenges, prevent widespread layoffs and limit firm bankruptcies effectively. Also, support should be provided to informal self-employed entrepreneurs and household enterprises through expansion of social protection programs. Looking ahead, measures to support viable firms need to be strengthened in the short-term while, as the crisis receded, focus should be on upgrading the services sector. Special focus should be on supporting firms that were most affected by the crisis and were viable before the crisis. Measures that risk propping up unviable firms should be avoided. Once the worst of the crisis is over, attention should turn to productivity, economic transformation and job creation. COVID-19 have exposed the vulnerability of a services sector dominated by small firms in commerce and retail sector. Digital transformation can offer a path to support this agenda. Digital technologies have proved to be game changers in times of crisis, including via sustaining education efforts. . 72 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.1. Services Growth Has Outstripped Manufacturing, But is Now Slowing a. Annual growth in manufacturing and services b. Growth in output per worker by sector 40 3.5 3.0 30 2.5 20 2.0 10 1.5 1.0 0 0.5 -10 0.0 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Agriculture Industry Services etc. Intersectorial -20 -0.5 Reallocation e ect Manufacturing, value added (annual % growth) 1997-2001 2002-2006 2007-2011 2012-2016 Services, value added (annual % growth) Source: WBG sta calculations based on data from WDI. Source: WBG sta calculations based on data from WDI. Note: Industry includes manufacturing. > > > Figure 4.2. Services Make Up a Large Share of Output, Led By Private Sector Services a. Services as a share of GDP in Mozambique, b. Sectoral contribution to growth by decade, LIC and SSA, 1994-2018 1991-2018 55 9.0% 8.0% 50 7.0% 6.0% 45 5.0% 4.0% 40 3.0% 35 2.0% 1.0% 30 0.0% 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Avg.1991-1999 Avg.2000-2009 Avg.2010-2018 Public Services Private Services Manufacturing Mozambique Low income Agriculture, animal production Extractive Industries Sub-Saharan Africa (excluding high income) hunting, fishing and silviculture Source: WB sta calculations based on data from Mozambique National Institute of Statistics (INE) and World Development Indicators (WDI) for sub-Saharan Africa (SSA) data. Note: Public services comprise education, health, social security and other activities of collective services, social and personal. Private services comprise electricity and gas distribution, water, construction, automobiles sale and repair, transport, hotels and restaurants, information and communication, financial activities, and real estate. Services firms have been highly geographically and Zambezia, while the five northern-most provinces concentrated in and around Maputo while the north has contain only 27 percent of all services firms. been lagging (Figure 4.3b). According to the 2014-15 firm census, 32 percent of services firms are in Maputo province The contribution of services to total employment is and Maputo city. Growth is also fastest in Maputo Cidade, increasing, though agriculture is still by far the largest where the number of services firms has increased almost source of jobs, employing 70 percent of the labor force six-fold since the 2002/03 census, while for the country as (Figure 4.4a). There has been some shift from agriculture to a whole, the number of firms increased by only 83 percent. services, with the industrial sector continuing to make up the Large firm concentrations can also be found in Gaza, Sofala smallest share at 7 percent for the last decade (2 percent MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 73 > > > Figure 4.3. Service Firms Account for the Largest Sectoral Share, and Are Concentrated Around Maputo a. Number of firms by sector, b. Number of firms by province, 2002-03 and 2014-15 2002-03 and 2014-15 50,000 60,000 700% 45,000 600% 40,000 50,000 500% 35,000 40,000 30,000 400% 25,000 30,000 300% 20,000 200% 15,000 20,000 100% 10,000 10,000 5,000 0% 0 0 -100% Agriculture, Extractive Manufacturing Private Cabo Delgado Gaza Inhambane Manica Maputo Maputo Cidade Nampul Niassa Sofala Tete Zambezia Mozambique animal Industries Services production, hunting, silviculture and fishing Total 2002-03 Total 2014-15 2002-03 2014-15 % Change between 2002-03 and 2014-15 Source: WB Sta calculations based on data from Mozambican Source: WB Sta calculations based on data from CEMPRE, Enterprise Census (CEMPRE), 2002-03 and 2014-15. 2002-03 and 2014-15. Note: The number of firms includes firms plus establishments. An establishment as an economic unit at a single physical location while a firm consists of one or more establishments. from 1991-1999). Over this period, services firms have been The services sector has seen the largest per-capita the largest source of employment growth (Figure 4.4b). productivity increases since 2002, followed by agriculture, Comparing employment of both full-time and temporary with industry coming last (Figure 4.4). However, the pattern workers in larger formal firms (5 or more employees) across is not consistent across the different subperiods. Instead, the World Bank’s Enterprise Surveys in 2006 and 2017,84 within-sector gains in agriculture have tended to decrease the growth significance of services jobs becomes even over time, whereas those in services have generally tended more striking. In 2006 more workers were employed in to increase to become the largest source of within-sector manufacturing than in services; by 2017 services firms had productivity gains in the last decade, outpacing equivalent more than twice as many temporary and full-time workers. gains in the other two major sectors. Per-capita productivity Growth was fastest among temporary workers, which were gains within industry were largest between 2002 and negligible in number in 2016. 2006, but have tended to decrease since. The gains from the intersectoral reallocation of labor are larger than Services’ Productivity Growth Has Started to any individual within-sector gains, and largely reflect the Decline movement of labor from agriculture to services (Figure 4.6). Over the last three decades, productivity growth has Looking within the services economy, productivity levels in explained the bulk of the growth in per capita value added some service sub-sectors are higher or similar to capital- in Mozambique (Figure 4.5). Within-sector productivity intensive manufacturing industries. There is substantial increases account for the largest part of the overall per- variation in the sales per worker in different sectors across capita productivity gains across the whole analysis period, manufacturing and services. Sales per employee in formal as well as in most sub-periods (except 1991-1996). The firms are lowest for hotels and restaurants and food next largest share is productivity increases stemming from manufacturing (Figure 4.729a). The category of “other static reallocation of labor across sectors.85 The influence services”, which includes motor vehicle repair, wholesale, of demographic change has been lower overall, but still transport, and ICT, had both the fastest growth in sales per particularly important in two sub-periods: 1991-1996 and worker from 2006-17 and the highest sales per employee 2012-2017. (Figure 4.729b). 84 World Bank (2006; 2017). Micro enterprises (fewer than 5 employees) are not included in this comparison because the sample frame for these in 2006 was different than in 2017. 85 Static gains (or losses) in productivity occur due to labor shifts from below- to above-average productivity level sectors (or vice versa) while dynamic gains (or losses) in productivity reflect to relocation of workers from below- to above-average productivity growth sectors (or vice versa). 74 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.4. Agriculture is the Largest Source of Jobs, Despite the Growth of Services a. Employment, average sectoral shares b. Full-time and temporary workers in manufacturing (1990s, 2000s and 2010s) and services firms in 2006 and 2017, and compound annual growth rate (CAGR; right axis) 100% 120,000 40% 35% 80% 100,000 30% 80,000 25% 60% 20% 60,000 40% 15% 40,000 10% 20% 5% 20,000 0% 0% 0 Full-time Temporary Full-time Temporary Avg.1991-1999 Avg.2000-2009 Avg.2010-2018 Manufacturing Services Agriculture Industry Services 2006 2017 CAGR (right axis) Source: WB sta calculations based on data from World Source: WB sta calculations based on WB Enterprise Survey Development Indicators (WDI). Mozambique 2007 and 2018 data. > > > Figure 4.5. Productivity Growth Explains the Bulk of Growth in Per Capita Value Added Shapley decomposition of per-capita value-added growth 8.0 Annual Change (percentage points) 6.0 Whith in-sector Productivity 4.0 Participation Rate Static Reallocation Total Period 2.0 Dynamic Reallocation Demographic Change 0.0 Employment Rate -0.2 1991-2017 1991-1996 1997-2001 2002-2006 2007-2011 2012-2017 y=4.23 % y=3.49 % y=5.57 % y=5.06 % y=3.68 % y=2.71 % Source: WBG sta calculations based on WDI data. However, productivity growth has been declining in rate of growth is clearly declining more quickly than for services faster than other sectors. Services are still the other sectors. Given the importance of services firms both largest sectoral driver of productivity growth, and unlike in providing jobs, and through their linkages to the rest of for manufacturing firms, sales per worker have been the non-services economy, it will be important to reverse increasing for most services sub-sectors. However, the this trend. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 75 > > > Figure 4.6. Services Have Seen the Largest Per-Capita Productivity Increases Shapley decomposition of per-capita value-added growth, productivity by major sector 8.0 Annual Change (percentage points) 6.0 Productivity change, industry sector Productivity change, Agricuture sector 4.0 Productivity change, Services sector Employment Rate Demographic Change 2.0 Productivity change, Intersectorial reallocation Participathic rate Total Period 0.0 -0.2 2002-2017 2002-2006 2007-2011 2012-2017 y=3.79 % y=5.06 % y=3.68 % y=2.71 % Source: WBG Sta Calculations based on WDI data. 4.2 Can Services Become a Driver of especially if retail is counted. Among formal firms, 74 per cent of workers are in retail and other services while about Economic Growth? one-quarter are in manufacturing (Table 4.2). As for the informal sector, most workers are in retail (70 per cent) with Services are a Significant Source of Jobs just 17% employed in manufacturing. Mozambique’s services firms have increasingly become Services sectors also provide a larger share of jobs to significant sources of jobs. Among services sub-sectors, the female workers than manufacturing (23.8 vs. 11.5 percent). largest share of jobs in both formal and informal firms is in In manufacturing, female workers account for 17 percent of retail (Figure 4.8). This is especially true for informal firms, jobs in food production, but only 1 percent in other sectors. where retail firms account for 70 percent of all jobs. Among Among services sectors, 49 percent of hotel and restaurant formal firms, retail (21 percent), construction (20 percent) workers are female, while this share is only approximately 10 and transport (14 percent) are the largest services sector percent for construction and other services. employers. The fastest growth in number of firms between 2002/03 and 2014/15 was in more complex services sectors, Services Play a Vital Supporting Role Across such as ICT (109 percent increase per annum), real estate the Economy (66 percent) and professional services (71 percent), followed by construction (56 percent) and transport (30 percent). Since services are essential inputs into most products, The impacts of this employment growth was mixed. productivity growth in services supports growth in the rest Between 2007 and 2018, annual growth in total employees of the economy. The services sector is intrinsically linked (both permanent and temporary) in services sector firms to the overall activity of an economy through value chains. (construction, hotel and restaurants and other services) This includes both forward linkages (the contribution of a with more than five employees exceeded all manufacturing particular sector as an input to other sectors’ exports) and categories. For “other services” (i.e. services of motor backward linkages (the contribution of all other sectors to vehicles, wholesale, transport, and IT) employment grew by a particular sector’s exports).86 These value chain linkages 68 percent a year. not only capture the full (direct and indirect) contribution of the services sector to a country’s exports, but also inform a As a result, there are now far more services workers in both comprehensive strategy to improve export competitiveness formal and informal enterprises than in manufacturing, of both goods and services. 86 Backward linkage refers to the interconnectedness of a particular industry with the “upstream” sectors from which it purchases its intermediate inputs directly and/or indirectly. For instance, services (such as ICT, transport and professional services) are an important source of value added in the manufacturing and agriculture sectors. Forward linkage is the interdependency of an industry with “downstream” sectors, to which it sells its intermediate inputs directly and/or indirectly. For instance, services sectors (including transport) use manufacturing output as inputs. 76 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.7. Productivity Levels in Some Service Sub-Sectors Compare Well with Manufacturing a. Sales per employee in USD (for formal firms) b. Annual change of mean sales per worker (2006-17) 30,000 20% 15% 23,870 25,000 21,945 10% 20,000 5% 0% 15,000 -5% 10,480 9,641 10,000 7,606 -10% 6,945 -15% 5,000 -20% 0 -25% Food Ohter Metal Hotel and Construction Other Ohter Food Metal Construction Hotel and Other Manufacturing Restaurants services Manufacturing Restaurants services Manufacturing Services Manufacturing Services Source: WB Enterprise Survey Mozambique 2007 and 2018. Note: Metal includes basic metals, machinery and equipment, electronics. Other manufacturing includes tobacco products, textiles, leather, garments, wood, paper, publishing, printing and recorded media, refined petroleum products, chemicals, rubber and plastics, non/metallic mineral products, fabricated metal products, transport machines, furniture. Other services include services of motor vehicles, wholesale, transport, and IT. Sales in constant 2011 international U$S adjusted by PPP 2011 and GDP deflator. Growth is calculated using the sales reported by each firm for the ES year and some years before (2007 ES asks for 2003 sales and 2018 ES asks for 2015 sales). It is computed as the average over all firms of the following variable (Sales ES i,t - SalesES i,t-1)/ SalesES i,t-1 where i corresponds to a firm. Growth of mean sales is calculated as: (MeanSales 2017 - MeanSales2006)/MeanSales2006. An analysis based on multipliers from Mozambique’s the highest multiplier besides “other services” (16.28) is for social accounting matrix (SAM) (van Seventer 2019) shows wholesale and retail trade (4.21). that Mozambique’s services sectors have substantial backward and forward linkages within the economy. This Sectors vary considerably in their backward and forward analysis shows that for most indicators of interest – output, linkage multipliers. In terms of output multipliers, backward employment, GDP – services have a higher multiplier for linkages are most significant for finance and insurance backward and forward linkages (Table 4.1) than manufacturing and accommodation and food, while forward linkages (Annex Table 1). Specifically, the weighted average of the are greatest for ICT and water supply/sewage. For rural services sector is slightly less than manufacturing for output employment, most backward linkages are from “other (1.46 vs 1.45), slightly more for GDP (0.76 vs. 0.65), less for services and accommodation food” – for forward linkages rural employment (1.80 vs 1.49), but significantly more for this is through “other services” and wholesale and retail. urban employment (4.36 vs. 1.57). The services sector creates Finally, for GDP the biggest impact is from real estate through considerably more jobs in cities than manufacturing, but backward linkages, and from transport and storage through fewer in rural areas. Interestingly, across all these indicators, forward linkages. the impact from services output, employment and GDP multipliers is larger via forward linkages (downstream impacts Trade and Investment are Key Growth Areas in subsequent stages of production), than for backward for Services linkages (impacts on downstream production), linking to the previous discussion about the large impact of backbone The expansion of its services trade is a major opportunity services on the economy. for Mozambique to diversify exports, improve the overall competitiveness of the economy and promote exclusive Services sub-sectors vary in their multiplier effects on the growth. The services sector—from telecoms, to banking, IT economy (Table 4.1). In terms of output multipliers, finance and business processing outsourcing (BPO)—is increasingly and insurance services have a significantly higher multiplier recognized as part and parcel of any trade strategy, both as (1.88) than real estate (1.10). In terms of GDP, higher-than- a source of export diversification in its own right as well as average sub-services multipliers are recorded for real estate, an input to manufacturing and other services production. As health and education. Among the fastest-growing services a result of technological changes, trade in services is now sectors, such as ICT, real estate and professional services, more possible than before and has become an option for multipliers tend to be relatively low for most variables. The export diversification. Services currently contribute to about highest rural employment multiplier (besides “other services”) half of global GDP. The WTO’s 2019 World Trade Report is for accommodation and food (1.24) and wholesale and predicts that global services trade will grow by 50 percent retail trade (1.17). For urban employment, job multipliers are in the next 20 years, and that the share of developing on the whole higher (4.36 vs. 1.49 for rural employment) and economies in global services trade will grow by 15 percent.87 87 World Trade Organization (2019). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 77 > > > Figure 4.8. The Largest Share of Services Jobs is in Retail Composition of services jobs (2018) 80% 70% Formal Firms Informal Firms 60% 50% 40% 30% 20% 10% 0% Construction Services of motor vehicles Wholesale Hotel and restaurants Transport Professional Services Other personal service activities Repair of computers, personal and house Other services Retail IT Source: Mozambique Enterprise Survey (World Bank 2018). Table 4.1. Services Have Better Links to the Economy than Manufacturing Services sector: combined backward and forward economy-wide SAM multipliers for metical 1 million increase in exogenous demand Sub-sector Output Rural employment Urban employment GDP Comb. Bwd Fwd Comb. Bwd Fwd Comb. Bwd Fwd Comb. Bwd Fwd Overall 1.45 1.40 1.50 1.49 1.46 1.53 4.36 4.34 4.38 0.76 0.74 0.78 Accommodation and food 1.44 1.64 1.25 1.24 1.98 0.51 2.23 2.66 1.80 0.58 0.70 0.47 Business services 1.42 1.31 1.53 0.33 0.14 0.52 2.81 2.52 3.10 0.87 0.84 0.90 Construction 1.34 1.60 1.07 0.57 0.72 0.42 1.25 1.65 0.84 0.43 0.56 0.29 Education 1.08 1.16 1.00 0.88 0.93 0.84 1.58 1.69 1.47 0.85 0.89 0.81 Electricity, gas and steam 1.53 1.26 1.81 0.42 0.31 0.53 1.42 1.30 1.54 0.94 0.85 1.03 Finance and insurance 1.88 1.67 2.08 0.56 0.35 0.77 1.94 1.74 2.13 0.61 0.52 0.69 Health and social work 1.13 1.27 1.00 0.62 0.69 0.56 1.76 1.94 1.57 0.75 0.81 0.68 Information and comm. 1.74 1.41 2.08 0.58 0.35 0.82 2.70 2.05 3.36 0.81 0.68 0.95 Other services 1.68 1.48 1.88 4.51 4.40 4.61 16.28 16.03 16.52 0.73 0.65 0.80 Public administration 1.27 1.49 1.05 0.35 0.48 0.22 1.32 1.64 0.99 0.52 0.63 0.41 Real estate activities 1.10 1.09 1.11 0.07 0.07 0.07 0.18 0.15 0.22 0.98 0.97 0.99 Transportation/storage 1.66 1.25 2.07 0.66 0.32 1.01 1.89 1.22 2.56 0.89 0.71 1.07 Water supply and sewage 1.90 1.60 2.19 0.64 0.50 0.79 1.78 1.59 1.97 0.75 0.62 0.88 Wholesale and retail 1.51 1.53 1.48 1.17 1.10 1.24 4.21 4.30 4.12 0.65 0.68 0.62 Source: van Seventer (2019). Note: SAM: Social accounting matrix. Averages for services and manufacturing output are weighted by output. Average employment multipliers are weighted by their respective employment shares in services employment. Average GDP multipliers are weighted by the shares in services GDP. All multipliers indicators exclude the household income-expenditure loop. 78 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.9. Mozambique’s Low Share of Commercial Services Exports Reflects its Income Level Mozambique and peers’ commercial services exports (% GDP) vs. per capita GDP, 2005-07 and 2015-17 Avg. 2005-2007 Avg. 2015-2017 80 80 Commercial Services Exports to GDP (%) Commercial Services Exports to GDP (%) 60 60 40 40 MUS MUS 20 20 MDG MDG GHA BWA TZA ETH BWA MOZ ZAF MOZ ETH AGO MWI 0 NGA 0 COD NGA AGO 4 6 8 10 12 6 7 8 9 10 11 Log of GDP per capita (Current USD) Log of GDP per capita (Current USD) Source: WB Sta calculations based on data from WDI. Currently in Mozambique, services are mainly inputs into exports are not only very small relative to the country’s overall other economic activities and into goods exports, rather exports, but are also dominated by traditional services (travel than an export activity in their own right. However, services and transport, Figure 4.10), whereas the exports of modern can become export activities in their own right and are already services (including ICT and professional services) are below playing a growing role in trade. Trade in services encompasses what would be expected for the country’s income level. This four modes of supply: mode 1, cross-border trade in services, distribution is also unusual relative to peers: only Ethiopia has is analogous to trade in goods and involves delivering services a larger share of transport services exports in its total services from one country to another; mode 2, consumption abroad, exports (Figure 4.10). Other business services, which used refers to consumers (e.g., tourists or students) traveling across to make up a much larger share Mozambique’s commercial borders; mode 3, commercial presence, occurs through the services, have declined by half. Its exports of services are establishment of a commercial presence by the producer dominated by traditional services to a larger extent than (e.g., a subsidiary or branch of a bank) in the country of the regional and income group averages – this is also the case consumer; and mode 4, movement of natural persons, takes for most of its peers (see Chapter 1). Whereas Mozambique place when the producer (e.g., a mining engineer) travels exported a larger share of services than the sub-Saharan across borders. The share of services trade in GDP increased Africa (SSA) average in 2007, by 2017 it had fallen behind it. from 11.6 percent in 2006 to almost 35 percent in 2018, while services exports more than doubled over that time Investment is closely linked to trade in services and can be period. The share of commercial services exports in GDP is an important channel for technology transfer. FDI does not low in Mozambique, but around the level expected given the automatically improve firm outcomes, but it can generally country’s income level (Figure 4.9). contribute to technology transfer, boost economic growth, reduce poverty and help countries integrate into global value Moreover, services firms are close to surpassing chains (GVCs) (Alfaro et al., 2004; Lall, 2002). Mozambique manufacturing firms in terms of their likelihood of stands out as one of the largest recipients of FDI inflows in exporting. According to the 2018 Enterprise Survey, Africa, particularly in the extractives and transport sectors. approximately 16.3 percent of services firms export, only This has been the main factor in maintaining double-digit slightly lower than for manufacturing firms (18.6 percent). output growth. However, FDI in services has been increasing There is considerable variation across services sectors, dramatically in recent years. According to Enterprise Survey ranging from 9 percent of formal enterprises in construction, data, 25% of services firms are owned totally or partially to 14 percent in hospitality and 21 percent in other services. by private foreign individuals, companies or organizations, This represents substantial growth from the previous survey compared to 17% of manufacturing firms; 11% are entirely (2007), when not a single sampled firm exported (though foreign owned (5% for manufacturing). This is also reflected among a much smaller sample frame). in FDI project data. Since 2015, the vast majority of FDI projects listed on the Financial Times fdi Markets database88 Mozambican commercial services exports represent a small have been in services, though total investment volumes have share of the country’s total exports. Commercial services been higher in the extractives sector. 88 https://www.fdimarkets.com/ MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 79 > > > Figure 4.10. Mozambique’s Services Exports are Dominated by Traditional Sectors Composition of services exports, 2005-2017 100% Personal, cultural, 90% and recreational services Other business services 80% Telecommunications, 70% computer, and information services 60% Charges for the use of 50% intellectual property n.i.e. Financial services 40% Insurance and 30% pension services Construction 20% Travel 10% Transport 0% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Source: WB Sta calculations based on data from UNCTAD. Note: n.i.e stands for not included elsewhere. > > > Figure 4.11. Most of Mozambique’s Peers Have a Lower Share of Traditional Services Exports Mozambique and peers’ composition of services exports, 2017 100% Personal, cultural, 90% and recreational services 80% Other business services 70% Telecommunications, computer, and 60% information services 50% Charges for the use 40% of intellectualproperty n.i.e. 30% Financial services 20% Insurance and pension services 10% Construction 0% Travel Mozambique Angola Botswana Burkina Faso Côte d'Ivoire Dem. Rep. of the… Ethiopia Ghana Kenya Madagascar Malawi Mauritius Namibia Nigeria Rwanda Senegal South Africa Uganda United Republic of… Zambia Zimbabwe Low-income… Sub-Saharan Africa Transport Goods-related services Source: WB Sta calculations based on data from UNCTAD. This has important policy implications. First, low-cost, high- Weak Innovation and Technology are Holding quality services should be a priority for any diversification Back Services Growth and competitiveness strategy in Mozambique. Second, this means that assessing and understanding constraints The government has made efforts to improve the country’s in services development are critical to their contribution connectivity through infrastructure and regulatory reform. to economic diversification and competitiveness in goods In 2003, the mobile market opened to competition, and industries. These issues are further explored in Section 4.3. there are currently three operators in Mozambique (ITU, 80 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.12. Mozambique is Among the Countries with the Most People O ine 0 -25 25-50 50-75 75-100 Data not available Source: (ITU, 2019). 2018). Following a historical lack of global connectivity, among the countries with the most people offline ( Figure Mozambique connected to the SEACOM undersea cable 4.11). It is slightly below the regional average in internet in 2009 and the Eastern Africa Submarine System in 2010. use but far below the global average of 53 percent of Although the Telecommunications Law of 2016, along with individuals online. In terms of broadband communications, the 2018 regulations, call for infrastructure sharing, this has Mozambique is among the laggards in the region, with not yet been enforced. There is currently duplication of only 0.07 broadband subscriptions per 100 people, behind backbone networks in certain parts of the country, while Zambia (0.1), Tanzania (0.2), Zimbabwe (1.0) and South Africa others are left unserved. If adequately implemented, this (5.2) (ITU, 2019). While there has been an increase in the use could improve affordability and, consequently, connectivity of telecoms by individuals in recent years, these figures speak across Mozambique (see Section 4.3 for more details). of an extremely challenging environment for IT services to build on. Despite these advances, digital connectivity in Mozambique remains poor and expensive. Mozambique ranks among Compared to manufacturing firms, services firms innovate the lowest countries in the region in the World Economic less, though the gap is relatively small. Approximately one- Forum’s (WEF) Networked Readiness Index, which measures third of services firms have innovated in new products in the a country’s use of ICT to increase competitiveness and well- last three years (32% compared to 37% for manufacturing) being. The WEF ranked Mozambique 112th in the world for and the share firms spend on research and development the extent to which the national legal framework enables is almost as small as manufacturing firms (8% compared ICT penetration and a safe development of business activity. to 10%). However, in terms of connectivity, services firms generally perform better than manufacturing firms with Mozambique’s telecommunications sector is below the 82% using a computer (58% for manufacturing), 73% using global and regional averages in the World Bank’s Investing the internet for business purposes (47% for manufacturing) Across Borders indicators.89 With only one-fifth of the and 44% having their own website (26% for manufacturing) population currently using the internet, Mozambique is (Table 4.2). 89 http://iab.worldbank.org/ MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 81 Table 4.2. Services Firms are Less Innovative than Manufacturing Firms, But More Connected % firms that innovate in … Manufacturing Services New products or services during the last three years 37% 32% New products or services for the main market during the last three years 21% 17% New or improved process during the last three years 16% 13% Firm spends on research and development activities 10% 8% % firms that use … Computer 58% 82% Internet for business purposes 47% 73% Firm has its own website 26% 44% Source: WB Enterprise Survey Mozambique 2018. Table 4.3. Services Firm Managers are More Educated than Most Manufacturing Managers Formal firms (with 5 or more employees) University degree Secondary school or above 2007 2018 2007 2018 Manufacturing Food 12% 34% 84% 75% Metal 55% 74% 100% 100% Other manufacturing 11% 35% 61% 73% Services Hotel and restaurants 14% 51% 100% 92% Construction 42% 65% 100% 100% Other services 0% 60% 100% 97% Source: World Bank Enterprise Surveys. Mozambique’s weak education system is an important the share of services firm managers who have completed obstacle to skills accumulation and competitiveness. secondary school, but this remains above 90% across sub- Despite an upward trend in recent decades, Mozambique’s sectors. Human Development Index (HDI) has been the lowest among neighboring countries for the past 25 years (UNDP, The quality of management practices varies among 2017). The lack of skills can be seen at all education levels sectors. Management quality is measured in the World Bank and across all sectors (UNCTAD, 2015). According to the Enterprise Surveys by four aspects: (i) addressing process World Bank’s Knowledge Economy Index (KEI), which problems such as machinery break-down, human errors, measures a country’s readiness to compete in the knowledge or failures in communication; (ii) monitoring performance economy, Mozambique ranked 90 out of 144 countries in indicators; (iii) setting and being aware of targets in service 2012. Mozambique’s KEI has dropped by almost 30 percent provision; and (iv) providing incentives such as performance since 1995, consistently lagging behind all of its neighboring bonuses. Overall the construction sector performs best, with countries (World Bank, 2012). managers performing more than one standard deviation above the mean (Figure 4.12). Management practices are Management Skills and Practices are worst in the hotels and restaurants sector. Improving 4.3 Strengthening the Regulatory The quality of management skills and practices is an Environment for Services is Key important determinant of firm productivity. Formal services firms with more than five employees have seen a significant An open trade and investment regime, supported by a increase in the number of managers who have completed transparent and effective regulatory framework, is key to university. Between 2007 and 2018, this increased from 14% developing the services sector. Creating such a framework to 51% for hospitality, 42% to 65% for construction and from requires a comprehensive understanding of the current laws, 0% to 60% for other services (Table 4.3. Services F). This is regulations, and practices affecting trade and investment in in stark contrast to manufacturing sector managers: only services. Mozambique’s legal framework for services trade 34% and 35% of managers in food production and other and investment is largely open and non-discriminatory. The manufacturing have university degrees (though this is much general regulatory framework allows for the establishment higher for metals, at 74%). There has been a slight decline in and operation of foreign services suppliers under equal 82 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.13. Management Quality is Lowest in the Hospitality Sector Organizational and managerial z-scores by sub-sector (2018) 2.0 Operations Monitoring Targets Incentives 1.5 1.0 0.5 Other manufacturing Other services 0.0 Food Metal Construction -0.5 -1.0 Hotel and restaurants -1.5 Source: WB Enterprise Survey (World Bank, 2018). conditions as Mozambican national and firms, in terms of Blanket Regulations Constrain Growth equity participation, licensing conditions, taxation, as well as general operating conditions. In addition, a specific regime The regulatory framework for services includes specific for investment projects, designed largely to benefit large laws and regulations and those that affect a wide range foreign investments but open to any kind of investment, of sectors (“horizontal” measures).90 Some horizontal laws provides additional regulatory and fiscal benefits. and regulations, in particular for the labor market, land use, and cross-border services, contain substantial limitations to The main weaknesses stem from poor implementation business operations. These administrative and regulatory of existing rules rather than restrictive regulations per barriers raise firms’ operating costs in Mozambique, limiting se. In recent years, the GoM has taken important steps to the benefits of market opening, especially for SMEs. Besides, improve the regulatory framework for the services sector by some sectors, such as telecommunications, do not have an reducing barriers to establishment and facilitating transfers adequate regulatory framework, giving rise to an informal of funds. However, applying these laws and regulations sector with low-quality services. remains discretionary and inconsistent, curbing the benefits of the regulatory reform. Some key laws also still pose Examples of horizontal measures that restrict services supply major restrictions on the services sector. The labor quota include the following, each discussed below: regime is an important restriction to growth by foreign firms, especially SMEs. Additionally, land access and use rights are • Cross-border services: measures on the transfer of complicated to navigate, often leaving investors with weak funds, restrictions on access to foreign currency and uncertain contractual rights over the real estate they • Consumption abroad: exit visas for nationals, restrictions control. Lack of clear guidance in the administration of laws on access to foreign currency and regulations contributes to a disconnect between the • Establishment and licensing: a framework for law and practice, leading to lengthy and costly processes establishing both foreign services firms as well as and inconsistent results that reduce certainty for services domestic companies suppliers. Lack of co-ordination between government • Employment of foreigners: domestic employment agencies also adds to inconsistent application of regulations, requirements for foreign companies as different agencies may interpret regulations in different • Land use: requirements for foreign businesses that want ways. This general weak regulatory guidance often leads to acquire land. to contradictory information from different agents within one agency and between the different agencies regarding Cross-border services requirements, raising the costs for regulatory compliance for services firms, especially SMEs. The regulation of cross-border services is inherently challenging due to the technical difficulties of controlling 90 Telecom regulations, for example, obviously affect services providers in the subsector, but so do other laws and regulations, such as regulations on buying and selling foreign currency, laws on entry and stay of foreigners, and procedures related to the establishment of firms. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 83 electronic traffic and the location of the services supplier, by in commercial banks. The Foreign Exchange Control definition, outside the country’s jurisdiction. It is, therefore, regulations provide the conditions under which foreign rare to find restrictions on cross-border trade that affect all firms operating in Mozambique may transfer funds: (i) the services sectors. investment project is registered with the Central Bank; (ii) all the funds are registered with the Central Bank; and Mozambique’s local establishment requirement is (iii) all tax obligations in relation to the investment project outdated, restrictive, and not suitable for the digital era. are fulfilled. Practice suggests that the first requirement is The Commercial Code requires that any service supplier traditionally met by submitting proof of an APIEX Investment who wishes to perform activities for more than one year in Authorization (see below). However, under the Investment Mozambique must open an establishment in the country and Law, obtaining an investment project authorization is only a appoint a resident in Mozambique with powers to receive voluntary procedure meant to unlock additional benefits – communication and notifications. Such local presence including the repatriation of invested capital – which means requirements represent a de facto prohibition of the provision that foreign-invested firms may or may not have pursued this of cross-border services. While they may play a role in tightly (see section below for further details). Yet the regulations regulated sectors, such as financial or professional services, of the Central Bank de facto require an investment license they are at odds with the expansion of the global services to register as a foreign firm and be able to enjoy any type trade. Technology increasingly allows for new and more of international transfer – not only repatriation of capital services to be performed remotely through digital means, as suggested in the Investment Law. As a result, firms that including a range of business services that typically offer few may have not opted for an investment authorization (or regulatory concerns (e.g. business outsourcing services) or whose investment authorization has expired) are forced to digital platform services that bring parties together, but do apply for this authorization only for the purpose of making not provide the services or good themselves (e.g. ride-hailing payments abroad, despite the flexibilities provided in the services; hotel booking services, etc.). Foreign Exchange Decree of 2017. However, while the measure remains valid, it appears The 2017 decree allowed for commercial banks, in addition that it is not actively enforced. If it were, Mozambican to the Central Bank, to process requests for registration citizens would not be able to use Google’s search engine and to process international current transfers. However, unless this company had a local presence in Mozambique, this has not been an immediate success. Stakeholders and similarly for all other online platforms available in the note that clerks in commercial banks are not familiar country. The GoM should consider amending Article 85 of with the regulations and requirements for registration the Commercial Code to allow for non-established services and transfers. Further, commercial banks may require suppliers in the country, or limiting the scope to the local additional documentation from applicants, which are not establishment requirement to specific services activities of required in the regulation or at the Central Bank, thus often particular regulatory sensitivity. unnecessarily impairing transfers. Ultimately, commercial banks defer applicants to the Central Bank, hence nullifying Transfer of Funds the advances of the decree. In addition to streamlining the registration process, as suggested above, the Central Bank Transferring money out of Mozambique remains a could consider disseminating clear guidelines and protocols, burdensome procedure that raises substantial operating and implementing brief training programs for commercial costs for firms. This is often singled out as a key regulatory banks to process registrations and payment requests to challenge by services firms, especially SMEs. The GoM has ensure compliance with the provisions under the Investment engaged in several reforms in recent years to improve these Law and the Decree of Foreign Exchange of 2017. processes. The Foreign Exchange Decree of 2017 allowed for international transfers related to current payments, and Establishment and licensing it generally facilitated the regime by introducing electronic registration for foreign exchange operations through Mozambique’s legal framework for the establishment of the commercial banks and eliminating the requirement foreign services firms, as well as domestic companies, is to convert half of export proceeds into local currency. largely open and non-discriminatory. Mozambique does However, weak implementation within the Central Bank not impose limitations on foreign establishment except and poor guidance to commercial banks mean that much in a few, selected, services (Table 4.4). The Commercial of the advances of the Foreign Exchange Decree are yet to Code does not distinguish based on nationality: a business materialize in practice. wishing to operate in Mozambique must publish its articles of association in the government gazette, register in the Two procedures relating to the transfer of funds appear Commercial Register and with the local tax authorities, and to limit the effective implementation of the Foreign obtain an operating license from the Ministry of Commerce. Exchange Decree: (i) registration of foreign investors with The Law on Investment guarantees equal treatment of all the Central Bank and (ii) registration and foreign transfers investors, including international investors. 84 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 4.4. Some Services Impose Limitations on Foreign Establishment Sector Limitation Real estate Nationalized real estate may not be transferred to foreign individuals or companies where the share capital is majority owned by foreign capital Public construction To obtain a permit for public construction, construction companies’ capital must be majority owned by nationals (individuals or companies) Tourism Travel agencies can only be owned by national shareholders Gambling Foreign entities may not hold more than 49 percent of the share capital in casino activities Private security Private security companies can only be owned by nationals and the share capital must be majority owned by nationals. However, this was ruled unconstitutional in 2008 and thus today it is not binding In recent years, the GoM has introduced reforms to make authorization, as it may grant them additional flexibility in it easier and faster to establish foreign firms. In 2017, it employing foreigners. In addition, foreign firms with a valid merged the agency in charge of approving investment investment authorization benefit from certain options in the projects and the agency responsible for special economic use of land that are otherwise closed to foreign-owned firms. zones and industrial free trade zones into one single entity, This means that, in practice, any foreign investment firm, the Agency of Investment and Export Promotion (APIEX). even SMEs, considers an investment authorization essential Even so, foreign firms wishing to become established for operating in Mozambique. in Mozambique face a strategic, if unfortunate, choice: either incur additional regulatory compliance costs during Employment of foreigners establishment by obtaining an investment authorization, or avoid seemingly unnecessary requirements and face Labor regulations are one of the main regulatory restrictions hefty regulatory compliance costs during the operation to the services sector, in the form of a strict labor quota of the business. This choice is reflected in two alternative system for foreign employees. The Labor Law of 2007 and regimes for establishment: the general framework, and the Foreign Employees Contracting Regulation provide the the “investment project authorization” channel provided by framework for employing foreigners, setting out specific APIEX. quotas for foreign employees, based on the company’s total workforce (Table 4.5.). There are some exceptions to The “general framework” involves following the standard these quotas, including for companies operating in Special rules for establishment and licensing and being subject Economic Zones and companies with an investment to the standard regulatory framework. This route entails authorization. To employ foreigners beyond the quota following the regulatory procedures for business registration allowance, the applicant must face a labor market test, and and then obtaining the specific business license at the show that the foreign candidate has the necessary academic relevant ministry or agency. Except for a few activities that or professional qualifications and that no Mozambican limit foreign participation, such as public construction and citizens have such qualifications. Services depend greatly on cultural services, establishment is open to both domestic specialized expertise, in particular high-value-added activities and foreign firms. The advantage of this channel is that it such as professional services and business services. These does not involve any additional procedures or requirements, are also typically provided by relatively small firms. As a result, thus reducing steps and costs. On the other hand, firms this restrictive and inefficient labor regime poses a particular under this regime are not eligible to obtain any flexibilities challenge to investment and growth in the services sector. to the general rules and regulations, including those on taxes, customs duties and procedures, transfers of fund, and The system is rife with discretionary decisions and poor employment of foreigners. administrative practices that weigh heavily on foreign firms, especially in the services sector. In practice, the The APIEX investment authorization91 is a de facto employment of foreigners is severely restricted even within requirement for any firm that wishes to transfer funds the quota, with firms required to submit an application to the abroad – not just to repatriate capital, but also to make Ministry of Education to recognise the person’s education current payments above the retail threshold. Firms that and expertise. This procedure often entails requesting and may need to employ foreign expertise beyond the standard translating foreign education certification, which adds time quota (see next section) also see a benefit in the investment and costs. Practitioners note that this step alone can take 91 The APIEX investment authorization channel is meant to provide exceptions to the main regulatory restrictions featured in the general framework – especially those relating to labor, taxes, and transfer of funds. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 85 Table 4.5. Companies are Subject to Labor Quotas when Hiring Foreign Employees Company size Foreign labor quota Small (10 employees or fewer) 10% Medium (11-99 workers) 8% Large (100 employees or more) 5% up to three months, severely delaying the hiring process beneficial to the growth of both foreign and domestic firms, and often missing out on qualified candidates. An additional especially in the services sector where skills and expertise constraint is the process for legalizing foreign documents, are fundamental. Alternatively, the Ministry of Labor could including diplomas and theses. In countries where the local apply beyond-quota labor market tests according to the consulate does not provide these services, or where there law, not as a strict prohibition to foreign employment. In the is no Mozambican consulate, foreigners must travel to the meantime, quota percentages could be applied as a share of nearest consulate that can assist with this. effective number of employees, instead of basing them on categories defined by the size of the company. Beyond the allowed quota, employment is subject to an increasingly burdensome labor market test. Labor market Land use tests, if not adequately and transparently regulated, often result in a highly discretionary process in which the regulator The regime for land use is another regulatory limitation has few strict guidelines for when to allow or reject an to the operation of foreign firms, including in the services application. Practitioners point out that political pressures sector. According to the Constitution and the Land Law to employ Mozambican nationals have made it increasingly of 1997, all land in Mozambique is the property of the challenging to obtain permission for employing foreigners, government and it may not be sold, mortgaged, or otherwise even in areas where Mozambican expertise is not readily alienated. The state determines the rights and conditions to available. use and develop the land. The main land right is a Right of Use and Development of Land (DUAT), conferred by the Replacing the quota system and the labor market test with state for renewable periods of 50 years. Article 11 of the knowledge transfer schemes could allow for hiring more law declares that to obtain a DUAT foreigners must have an skilled foreigners where local capacity is scarce. The GoM investment authorization (see above) and comply with certain could foster the implementation of training and capacity- conditions: (i) foreign individuals must have resided for five or building programs as part of new investment projects, more years in Mozambique; and (ii) foreign companies must encouraging knowledge transfer from foreign skilled be incorporated or registered in Mozambique. Titleholders personnel to the domestic workforce. The training program of a DUAT may transfer infrastructure, structures, and land implemented by Mozal is often regarded as a success by the improvements with the proper authorization. When a Mozambican business community (Box 4.1). This would be building on urban land is transferred, the land accompanies Box 4.2. Labor Market Quotas: Approaches and Alternatives Developing economies that wish to protect their domestic workforce through a labor quota regime should make decisions based on periodic reviews of the domestic labor market. Public-private consultations, such as those carried out by the Dominican Republic’s Migration Council (OECD & ILO, 2018), can help set annual quotas based on the sector’s specific needs. In Ghana, on the other hand, a committee grants employment authorization based on the amount of investment by the prospective employer. This can help ensure the benefits of labor migration for the domestic economy. An alternative to a labor market quota system is a knowledge transfer program. In 1997, aluminum smelter company Mozal set up the first major foreign investment program in Mozambique. Its training and development initiatives include a graduate development program to expose graduates to its facilities and the heavy industry environment. Additionally, Mozal developed the Mozal Community Development Trust (MDCT) and the SME Empowerment Linkage Program (SMEELP), so that local SMEs and the local community could benefit from its resources. MDCT’s main objective was to carry out programs including small business development, education, and training (Mitsubishi Corporation, 2019). SMEELP helped SMEs seeking to bid for contracts for the Mozal construction phase by providing capacity building, training, and technical assistance (IFC, 2003). 86 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 4.14. Mozambique’s Governance Indicators have been Trending Downwards in Recent Years 0 -0.2 -0.4 -0.6 -0.8 -1 -1.2 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Government e ectiveness Regulatory quality Rule of law Source: (World Bank, 2018) it automatically. However, when a building on rural land is formulate and implement sound policies and regulations transferred, land does not go with it unless the authorities that permit and promote private sector development. It approve. declined slightly between 2015 and 2016, and has remained stagnant ever since (Figure 4.1332). Perceptions of the rule These burdensome requirements lead to investors seeking of la, related to adherence to rules on contract enforcement, costly alternatives in order to operate in Mozambique. property rights, the police, and the courts, as well as the For example, in the telecommunications sector, majority likelihood of crime and violence, improved over 2009- foreign-owned service suppliers tend to resort to creative 2010 but have declined since then. Perceptions of the solutions to obtain land, often under uncertain legal quality of public services, the civil service, and its degree of coverage (Box 4.2). This increases risks and operating costs, independence from political pressure in policy formulation, reducing the country’s competitiveness and attractiveness to as captured by government effectiveness measures, steadily foreign investors, especially in areas, like services, that are declined between 2013 and 2015, and remained stagnant the not directly dependent on Mozambique’s natural resources. past few years. These observations not only contrast with The GoM should adopt a legal regime for a land lease of the progress achieved in other areas, such as liberalization up to 99 years, applicable to foreign firms and citizens. As of investment and the simplification of business procedures, a temporary measure, the government could consider but also help explain the endurance of some regulatory extending the flexibilities on the use of urban land to rural challenges listed above, despite the reforms undertaken to areas. reduce regulatory restrictions. Poor Governance is Undermining Services Implementation of regulations is often discretionary because the lower levels of the administration lack Studies show that the quality of institutions matters in guidance. Laws and regulations are published in the official the design of policies for the services sector.92 Services gazette; some government agencies also publish their laws export growth is associated with a transparent and efficient and regulations on their websites in Portuguese. However, institutional environment.93 The quality of institutions plays lower regulatory instruments, such as decrees and ministerial a greater role for trade in services than for trade in goods.94 guidelines, such as those setting out licensing requirements The institutional capacity to design adequate policies, and criteria by each ministry, are often not available or not regulate different services sectors, and ensure compliance easily accessible. with laws and regulations is thus a key driver of success in the services sector. Policies and regulations targeted at the services sector in Mozambique follow the general patterns described above: A weak governance framework is hampering regulatory open and non-discriminatory regimes are hampered by reforms needed for services sector growth. Regulatory weak governance and poor implementation. This is specially quality is understood as the ability of the government to the case in key backbone services, such as telecoms and IT 92 Goswami, Mattoo, and Sáez (2012). 93 Grunfeld & Moxnes (2003). 94 WTO (2019). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 87 services, as well as transport and logistics. Regulation in these Telecom infrastructure remains weak in Mozambique, sectors has advanced in recent years towards creating open especially for business purposes. While the GoM has and competitive markets, and ensuring minimum quality tried to improve regulation of the telecoms sector, much standards. However, oversight and enforcement have been remains to be done. Mozambique’s National ICT Policy largely ineffective, resulting in market conditions that do and Implementation Strategy was among the first in Africa, not reflect the regulatory guidelines. This not only hampers and led to a new telecommunications law in 2016. This the expansion of these sectors by swelling investment and aims to revamp the institutional and legal framework for the operating costs, it also reduces economic competitiveness operation of telecom services. Recognizing the importance for trade in goods and other services by offering expensive of attracting investment in the sector, it allows domestic and and/or low-quality services, despite these being central to foreign companies to apply for telecom licenses. The law productivity across the board. The sections below describe also introduces regulations on regulatory fees, licensing, the bottlenecks in these backbone services sectors. and interconnection of telecommunications networks. Regulations aim to increase the standards and efficiency of Telecommunications and IT-enabled Services Need the services, in order to promote investments in the sector. to be Strengthened However, weaknesses in governance and regulation add hurdles to the expansion of telecom infrastructure (Box 4.2). While mobile connectivity has advanced substantially, telecom infrastructure and the IT services sector generally Trade in IT-enabled services (ITES)—such as back-office remain greatly underdeveloped in Mozambique. A strong services, IT maintenance, and software development— telecommunications infrastructure, including broadband generally benefits from open regulatory regimes in connectivity based on fiber optic network, is crucial to Mozambique. But as cross-border trade in ITES relies on the information technology (IT) sector. Manufacturing telecommunications platforms, it is difficult for regulators and primary sectors also depend on telecommunications, to monitor and therefore regulate it. The most important especially as modern production increasingly relies on barriers in this sector are thus found in supply-side constraints services often performed remotely (such as research and and limitations that affect the telecom infrastructure and development and ICT services), which requires sharing other services sectors that serve as input for ITES, such as heavy loads of digital information. These IT-enabled services accounting, legal, and financial services (Engman 2010). include traditional communication as well as digital services. Some constraints reflect the inherent difficulty of supplying For these IT-enabled services to grow, they must rely on a services remotely;A others are linked to the enabling solid telecommunications market overseen by a sound and infrastructure for ITES, shortages of human capital, or transparent regulatory framework. regulatory and institutional conditions. Box 4.3. Removing the Telecommunications Bottlenecks in Mozambique While the government has enacted regulations on infrastructure sharing, mobile telecom operators continue to deploy proprietary fiber optics networks and build proprietary antennas, thus multiplying the sunk investment costs. Operators have decided to incur these additional costs on the grounds that other telecom networks are not up to the standard required to ensure their services, or because they may not access other operators’ networks under fair conditions. This is despite ARECOM (Autoridade Reguladora das Comunicações – Mozambique’s telecom regulator) issuing rules regulating the sharing of infrastructure, as well as mandating quality standards in the construction of infrastructure. These regulations, however, appear to be poorly enforced, thus creating the conditions for the current multiplicity of network infrastructure. A review of the regulations on infrastructure building and sharing and the current practices for enforcement would be essential to foster the expansion and quality improvement of mobile telecom services. The current regulation for land use further increases costs of infrastructure deployment, as operators need to access land, including in rural areas, for the building of antennas. Although telecom operators are eligible for a special regime for acquiring land under the Telecommunications Law of 2016, this has not been implemented by secondary regulation. As a result, telecom firms end up falling under the DUAT regime (see above). Since the DUAT regime does not allow foreign firms to acquire land in rural areas, telecom operators, especially those controlled by foreign capitals, must make ad hoc arrangements with property owners to deploy infrastructure, increasing the risks of the investment. GoM should take steps to implement the special land regime set out in the Telecommunications Law in order facilitate the growth of the sector and to increase competitiveness in this key backbone service. 88 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 4.6. Mozambique: Certified Professionals are Falling Behind Neighboring Countries Country Certificates Sites CISPP members Botswana 2 4 8 Malawi NA 3 4 Mauritius 16 22 24 Mozambique NA 2 3 Rwanda 1 3 1 South Africa 53 134 483 Tanzania 3 7 8 Zambia NA 3 11 Sources: (ISO, 2019), ((ISC)², 2020). Note: CISSP: Certified Information Systems Security Professional. Some horizontal restrictions are particularly harmful to Index95 ranked Mozambique 136th out of 137 economies in the IT sector, especially restrictions on the employment 2018, highlighting access to finance as one of the country’s of foreign experts. IT services providers typically depend on most problematic factors. The IT sector is particularly experts who conduct discrete, highly specialized, often time- vulnerable, as IT companies generally have few physical sensitive tasks. In the absence of such skills in Mozambique, assets or facilities that can be used as collateral. In recent it is crucial for these companies to be able to import years there has been a surge in the digital entrepreneurial talent, which will in turn increase local capacity. An open ecosystem, including incubators and innovation spaces, environment for the mobility of individual suppliers is thus as well as competitions.96 This strengthens the capacity of key to the development of the sector. Yet Mozambique’s digital entrepreneurs in preparing business plans that can quota regime severely restricts the country’s ability to tap into help support financial loans. To complement this, regulations international expertise where needed, even though the legal on banking and other financial services should also provide and regulatory framework for IT services in Mozambique is for the ability to include intangible assets, such as intellectual largely open to competition, including foreign participation. property rights or contracts for services supply. The restrictive regime for employing foreign experts The GoM could implement further policies to enhance compounds the Mozambican workforce’s weak capacity, the digital entrepreneurship ecosystem, such as providing limiting the expansion of the IT sector. IT skills remain support through innovation hubs. It could promote scarce, with Mozambique falling behind its neighbors in the the establishment of export-oriented IT firms in Special number of software certifications (Table 4.6). For example, Economic Zones, where they have access to certain there are only three Certified Information Systems Security additional benefits, including duty-free imports and a more Professional (CISSP) credential holders in the country liberal quota for foreign employment. The policy agenda ((ISC)², 2020). CISSP is one of the most globally demanded should also include digital literacy at all education levels, certifications in the IT sector, granted to experienced to ensure the improvement of digital skills and access security professionals. Similarly, as of 2018, Mozambique to technology. Finally, the GoM must ensure capacity had no professionals with these certificates. The fact that building within the relevant government agencies, training there are currently no certification programs for IT skills government officials as well as connecting with technical means providers must travel abroad to obtain certifications, schools and engineering universities. and keeps the numbers of certified IT professionals low. The Mozambican Association of ICT companies (AMPETIC) has Transport and Logistics are Undermined by engaged in several activities to increase capacity, including Regulatory Constraints training sessions, workshops, and conferences (AMPETIC, 2019). A pilot incubator program for IT firms and a pilot The main challenges in the transport and logistics certification program could boost the exports of IT-enabled subsectors in Mozambique are also related to a weak services in Mozambique. governance framework. Transport is the second largest employer in the services sector and has the highest sales Access to finance is particularly challenging in the IT sector. per employee, making it a very important sector for The World Economic Forum’s Global Competitiveness Mozambique.97 Low regulatory capacity results in opaque 95 World Economic Forum (2018). 96 World Bank, Digital Economy Memo (2019). 97 World Bank (2018). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 89 regulatory procedures, increasing operating costs and de Institutional reforms were carried out in the early 2000s facto market access restrictions. Improving the governance to co-ordinate government efforts and develop sector of regulations of carriers, vehicles, drivers, service types, and policy. The development of customs and primary roads service quality is vital to ensure the efficient functioning of has improved the logistics sector. Thirty percent of the road commercial enterprises and attract foreign investment. network was built in the last decade. However, the quality of overall infrastructure has continued to decline over the past The Ministry of Transport and Communication is the main ten years, consistently lagging behind the regional average body to execute transport policies and communicate (Figure 4.15). In 2018, Mozambique was ranked 126th out of within public and private domains. Under the Transport in 137 countries for the quality of transport infrastructure. The Automobiles Regulation of 2019 there are no foreign equity country’s score is even lower for the quality of roads, where participation limits or a minimum share capital for foreign it ranked 129th. equity participation. Firms incorporated in Mozambique with foreign ownership and/or control are allowed to carry out The development of the transport sector focuses mainly the following activities: domestic and international transport; on the extractive industries, particularly the Nacala freight; goods transport; and rent services. Non-established Logistics Corridor project, while other networks are foreign services providers are free to solicit business from neglected.99 Vale Logistics Limited has invested over USD domestic customers on a cross-border basis and licensing 3 million since 2013 in infrastructure improvement along the requirements do not discriminate on the basis of nationality. corridor.100 Meanwhile, less than one-third of Mozambique’s However, several regulatory constraints, including roads are paved.101 Poor road connections, even within the checkpoints and customs regulations, limit the transit of Nacala Corridor, lead to accessibility problems for the rural goods.98 population and higher market access costs for producers. Furthermore, an assessment of the Nacala Corridor carried Poor governance coupled with a lack of logistics out by USAID highlights the poor implementation of existing infrastructure result in low performance of the sector regulations, and informal fees. (Figure 4.14). There are significant differences between regulatory procedures as stated in regulations and as In 2017, the government repealed a regulation that required practiced. Different authorities enforce the same regulations the use of the Nacala Special Economic Terminal (TEEN), but apply and interpret the laws differently. There is also a allowing exporters to use any legally recognized terminal lack of co-ordination between ministries, leading to the for customs clearing procedures. This move was praised need for numerous different permits and lengthy and costly by the private sector; however, the government has not yet processes to acquire them. issued guidance on how to proceed after the repeal, and > > > Figure 4.15. Mozambique’s Logistics Performance Could be Stronger a. Mozambique’s logistics performance b. Logistics performance compared with neighboring countries Customs 4 3 Timeliness Infrastructure 2 1 0 Tracking & International Tracing shipments Malawi Mozambique South Africa Tanzania Zambia Zimbabwe Logistics competence Source: (World Bank, Logistics Performance Index, n.d.). 98 USAID (2018). 99 World Bank (2019). 100 USAID (2018). 101 WTO (2017). 90 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM stakeholders are concerned about arbitrary costs imposed As part of the South African Development Community on exports.102 The government needs to provide clear rules (SADC), Mozambique is a signatory of the Protocol on for post-TEEN procedures for customs officers and ensure Transport, Communications and Meteorology. Under that adequate enforcement. framework, GoM agreed to assist in the development of a safe, secure, and reliable road network. Notably, it agreed to Arbitrary enforcement of rules leads to the decreased collaborate on a harmonized regional road policy to support competitiveness of the sector. The authorities, as well as commercial transport operators through equal treatment, informal transporters, often ignore axle load restrictions nondiscrimination, reciprocity, and fair competition. which are meant to limit truck freight capacity to 30 tons.103 However, many obligations under this agreement remain Although weighbridges along the corridor are reportedly unimplemented. For example, vehicle dimensions and broken, operators are still charged fees. These measures not weight regulations in Mozambique are stricter that in only make it harder for services providers to compete in the neighboring countries, hindering competitiveness.105 GoM market, they exacerbate the deterioration of roads. The GoM should set targets for implementing the protocol, including needs to work to ensure the transparent application of the recognition of licenses among border countries and an existing regulations through the dissemination of guidelines update of the safety regulations. and provide trainings to inspectors to reduce corruption among officials. Additionally, it should establish transparency While rail transport would be a cheaper alternative to road procedures and a mechanism for administrative appeal. transport, weak governance and infrastructure prevent the Finally, it must ensure proper calibration of weighbridges. sector from advancing.106 Inefficiencies in the sector lead to losses of over USD 200 million a year.107 Slow wagon speed The Beira corridor is another important transport corridor and long loading and unloading times make rail transport in Mozambique. It stretches from the Beira port to uncompetitive, even within the Nacala corridor.108 An destinations in the Democratic Republic of Congo, Zambia, additional hindrance to rail transport is the domination of the Zimbabwe, and Malawi. Among competitors in the region, sector by Caminhos de Ferro Mocambique (CFM), with low this corridor has the highest market share for international participation of the private sector. This also translates into a transit cargo.104 Of the bottleneck types reported, regulatory conflict of interest, as CFM, a state-owned enterprise, acts bottlenecks seem to be the most prevalent. These are simultaneously as regulator and operator. Informal costs in mainly high charges and fees resulting from corruption or the form of check points can increase the cost of a container lack of risk management. by up to 30 percent. As a result, 80 percent of all freight on the corridor is transported by road. The construction of > > > a new line connecting the province of Tete to a new coal Figure 4.16. Infrastructure Quality is Declining terminal at the Nacala port should bring new opportunities and is Below the Regional Average for development. The government is also considering investing in two additional rail lines linking Mozambique to South Africa and Zimbabwe.109 105 4.4 Recommendations 110 There is significant potential for growth in services. The sector needs to transition to a more complex service sector 115 to bolster its contribution to growth and strengthen its role 120 as a backbone to the economy. Growth in services—in particular in backbone services, such as telecoms, transport 125 and logistics—has the potential to spur growth in other export-oriented activities, including agribusiness and mining. 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Development of these sectors, which are still severely Sub-Saharan Africa median Mozambique underdeveloped, should remain Mozambique’s top policy priority regarding trade and investment in services. Source: (World Economic Forum, The Global Competitiveness Report, 2018). 102 USAID (2018). 103 USAID (2018). 104 USAID (2018). 105 USAID (2018). 106 World Bank (2019a). 107 World Bank (2019a). 108 USAID (2018). 109 World Bank (2019a). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 91 Table 4.10 summarizes key recommendations for moving forward. They are divided into recommendations for horizontal and sector-specific regulations. Chapter 4. Can Services Drive Growth in Mozambique: Summary of recommendations Recommendations Responsibility Timeline for Scale up access to public services, notably infrastructure, in rural districts Ensure all foreign firms, notably SMEs, obtain access to foreign transfers, even in the Banco de absence of Investment Authorization by APIEX, in compliance with the Investment Short term Moçambique Law. Provide commercial banks with clear guidelines and trainings to allow international Banco de Short term transfers by already registered foreign firm -Decree 49/2017 + Notice 20/GBM/2017 Moçambique Labor regime Ideally, abolish quota system and labor market tests (LMT). Replace by knowledge Assembleia da Medium term transfer schemes to increase domestic capacity República If not, apply beyond-quota LMTs according to law – not as strict prohibition to Ministério do Short term foreign employment- in Min of Labor Trabalho Meantime, apply quota percentages as a share of effective number of employees, Ministério do Short term not based on categories defined by size of company Trabalho Land use Assembleia da Adopt legal regime for land lease up to 99 years – including by foreign firms/citizens Medium term República Ministério da Meantime, consider extending flexibilities on use of urban land to rural areas Short term Terra e Ambiente Cross-border services Amend Article 85 of the Commercial Code to allow for non-established services Assembleia da suppliers in the country, or limit the scope to the local establishment requirement to Medium term República specific services activities of particular regulatory sensitivity Telecom Ministério da Implement “special regime” for access to land by telecom Medium term Terra e Ambiente Autoridade Update infrastructure sharing regulations and enforce compliance Reguladora das Short term Comunicações IT Promote establishment of export-oriented IT firms in SEZ AIPEX Medium term Autoridade Ensure transparent and predictable regime for access to top level domains Reguladora das Short term Comunicações Road Transport Ministério dos Update safety regulations, in coordination with SADC Protocol on Transport, Transportes e Short term Communications and Meteorology and also South African standards Comunicações Ministério dos Improving truck driver licensing and trainings, including by introducing an Transportes e Short term accreditation system as well as a database of driving infractions Comunicações 92 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM References Alfaro, L., Areendam, C., Sebnem, K.-O., Selin, S. (2004). USAID. (2018). Assessment of Beira Development Corridor. FDI and Economic Growth: The Role of Local Financial Washington, DC: United States Agency for International Markets. Journal of International Economics, 64: 89–112. Development. AMPETIC. (2019). Plano de Actividades 2019. Maputo: USAID. (2018). Assessment of Nacala Development Corridor. Associação Moçambicana de Profissionais e Empresas USAID. (2018). Nacala Corridor & Port Performance de Tecnologias de Informação. Assessment: November 2018 Final Draft Report. Goswami, A. G., Mattoo, A., & Saez, S. (2012). Service Exports: Washington, DC: United States Agency for International Are the Drivers Different for Developing Countries? Development. Washington, DC: World Bank. WITS. (2019). World Integrated Trade Solution. Retrieved Grunfeld, L., & Moxnes, A. (2003). The Intangbile from https://wits.worldbank.org/CountryProfile/en/ Globalisation: Explaining Patterns of International Trade Country/MOZ/Year/2017 in Services. Norwegian Institute of International Affairs World Bank. (2006). Mozambique Enterprise Survey. Paper No. 657. Washington, DC: The World Bank. IFC. (2003). Mozambique- IFC helps Mozal expand successful World Bank. (2012). Knoema. Retrieved from https://knoema. SME linkages program to daily operations. Retrieved from com/atlas/Mozambique/topics/World-Rankings/World- ISO. (2019). The ISO Survey of Management System Rankings/Knowledge-economy-index Standard Certifications. International Organization for World Bank. (2018). Mozambique Enterprise Survey. Standardization. Washington, DC: The World Bank. ITU. (2019). Measuring digital development: Facts and figures. World Bank. (2018). Worldwide Governance Indicators. Geneva: International Telecommunications Union. Retrieved from http://data.worldbank.org/data-catalog/ Lachler, Ulrich;  Walker, Ian.  2018.  Mozambique  Jobs worldwide-governance-indicators Diagnostic. Jobs Series;No. 13. World Bank, Washington, World Bank. (2019a). CPSD Sector Scan. World Bank. DC. © World Bank World Bank. (2019b). Digital Economy Memo. Lall, S. (2002). Linking FDI and Technology World Bank. (2020). Doing Business. Washington DC: World Development for Capacity Building and Strategic Bank Group. Competitiveness. Transnational Corporations, 11. World Bank Group. (2019). Mozambique Digital Economy Mitsubishi Corporation. (2019). Retrieved from https://www. Assessment. Background Paaper Series. Digital mitsubishicorp.com/jp/en/csr/management/business/ Infrastructure. Washington, DC: World Bank. sustainability06.html World Bank. (n.d.). Logistics Performance Index. Retrieved OECD, & ILO. (2018). How Immigrants Contribute to from https://lpi.worldbank.org/ Developing Countries’ Economies. OECD/ ILO. South32. World Economic Forum. (2016). Global Information (2019). Mozal Aluminum. Technology Report. The Local. (2018, January 9). One quarter of EU Blue Card World Economic Forum. (2018). The Global Competitiveness holders in Germany from India: report. https://www. Report. World Economic Forum. thelocal.de/20180109/one-quarter-of-eu-blue-card- WTO. (2017). Trade Policy Review- Mozambique. Geneva: holders-in-germany-from-india-report World Trade Organization. UNCTAD. (2015). Revised Diagnostic Trade Integration Study WTO. (2019). World Trade Report. Geneva: World Trade for Mozambique. Geneva: UNCTAD. Organization. UNDP. (2017). Human Development Reports. Retrieved from http://hdr.undp.org/en/data MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 93 5. 94 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM How Does Corruption Affect Firm Performance? This chapter explores the challenges posed to private sector growth by weak governance and rising corruption in Mozambique. It puts the country’s performance in historical and comparative perspective, revealing a worsening of the governance setting. It then presents empirical analysis on what corruption looks like in terms of the firms affected, the size of bribes and the impact on firm performance. It finds that corruption has a negative effect on performance, with the highest-performing firms hit the hardest. The chapter concludes by summarizing what is being done, and what more can still be done, to alleviate the problem. 5.1 Introduction110 corruption. First, the government plays a more prominent role in the economy to combat the pandemic and Weak governance and corruption pose key challenges to provide economic support to firms and households. private sector growth in Mozambique, and are exacerbated This expanded role is critical, but it also increases by a lack of transparency.111 As noted in the previous chapters opportunities for corruption. Second, as public finances of this CEM, the private sector could become an engine worsen, Mozambique needs to pay special attention to of economic transformation and job creation, but is being preventing tax evasion and the waste and loss of funds held back by several binding constraints, including weak caused by corruption in public spending. Finally, the governance and corruption (see Box 5.1). Mozambique’s COVID crisis is testing citizens’ trust in public institutions, scores on the Worldwide Governance Index112 have fallen and ethical behavior becomes more pressing when in all indicators in recent years, and lag behind sub-Saharan medical services are in high demand. Curbing corruption Africa (SSA) averages. According to Transparency International requires government ownership of reforms, international (2019), Mozambique is ranked among the most corrupt cooperation, and a joint effort with civil society and the countries in the world, below Kenya and on a par with Angola private sector. It also involves political will and timely and Nigeria.113 World Bank (2019a) identifies corruption as the implementation of reforms over time. number one obstacle to doing business in Mozambique, both for domestic and foreign investors and for all sizes of firm. A This chapter provides an overview of corruption and weak regulatory framework also means that fiscal authorities broader governance challenges in Mozambique. It charge more than they should, which further increases the presents new empirical analysis on firms’ perceptions of cost of doing business, notably for small-size investors. High corruption and the impact of corruption on productivity and contracting costs that result from a poor legal and regulatory performance. This provides further evidence to inform and framework discourage investors, especially smaller firms. support the government in implementing its anti-corruption reform measures contained in the recent Diagnostic Report The COVID-19 pandemic has heightened the importance on Transparency, Governance and Corruption published by of more robust governance and the need to tackle the GoM.114 110 This chapter is based on a background paper prepared by Halfdan Lynge (Senior Lecturer, Wits School of Governance, University of the Witwatersrand) with guidance from Shireen Mahdi (Senior Economist, World Bank). 111 World Bank (2020a). 112 Worldwide Governance Indicators. Available at https://info.worldbank.org/governance/wgi/. 113 Mozambique ranked 146 out of 198 countries on Transparency International’s 2019 Corruption Perception Index, though it has jumped 7 places since 2018. 114 GoM (2019). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 95 5.2 What are the Broader Governance services, Mozambique ranks in the 17.8th percentile, lagging behind Côte d’Ivoire (31.7), Ghana (34.1), Tanzania (21.2), and Challenges in Mozambique? Uganda (29.8). In rule of law, which captures perceptions of the extent to which public agents have confidence Mozambique’s governance indicators have been in and abide by the rules of society, Mozambique ranks progressively deteriorating. Various measures reveal a among the lowest (in the 14.4th percentile), far behind Côte weak governance and institutional setting that has only d’Ivoire (30.8), Ghana (56.7), Kenya (37.98), Tanzania (31.3), worsened in recent years, according to the Worldwide and Uganda (43.3). Mozambique scored the lowest on the Governance Indicators (Figure 5.1a). While Mozambique still World Bank’s Country Policy and Institutional Assessment fares better than the SSA average for two indicators (voice (CPIA) transparency, accountability, and corruption indicator and accountability, and regulatory quality), it continues to fall (Figure 5.1b). In addition, the country has also shown limited behind neighboring countries on government effectiveness, progress compared to its peers on the quality of political control of corruption, and rule of law. In the government institutions (Figure 5.1c and d). effectiveness indicator, which measures the quality of public > > > Figure 5.1. Mozambique’s Corruption and Governance Indicators are Getting Worse a. Governance indicators: (lower = worse) b. CPIA: Transparency, accountability, and corruption, 2019 0 4.0 3.5 3.0 2.5 2.0 -1 1.5 1.0 0.5 0.0 SSA Mozambique Guinea Uganda Tanzania Ghana Cote d’Ivoire 2000 2003 2005 2007 2009 2011 2013 2015 2017 -2 Government e ectiveness Regularity quality Rule of law Control of corruption Source: World Governance Indicators. Source: World Bank Country Policy and Institutional Assessment. c. Indicators of political institutions (2000-2018) d. Peer Comparison: Indicators of political nstitutions, (2000-2018) 6 10 8 4 6 4 2 2 0 Mozambique Ghana Guinea Tanzania Uganda Cote d’Ivoire 0 -2 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Regulation of Executive Executive Constraints Executive Constraints Regulation of Executive Polity score Polity score Source: Polity IV Project. Source: World Development Indicators. Note: The polity score measures the degree of political constraints, political competition, and executive recruitment. It ranges between -10 to 10, with higher values denoting more democratic institutions. 96 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM These findings echo those from other global and regional trucks121) and informal checkpoints and charges such as road indicators, including the Corruption Perception Index115 fees (see Chapter 2). In addition, there is no independent and the survey of business leaders by the World Economic regulator for ports. A Sub-National Doing Business analysis Forum116. Mozambique also performs poorly in other found that tariff books for all three main ports were missing measures of governance and corruption, including the information, and that ports “regulate themselves in terms of Ibrahim Index of African Governance (IIAG)117, where it the scale and structure of tariffs”.122 comes 25th out of 54 countries, with the overall governance trend also showing an accelerating decline. The hidden debt What are the Impacts? revelations of 2016 and other high-profile corruption cases have further exacerbated perceptions of corruption. In a Weak governance and corruption constrain growth and 2018 survey of Mozambique, 37 percent of respondents said public service delivery through various transmission “all” or “most” government officials are corrupt, up from 19 channels. A study by the Centre for Public Integrity percent in 2006.118 estimated the costs of corruption to Mozambique during the period 2002 to 2014 to be USD 4.9 billion (approximately Corruption is on the rise across sectors, and is identified 30 percent of GDP in 2014).123 This amounts to around by firms as their biggest obstacle in the 2018 World Bank USD 500 million a year, a figure higher than the state Enterprise Survey. It was also the second most problematic budget annual allocations to social spending on health factor to doing business according to the Global and education.124 The study highlighted that the impact of Competitiveness Report (2019). It appears to be a problem in these costs is widespread, affecting firms in the private and many sectors. For instance, the World Bank (2020a) finds that financial sectors, taxpayers, public service providers, as well the customs administration presents businesses with a high as Mozambique’s international reputation. The International risk of corruption. According to the Global Enabling Trade Monetary Fund has identified three main channels through Report (2016), corruption is more pervasive when dealing which corruption affects the economy. First, corruption with imports than with exports, owing to rent-seeking, non- weakens macroeconomic stability and lowers economic tariff barriers and burdensome import procedures. The 2018 growth. Corruption affects tax collection efficiency and can Enterprise Survey indicates a rise in rent-seeking practices lower spending efficiency through higher costs, reducing since the previous survey in 2007, especially for smaller firms the quality of public goods and services. Second, corruption in the services sector.119 Corruption is especially rife in the poses fiscal risks and lowers the performance of SOEs.125 transport sector.120 The sector is characterized by widespread Third, corruption limits financial sector development; it rent seeking, opaque operations of state-owned enterprises fosters the use of cash, can facilitate money laundering, and (SOEs), and violations of public procurement laws, among increases the costs of financial intermediation.126 others. Similarly, according to the Enterprise Survey, in 2018 about 30 percent of firms were asked to pay bribes to get Lack of transparency, enforcement challenges and electricity connections. persistent corruption affect fiscal governance. Revenue collection falls below potential as a result of rent- The Country Private Sector Diagnostic (CPSD, 2020a) seeking practices surrounding high-profit businesses. identified governance issues and corruption as cross- Tax expenditures, in the form of extensive and complex cutting constraints to private investment, hampering VAT exemptions, create opportunities for evasion. Private economic development and job creation. The diagnostic companies also report significant delays in receiving VAT found corruption to be pervasive in the transport and logistics refunds. Only a third of central and provincial procurement sector, which poses added indirect costs to firms operating is awarded through competitive tendering procedures and in other sectors of the Mozambican economy such as procurement processes are inefficient. Public investment extractives, financial sector, forestry, and fisheries through continues to be affected by factors such as weak appraisal forward linkages. For example, the lack of implementation of and monitoring capacity, and delays in disbursements, regulations is facilitated by corruption (e.g. on overloading of amongst others.127 115 Transparency International. https://www.transparency.org/en/cpi 116 World Economic Forum (2019. 117 Ibrahim Index of African Governance (IIAG). Available at https://mo.ibrahim.foundation/iiag. 118 Afrobarometer (2007, 2019). 119 World Bank (2019a). 120 World Bank (2020a). 121 According to the report, “Overloading impacts road quality while making road artificially cheaper than rail”. 122 Cited in the World Bank’s Country Private Sector Diagnostic (World Bank, 2020a). 123 Centro de Integridade Pública & Chr. Michelsen Institute (2016). 124 Centro de Integridade Pública & Chr. Michelsen Institute (2016). 125 The volume of guarantees and contingent liabilities to SOEs has grown in recent years, which was at the heart of the hidden debt crisis that pushed Mozambique into debt distress. 126 IMF (2018). 127 GoM (2019). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 97 Box 5.1. What is corruption and why does it matter? Corruption is broadly defined as the abuse of public power for private gain.a The term ‘corruption’ is mainly used in this chapter to refer to ‘extractive corruption’ or ‘corruption from above’, where the state is seen as the stronger part in its relationship with the private sector. According to this perspective, public officials seek rents from their control over licenses, contracts, and public services, while firms are relatively passive players. Corruption is thus seen as an agency problem, where public officials entrusted with carrying out tasks on behalf of the state engage in illegal activities for personal enrichment. However, other forms of corruption exist. These include state capture, where firms manipulate the formulation of laws and other government policies to their own advantage through illicit or non-transparent means, and corruption where consumers use payments and connections to gain preferential access to public goods/ services or to circumvent formal procedures, such as access to education and health services.b There are two competing arguments about the effects of corruption. The first is that corruption “sands the wheels of growth”.c This is based on the idea that public officials sometimes deliberately slow down or derail administrative processes in order to get an opportunity to extract a bribe. For firms, this works as a tax, only worse as bribes do not go into state coffers. By acting as a tax on certain factors of production, corruption can change relative factor prices and lead to sub-optimal use of inputs. In addition, firms may be less inclined to invest in productivity-enhancing technologies because of the additional regulatory scrutiny that such actions may attract. The second argument is that corruption “greases the wheels of growth”.d Some contend that corruption actually speeds up administrative processes in economies that are plagued by inefficiencies in the regulatory system. For firms, this represents the second-best option: paying bribes is not ideal, but not getting anything done is worse.e Firm-level research has typically focused on the effect of corruption on productivity, sales, and profits.f The evidence is mixed. Some papers have found that corruption has a negative effect by slowing down productivity, investment, sales and employment growth.g Others have been inconclusive or have found evidence indicating that corruption can have a positive effect.h In Africa, most country studies find a negative effect, while some cross-country studies suggest a positive effect.i Overall, the firm-level literature indicates that the relationship between corruption and economic performance depends on the context, hence pointing to a possible heterogeneous effect. a. Cuervo-Cazurra (2006); Rodriguez et al. (2006); Transparency International broadens the definition to include any “abuse of entrusted power for private gain” (Transparency International, 2020, author’s emphasis). b. Williams and Onoshchenko (2014b, 2014a, 2015). c. Dridi (2013), Guetat (2006), and Mauro (1995). d. Lui (1985); Méon and Sekkat (2005). e. The idea can be traced back to Huntington (1968), who argues that “in terms of economic growth, the only thing worse than a society with a rigid, over-centralised, dishonest bureaucracy is one with a rigid, over-centralised, honest bureaucracy (Huntington, 1968, p. 69). According to Huntington, the positive effect of corruption was demonstrated in the US in the 1870s and 1880s, where bribery in railroad, utility, industrial corporations produced faster growth. Other early proponents of the argument include Leff (1964) and Leys (1965). f. An exception is Bai et al. (2017) who focus on the effect of firm growth in corruption. g. Aterido et al. (2007). h. Ayaydın and Hayaloglu (2014); Seker and Yang (2012); Wang and You (2012). i. Williams and Kedir (2016). The impacts of corruption require deeper understanding. Bank Enterprise Surveys that “establishments like this one” Box 5.1 explores the literature on corruption, showing that give gifts or make informal payments to public officials to its impact is not clear cut. The next section presents analysis “get things done”. It focuses on firms doing business with of firms’ experiences of corruption in Mozambique to government to access public services or contracts. The understand the impacts more clearly. analysis examines the key determinants of exposure to corruption, as proxied by the share of total annual sales that 5.3 How Do Firms in Mozambique firms spend on bribes. It estimates the effect of corruption on sales and productivity growth.128 Experience Corruption? The analysis draws on data from the World Bank’s 2007 This section looks specifically at firms’ experiences of and 2018 Mozambique Enterprise Surveys. The dataset corruption in Mozambique. The analysis draws on a narrow covers 1,311 firms in three industries (manufacturing, retail definition of corruption and considers that firms that are services, and other services) across 8 of Mozambique’s 11 ‘exposed to corruption’ are those that agreed in the World 128 The analysis uses year-location-industry means as an instrument to address the potential problems of endogeneity and measurement error. 98 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM provinces (Maputo City, Maputo Province, Sofala, Manica, were most exposed, followed by firms in other services; Tete, Zambezia, Nampula, and Cabo Delgado).129 Only formal while in the north, firms in manufacturing were most firms were included. Data on informal firms were collected exposed, closely followed by firms in other retail services as part of the 2018 Mozambique Enterprise Survey but were (Figure 5.2). Firms that are ‘exposed to corruption’ are those excluded from the analysis due to weighting problems.130 that reported in the Enterprise Surveys that, “establishments Box 5.2 summarizes the methodology, while Table D3 in like this one” give gifts or make informal payments to public Appendix D provides an overview of the sample distributions officials to “get things done” (Box 5.2).131 The estimates across firm location, size, and industry. The analysis tests if should be interpreted cautiously as they do not address the the widely held perceptions of corruption are grounded in possible bias described in Appendix D1.. Also, particularly reality and examines the extent to which these perceptions in the center of the country, the estimates are based on affect firm performance. As the analysis compares two data relatively few observations, which means the margin of error points from the Enterprise Surveys, the results also provide is high. Nonetheless, there appear to be important regional an indication of direction of travel on some corruption and industry variations. metrics by province and sector. Who Pays the Biggest Bribes? Who is Exposed to Corruption? Older firms are likely to pay bigger bribes. To answer this The analysis shows that in 2018, firms in the center and question, we ran a series of regressions with the share of north of the country were more exposed to corruption total annual sales firms spend on bribes as the dependent than firms in the south. In the center, firms in retail services variable. As the dependent variable is censored (with values Box 5.2. Data, assumptions and sampling Data: The dataset drawn from the World Bank’s 2007 and 2018 Mozambique Enterprise Surveys contains more than 300 variables. The surveys included questions about ‘gifts’ and ‘informal payments’ made to public officials to access public services, secure government contracts, and ‘getting things done’. Note that most corruption questions were framed in general terms (e.g. “when establishments like this one do business with government, what percent of the contract value would be typically paid in informal payments or gifts to secure the contract?”). Presumably, this was an attempt to reduce social desirability bias. Assumptions: The analysis in this chapter assumes firms respond on the basis of their own experiences. When firms say “establishments like this one” typically pay 10 percent of the contract value when they do business with governments, we will assume they themselves pay 10 percent. This follows the interpretation used in other similar studies of firm corruption drawing on World Bank Enterprise Survey data.a However, the construction of the question can complicate the interpretation. Some firms may generalize from “establishments like this one” to firms in the same location or industry, while others may generalize to firms of similar size and age. Sampling and weighting: Firms were selected using stratified random sampling and weighted across province, industry, and firm size. This means we have unbiased estimates both for the population of firms and for different subsets, with a known level of precision. The sample frame (Appendix D, Table D3) consisted of listings of firms from two sources: for panel firms (i.e. firms surveyed in 2007 and 2018), the list of 479 firms from the 2007 Mozambique Enterprise Survey was used; for new firms (i.e., firms not covered in 2007), firm data from the National Institute of Statistics (INE) were used. Measures were taken to ensure the quality of the sample frame. However, it is not immune to the typical problems found in firm surveys (positive rates of non-eligibilityb, repetition, non-existent firms, etc.). Only observed non-eligible firms were excluded from the sample frame.c Box D1 in Appendix D provides the econometric specifications. a. Fisman & Svensson (2007). b. The weights are computed based on three sets of criteria: strict, media, and weak. Depending the eligibility criteria, some observations drop out, giving “positive rates of non-eligibility”. c. The 2018 Mozambique Enterprise Survey contains three different weights, based on three different sets of eligibility criteria (weak, median, and strict). Throughout the analysis, we use the ‘weak’ set, which comes closest to the single weight contained in the 2007 Mozambique Enterprise Survey. 129 Maputo City and Maputo Province were merged into Greater Maputo. Gaza, Inhambane, and Niassa were not included, and inferences can be made only to the remaining seven provinces plus Maputo City. 130 The population of informal firms is unknown. 131 We aggregate at regional rather than provincial level to get a reasonable number of observations in each ‘cell’ and estimates with a reasonable error margin. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 99 between 0 and 100 percent) and the values are clustered affects firm age. In other words, as firms age, the pressure below 20 (few firms spend over 20 percent), the analysis to pay more increases, and unless the firms comply, they go uses a logit model. The results are reported in Appendix out of business. Firms with at least partial foreign ownership D, Table D5. The results suggest a strong correlation also pay bigger bribes. However, the analysis excludes between firm characteristics and the size of bribes paid. firms that do not pay bribes. Accordingly, the results do not They also indicate that older firms spend more. This suggest that firms with foreign ownership are more likely seems counterintuitive, as one may expect older firms to to engage in corruption, but that when they do engage, have more established linkages with public officials that they are more likely to pay bigger bribes. It could be that would allow them to circumvent direct payments. One firms with foreign ownership are assumed to have access interpretation could be that it is not firm age that affects to finance and that the corrupt, profit-maximizing public how much they spend, but that how much they spend officials require bigger bribes. > > > Figure 5.2. Exposure to Corruption Varies by Region, Year and Industry Firm exposure to corruption across year, region, and industry 30 25 20 15 10 5 0 2007 2018 2007 2018 2007 2018 2007 2018 Manufacturing Retail services Other services All industries North Centre South All regions Source: The underlying data are provided in Annex 5A. > > > Figure 5.3. Most Bribes Represent a Small Share of Annual Sales a. 2007 b. 2018 50 50 Number of observations Number of observations 40 40 30 30 20 20 10 10 0 0 0 10 20 30 40 50 0 10 20 30 40 50 % of total annual sales % of total annual sales 100 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Overall, the figures suggest a link between region and 30 percent in 2018. This confirms the deteriorating trends in industry, and that the capacity of public officials to extract corruption discussed in Section 5.2. It also appears to indicate rents depends on the local context. In some regions, the salient corruption issues in the transport and logistics sector configuration of markets and local patronage networks identified in the recent Country Private Sector Diagnostic allows them to (potentially) demand bribes; in others, it (CPSD) (World Bank, 2020a), as construction contracts are does not. We return to this below, when we describe the tightly linked to the sector. econometric specifications in the main analysis. There is no evidence that corruption speeds up service How Big are the Bribes? delivery. Appendix D2 contains an analysis of whether corruption speeds up service delivery; i.e. whether firms About 13 percent of firms said they were exposed to who said an informal gift or payment had been expected corruption in 2018, down from 17 percent in 2007. Most or requested in connection with their application for an firms have not been exposed to corruption, therefore, and operating license, an electrical or water connection, a those that (indirectly) indicated that they were exposed, say construction-related permit, an import license, received the they spend relatively little. Figure 5.3 plots the amounts that public services quicker than other firms. In short, we find no firms say they spend on bribes as a share of total annual robust evidence that this is the case. sales.132 The vast majority spend less than 10 percent. Only 18 firms said they spend more than 20 percent. This raises Is Corruption Seen as an Obstacle to Firm the question: what type of firms are spending a large share Performance? on bribes? To answer this question, we ran a series of regressions with the share of total annual sales firms spend Corruption may impede new firms from entering a on bribes as the dependent variable. The results are reported market.133 This is suggestive of a negative relationship in Appendix D. Firms in the south and in manufacturing between firm size and perceptions of corruption. However, spend more than firms in the north and in other industries, considering that our dataset includes micro firms that all else being equal. Combined with Figure 5.2, the findings possibly operate below the radar, it seems likely that the indicate that fewer firms in the south and in manufacturing relationship is curvilinear. To test this possibility, we ran a are exposed to corruption, but that those that are exposed, series of regressions. The analysis uses two different variables tend to pay bigger bribes. as our main regressors: firm sales and number of employees. In addition, we include the control variables used in other Why Are Bribes Paid? sections of this chapter. The results are reported in Table D9, Appendix D3. The results suggest that the effect of sales is, Many firms reportedly paid bribes to access public services. indeed, not constant. For smaller companies, the relationship They consist of gifts or informal payments to get licenses, between sales and perceptions that corruption is an obstacle permits, utility connections, and to pay taxes. This is referred is positive, while for larger companies it is negative. Figure to as extortion, although no coercion is usually involved. In D1 in Appendix D3 demonstrates that for smaller firms, sales both 2007 and 2018, firms were asked whether they had have a positive but diminishing effect, whereas for larger applied for an operating license, an electrical, water, or firms it has a negative and diminishing effect, all else being phone connection, a construction-related permit etc., or equal. This suggests that: (i) firms can operate below the whether they had tried to get import or export clearances or radar but that they draw attention as soon as their annual pay taxes, over the last two years, and whether an informal sales start to increase; and (ii) firms can reduce the negative gift or payment had been expected or requested (Figure 5.4). effect of corruption by employing more people. The analysis On average, 15-20 percent of all applications or attempts also finds that corruption appears to be a smaller obstacle to clear goods or pay taxes came with the expectation of to firm operations in manufacturing than in retail and other or a request for a bribe. Bribes to get an electrical or water services. Second, corruption also appears to be a smaller connection or a construction-related permit were the most obstacle in the north than in the south. common, followed by bribes to get an import or operating license. Another type of corruption faced by many firms is the need to pay bribes to secure government contracts. In 2018, two- Overall, bribes to access public services appear to be thirds of firms said they do business with government. Yet, more common today than 10 years ago (Figure 5.4). This is only 19 percent said they “secured or attempted to secure particularly the case for construction-related permits, where a government contract” in the last year, and, of those, only the share of applications that came with the expectation or 15 percent said “informal payments or gifts” are “typically a request for a bribe has increased from 7 percent in 2007 to paid […] when establishments like this one do business with 132 Some firms provided local currency figures. In these cases, we divided the figures with the firms’ total annual sales in the last financial year. 133 Campos et al. (2010). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 101 > > > Figure 5.4. Bribes for Public Services are on the Increase 30 % firms who said yes 20 10 0 Getting an operating license Getting an electrical connection Getting a water connection Getting a phone connection Getting a construction- related permit Getting an import license Getting an import cleared Getting an export cleared Paying taxes 2007 2008 Notes: The figures are calculated by dividing the number of firms who reported that “an informal gift or payment [was] expected or requested” by the number of firms who reported having applied in the last two years. The underlying data are provided in Appendix D. the government”. In other words, only around 3 percent 20 percent of total annual sales. The analysis in this section of all firms (indirectly) admitted they paid bribes to secure supports this estimate. Figure 5.5 plots the size of the bribes government contracts in the last year. This figure is lower as a share of total contract value. It shows that, in both 2007 than expected. A recent report by the Centro de Integridade and 2018, only three firms said that bribes paid to secure Pública (CIP) and Chr. Michelsen Institute, an independent government contracts that exceed 20 percent of the total non-profit research foundation, claims that “there is a large contract value. Fewer firms paid bribes in 2018. degree of corruption in the tendering for public construction contracts”134 and that “more often than not tender processes A combination of three factors probably explains why get suspended midway only for the contract to be awarded perceived levels of corruption are higher than actual to a particular company”.135 Along the same lines, the levels of corruption measured in this analysis. The first Business Anti-Corruption Portal, an often-referenced website factor is presumably a consequence of the ‘hidden debt’ funded by the European Commission, notes that “public revelations.137 Second, informal payments may be made procurement is a high-risk sector” in Mozambique and that in ways that are not captured by the variable; for example, “companies perceive favoritism towards well-connected by including public officials as well-paid, non-executive firms to be widespread among procurement officials directors or among the firm shareholders. Finally, there and report that funds are often diverted to companies or may be significant under-reporting in the data, due to social individuals due to corruption”.136 desirability bias. This possibility should be explored further, using two-step bias correction models.138 Only few firms paid bribes that exceed 20 percent of total contract values. Above we saw that bribes rarely exceed 134 CIP (2016), p. 6. 135 CIP (2016), p. 56. 136 GAN Integrity Solutions (2020). 137 This claim is supported by Afrobarometer data (2019), available at  http://www.afrobarometer.org. In 2018, 49 percent of respondents said the level of corruption in Mozambique had increased “somewhat” or “a lot”. Yet, compared with 2015, there was no (or no significant) increase in how often firms reported to have paid bribes for documents and permits, treatments in public clinics and hospitals, school services, water and sanitation services, or to avoid problems with the police. 138 C.f. Bushway et al. (2007); Heckman (1979). 102 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM > > > Figure 5.5. Bribes to Secure Government Contracts are Decreasing in Size and Number a. Size of bribes paid to secure government b.v Size of bribes paid to secure government contracts (2007) contracts (2018) 15 15 Number of observations Number of observations 10 10 5 5 0 0 0 10 20 30 40 50 0 10 20 30 40 50 % of contract value % of contract value > > > Figure 5.6. Bribes Reduce Both Sales and Productivity, Especially for High-Performing Firms a. Estimated e ect of corruption on sales b. Estimated e ect of corruption on productivity 20 20 Productivity growth (%) 10 10 Sales growth (%) 0 0 -10 -10 -20 -20 0 5 10 15 20 0 5 10 15 20 Bribes (% of annual sales) Bribes (% of annual sales) High performance firms Medium performance firms High performance firms Medium performance firms Low performance firms Low performance firms Note: Low, medium, and high-performance firms are firms with sales or productivity growth rates in the first, second, and third quartile respectively 5.4 Corruption is Bad for Productivity means every additional percentage point spent on bribes reduces productivity growth by 8.8 percent. To get a better and Growth understanding, let us consider an example. The median sales growth rate in our data is 11.2 percent, and the median Corruption has a negative effect on firm performance, productivity growth rate is 7.1 percent. This means a “typical” measured both as sales and productivity growth. Having firm improves its sales by about 11.2 percent every year and described some of the general patterns of firm corruption improves its productivity by about 7.1 percent. Suppose this in Mozambique, this section turns to the main analysis of firm increases its bribe rate from 0 to 1 percent of annual whether corruption affects firm performance. The results are sales. This means its sales growth rate would drop to 10.6 reported in Table D1 in Appendix D1, and the robustness of percent, while its productivity growth rate will drop to 6.5 the results is discussed in Box 5.3. percent. If the firm increases its bribe rate further to 2 percent of annual sales, its sales growth rate will drop to 10 percent The results suggest that every additional percentage point (by 5.5 percent of 10.6), and its productivity growth rate will spent on bribes reduces sales growth by 5.5 percent. drop to 5.9 percent (by 8.8 percent of 6.5). A coefficient of -0.088 (Model 3 in Appendix Table D) MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 103 Box 5.3. Sensitivity analysis shows that the main findings are robust As a robustness check, the analysis excluded a few outliers that could potentially influence the results. To test this possibility, the analysis reran models (1) and (3) in Table D1 (Appendix D) on three subsets of the dataset, where we excluded observations that were more than three standard deviations above or below the mean of Sales Growth, Productivity Growth, and Bribe. This eliminated 1-2 percent of the observations, which is slightly more than expected, assuming normal distribution. The results are reported in Table D2 of Appendix D. Models (1) and (3) eliminate observations that are more than three standard deviations above or below the mean of Sales Growth and Productivity Growth. Models (2) and (4) eliminate observations that are more than three standard deviations above the mean of Bribe. The results indicate that the outliers pull the effect towards zero. In other words, if anything, corruption is more damaging than the baseline results indicate, once outliers are taken into account. Sensitivity analysis showed that the findings in Table D2 are robust to the choice of robust standard errors. To test whether the choice of standard errors has any effect on the results, the analysis reran models (1) and (3) in Table D2 with robust standard errors, allowing for clustering by province, industry, and province-industry. So far, the analysis used standard errors. However, a Breusch-Pagan test suggest the possible use of (heteroskedasticity-consistent) robust standard errors. We also experimented with several alternative control variables, including firm capacity proxies (years of education and experience of the top manager) and year, location (region), and industry fixed effects. As a final robustness test, we changed the dependent variable. Firms may not know exactly how much they spent on gifts and informal payments, but they certainly know whether they spent anything at all. The results are reported in Appendix D. The variables of interest are largely unaffected by the changes. When we rerun the models with robust standard errors, BRIBE_PER has a consistently negative and statistically significant on productivity growth. When the effect on sales growth is estimated, it is also consistently negative but only statistically significant when we allow for clustering by province. When we allow for clustering by industry and province-industry, the coefficient falls just outside the area of statistical significance. The analysis shows a negative effect of corruption on both (those with high sales and productivity growth rates) that pay sales and productivity growth. Figures 5.6a and b illustrate relatively small bribes. For low-performance firms and firms this for low, medium, and high-performance firms. The that pay larger bribes, the effect gradually approaches zero. effect of bribe depends on the level of firms’ performance. For high-performance firms, the effect of a bribe increase of 5.5 Recommendations: What is Being 1 percent of annual sales translates into a sales growth loss of 1.4 percentage points and a productivity growth loss of 1.9 Done and What More Can Be Done? percentage points, whereas for medium-performance firms, Despite recent efforts to strengthen governance and it translates into a sales growth loss of 0.6 percentage points reduce corruption, further efforts are needed to reap and a productivity growth loss of 0.6 percentage points. For firms’ performance potential in Mozambique. The GoM low-performance firms, the effect is limited (and it should has identified anti-corruption efforts as a critical priority. be interpreted cautiously, due to the negative starting point). Following the hidden debt crisis, it adopted an anti- corruption plan in November 2016. In 2019, the GoM The size of the bribe also plays a role. High-performing prepared a diagnostic report on transparency, governance, firms that pay relatively small bribes lose 1-2 percentage and corruption with support from the IMF, which established points of sales and productivity growth when they increase concrete reform measures and identified the entities their bribe payments from 0 to 1 percent of total annual responsible for implementing them (GoM, 2019). These sales. For medium-performance firms the loss is around measures include, among others, putting in place institutional 0.6 percentage points. Bribe payments have a particularly frameworks to address governance and corruption across damaging effect on high-performance firms, and firms areas such as administration of justice and anti-money that generally pay little bribes (as a share of total sales). laundering/countering financing of terrorism (AML/CFT).139 In other words, medium-performance firms that already The authorities adopted regulations to strengthen debt pay significant bribes lose relatively less from paying an and guarantees management and increase transparency in additional percent of their sales. The effect declines as firm 2017. GoM also passed a new legal framework for SOEs to performance drops and corruption increases. This means strengthen oversight, transparency, and risk management corruption is most damaging to high-performance firms 139 GoM (2019). 104 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM in 2018. The overarching PFM legislation is currently being performing ones. These objectives are central to inclusive revised, and a government-wide financial management growth in Mozambique. Corruption through bribes to access system (e-SISTAFE) has been implemented progressively. It public services and government contracts subdues growth is vital that the government stay on track concerning these by constraining sales and productivity. Reducing corruption, reforms and others in the 2019 report. Overall, more needs which acts as an additional tax on economic activity, is key to be done to implement the key reform measures identified to sustaining and enabling high-performing firms, which in the IMF-GoM diagnostic report. This is even more pressing suffer most from corruption,. The growth potential of firms as GoM combats Covid-19, considering that its role in the in the services sector is discussed in Chapter 4. economy has become more prominent. Finally, the government should place emphasis on The country could further strengthen the regulatory formalizing firms. Informal firms still represent about 90 framework and the business environment in which firms percent of the enterprises in Mozambique. The sizeable operate. A critical measure relates to consolidating licensing informal economy with limited financial inclusion means for all economic activities. This involves unifying licensing that a high share of transactions is in cash, making it difficult for economic activities under a single entity and mandating to track and control illicit activities. Capturing these firms centralized payment of licensing fees. This unification in the formal sector would significantly reduce unlawful may prioritize firms that apply for import/export licenses, transactions and related business activities and increase enhancing oversight of construction-related permits. It the government’s revenue base. The additional fiscal space would help to reduce the cost of starting a business and the generated by bringing firms to the formal sector would allow opportunities for extracting bribes. The government could the government to finance the policy recommendations also digitize the company registry and make it available to proposed in this chapter more comfortably. the public, which would enhance transparency and allow for easy access by ministries and civil society. The registry would strengthen the capacity to inspect and increase transparency and accountability. These measures could be most impactful in the transport and logistics sector. This is because corruption issues are salient in this sector, and it has strong forward linkages with other sectors such as extractives, financial sector, forestry, and fisheries (World Bank, 2020b). The authorities should consider intensifying efforts to enforce laws and regulations involving firms. GoM has strengthened the administration of justice and put in place an AML/CFT framework. Further efforts should seek to increase institutional capacity to implement and enforce these regulations with a focus on reducing the costs associated with contract enforcement by firms, observing public procurement rules, and removing undue price controls for some goods. With regards to the AML/CFT framework, efforts include developing and effectively implementing regulations and procedures for designated non-financial businesses and professions registration and filing of suspicious transaction reports. Also, the authorities should consider prioritize digitizing judicial resources and recording systems. The government should improve concession regulations, including criteria for the award of concessions. It is also crucial to compile an electronic list of those defined as “public servants” as defined in the Public Probity Law in each ministry and public institution. These measures combined would increase transparency and reduce human errors and graft, increasing the rule of law’s effectiveness under which firms conduct their business (GoM, 2019). The government needs to keep high-performing firms successful and increase sales and productivity of weaker MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 105 References Afrobarometer. (2007). Mozambique Round 3 Data. Government of Mozambique (2019). Diagnostic Report on Afrobarometer. Transparency, Governance and Corruption. Prepared Afrobarometer. (2019). Mozambique Round 7 Data. with the Assistance of the Legal and Fiscal Affairs Afrobarometer. Departments of the International Monetary Fund. IMF. Asiedu, E., & Freeman, J. (2009). The Effect of Corruption Grzymala-Busse, A. (2008). Beyond Clientelism: Incumbent on Investment Growth: Evidence from Firms in Latin State Capture and State Formation. Comparative Political America, Sub-Saharan Africa and Transition Countries. Studies, 41(4/5), 638–673. Review of Development Economics, 13(2), 200–214. Guetat, I. (2006). The Effects of Corruption on Growth Aterido, R., Hallward-Driemeier, M., & Pagés, C. (2007). Performance of the MENA Countries. Journal of Investment Climate and Employment Growth: The Economics and Finance, 30(2), 208–221. Impact of Access to Finance, Corruption and Regulations Hallward-Driemeier, M., Wallsten, S., & Xu, L. C. (2006). Across Firms (No. 3138; IZA Discussion Paper). Ownership, Business Climate and Firm Performance: Athanasouli, D., Goujard, A., & Sklias, P. (2012). Corruption Evidence from Chinese Firms. Economics of Transition, and Firm Performance: Evidence from Greek firms. 14(4), 629–647. International Journal of Economic Sciences and Applied Heckman, J. J. (1979). Sample Selection Bias as a Specification Research, 5(2), 43–67. Error. Econometrica, 47(1), 153–161. Ayaydın, H., & Hayaloglu, P. (2014). The Effect of Corruption Hellman, J. S., Jones, G., & Kaufmann, D. (2003). Seize on Firm Growth Growth: Evidence from Turkey. Asian the State, Seize the Day: State Capture and Influence Economic and Financial Review, 4(5), 607–624. in Transition Economies. Journal of Comparative Bai, J., Jayachandran, S., Malesky, E. J., & Olken, B. A. (2017). Economics, 31(751–773). Firm Growth and Corruption: Empirical Evidence from Huntington, S. P. (1968). Political Order in Changing Societies. Vietnam. The Economic Journal, 129(1), 651–677. Yale University Press. Black, S. E., & Lynch, L. M. (1996). Human-Capital Investments IMF (2018). Selected Issues Paper, The Governance Challenge and Productivity. American Economic Review, 86(2), in Mozambique. IMF Country Report No. 18/66, March 263–267. 2018. International Monetary Fund: Washington DC. Bushway, S., Johnson, B. D., & Slocum, L. A. (2007). Is the Keefer, P., & Knack, S. (1997). Why Don’t Poor Countries Catch Magic Still There? The Use of the Heckman Two-Step Up? A Cross-national Test of an Institutional Explanation. Correction for Selection Bias in Criminology. Journal of Economic Inquiry, 35(3), 590–602. Quantitative Criminology, 23(151–178). Kimuyu, P. (2007). Corruption, Firm Growth, and Export Campos, N. F., Estrin, S., & Proto, E. (2010). Corruption as Propensity in Kenya. International Journal of Social a Barrier to Entry: Theory and Evidence (No. 5243; IZA Economics, 34(3), 197–217. Discussion Paper). Knack, S., & Keefer, P. (1995). Institutions and Economic Centro de Integridade Pública & Chr. Michelsen Institute. Performance: Cross-country Tests Using Alternative (2016). The costs of corruption to the Mozambican Institutional Measures. Economics and Politics, 7(3), economy: Why it is important to fight corruption in a 207–227. climate of fiscal fragility (No. 6; CMI Report). Krueger, A. B., & Angrist, J. (2001). Instrumental Variables and Cuervo-Cazurra, A. (2006). Who Cares about Corruption? the Search for Identification: From Supply and Demand to Journal of International Business, 37(6), 807–822. Natural Experiments. Journal of Economic Perspectives, Dridi, M. (2013). Corruption and Economic Growth: The 15(4), 69–85. Transmission Channels. Journal of Business Studies Krumpal, I. (2013). Determinants of Social Desirability Bias Quarterly, 4(4), 121–152. in Sensitive Surveys: A Literature Review. Quality and Faruq, A., Webb, M., & Yi, D. (2013). Corruption, Bureaucracy Quantity, 47(4), 2025–2047. and Firm Productivity in Africa. Review of Development Leff, N. (1964). Economic Development through Bureaucratic Economics, 17(1), 117–129. Corruption. American Behavioral Scientist, 8(3), 8–14. Fisman, R., & Svensson, J. (2007). Are Corruption and Leys, C. (1965). What is the Problem about Corruption? Taxation Really Harmful to Growth? Firm Level Evidence. Journal of Modern African Studies, 3(2), 215–230. Journal of Development Economics, 83, 63–75. Lui, F. (1985). An Equilibrium Queuing Model of Bribery. GAN Integrity Solutions. (2020). Mozambique Corruption Journal of Political Economy, 93(4), 760–781. Report. Business Anti-Corruption Portal. https://www. Mauro, P. (1995). Corruption and Growth. Quarterly Journal ganintegrity.com/portal/country-profiles/mozambique/ of Economics, 110(3), 681–712. Gbetnkom, D. (2017). Corruption and Small- and Medium- Medina and Schneider. (2018). Shadow Economies Around Sized Enterprise Growth in Cameroon. In S. Kayizzi- the World: What Did We Learn Over the Last 20 Years? Mugerwa, A. Shimeles, A. Lusigi, & A. Moummi (Eds.), IMF WP/18/17. Inclusive Growth in Africa: Policies, Practice, and Lessons Méndez, F., & Sepúlveda, F. (2006). Corruption, Growth and Learnt. Routledge. Political Regimes: Cross Country Evidence. European 106 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Journal of Political Economy, 22, 82–98. https://doi. Surveys Data Set. World Bank. org/10.1016/j.ejpoleco.2005.04.005 World Bank. (2019a). The Mozambique 2018 Enterprise Mendoza, R. U., Lim, R. A., & Lopez, A. O. (2015). Grease or Surveys Data Set. World Bank. Sand in the Wheels of Commerce? Firm Level Evidence World Bank. (2019b). The Mozambique 2018 Micro- on Corruption and SMEs. Journal of International Enterprise Survey Data Set. World Bank. Development, 27(415–439). World Bank (2020a). Mozambique Country Private Sector Méon, P.-G., & Sekkat, K. (2005). Does Corruption Grease Diagnostic. World Bank Group and SIDA. or Sand the Wheels of Growth? Public Choice, 122(1), World Bank (2020b) (Forthcoming) Mozambique World Bank 69–97. Group. Mo, P. H. (2001). Corruption and Economic Growth. Journal World Economic Forum (2019). The Global Competitiveness of Comparative Economics, 29(1), 66–79. Report. World Economic Forum: Geneva. Rodriguez, P., Siegel, D. S., Hilman, A., & Eden, L. (2006). World Economic Forum and Global Alliance for Trade Three Lenses On The Multinational Enterprise: Politics, Facilitation (2016). Global Enabling Trade Report. World Corruption, and Corporate Social Responsibility. Journal Economic Forum: Geneva. of International Business StudiesBusiness Studies, 37(6), 733–746. Rose-Ackerman, S. (1999). Corruption and Government. Cambridge University Press. Seker, M., & Yang, J. S. (2012). How Bribery Distorts Firm Growth: Differences by Firm Attributes (No. 6046; World Bank Policy Research Working Paper). Svensson, J. (2005). Eight Questions about Corruption. Journal of Economic Perspectives, 19, 19–42. Transparency International. (2019). Corruption Perceptions Index 2019. Transparency International. Transparency International. (2020). How do you define corruption? What Is Corruption? https://www. transparency.org/what-is-corruption#define Treisman, D. (2007). What Have We Learned About the Causes of Corruption From Ten Years of Cross-National Empirical Research? Annual Review of Political Science, 10, 211–244. Wald, A. (1940). The Fitting of Straight Lines if Both Variables Are Subject to Error. Annals of Mathematical Statistics, 11(3), 284–300. Wang, Y., & You, J. (2012). Corruption and Firm Growth: Evidence from China. China Economic Review, 23(2), 415–433. Williams, C. C., & Kedir, A. M. (2016). The Impacts of Corruption on Firm Performance: Some Lessons From 40 African Countries. Journal of Developmental Entrepreneurship, 21(4), 1–19. Williams, C. C., & Onoshchenko, O. (2014a). Evaluating the prevalence and nature of blat in post-Soviet societies: A case study of the education sector in Ukraine. International Journal of Social Economics, 41(9), 747– 759. Williams, C. C., & Onoshchenko, O. (2014b). Evaluating the role of blat in finding graduate employment in post-Soviet Ukraine: The ‘dark side’ of job recruitment? Employee Relations, 36(3), 254–265. Williams, C. C., & Onoshchenko, O. (2015). An evaluation of the persistence of blat in post-Soviet societies: A case study of Ukraine’s health services sector. Studies in Transition States and Societies, 7(2), 46–63. World Bank. (2007). The Mozambique 2007 Enterprise MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 107 Annexes 108 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Annex A1. Lessons from fast-growing economies Table 1. Growth Commission recommendations and Mozambique’s experience Common characteristics of fast-growing economies Mozambique strategy and experience They fully exploited the world economy. Has been increasingly open to international trade with, and FDI from, third countries. They maintained macroeconomic stability. Relatively stable macro environment, despite the instability that ensued the hidden debt crisis; Limited central bank financing of fiscal deficits; High public debt burden. They mustered high rates of saving and investment. Savings are low. Gross investment rate higher compared to regional and structural peers. They let markets allocate resources. Low market competition in some key sectors (including transport) are obstacles to private sector development. They had committed, credible, and capable Yes despite major setbacks in recent years. governments. Policy ingredients Technology transfer. FDI is high but domestic market dominated by small and informal firms. Competition and structural change. Limited. Efficient labor markets. Weak. Export promotion and industrial policy. Nascent stage. Capital flows and financial market openness. Not open capital account. Financial sector development. Financial market growing, but lags behind those of other middle-income SSA countries, with poor saving culture. Urbanization and rural investment. Largely unmanaged urbanization. Weak rural infrastructure. Equity and equality of opportunity. Significant income inequality. Challenges remain in equality of income and opportunity. Regional development. Urban population is concentrated in and around of Maputo Uneven regional development. The quality of the debate. Limited but improving policy debates. Examples of sub-optimal growth policies Subsidizing energy except for very limited subsidies Limited subsidies through VAT for electricity targeted at highly vulnerable sections of the population. Providing open-ended protection of specific sectors, Considerable protection of specific industries. industries, firms, and jobs from competition. Imposing price controls to stem inflation, which is Limited much better handled through other macroeconomic policies. Banning exports for long periods of time to keep No domestic prices low for consumers at the expense of producers. Poor regulation of the banking system combined with excessive direct control and interference. Resisting urbanization and as a consequence No underinvesting in urban infrastructure. Source: World Bank staff elaboration based on Commission on Growth and Development (2008). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 109 > > > Figure 1. Sales per employee by sector (permanent + temporary employees) a. Formal: Sales per employee in constant 2011 international $ 70,000 60,000 2007 2018 50,000 40,000 30,000 20,000 10,000 0 Metals, Machinery, Food Other Construction Retail Tourism Other Services Computer, Manufacturing and Electronics b. Sales per employee in constant 2011 international $ (adjusted by 2011 PPP) 250 Services (average growth) 40% % change 2007 - 2018 35% 200 30% Thousands Manufacturing (average growth) 150 25% 20% 100 15% 10% 50 5% 0 0% Transport Service of motor vehicles Hotel and restaurants Publishing, printing and recorded... Transport machines Machinery and equipment Tobacco Refined petroleum product Construction Food Non-metallic mineral products Furniture Garments Series 1 Series 3 Source: World Bank Sta estimates based on data from the ES. > > > Figure 2. Mozambique: Public and private capital to output ratios Mozambique’s private capital-output ratio is one of the lowest in Sub-Saharan Africa (2015-2017) 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Guinea-Bissau Djibouti Rwanda Burundi Mozambique Ethiopia Cote d'Ivoire Mali Malawi Nigeria Guinea Kenya Sierra Leone Gambia Burkina Faso Chad Namibia Madagascar Comoros Senegal Congo, DR Uganda Sudan Zimbabwe Togo Lesotho Mauritius Cameroon Botswana Benin Eq. Guinea South Africa Zambia STP Tanzania Niger Angola Switzerland Gabon Liberia Mauritania Seychelles Ghana Cabo Verde Source: IMF (2020) Investment and Capital Stock Dataset. 110 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Annex A2. Mozambique’s LNG Sites Significant progress has already been made in the Anadarko Petroleum Company, which announced development of several LNG sites. Currently, four different an FID in June 2019. In September 2019, Anadarko’s sites are being developed in Area 1 and 4 of the Rovuma share was acquired by Total. The project involves Basin (Table 1). In total, five trains will be developed with the development of two onshore trains with a total a combined liquefication capacity of 31.5 million tons per liquefication capacity of 12.9 MTPA with production annum (MTPA) in Coral South, Golfinho and Rovuma LNG 1 from the first train starting by 2024. with the likely addition of more trains in Rovuma LNG 2. Full • Rovuma LNG 1 (Mamba): Located in Area 4, this project capacity of those five trains is expected to be reached by is operated by Mozambique Rovuma Venture, led by 2027. The sites under development are the following: Exxon Mobil and Eni, and plans the development of two onshore trains with a total liquefication capacity of • Coral South: Area 4’s Coral South offshore development 15.2 MTPA. Exxon Mobil plans to invest USD 500mn in project is a floating LNG project (FLNG). It is led by Eni the initial construction phase and an FID is expected in and will have an offshore floating train with a liquefication 2020. Production is expected to start in 2025. capacity of 3.4 MTPA. A final investment decision • Rovuma LNG 2 (Prosperidade): The development of (FID) was announced in June 2017 and production is this project, expected to add around 10 MTPA with expected to start in mid-2022. onshore trains, is forecasted to follow with a production • Golfinho (Atum): This project in Area 1 was led by the start in 2030. Table 2. Estimated investment amount and expected capacity by LNG site Block Facility type LNG capacity Production start (planned) Coral South Area 4 FLNG 3.4 MTPA 2022 (1 train) Golfinho Area 1 Subsea system 12.9 MTPA 2024 (2 trains of 6.44 MTPA) Rovuma LNG 1 (Mamba) Area 4 Subsea system 15.2 MTPA 2025 (2 trains of 7.6 MTPA) Rovuma LNG 2 (Prosperidade) Area 1 Subsea system about 10 MTPA 2030 Total 41.5 MTPA Source: World Bank Staff Estimates. Note: These are only estimated investment figures, which can vary from the actual invested amount due to various factors. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 111 Annex A3. Assumptions and Scenarios of the MAPTIO and Multiplier Models Several assumptions have been made for the Assumptions underlying the multiplier analysis baseline scenario of the MAPIO model: • The social accounting matrix (SAM) for Mozambique is • LNG production is restricted to the four identified sites presented in 2015 local currency (MZN). Expenditures and it is assumed there will be no delays in starting are provided in 2019 USD and the analysis is conducted production. in USD. A fixed exchange rate of MZN 60 to the USD • Estimates on the implementation and production is used throughout the period of observation, where timeframe, with associated revenues and costs necessary. (investments and operating costs), have been obtained • In addition to the SAM, the model uses employment from the World Bank energy team. data in order to examine multiplier impacts on the • The majority of the program is assumed to be demand for labor. To do so, it is necessary to adjust implemented by foreign firms and labor; only 10 percent employment-output ratios that match the 2015 SAM is provided by local firms and labor. with the exchange rate. • The increased demand (due to implementation of the • Estimates of employment-output elasticities are few and program) has only a limited initial impact on supply, far between, in particular at the industry and occupation rising over the implementation period. Supply effects level. Estimates by Moolman (2003) for South Africa are are not assumed to persist after the implementation now more than 15 years old. Kapsos (2005) estimates period is completed. are also rather dated but they cover a range of countries • All LNG produced is sold at the prevailing international and a limited number of sectors (agriculture, industry LNG price, which, after projected to fall between 2019 and services) although no distinction is made between and 2021, is expected to grow by 2.6 percent annually occupation groups or levels education attainment. over the period 2021-33. Given that there is not yet a Mozambique is covered by Kapsos (2005) and we map defined policy in place for domestic consumption, these estimates broadly to the activities identified in the possible subsidized pricing for domestic consumption model.140 has not been considered. • An additional assumption is made by the WB input data • Similarly, revenues and costs of the participation of ENH that the share of wage earnings of foreign workers on in the project development have not yet been identified, the project that is spent locally is 15 percent (as opposed therefore, these are excluded from the model. to 100 percent for local workers). If we assume that • The model does not consider multiplier effects or foreign workers earn more than their local counterparts possible externalities, positive or negative, such as by a factor of 6.66 and the distribution across labor environmental costs. type is the same, then the earnings and the income expenditure loop associated with foreign workers is Scenarios of the MAPIO model: about the same as for the local counterparts.141 In doing so, we use the full complement of foreign workers Scenario 1 – baseline scenario: assumptions as outlined based locally as contributing to local employment, above including their particular income and expenditure loop. Scenario 2 – low production: In the low production scenario, • The data used for the construction of the 2015 SAM investment in LNG Rovuma 2 (Prosperidade), which is still at here are based on unpublished INE142 industry-level a concept stage, would not go ahead, implying a production production accounts, commodity-level supply-demand of about 10 MTPA less than in the baseline. balances and a supply matrix for 2015 together with Scenario 3 – price shock: A negative price shock in the National Directorate of Planning and Budget (DNPO) magnitude of a double standard deviation of the LNG price government statistics and IMF balance of payment is assumed to affect LNG prices in 2025 and persist to 2033. statistics for 2015. To this is added unpublished INE Despite this price shock, LNG production is assumed to household and labor market survey data for 2014–15 and remain unchanged. a use matrix from a 2007 Mozambique SA (Arndt and Thurlow, 2014). The 2015 SAM for Mozambique identifies 140 The employment elasticity for Mozambique’s services is estimated to be higher than unity while estimates for Sub-Saharan African countries are typically around 0.75. The latter is used in the analysis. The elasticities are used indiscriminately to all types of labor that are identified in the employment data. 141 Adjustment factors to change this assumption can be introduced. 142 The Mozambique Statistical Agency. 112 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 55 industries (entities responsible for the production of and along rural and urban residence. The government goods and services process) as well as commodities represents the main government expenditure account (the goods and services produced by the activities). and various tax revenue accounts. Finally, a savings = There are also three production factors identified: investment account shows how savings are retained capital, land and labor with the latter disaggregated (collected) by the four institutions mentioned above into education attainment levels and by urban and rural (including the current account balance as the negative areas. Four main institutions (enterprises, households, of savings by the rest of the world) which are made government, and the rest of the world) are distinguished available for investment demand (including change in where households are disaggregated by income levels stocks). Box 1. The implications of different approaches to achieving inter-generational equity A modified version of the PIH would allow for frontloaded investment but might not be advisable for Mozambique. In Mozambique and other developing countries capital is scarcer than labor. If the rate of return on investment in physical and human capital is higher than the rate of return on savings abroad, it may be desirable to scale up such investment in a context of capital scarcity. However, frontloaded investment does not always lead to higher growth. Hence, frontloaded spending must be fully compensated by fiscal adjustment in the medium term. Future fiscal adjustment is needed to rebuild the level of financial wealth to the same level as in the standard PIH framework. If scaling up does have a positive impact on growth, less future fiscal adjustment would be needed, but no such growth impact is assumed ex ante. Given Mozambique’s current need for fiscal consolidation and the fiscal risks associated with borrowing against future revenues, the benefits and risks of this approach would have to be weighed carefully. A comprehensive fiscal sustainability framework (FSF) would allow for a lower level of net financial wealth to enable higher investment in physical and human capital. Frontloaded investment is assumed to have a positive impact on non-resource growth and tax revenues, but also higher operations and maintenance expenditure to maintain a higher capital stock. Financial wealth is stabilized at a lower long-term level than in the PIH or modified PIH frameworks. Instead the country accumulates more physical and human capital. These approaches lead to different conclusions regarding the time profile for fiscal space in Mozambique and highlight the risks related to frontloading investment. Segura-Ubiergo et al. (2013) provide illustrations for the case of Mozambique using the non-resource primary balance as a share of non-LNG GDP as a measure of fiscal space. With the Norwegian approach, fiscal space would increase from zero in the pre-production phase to almost 7 percent in 2050. The PIH provides a constant fiscal space of less than 3 percent. In contrast, the modified PIH and FSF frameworks allow for a gradual increase in fiscal space to almost 8 percent towards the end 2020s followed by a tighter fiscal stance. In the modified PIH framework, it even becomes necessary to have a surplus for some years to rebuild fiscal sustainability following the period of frontloaded investment. This illustrates the longer-term risks involved in expanding public investment in a context of low public investment efficiency. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 113 Table 3. Headline result for the multiplier analysis Direct Employment % of 2019 Direct Employment % of 2019 Version, 2019-2040 Benchmark GDP Version, 2019-2033 Benchmark GDP (in USD million) Local Capital Expenditure 5,347 5,347 Local Operational Expenditure 4,113 2,144 All Local Expenditure 9,460 62.2% 7,491 49.3% Direct and Indirect Gross Value 13,214 10,428 of Production for All Local Expenditure Output Multiplier 1.40 1.39 Direct Impact on GDP 1,176 7.7% 997 6.6% Indirect Impact om GDP 3,848 25.3% 3,065 20.2% Direct and Indirect Impact on GDP 5,025 33.1% 4,062 26.7% GDP Multiplier 0.53 0.54 Impact on Agriculture GDP 984 782 Industry GDP 630 527 Construction GDP 416 396 Transport GDP 731 627 Communication GDP 269 214 Accommodation and Good 396 216 Services GDP Trade Services GDP 442 235 Financial and Business Services 866 719 GDP Services Gross Output 290 346 Household Income 4,058 3,309 Low Income Household Income 809 642 (bottom 60%) High Income Household Income 3,249 2,666 (top 40%) Total tax revenues 1,232 932 (in millions of worker years) Direct Impact on Employment 0.140 0.131 Direct and Indirect Impact on 2.001 1.603 Employment Not completed primary 1.100 0.869 Completed primary 0.600 0.485 Completed secondary 0.194 0.159 Completed tertiary 0.107 0.090 Source: World Bank Staff Estimates. 114 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Annex 5A: Panel Growth Regression Model: Data Sources and Estimates Box 1. Determinants of Growth: Panel regression model The analysis uses a panel regression to explore growth determinants. As described by Islam (1995), panel data methods allow for differences in the aggregate production function, thereby identifying the individual “country effect”. This approach addresses the omitted variable bias problem often detected in the cross-sectional regressions by controlling for the omitted variables that are constant over time in the form of individual effects. The model, covering two time periods between 2008-2012 and 2013-2017 is specified as follows: Where is the growth rate between t-τ (first panel 2008-2012) and t (second panel 2013-2017), is per capita GDP for district i at time t and εi are error terms. The model parameters include an intercept a, dependence on the level in the initial period given by β and influence of selected explanatory variables contained in the matrix, Wi,t-T, denoted by δ. A series of models are estimated that begin with a core set of variables that are correlated with GDP per capita in the initial period. These include labor supply, human capital, physical capital, and access to technology, public investment and private sector activity (model 1). The core model is subsequently augmented to capture the effects of spatial characteristics using binary variables covering the regions; level of urbanization (city, town or rural districts); and variables indicating proximity to extractives, growth corridors or cities (models 2-8). The models are estimated separately for agriculture (AG) and, industry and services (IS). For the AG regressions, only rural districts are used for estimating the models as the analysis primarily seeks to understand the variations between rural districts and impacts of spatial drivers. For the IS regressions, the full set of districts (rural, towns and cities) are all utilized in order to investigate the role structural change. The core variables illustrated in the table above are augmented for AG by also taking weather conditions into account. Electricity is removed from the IS regressions due to the association with night-time lights. Furthermore, the potential impact of structural change on the economy is accounted for by regressing on the share of GDP contributed by AG and IS respectively. Ideally, the analysis would have been supplemented with a fixed-effects panel estimation model to take into account the possibility of unobserved heterogeneity and control for the effect of variables that are invariant over time (Islam, 1995). However, data limitations that restrict panel analysis to two periods constrain this possibility. Due to these data limitations the growth rates are modelled using an ordinary least squares (OLS) regression. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 115 Table 4. Data Description Indicator Description Period Source Panel regression Labour Working age population, which correspond to the total 2007 & 2017 INE: 2007 and 2017 population with ages between 15 and 64 Census Education Grade 7 graduates, which is equivalent to students with 2007 & 2017 Mozambique Ministry of the full primary cycle completed education Roads Length of road per area (km/area) 2012 & 2016 National Roads Administration Mobiles Number of people that own a mobile 2007 & 2017 INE: 2007 and 2017 Census Electricity Access to electricity given by the proportion of total 2007 & 2017 INE: 2007 and 2017 population that use electricity as the source of energy Census Water Access to water given by the proportion of total 2007 & 2017 INE: 2007 and 2017 population that has access to piped water (in the tap and Census yard) PubInv Public Investment 2009-2017 World Bank spatial estimated using Boost PriInv Private Investment 2003-2016 FdIMarkets Rainfall Annual rainfall 2000-2017 University of Delaware Rain shock A rain shock calculated by the following formula: 2000-2017 University of Delaware Rain(2017)/Average Rain (2013-2016)-1 Firms Number of firms 2002/3 & Mozambique Enterprise 2014/5 Census (CEMPRE) Money Agents Number of representants of mobile money agencies 2016-2018 Bank of Mozambique North Districts located in the north region of the country N/A Centre Districts located in the center region of the country N/A City Districts with total population >=50000 and population N/A density > 1500 km^2 Town Districts with total population] 50000;5000] and N/A Population Density ]1500; 250] km^2 Extractives & Districts with operational formal extractive activities and/ N/A World Bank classification Megaprojects or megaprojects based on data from Bank from the Bank of Mozambique and UN Comtrade (https:// comtrade.un.org/) Near Extractives & Districts that border others with operational formal N/A World Bank classification Megaprojects extractive activities and/or megaprojects based on data from Bank from the Bank of Mozambique and UN Comtrade (https:// comtrade.un.org/) Near Cities & Large Districts that border cities and large town following the N/A World Bank Towns classifications provided above On Development District located along one of the main transport corridors N/A World Bank transports unit Corridor (Beira, Nacala, and Maputo) Additional variables used in the Cross section regression PLM Project Last Mile (PLM) index, that categorizes roads 2018 PLM accessibility to health centers using classification from 1 to 3, for higher to lower quality 116 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 5. Panel regression for AG GDP per capita Variable \ Model M1 M2 M3 M4 M5 M6 M7 M8 Intercept -2.99** -2.84** -2.99** -2.95** -2.84** -2.95** -2.92** -2.82** Base -0.66*** -0.63** -0.66*** -0.66** -0.64** -0.66*** -0.67*** -0.64** Labour -0.01 -0.02 -0.01 -0.01 -0.01 -0.04 -0.01 -0.01 Education -0.02 -0.01 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 Roads -0.17 -0.17 -0.17 -0.16 -0.15 -0.13 -0.16 -0.17 Mobiles -0.02 -0.03 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 Electricity 0.03** 0.03* 0.03** 0.03** 0.03** 0.03** 0.03* 0.03** Water 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 PubInv 0.02 0.01 0.02 0.02 0.02 0.02 0.02 0.02 PriInv -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 Firms 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 Rainfall -0.03 -0.04 -0.03 -0.02 -0.03 -0.03 -0.03 -0.02 gAG gIS -0.00 0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 North 0.36 Centre 0.35 City 0.00 Town 0.00 Extractives & -0.19 Megaprojects Near Cities & Large -0.27 Towns Near Extractives & -0.32 Megaprojects On Development -0.16 Corridor Near Development -0.13 Corridor R2 0.08 0.07 0.08 0.07 0.08 0.07 0.07 0.07 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 117 Table 6. Panel regression for IS GDP per capita Variable \ Model M1 M2 M3 M4 M5 M6 M7 M8 Intercept -11.12*** -11.26*** -11.38*** -11.61*** -10.91*** -11.40*** -12.55*** -9.52*** Base -4.83*** -4.64*** -4.84*** -4.95*** -4.79*** -4.85*** -4.96*** -4.95*** Labour -0.60 -0.41 -0.67 -0.61 -0.61 -0.58 -0.66 -0.75* Education -0.18 -0.10 -0.18 -0.18 -0.17 -0.18 -0.20 -0.17 Roads 2.66** 2.30** 2.54** 2.71** 2.65** 2.65** 2.58** 2.45** Mobiles -0.02 0.06 -0.01 -0.04 -0.01 -0.02 -0.02 -0.05 Electricity Water 0.12* 0.16*** 0.12** 0.12** 0.13** 0.12* 0.13** 0.12* PubInv -0.54*** -0.28* -0.54*** -0.54*** -0.52*** -0.54*** -0.52*** -0.44*** PriInv -0.02 -0.01 -0.03 -0.02 -0.02 -0.02 -0.03 -0.04 Firms -0.00 -0.02 -0.01 -0.01 -0.01 -0.01 -0.01 0.00 Rainfall gAG -0.82 -0.82 -0.76 -0.82 -0.89 -0.80 -0.80 -0.86 gIS North -1.87 Centre -7.39*** City -0.79 Town 2.12 Extractives & 1.74 Megaprojects Near Cities & Large -1.76 Towns Near Extractives & 1.03 Megaprojects On Development 1.77 Corridor Near Development -3.76** Corridor R2 0.50 0.55 0.49 0.50 0.50 0.50 0.50 0.52 118 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Annex 5B. Convergence analysis results Box 1. Measuring Growth Convergence Perhaps the simplest measure of convergence, known as σ-convergence, seeks to understand whether an income gap between economic zones is decreasing over time. One approach to test for σ-convergence is to analyse the coefficient of variation (CV), which is defined as standard deviation divided by the mean for GDP per capita (Gömleksiz et al., 2017). A decline in the CV suggests that the spread of the income distribution is shrinking. Another approach is to test for β-convergence, which seeks to quantify whether or not an economy is converging towards its steady state income level. Following Barro and Sala-i-Martin (2004) and Gömleksiz et al. (2017), β-convergence can be tested by running the following regression: where is the speed of convergence. If convergence occurs (α1 < 0) then higher initial income levels have a negative effect on final growth. Thus, β measures the annual convergence rate of an economy towards its steady state income level. The Gini index is the third measure applied to gage statistical dispersion. It is often applied to the income or wealth distribution of a nation’s residents, thereby offering a means of quantifying inequality (Gini, 1909). Inequality on the Gini scale is measured between zero, where everybody is equal, and one, where all the country’s income is earned by a single person. An alternative approach to testing for convergence is therefore to estimate the Gini index for the distributions of GDP by district for the two panel periods. Tests of σ-convergence and β-convergence were undertaken for all sectors and using all districts and only rural districts. Evidence for β-convergence was found in all cases apart from agriculture (all districts). In contrast, σ-convergence was not found for agriculture (all districts and rural districts) and excluding extractives (rural districts). Progress towards improving equality (a decreasing Gini coefficient) was found for industry & services and excluding extractives (all districts) and for industry & services (rural districts). Table 7. Testing for σ-convergence by sector using all and rural districts Districts Sector CV 2008-2012 CV 2013-2017 Change All Agriculture 0.378 0.383 1.49% All Industry & Services 1.577 1.352 -14.28% All Excluding Extractives 0.984 0.924 -6.10% Rural Agriculture 0.211 0.216 2.17% Rural Industry & Services 1.629 1.445 -11.32% Rural Excluding Extractives 0.884 0.906 2.49% MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 119 Table 8. Testing for β-convergence by sector using all and rural districts. Districts Sector α0 α1 β R2 All Agriculture 0.0252 0.0037 -0.36% 5.1% All Industry & Services -0.1409 -0.0453 5.13% 35.3% All Excluding Extractives -0.0079 -0.0108 1.11% 4.2% Rural Agriculture -0.0238 -0.0063 0.64% 6.0% Rural Industry & Services -0.1424 -0.0457 5.19% 36.0% Rural Excluding Extractives 0.0145 -0.0061 0.62% 1.1% Table 9. Analysis of Gini indices by sector using all and rural districts Districts Sector Gini 2008-2012 Gini 2013-2017 Change R2 All Agriculture 0.44 0.47 5.1% 5.1% All Industry & Services 0.73 0.69 -5.1% 35.3% All Excluding Extractives 0.55 0.53 -4.2% 4.2% Rural Agriculture 0.34 0.36 7.4% 6.0% Rural Industry & Services 0.58 0.56 -4.2% 36.0% Rural Excluding Extractives 0.35 0.36 3.3% 1.1% 120 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Annex 5C. Consumption/ poverty elasticities to growth Table 10. Consumption/poverty elasticities to growth Income/Consumption Income/Poverty 09-03 15-09 15-03 09-03 15-09 15-03 Cabo Delgado -0.5 0.4 0.0 0.4 -0.2 0.0 Cidade de Maputo 7.1 6.8 8.6 1.4 -3.7 -2.7 Gaza 0.5 1.2 0.7 -0.2 -0.6 -0.3 Inhambane 0.5 1.9 0.9 -0.2 -1.0 -0.3 Manica 2.1 2.3 2.4 -1.1 -1.5 -1.1 Maputo -1.0 20.3 11.1 5.3 -8.5 -4.6 Nampula 2.3 -1.0 0.6 -0.3 0.3 -0.1 Niassa 1.9 -1.1 0.2 -0.7 1.4 0.0 Sofala -0.8 2.3 0.2 1.2 -1.0 0.1 Tete 1.0 1.4 1.2 -0.4 -0.9 -0.4 Zambezia -0.1 2.1 0.7 0.5 -0.9 0.0 National 1.29 3.44 2.60 -0.08 -0.69 -0.30 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 121 Annex 5D. Urbanization data Table 11. Districts per Urban Level Classification Total Number Districts Cities 6 Cidade Da Matola, Cidade De Chimoio, Cidade De Maputo, Cidade De Nampula, Cidade De Pemba and Cidade De Quelimane Town 10 Boane, Cidade Da Beira,Cidade De Inhambane,Cidade De Lichinga Cidade de Maxixe, Cidade de Nacala-Porto,Cidade De Tete, Cidade De Xai-Xai, Ilha De Mocambique and Marracuene Rural 126 Alto Molocue, Ancuabe, Angoche, Angonia, Balama, Barue, Bilene, Buzi, Cahora Bassa, Caia, Changara, Chemba, Cheringoma, Chibabava, Chibuto, Chicualacuala, Chifunde, Chigubo, Chinde, Chiure, Chiuta, Chokwe, Cuamba, Dondo, Erati, Funhalouro, Gile, Gondola, Gorongosa, Govuro, Guija, Guro, Gurue, Homoine, Ibo, Ile, Inharrime, Inhassoro, Inhassunge, Jangamo, Lago, Lalaua, Lichinga, Lugela, Mabalane, Mabote, Macanga, Machanga, Machaze, Macomia, Macossa, Maganja Da Costa, Magoe, Magude, Majune, Malema, Mandimba, Mandlakaze/Manjacaze, Manhica, Manica, Maravia, Maringue, Marromeu, Marrupa, Massangena, Massinga, Massingir, Matutuine, Maua, Mavago, Mecanhelas, Meconta, Mecuburi, Mecufi, Mecula, Meluco, Memba, Metarica, Milange, Moamba, Moatize, Mocimboa Da Praia, Mocuba, Mogincual, Mogovolas, Moma, Monapo, Montepuez, Mopeia, Morrumbala, Morrumbene, Mossuril, Mossurize, Muanza, Muecate, Mueda, Muembe, Muidumbe, Murrupula, Mutarara, Nacala-A-Velha, Nacaroa, Namaacha, Namacurra, Namarroi, Nampula, Namuno, Nangade, Ngauma, Nhamatanda, Nicoadala, Nipepe, Palma, Panda, Pebane, Pemba, Quissanga, Ribaue, Sanga, Sussundenga, Tambara, Tsangano, Vilankulo, Xai-Xai, Zavala, Zumbu Table 12. Growth node districts defined as rural districts with IS GDPpc greater than 20% District Province Region Growth Growth AG GDPpc IS GDPpc AG GDPpc IS GDPpc (2008-2012) (2008-2012) Panda Inhambane S 1.2% 87.5% 8.1 0.4 Maua Niassa N 1.1% 64.8% 8.3 1.0 Nacaroa Nampula N 1.1% 47.6% 8.0 0.3 Lichinga Niassa N 1.3% 47.4% 7.8 0.0 Metarica Niassa N 1.1% 47.1% 8.3 0.9 Mecuburi Nampula N 0.8% 43.4% 7.6 0.1 Mecanhelas Niassa N 1.0% 41.9% 7.4 0.3 Meluco Cabo Delgado N 1.0% 39.8% 8.2 0.9 Chigubo Gaza S 1.3% 36.9% 7.7 0.7 Quissanga Cabo Delgado N 1.2% 36.3% 7.5 0.3 Muidumbe Cabo Delgado N 0.8% 36.2% 8.1 0.4 Sanga Niassa N 1.6% 36.2% 7.2 0.2 Massangena Gaza S 0.1% 36.1% 5.3 0.5 Nangade Cabo Delgado N -0.8% 35.8% 8.0 1.8 Ibo Cabo Delgado N 3.8% 35.3% 0.2 0.7 Palma Cabo Delgado N 0.2% 31.5% 8.5 2.3 Macomia Cabo Delgado N 1.3% 27.2% 7.4 3.2 Chiuta Tete C 1.2% 27.2% 8.1 1.3 Balama Cabo Delgado N 0.8% 25.7% 6.5 0.2 Marrupa Niassa N -5.5% 21.7% 8.2 2.6 Ancuabe Cabo Delgado N 1.3% 21.4% 7.2 1.6 Pemba Cabo Delgado N 1.0% 20.5% 7.6 2.2 122 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Annex 6. Table 13. Manufacturing’s Links to the Economy are Weaker than Services Manufacturing sector: combined backward and forward economy-wide SAM multipliers for M1 million increase in exogenous demand Sub-Sector Output Rural employment Urban employment GDP Comb. Bwd Fwd Comb. Bwd Fwd Comb. Bwd Fwd Comb. Bwd Fwd Overall 1.46 1.56 1.36 1.80 2.86 0.74 1.57 1.89 1.26 0.65 0.70 0.59 Beverages & tobacco 1.41 1.72 1.10 1.02 1.92 0.12 1.07 1.76 0.38 0.44 0.62 0.27 Cassava 1.36 1.42 1.31 12.60 12.62 12.59 3.55 3.63 3.48 0.90 0.93 0.88 Cattle 1.19 1.16 1.23 7.43 7.45 7.41 1.20 1.16 1.24 0.92 0.91 0.93 Cereal and veg. processing 1.19 1.20 1.19 1.08 1.66 0.49 0.54 0.66 0.42 0.51 0.53 0.48 Chemicals 1.73 1.50 1.97 0.44 0.23 0.66 1.34 1.17 1.51 0.65 0.56 0.74 Clothing 1.24 1.42 1.06 0.31 0.30 0.31 0.90 1.18 0.62 0.59 0.67 0.52 Coal and lignite 1.30 1.30 1.30 0.13 0.18 0.08 0.45 0.61 0.28 0.67 0.70 0.65 Coffee and tea 2.02 1.79 2.25 2.28 3.08 1.49 1.87 1.68 2.06 0.55 0.54 0.56 Cotton and fibres 1.70 1.14 2.25 7.23 6.73 7.74 2.16 1.95 2.36 1.18 0.96 1.40 Fishing 1.23 1.45 1.00 1.25 1.43 1.07 0.88 1.29 0.46 0.52 0.63 0.42 Forestry 1.18 1.22 1.14 7.11 7.12 7.10 1.65 1.74 1.56 0.78 0.80 0.76 Fruits and nuts 1.04 1.05 1.02 10.79 10.80 10.79 2.07 2.09 2.04 0.95 0.96 0.94 Groundnuts 1.04 1.08 1.01 14.53 14.54 14.51 2.79 2.84 2.73 0.92 0.93 0.90 Leather and footwear 1.17 1.34 1.01 0.82 1.63 0.01 1.81 2.02 1.60 0.79 0.91 0.66 Machinery and equipment 1.39 1.56 1.23 0.21 0.27 0.16 0.68 0.91 0.45 0.57 0.64 0.49 Maize 1.05 1.05 1.05 13.64 13.59 13.68 3.73 3.74 3.72 0.98 0.98 0.99 Meat 1.22 1.31 1.13 1.45 2.25 0.64 1.58 1.71 1.45 0.86 0.94 0.77 Metals and metal products 1.44 1.72 1.16 0.18 0.30 0.06 0.60 1.01 0.19 0.45 0.61 0.29 Natural gas 1.13 1.25 1.00 0.05 0.10 0.00 0.21 0.41 0.01 0.71 0.77 0.64 Non-metal minerals 1.71 1.32 2.09 0.38 0.26 0.51 0.79 0.58 1.01 0.86 0.76 0.95 Other cereals 1.61 1.03 2.18 13.63 13.39 13.86 3.88 3.70 4.06 1.17 0.94 1.40 Other crops 1.15 1.11 1.20 12.51 12.48 12.53 3.49 3.43 3.56 0.95 0.94 0.96 Other foods 1.52 1.66 1.38 2.63 4.30 0.97 2.42 3.14 1.70 0.75 0.90 0.60 Other livestock 1.11 1.08 1.15 7.76 7.76 7.76 1.13 1.08 1.17 0.97 0.96 0.99 Other manufacturing 1.29 1.49 1.10 0.19 0.26 0.12 0.87 1.16 0.57 0.59 0.69 0.49 Other mining 1.47 1.33 1.60 3.39 3.41 3.37 2.43 2.53 2.32 0.69 0.67 0.71 Other oilseeds 1.27 1.28 1.26 12.83 12.85 12.82 3.77 3.84 3.70 0.89 0.90 0.88 Other roots 1.03 1.06 1.01 10.47 10.48 10.46 2.89 2.93 2.86 0.93 0.94 0.92 Poultry 1.20 1.28 1.12 6.82 7.51 6.14 1.14 1.36 0.93 0.84 0.90 0.78 Pulses 1.02 1.03 1.01 12.14 12.15 12.14 2.31 2.33 2.29 0.95 0.96 0.95 Rice 1.63 1.33 1.93 12.98 12.85 13.10 3.65 3.57 3.73 1.07 0.95 1.19 Sorghum and millet 1.02 1.03 1.01 14.53 14.53 14.52 3.97 3.99 3.96 0.97 0.98 0.97 Sugar cane 1.71 1.05 2.38 11.89 11.42 12.36 3.99 3.18 4.80 1.28 0.99 1.57 Textiles 1.44 1.70 1.18 2.19 3.32 1.05 1.07 1.60 0.54 0.68 0.87 0.48 Tobacco 1.15 1.05 1.25 13.24 13.23 13.24 3.68 3.68 3.69 1.01 0.99 1.03 Vegetables 1.01 1.03 1.00 15.19 15.20 15.19 2.88 2.90 2.86 0.97 0.97 0.96 Wood and paper 1.64 1.39 1.89 0.72 0.76 0.67 1.64 1.31 1.97 0.87 0.74 0.99 Source: van Seventer (2019). Note: SAM: Social accounting matrix. Averages for services and manufacturing output are weighted by output. Average employment multipliers are weighted by their respective employment shares in manufacturing employment. Average GDP multipliers are weighted by the shares in manufacturing GDP. All multipliers indicators exclude the household income-expenditure loop. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 123 Annex 7. Impacts of Corruption on Firm Growth Models (1) and (3) use year-location-industry averages as an instrument. Models (2) and (4) change the instrument to a dummy variable indicating whether there are women among the firm owners. An F-test suggests both are strong instruments, meaning they have a high correlation with BRIBE_PER.143, Our preference are models (1) and (3), as they produce better test results. Table 13. Estimated effect of corruption on firm growth, main analysis Sales growth Productivity growth Model (1) Model (2) Model (3) Model (4) BRIBE_PER -0.055*** -0.027* -0.088*** -0.043** (0.020) (0.016) (0.024) (0.017) LOG_SALES -0.079*** -0.075*** -0.102*** -0.094*** (0.010) (0.009) (0.012) (0.009) LOG_EMPLOYEES 0.131*** 0.132*** 0.160*** 0.161*** (0.017) (0.015) (0.021) (0.017) LOG_YEARS_OPERATIONAL 0.011 -0.005 0.037 0.011 (0.025) (0.022) (0.031) (0.024) TRADE_BIN 0.050 0.010 0.129** 0.063 (0.050) (0.043) (0.062) (0.047) FOREIGN_OWNERSHIP_BIN 0.002*** 0.001*** 0.002*** 0.001** (0.001) (0.001) (0.001) (0.001) Constant 1.008*** 0.955*** 1.213*** 1.125*** (0.123) (0.108) (0.151) (0.117) N 969 969 969 969 R2 -0.261 0.008 -0.779 -0.096 Adjusted R2 -0.269 0.001 -0.790 -0.103 *p < .1; **p < .05; ***p < .01 Notes: Models (1) and (3) use year-location-industry averages as an instrument. Models (2) and (4) use a dummy variable indicating whether there are women among the firm owners. 143 We also experimented with other instruments, including the share of firm ownership held by women and an ordinal variable indicating whether firms trust the courts. An F-test indicates both are weak instruments. 124 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 14. Estimated effect of corruption on firm growth, excluding outliers Sales Growth Productivity Growth (1) (2) (3) (4) BRIBE_PER -0.042*** -0.065*** -0.076*** -0.105*** (0.016) (0.024) (0.021) (0.029) LOG_SALES -0.042*** -0.079*** -0.072*** -0.101*** (0.008) (0.010) (0.011) (0.012) LOG_EMPLOYEES 0.072*** 0.118*** 0.112*** 0.138*** (0.014) (0.018) (0.019) (0.022) LOG_YEARS_OPERATIONAL 0.007 0.008 0.028 0.032 (0.020) (0.024) (0.027) (0.030) TRADE_BIN 0.040 0.061 0.113** 0.142** (0.042) (0.052) (0.054) (0.063) FOREIGN_OWNERSHIP_BIN 0.001** 0.002*** 0.002*** 0.003*** (0.001) (0.001) (0.001) (0.001) Constant 0.570*** 1.024*** 0.870*** 1.236*** (0.103) (0.124) (0.137) (0.151) N 946 957 949 957 R2 -0.261 0.008 -0.779 -0.096 Adjusted R2 -0.269 0.001 -0.790 -0.103 *p < .1; **p < .05; ***p < .01 Notes: Model (1) eliminate observations that are more than three standard deviations above or below the mean of SALES_GROWTH. Model (3) eliminate observations that are more than three standard deviations above or below the mean of SALES_GROWTH. Models (2) and (4) eliminate observations that are more than three standard deviations above the mean of BRIBE_PER. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 125 Table 15. Sample distribution across firm location, size, and industry Firm location Firm size Firm industry Total Manufacturing Retail services Other services Maputo (greater) Large 33 11 13 57 Medium 136 28 35 199 Small 228 94 33 355 Micro 40 80 19 139 Sofala Large 8 0 16 24 Medium 19 3 11 33 Small 46 10 12 68 Micro 8 23 9 40 Manica Large 3 1 3 7 Medium 7 1 3 11 Small 6 16 7 29 Micro 3 8 3 14 Tete Large 1 1 10 12 Medium 7 2 4 13 Small 6 2 4 12 Micro 2 8 2 12 Zambezia Large 1 0 1 2 Medium 4 4 12 20 Small 8 6 11 25 Micro 3 6 6 15 Nampula Large 15 0 6 21 Medium 15 4 5 24 Small 47 13 21 81 Micro 13 6 11 30 Cabo Delgado Large 2 0 4 6 Medium 4 4 10 18 Small 4 8 16 28 Micro 4 8 4 16 Total 673 347 291 1311 126 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Box 1. Econometric specifications Addressing endogeneity bias There are two problems when estimating the effect of corruption on firm performance: (i) endogeneity and (ii) measurement error.a Public officials can ‘dose’ the amount of harassment. This means two firms in the same location and industry may have to pay different bribes for the same ‘service’, and the difference may be correlated with (unobservable) factors influencing growth. If public officials target well performing firms, with the expectation that they can (and are willing to) pay bigger bribes, this will cause endogeneity. Following the literature (Fisman and Svensson, 2007; Mendoza et al., 2015), the analysis uses instrumental variables (IV) estimation. Year-location-industry means is used as an instrument.b This is based on the idea that the ‘default’ bribes that firms pay is determined by year, location, and industry characteristics—the capacity of public officials in a specific location to extract bribes and the extent to which firms in a specific industry require licenses, permits, and services. If this assumption holds, year-location-industry will be a valid instrument as it is exogenous to the firms and uncorrelated with firm performance. As a robustness test, we include female firm co-ownership as an additional instrument. Female firm co-ownership has been used in other studies of the effect of corruption on firm performance (Williams and Kedir, 2016) and is based on the idea that women are less likely to engage in risky behaviour (Black and Lynch, 1996). Data from the 2018 Mozambique Enterprise Survey support this argument. Firms without female co-ownership are 24 percent more likely to pay bribes than firms with female co-ownership, and 44 percent more than more likely to pay bribes than firms with full female ownership. Model Variables Two alternative dependent variables are used: sales growth and productivity (sales per worker) growth (Fisman and Svensson, 2007; Goedhuys et al., 2016). The analysis uses average annual growth over a two-year period.c Although we would ideally look at growth over a longer period, this was not possible due to data limitations. The analysis includes the same control variables that are used in other, similar studies (see Box 1). Since firm size may be correlated with bribe payments (larger firms may be more visible to corrupt public officials) and may also affect future growth, we include the natural logarithms of firm sales and the number of employees in the last complete financial year minus two (T-2). We use firm sales and the number of employees in T-2, treating it as a signal to corrupt public officials how much they can extract in T. Similarly, we include the number of years the firm has been operational, which many previous studies have found to be correlated with growth, and may be correlated with bribes if older firms have better access to finance and better networks within government. Firms involved in trade, either exporting or importing, may be more vulnerable for rent extraction and subject to greater bureaucratic scrutiny and regulation than firms with only local sales. The analysis includes a dummy variable that reflects whether a firm either exports or imports directly (TRADE_BIN). Finally, we include a variable representing the share of firm ownership held by foreigners, either firms or individuals. Such firms may grow more quickly due to greater resources, access to markets, and technical expertise, while they may be exempt from bureaucratic harassment as an inducement to locate their operation in Mozambique. a. Social desirability bias is the bias that emerges when “survey respondents underreport socially undesirable activities and overreport socially desirable ones” (Krumpal, 2013, p. 2025). It has been noted in the context of research on violence, drug abuse, prostitution, and corruption. b. Using year-location-industry means as an instrument addresses a second problem: measurement error. IV estimation produce consistent estimates even in the presence of measurement error, provided the instrument is valid, meaning it is correlated with the correctly measured explanatory variable but uncorrelated with the measurement error and the estimation error (i.e. from the model with the correctly measured data) (Krueger & Angrist, 2001). Year-location-industry means appear to meet these criteria. They are correlated with the probability that firms pay bribes, as we saw in section 4.1, but uncorrelated with the probability that they lie about their corrupt payments. c. In fact, for 2007, the variables measure the average annual growth over a three-year period, given how the questions are framed. Accordingly, we divide not by three instead of two. MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 127 Table 16. Estimated effect of key firm characteristics on the size of bribes paid Dependent variable: The share of total annual sales that firms spend on bribes (1) (2) SALES_LOG -0.010 -0.041 (0.056) (0.052) EMPLOYEES_LOG -0.011 0.011 (0.111) (0.104) YEARS_OPERATIONAL_LOG 0.285** 0.359*** (0.112) (0.105) FOREIGN_OWNERSHIP 0.004* 0.005** (0.002) (0.002) INDUSTRY_RETAIL -1.381*** -1.352*** (0.250) (0.243) INDUSTRY_OTHER -0.527** -0.563*** (0.215) (0.210) REGION_NORTH -0.782*** -0.790*** (0.282) (0.266) REGION_CENTRE 0.211 0.232 (0.276) (0.258) 2018 0.348 (0.328) Constant -2.295*** -2.429*** (0.749) (0.682) N 81 148 AIC 330.889 398.375 *p < .1; **p < .05; ***p < .01 Notes: Model 1 is limited to 2018 data. Model 2 pools 2018 data with data from 2007 and introduces a year dummy variable. The variables include: the natural logarithm of total annual sales in the last financial year (SALES_LOG); the natural logarithm of the total number of employees in the last financial year (EMPLOYEES_LOG); the natural logarithm of the total number of years the firm has been operational (YEARS_OPERATIONAL_LOG); the share of firm ownership that is held by foreigners, whether firms or individuals (FOREIGN_OWNERSHIP); industry dummy variables, using manufacturing as the reference category; region dummy variables, using the South as the reference category; and a year dummy variable, using 2007 as the reference category. 128 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 17. Estimated effect of corruption on firm growth, robust standard errors Sales growth Productivity growth (1) (2) (3) (4) (5) (6) BRIBE_PER -0.055* -0.055 -0.055 -0.088** -0.088** -0.088* (0.029) (0.039) (0.036) (0.041) (0.042) (0.051) LOG_SALES_TMIN -0.079*** -0.079*** -0.079*** -0.102*** -0.102*** -0.102*** (0.021) (0.010) (0.031) (0.022) (0.009) (0.034) LOG_EMPLOYEES_TMIN 0.131*** 0.131*** 0.131*** 0.160*** 0.160*** 0.160*** (0.027) (0.012) (0.032) (0.030) (0.028) (0.032) LOG_YEARS_OPERATIONAL 0.011 0.011 0.011 0.037 0.037** 0.037 (0.049) (0.029) (0.038) (0.056) (0.017) (0.052) TRADE_BIN 0.050 0.050 0.050 0.129 0.129 0.129 (0.092) (0.050) (0.099) (0.192) (0.080) (0.185) FOREIGN_OWNERSHIP_BIN 0.002* 0.002* 0.002 0.002* 0.002 0.002 (0.001) (0.001) (0.002) (0.001) (0.002) (0.002) Constant 1.008*** 1.008*** 1.008*** 1.213*** 1.213*** 1.213*** (0.271) (0.123) (0.385) (0.220) (0.111) (0.388) N R2 Adjusted R2 *p < .1; **p < .05; ***p < .01 Notes: Models (1) and (4) allow for clustering by province; models (2) and (5) allow for clustering by industry; models (3) and (6) allow for clustering by province-industry Table 18. Estimated effect of corruption on firm growth, additional control variables (1) Sales Growth Productivity Growth (1) (2) (3) (4) BRIBE_BIN -0.030* -0.052*** -0.065*** -0.079*** (0.018) (0.019) (0.022) (0.022) Sales (ln) -0.076*** -0.079*** -0.099*** -0.102*** (0.009) (0.010) (0.011) (0.011) Employees (ln) 0.117*** 0.133*** 0.152*** 0.166*** (0.016) (0.017) (0.019) (0.021) Years_Operational (ln) 0.004 0.010 0.038 0.034 (0.023) (0.024) (0.030) (0.029) TRADE_BIN -0.017 0.044 0.082 0.111* (0.043) (0.049) (0.052) (0.058) Foreign_Ownership_BIN 0.002*** 0.002*** 0.002*** 0.002*** (0.001) (0.001) (0.001) (0.001) Manager_Years_Education (ln) 0.169*** 0.102* (0.048) (0.057) Manager_Years_Experience (ln) 0.009 -0.015 (0.028) (0.040) 2018 0.036 0.101 (0.071) (0.084) Constant 0.719*** 0.967*** 1.044*** 1.099*** (0.131) (0.139) (0.156) (0.164) N 956 969 956 969 R2 0.003 -0.221 -0.374 -0.606 Adjusted R2 -0.006 -0.229 -0.385 -0.618 *p < .1; **p < .05; ***p < .01 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 129 Table 19. Estimated effect of corruption on firm growth, additional control variables (2) SALES GROWTH PRODUCTIVITY GROWTH (1) (2) (3) (4) BRIBE_BIN -0.057** -0.049*** -0.097*** -0.073*** (0.023) (0.017) (0.030) (0.020) LOG_SALES_TMIN -0.081*** -0.079*** -0.107*** -0.102*** (0.011) (0.009) (0.014) (0.011) LOG_EMPLOYEES_TMIN 0.133*** 0.115*** 0.167*** 0.147*** (0.018) (0.018) (0.023) (0.021) LOG_YEARS_OPERATIONAL 0.015 0.031 0.051 0.057* (0.027) (0.026) (0.035) (0.030) TRADE_BIN 0.052 0.045 0.139** 0.108** (0.054) (0.047) (0.069) (0.054) FOREIGN_OWNERSHIP_BIN 0.002*** 0.002*** 0.002*** 0.002*** (0.001) (0.001) (0.001) (0.001) REGION_CENTRE 0.010 0.041 (0.038) (0.049) REGION_NORTH 0.028 0.120* (0.050) (0.064) INDUSTRY_RETAIL_SERVICES 0.007 0.055 (0.043) (0.050) INDUSTRY_OTHER_SERVICES 0.157*** 0.232*** (0.051) (0.059) Constant 1.004*** 0.926*** 1.193*** 1.064*** (0.134) (0.119) (0.171) (0.137) N 969 969 969 969 R2 -0.289 -0.160 -0.955 -0.446 Adjusted R2 -0.299 -0.170 -0.971 -0.458 *p < .1; **p < .05; ***p < .01 130 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM Table 20. Estimated effect of firm size on perceptions of corruption as an obstacle Model (1) Model (2) Model (3) Sales (ln) 0.087*** 0.151*** 0.151*** (0.025) (0.034) (0.034) Employees (ln) -0.122** 0.492** 0.508** (0.053) (0.231) (0.232) Years_Operational (ln) 0.381*** 0.379*** 0.378*** (0.051) (0.051) (0.051) Industry_Retail (0= Industry_Manufacturing) 0.250* 0.298** 0.266* (0.129) (0.129) (0.137) Industry_Other (0= Industry_Manufacturing) 0.225* 0.260** 0.227* (0.129) (0.130) (0.137) Region_Central (0=Region_South) 0.040 0.020 0.014 (0.095) (0.095) (0.095) Region_North (0=Region_South) -0.187** -0.218** -0.223** (0.089) (0.089) (0.089) Foreign_Ownership 0.010*** 0.010*** 0.010*** (0.001) (0.001) (0.001) 2018 (0=2007) 0.162 Sales(ln) X Employees (ln) -0.038*** -0.038*** (0.014) (0.014) Constant -0.652** -1.647*** -1.783*** (0.321) (0.485) (0.520) N 1,176 1,176 1,176 R2 0.146 0.152 0.152 Adjusted R2 0.141 0.145 0.145 *p < .1; **p < .05; ***p < .01 Note: The analysis uses two different variables as our main regressors: the natural logarithms of firm sales (SALES_LOG) and the number of employees (EMPLOYEES_LOG) in the last financial year. All models all OLS models. Model 1 is the baseline model without the interaction term. Models 2 and 3 introduces the interaction term and includes the base terms of SALES_LOG and EMPLOYEES_LOG, following Brambor Clark, and Golder (2006). MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM 131 > > > Figure 1. Simulated e ect on perceptions 4 Perception of corruption as an abstacle 3 2 1 0 0 1 2 3 4 5 Sales in last financial year (billion MZN) Employees 1 5 20 100 1051 Notes: The figure plots the e ect of Sales(ln) for di erent levels of Employees_Log Coe cients extracted from model 2 in Table 12, which produces the best model fit. 132 MOZAMBIQUE COUNTRY ECONOMIC MEMORANDUM