Ethiopia’s great transition: The next mile A Country Economic Memorandum Ethiopia Country Economic Memorandum Report No: AUS0002952 . Ethiopia’s great transition: the next mile Ethiopia - Country Economic Memorandum . June 17, 2022 . Macroeconomics, Trade, and Investment, Africa East Region The World Bank Group . 1 Ethiopia Country Economic Memorandum © 2022 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 23 22 21 20 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive directors, 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. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions Attribution—Please cite the work as follows: World Bank. 2022. Ethiopia’s great transition: the next mile. A Country Economic Memorandum Washington, DC: World Bank. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third- party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Example of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. Cover and interior art: Design and Development Minds. 2 Ethiopia Country Economic Memorandum Acknowledgements This study was prepared by a multisectoral team led by Miguel Eduardo Sanchez Martin (Senior Country Economist and Task Team Leader) under the guidance of Mathew Verghis (Practice Manager). The team is grateful for guidance and support from Ousmane Dione (Country Director), Asad Alam (EFI Regional Director), Marius Vismantas (Program Leader), and John Litwack (Lead Economist). The report benefited from excellent peer reviewer comments by Tom Bundervoet, Nora Carina Dihel, and Michael Geiger. Contributing authors by chapter are as follows: Chapter 1 on growth and convergence was prepared by Patrick McSharry, Konstantin Wacker, Gabriela Schmidt, Mike Nyawo, Samuel Mulugeta, Frehiwot Worku Kebede, and Miguel Sanchez Martin. Benjamin Stuart, Brian Blankespoor, Manex Yonis, Senidu Fanuel, Wondimagegn Mesfin Tesfaye, and Zerihun Getachew provided extremely valuable inputs and feedback. Chapter 2 on labor mobility was prepared by Obert Pimhidzai, Berhe Mekonnen, Mike Nyawo and by Justin Van Dijk. It benefited from inputs and feedback by Christina Wieser, Vikas Choudhary, Welela Ketema, Biruktayet Assefa Betremariam, Koen Maaskant, Victoria Strokova, and Stephen Muzira. Chapter 3 on manufacturing productivity was prepared by Hibret Belete Maemir and Kaleb Girma Abreha, and it benefited from contributions by Koen Maaskant, Mengistu Bessir Achew, Fantu Farris Mulleta, and Anuradha Ray. Mike Nyawo and Frehiwot Fanta assembled the firm-level panel dataset. Professor Margaret McMillan (Tufts University) and Mia Ellis (International Food Policy Research Institute) generously provided feedback and cross-checked the results of the productivity analysis with those in their latest research. Chapter 4 on trade was prepared by Ben Shepherd, Michele Ruta, and Mike Nwayo, with contributions from Paul Brenton, Charles Kunaka, Ankur Huria, Tilahun Esmael Kassahun, and Miguel Sanchez Martin. Chapter 5 on agriculture prices and policies was prepared by John Keyser and Zerihun Getachew Kelbore, with inputs from Wondimagegn Mesfin Tesfaye. It also benefited from the feedback provided by Vikas Choudhary, Elliot Mghenyi, Biruktayet Assefa Betremariam, and Manex Bule Yonis. The econometric analysis of the impacts of inflation was led by Professor Léonce Ndikumana (University of Massachusetts at Amherst) and Janvier Nkurunziza (UNCTAD). Chapter 6 on green resilience was prepared by Pablo Benitez, James Blignaut, Dawit Woubishet Mulatu, Tuukka Castrén, Jian Xie, and Ruxandra Floroiu. It benefited from contributions by Kanta Kumari, Yabei Zhang, Evariste Rutebuka, Shewakena Aytenfisu Abab, Diego Herrera, Sofia Elisabet Ahlorth, and Ross Hugues. Minna Hahn Tong provided editorial review, and Yohana Girma Wudneh provided support to the team. A word of appreciation to the Development Partners that contributed to the initial discussions on the concept and topics of the report: Mark George, JP Fanning, Fergus McBean (FCDO UK); Anne Bridget Masson, Lauriane Houbin (Embassy of France); Balineau Gaelle, Remy Ruat, Benjamin Amara (AfD); Christof Weigelmeier (Embassy of Germany); Diana Hedrich, Binyam Tadesse (KfW); Paul Mpuga (AfDB); Weyinmi Omamuli (UNDP); Giulia Zanvettor (EU); Fanaye Tadesse (UNICEF). During report preparation the team benefited from exchanges with the International Monetary Fund (Jules Leichter, Rahul Giri), Harvard Center for International Development (Tim O’Brien, Tim Cheston), and the International Growth Centre (Kathie Krumm, Seneshaw Tamru). The team is grateful to the Central Statistical Agency of Ethiopia for kindly facilitating access to the different waves of the Manufacturing Business Survey and the agriculture surveys. The team would also like to acknowledge numerous government officials for providing invaluable comments and suggestions during the thematic validation discussions: 3 Ethiopia Country Economic Memorandum Ministry of Finance: Ato Abebe Tadesse, Ato Jonse Gedefa, Ato Mezgebu Amha, W/z Abeba Alemayehu, Ato Brook Taye, Ato Andinet Tadesse, Ato Henok Ketema, Ato Tigistu Sebsibe, Ato Fitsum Abreham, Ato Adugna Nemera. National Bank of Ethiopia: Ato Melesse Tashu. Ministry of Planning and Development: Ato Mohammed Abas. Ethiopia Jobs Commission: Ato Dawit Dame, Ato Selamawit Adnew, Ato Tiumezgi Fikadu, Ato Alemtsehay Legesse. Ministry of Industry: Ato Paulos Berga, Ato Admasu Yifru, W/z Maryamawit Engdawork. Ministry of Trade and Regional Integration: W/z Yodit Alemayehu. Ethiopia Investment Commission: Ato Daniel Teressa. Ministry of Revenue: Ato Fekadu Bekele, Ato Kassaye Ayele. Ministry of Agriculture: Ato Esayas, Ato Sileshi Bekele, Ato Wondale Habtamu, Ato Germame Garuma, Ato Elias Awol, Ato Tefera Tadesse, Ato Berhanu Assefa. Agriculture Transformation Agency: Ato Eshetayehu Tefera, Ato Techane Adugna, Ato Manfredo Nigussie. Environmental Protection Agency: Prof. Fekadu Beyene, W/z Kibebework Getachew, Dr. Adefires Worku, Ato Muluneh G. Hedeto, Ato Bemnet Teshome, Ato Neguss Gebre. Ethiopian Biodiversity Institute: Dr. Melesse Maryo, Ato Tariku Geda. Ethiopian Wildlife Conservation Authority: Ato Kumera Wakjira. 4 Ethiopia Country Economic Memorandum Contents Acknowledgements ..................................................................................................................... 3 Acronyms .................................................................................................................................... 7 Part I: Inclusion ............................................................................................................................... 9 1 Setting the stage: growth and convergence in Ethiopia ........................................................ 10 1.1 Motivation: what went off-track? ................................................................................... 10 1.2 Context: robust growth at the cost of widening macroeconomic imbalances and recent shocks that threaten reform gains ............................................................................................. 11 1.3 The drivers of growth in Ethiopia .................................................................................. 14 1.4 Subnational economic activity and convergence ........................................................... 18 1.5 Conclusion and policy implications ............................................................................... 22 1.6 References ...................................................................................................................... 24 2 Improving access to off-farm jobs and labor mobility in Ethiopia. ...................................... 26 2.1 Motivation: what went off-track? ................................................................................... 26 2.2 Economic growth and the rural-urban divide ................................................................. 26 2.3 Rural off-farm employment and spillover effects of urban areas .................................. 29 2.4 Rural-urban migration: benefits, determinants, and constraints .................................... 41 2.5 Policy options ................................................................................................................. 48 2.6 References ...................................................................................................................... 53 Part II: Competitiveness ................................................................................................................ 55 3 Productivity dynamics in Ethiopian manufacturing ............................................................. 56 3.1 Motivation: what went off-track? ................................................................................... 56 3.2 Context: Ethiopia’s manufacturing ambition ................................................................. 57 3.3 Manufacturing productivity dynamics in Ethiopia......................................................... 61 3.4 Productivity dispersion and resource misallocation ....................................................... 70 3.5 Policy constraints and proposed solutions ..................................................................... 75 3.6 References ...................................................................................................................... 83 4 Reviving trade ....................................................................................................................... 87 4.1 Motivation: what went off-track? ................................................................................... 87 4.2 The quest for export diversification ............................................................................... 88 4.3 Service exports: growing potential ................................................................................. 93 4.4 Global value chain participation .................................................................................... 95 4.5 Trade policy in Ethiopia and reform prospects .............................................................. 99 4.6 Conclusion and policy implications ............................................................................. 107 5 Ethiopia Country Economic Memorandum 4.7 References .................................................................................................................... 113 Part III: Sustainability ................................................................................................................. 115 5 Food prices and policy distortions ...................................................................................... 116 5.1 Motivation: what went off-track? ................................................................................. 116 5.2 The long-run determinants of inflation in Ethiopia...................................................... 117 5.3 Demand, supply, and market functioning factors potentially affecting food prices .... 118 5.4 Policy distortions hindering agriculture potential in Ethiopia ..................................... 130 5.5 Conclusion and policy implications ............................................................................. 138 5.6 References .................................................................................................................... 142 6 Making the most of natural capital ..................................................................................... 145 6.1 Motivation: what went off-track? ................................................................................. 145 6.2 The importance of natural capital................................................................................. 146 6.3 The costs and drivers of land degradation .................................................................... 150 6.4 The potential of Ethiopia’s natural and productive landscapes.................................... 153 6.5 Policy options for rebuilding and leveraging natural capital to achieve sustainable and green growth ........................................................................................................................... 161 6.6 References .................................................................................................................... 169 6 Ethiopia Country Economic Memorandum Acronyms ACC Agriculture Commercialization Cluster ADLI Agriculture Development Led Industrialization AfCFTA African Continental Free Trade Area AgSS Agriculture Sample Survey ATA Ethiopian Agricultural Transformation Agency BAU Business-as-Usual CALM Climate Action through Landscape Management CBE Commercial Bank of Ethiopia CGS Credit Guarantee Scheme CLU Cluster Farm Model COM Commercial Farm Model COMESA Common Market for Eastern and Southern Africa CPI Consumer Price Index CRGE Climate Resilient Green Economy Strategy CSA Central Statistical Agency DALY Disability-Adjusted Life Year DBE Development Bank of Ethiopia DMSP Defense Meteorological Satellite Program DTA Deep Trade Agreements EABC Ethiopian Agricultural Businesses Corporation EAC East African Community EBI Ethiopian Biodiversity Institute ECCSA Ethiopian Chamber of Commerce & Sectoral Association ECWA Ethiopian Wildlife Conservation Authority ECX Ethiopian Commodity Exchange EOG Earth Observation Group EEP Ethiopia Electric Power EEU Ethiopia Electricity Utility Company EGTE Ethiopian Grain Trade Enterprise EIC Ethiopia Investment Commission EPA Environmental Protection Authority ERHS Ethiopian Rural Household Survey ESLSE Ethiopian Shipping and Logistics Services Enterprise ESS Ethiopia Economic Survey eSW Electronic Single Window ETBC Ethiopian Trading Business Corporation EU European Union EWCA Ethiopian Wildlife Conservation Authority FAM Family Farm FPC Farmer Production Clusters FSC Forest Stewardship Council FX Foreign Exchange GDP Gross Domestic Product GTP Growth and Transformation Plan GVC Global Value Chain HAP Household Air Pollution HGERA Homegrown Economic Reform Agenda IATA International Air Transport Association ID Identification IMF International Monetary Fund IP Industrial Park IPDC Industrial Parks Development Corporation IRR Internal Rate of Return K Potassium LMIS Labor Market Information System 7 Ethiopia Country Economic Memorandum LSMS Living Standards Measurement Study (Ethiopia Socioeconomic Survey) MAI Mean Annual Increment MoA Ministry of Agriculture MoPD Ministry of Planning and Development MUDI Ministry of Urban Development and Infrastructure MoWIE Ministry of Water, Irrigation and Electricity of Ethiopia MRS Marginal Rate of Substitution MTF Multi-Tier Framework MtCO2e Metric tons of carbon dioxide equivalent NBE National Bank of Ethiopia NDC Nationally Determined Contribution NIDP National Identity Program NILUPP National Integrated Land-Use Plan and Policy NTL Night-Time Light NTFP Non-Timber Forest Product NTM Non-Tariff Measure OECMs Other Effective Area-Based Conservation Measures OFWE Oromia Forest and Wildlife Enterprise OLS Ordinary Least Squares PEFC Programme for the Endorsement of Forest Certification PES Payment for Ecosystem Services PES Public Employment Services PPD Public-Private Dialogue PPAMS Public Project Administration and Management System Proclamation PPPDS Public Procurement and Property Disposal Service PSNP Productive Safety Net Programme PPML Poisson Pseudo-Maximum Likelihood REED+ Reducing emissions from deforestation and forest degradation R&D Research and Development RCA Revealed Comparative Advantage RWE Roundwood Equivalent SFE State Forest Enterprise SFM Sustainable Forest Management SME Small and Medium-Sized Enterprise SNNP Southern Nations, Nationalities, and Peoples' Region SOE State-Owned Enterprise SPS Sanitary and Phytosanitary STD Sexually Transmitted Disease TCCPA Trade Competition and Consumer Protection Authority TBT Technical Barrier to Trade TEV Total Economic Value TFA Trade Facilitation Agreement TFP Total Factor Productivity TFPR Revenue Productivity TRIPS Trade-Related Aspects of Intellectual Property Rights UPSNJP Urban Safety Net and Jobs Project US United States VIIRS Visible Infrared Imaging Radiometer Suite WTO World Trade Organization 8 Ethiopia Country Economic Memorandum Part I: Inclusion 9 Ethiopia Country Economic Memorandum 1 Setting the stage: growth and convergence in Ethiopia 1.1 Motivation: what went off-track? Ethiopia’s rapid growth over the past two decades has resulted in a surge in income per capita levels, with the country approaching fast the middle-income milestone. After thirty years of stagnating or declining income per capita, growth took off in the early 2000s (Figure 1, left panel). During 2004-2019, Ethiopia was the fastest-growing country in the world and the only economy averaging just over 10 percent growth. Output per capita increased eight-fold, from US$110 in 2003 to US$880 in 2020, and Ethiopia is expected to become a lower-middle-income economy by 2025. Ethiopia has experienced nearly two decades of fast convergence in income level with other countries, although there is still a long way to go: the income gap with the United States has narrowed but is still substantial—corrected for differences in countries’ price levels, Ethiopians only earn 4 percent of the income level in the United States (Figure 1, right panel). Figure 1. Ethiopia’s output level has surged since the early 2000s Real output per capita in Ethiopia Real output per capita relative to the United States 5.0% share of US level, % 4.0% 3.0% 2.0% 1.0% 0.0% 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018 GDP per capita GDP per worker Source: Authors’ calculations based on PWT10.0 and IMF WEO. Over the past decade, fast growth was driven by capital accumulation, but the extent to which this growth has been equally distributed is unclear. Public infrastructure spending accelerated dramatically in the first half of the 2010s, helping underpin fast economic growth. However, this approach seems to have had important shortcomings. Contrary to the findings of World Bank (2015) which examined an earlier period, total factor productivity (TFP) declined during 2011-2020, contributing negatively to growth. In addition, inequality at the household level increased between 2011 and 2016. Finally, macroeconomic imbalances have widened, a trend exacerbated by recent shocks. This chapter discusses the drivers of growth in Ethiopia and, in the absence of official subnational gross domestic product (GDP) figures, examines whether there has been convergence in economic activity at the subnational level. Section 1.2 introduces the macroeconomic trends observed during the past decade and discusses how high growth has led to growing imbalances, which have been accentuated by recent shocks. Section 1.3 presents the results of growth decomposition analysis since the early nineties. Section 1.4 uses a novel approach that leverages night-time light data to proxy growth in industry and services at the zone 10 Ethiopia Country Economic Memorandum administrative level, while agriculture activity is estimated by drawing from official surveys. Section 1.5 discusses policy implications and concludes. 1.2 Context: robust growth at the cost of widening macroeconomic imbalances and recent shocks that threaten reform gains While growth has slowed somewhat in recent years, it has remained high despite several shocks. Growth averaged 10.8 percent during the first half of the past decade, prior to easing as a series of events hit the country. These events have included civil unrest in 2016 and 2017, the COVID-19 pandemic, and the armed conflict that started in November 2020, devastating the Tigray region and affecting Amhara and Afar, as well. Despite these shocks, growth remained above 6 percent in 2019/20 and 2020/21. Over the past decade, growth has been driven by services and a boom in construction (associated with both housing and infrastructure projects), while the contribution of agriculture to GDP decreased during the second half (Figure 2, left panel). The contribution of manufacturing to growth has remained modest (see Chapter 3). On the demand side, growth has been driven primarily by public investment, as discussed in more detail in the next section. Private consumption has been a significant driver, as well, although to a lesser extent than in other countries (Figure 2, right panel). Figure 2. Growth in Ethiopia has eased in recent years, while remaining robust Supply-side contributions to growth 14 Demand-side contributions to growth 18 16 12 14 10 12 Percentage points 10 Percentage points 8 8 6 6 4 4 2 2 0 -2 0 -4 -2 -6 -8 Agriculture Manufacturing Private consumption Gov't consumption Construction Other industry Investment Net exports GDP Services GDP Source: Ministry of Planning and Development of Ethiopia. To facilitate cheap imports for large public investment projects, the exchange rate has remained overvalued, leading to a deterioration in the external balance. For most of the past decade, Ethiopia has kept a predictable crawling-peg arrangement, with the exchange rate depreciating at about 6 percent a year in nominal terms. However, as inflation has been in double digits, higher than in trading partners, the exchange rate has appreciated in real terms over time. The Birr was estimated to be overvalued by 25 percent as of June 2019 (IMF 2020). Overvaluation favored imports while hindering the competitiveness of exports, which declined markedly during the past decade. After peaking at above 30 percent of GDP in 2014/15, imports of goods and services started to decline, as well, as foreign currency became increasingly scanty (Figure 3, left panel). Several foreign exchange restrictions were introduced over time, leading to the surge of a parallel market premium. While foreign direct investment inflows have been robust, they have probably remained below 11 Ethiopia Country Economic Memorandum potential due to these constraints and a poor business climate. Recently, the armed conflict in the Northern part of Ethiopia has exacerbated inflation (by affecting crops and expectations) and resulted in a significant slowdown in official flows (Figure 3, right panel). This has resulted in dwindling reserve levels, at below one month of imports as of December 2021, and increasing pressures on the currency. Figure 3. Exports as a share of GDP have declined over the past decade, while the current account deficit peaked by the mid 2010s Current account deficit, FDI and official Evolution of exports and imports 14.0 inflows 35.0 12.0 30.0 10.0 25.0 8.0 20.0 6.0 4.0 % of GDP 15.0 % of GDP 2.0 10.0 0.0 5.0 -2.0 7.1 6.3 5.8 4.8 3.9 3.6 3.4 2.8 2.8 3.3 0.0 -4.0 -6.0 -8.0 -10.0 Services exports -12.0 Goods exports -14.0 Imports of goods and services Current account balance FDI Official inflows Source: National Bank of Ethiopia. To finance investments, external borrowing by the government and the SOEs increased significantly over the past decade, which—coupled with the decline in exports—placed the economy at high risk of debt distress. The Public and Publicly Guaranteed (PPG) external debt as a ratio of GDP jumped from 18 percent at the beginning of the past decade to nearly 30 percent in FY18. The authorities have since curbed non-concessional borrowing, and the stock of external debt to GDP has declined despite the recent shocks. Nevertheless, as the non-concessional loans accumulated in previous years started to mature and export performance deteriorated, Ethiopia’s risk of debt distress according to the debt sustainability analysis has remained at “high” risk since 2017. In February 2021, the authorities applied to receive debt treatment under the G-20 Common Framework, with the aim of reprofiling external debt service payments, but progress in the conversations has been limited to date. Debt vulnerabilities remain high, and limitations in access to external resources may make it more challenging to comply with upcoming debt repayment obligations. 12 Ethiopia Country Economic Memorandum Figure 4. The debt burden surged during the past decade, placing the economy at high risk of debt distress External public and publicly guaranteed debt and debt service 35 30 Central government SOEs Debt service ratio (RHS) % of export of goods and services 30 25 25 20 % of GDP 20 15 15 10 10 5 5 0 0 Source: Ministry of Finance, National Bank of Ethiopia. Financial repression has limited the development of the private sector in Ethiopia. On the domestic front, state-owned enterprises (SOEs) borrowed heavily from the state-owned Commercial Bank of Ethiopia. The share held by SOEs in bank credit increased to over 55 percent of total credit by 2014/15, prior to decreasing in recent years (Figure 5). The priority given to the SOEs meant that credit allocation to the private sector was limited. The share of the private sector in total bank credit was around just 30 percent for most of the past decade until started to improve in 2018/19. Meanwhile, the government sold treasury bills to captive investors such as pension funds and the Development Bank of Ethiopia at very low yields and also borrowed sizable amounts through direct advances from the central bank (at an average of 1 percent of GDP a year). While treasury bill auctions have been successfully introduced and regularly conducted since November 2019, real interest rates remain deeply negative against a backdrop of very high inflation. Figure 5. State-owned enterprises have dominated domestic credit allocation, crowding out the private sector Bank credit by borrower 100% 90% 80% 70% percent share 60% 50% 40% 30% 20% 10% 0% Central gov't SOEs Private sector Source: National Bank of Ethiopia. Recognizing the shortcomings of the state-led development model, the Government has initiated the implementation of a comprehensive reform program. In September 2019, Ethiopian authorities unveiled 13 Ethiopia Country Economic Memorandum a Homegrown Economic Reform Agenda (HGERA) comprising three sets of measures. The first set of structural reforms is aimed at fostering efficiency and introducing competition in key growth-enabling sectors (telecom, logistics, energy), improving the business environment, introducing a new framework for Public- Private Partnerships (PPPs), and aligning the regulations on civic engagement and governance of the SOEs with international best practices. The second set of reforms aims at addressing macroeconomic imbalances by reducing overvaluation and introducing a market-determined exchange rate, reducing the fiscal deficit by strengthening revenue mobilization and consolidating expenditure (including by SOEs), increasing government financing through market-based instruments, and removing distortions in the flow of credit to the private sector. The third set of reforms consists of identifying key sectors for which strategies are being revisited (agriculture, manufacturing, mining, tourism, ICT, and creative industries). Progress in macroeconomic reforms has been limited, as the economy has been affected by shocks. Despite the lapse of the IMF program, authorities kept pursuing a faster rate of currency depreciation during the first months of FY22; however, success in reducing real overvaluation has been limited due to the exceptionally high inflation rate (over 30 percent). This high inflation has been caused by supply-side factors: the impact of the armed conflict on agriculture production as well as the more recent surge in international commodity prices. This raises questions on whether the objective of eliminating real overvaluation and the parallel market premium as well as moving toward a market-determined exchange rate remains within reach. In December 2019, the NBE repealed a highly constraining rule on private banks which was in place since April 2011 and which required them to allocate 27 percent of their new loans to the purchase of NBE bills with a low interest rate and relatively long maturity (five years). Around that time, treasury bill auctions were introduced and have been conducted twice a month since then, allowing private sector actors to diversify their portfolio. In FY21, T-bill issuances reached a record high of 132 billion Birr, or 2.6 percent of GDP equivalent. However, direct advances from the central bank to the government have not been fully phased out as originally planned, since expenditure needs have increased due to the COVID-19 pandemic and the conflict. Authorities have kept making progress on the structural reform agenda, although significant risks threaten both implementation and potential development gains. Despite the shocks, on the structural reform side, the authorities adopted a new Commercial Code, e-Transactions proclamation, and new Capital Markets proclamation in 2021. In the telecom sector, a new commercial license has been issued, effectively ending the state monopoly; however, the partial privatization of the incumbent Ethio Telecom has been postponed, with the authorities citing the current turmoil in international markets as the reason why the proposals received have not been acceptable. Overall, reforms undertaken to date are expected to bring a significant improvement in the business environment in Ethiopia. However, their ultimate impact remains contingent on the evolution and outcome of the armed conflict in Tigray which, in addition to the devastation and suffering caused, has resulted in the suspension of preferential treatment under the African Growth and Opportunity Act and is likely to put off potential investors. 1.3 The drivers of growth in Ethiopia 1.3.1 Large capital formation and lagging productivity During the 1990s, growth in Ethiopia was highly volatile. An augmented Solow growth accounting framework was applied to decompose output growth for the periods 1993-2003, 2003-2011, and 2011-2020 (see Annex 1.1 for methodology and data details). In the first decade after the country’s transition to a market- based economy following the fall of the Derg regime (1993-2003), growth was supported by labor and capital accumulation and, to a lesser degree, by TFP growth. Swings in economic growth were frequent, with the 14 Ethiopia Country Economic Memorandum country registering negative growth in 1998 and 2003 against a backdrop of droughts and armed conflict. Per capita income levels declined during this period. Ethiopia was able to attain double-digit growth in the 2000s mainly thanks to improvements in productivity. The Agriculture Development Led Industrialization (ADLI) strategy initiated in the 1990s was reinforced in the early 2000s through a new emphasis on supply-side agriculture support (including extension services) and a widened scope that included the industry and urban sectors.1 This strategic shift presumably supported a growth takeoff driven by TFP improvements (both within agriculture as well as from reallocation of resources to modern sectors). On the other hand, contrary to previous findings which argued that capital accumulation had been one of the main drivers of the growth takeoff (Moller and Wacker 2017), the role of capital accumulation seems to have been more ambiguous in the 2003-2011 period: although capital accumulated at a fast pace, the capital-output ratio declined during those years. Labor and human capital accumulation continued contributing to growth at broadly similar levels as in the previous period and have generally been the most stable drivers over the three periods analyzed (Figure 6). The contribution of human capital to growth, at about 1 percentage points of GDP across the three decades, is similar to that in Kenya and Cambodia, but lower than the contribution of human capital in Rwanda, Uganda, or Vietnam. Figure 6. Growth accounting suggests that productivity has no longer contributed to growth (over the past decade) Growth accounting results: contributions to growth 14% 12% 10% Percentage points 8% 6% 4% 2% 0% -2% 1993-2003 2003-2011 2011-2020 -4% -6% productivity capital-output ratio human capital labor output growth Source: Authors’ calculations based on PWT10.0 and IMF WEO. Over the past decade, capital investment has been the main driver of growth, while productivity has declined. Over the past decade, the two Growth and Transformation Plans (GTP I and II) emphasized, among other objectives, significantly scaling up public infrastructure investment to support continued high growth. As a result, it is estimated that the accumulation of physical capital contributed 7.3 percentage points to growth per year during 2011-2020 (Figure 6). In contrast, TFP contributed negatively to growth during this period. This suggests that large investments, mainly by the public sector and the state-owned enterprises (SOEs), may have been partly ineffective—at least so far—in terms of boosting productivity. This remarkable finding is robust to alternative assumptions and checks (see section 1.1 in the Annex). 1 See: Industrial Development Strategy, 2002; Sustainable Development and Poverty Reduction Program (SDPRP), 2002/03-2004/05; A Plan for Accelerated and Sustained Development to End Poverty (PASDEP), 2005/06-2009/10. 15 Ethiopia Country Economic Memorandum The finding that Ethiopia has experienced a TFP decline is also remarkable because the growth and development accounting literature attributes the largest part of cross-country income differences to differences in TFP, with capital accumulation playing a less important role (e.g., Caselli, 2005; Hsieh and Klenow, 2010).2 It could be argued that Ethiopia had a particular gap in capital endowment and that the importance of capital accumulation over the last two decades was mainly a ‘catch-up’ effect. In fact, the capital- output ratio of Ethiopia was, if anything, among the higher ones in the region in 2003 already, and following large annual investments, by 2019 it was above the ratio of all peer countries except for Kenya (Figure 7 right panel). The figure also illustrates that the capital-output ratio during the 2003-2011 period declined. In addition to misallocation of the increasingly large public-sector-led capital investments, the constraints to private sector development in Ethiopia (e.g. in terms of accessing foreign exchange, finance, or entering certain sectors) may be one of the reasons why total factor productivity has not improved. Figure 7. Ethiopia’s gross capital formation over the past decade has stood out among peers Gross capital formation, average 2011-19 40 35 30 Percent of GDP 25 20 15 10 5 0 Source: World Development Indicators and authors’ calculations based on PWT10.0 and IMF WEO. Note: ‘L-MIC’ = Lower middle-income countries average. ‘SSA’ = Sub-Saharan Africa average. 1.3.1 Some, but limited, structural transformation Over the past decade, labor productivity stagnated in agriculture and services while it surged in industry, driven by construction. While there have been increases in labor productivity3 since 1993 in all sectors, the dynamics by sector have been somewhat different. Agriculture, which accounted for about 60 percent of value added in 1993 and just about 30 percent in 2020, has the lowest labor productivity of the three sectors. Agriculture experienced a relatively steady increase in labor productivity since 2004—coinciding with the emphasis on extension services—and up to the early 2010s (Figure 8, left panel). Meanwhile, services, which started out already at the highest labor productivity in 1993 (when it accounted for 29 percent of the economy’s value added and 16 percent of employment) experienced significant productivity increases in the 1990s and 2000s but not in the most recent decade. Around 2015, services was overtaken by industry as the 2 These studies are usually based on the same Cobb-Douglas production function as used in this note. Studies investigating other aggregate production functions generally find a higher importance of physical capital accumulation (e.g., Aiyar and Dalgaard 2009; Trenczek 2020). 3 “Labor productivity,” which is output per worker, is then defined as output (or value added) divided by respective employment: K/L. Note that this is a different concept than total factor productivity, which considers output conditional on all production factors. Labor productivity may increase in the absence of any technological or allocative progress if more capital is used in production: as long as the marginal product of capital is positive, this will increase output per worker. 16 Ethiopia Country Economic Memorandum sector with the highest productivity. The surge in the labor productivity of industry over the past decade was driven by huge increases in the productivity of the construction sector, presumably including public works (Figure 8, right panel). Figure 8. A surge in labor productivity growth in industry is mostly explained by the construction sector Ethiopia, value added by sector Ethiopia, value added in industry 14,000 25,000 12,000 20,000 US$ (2017 prices) US$ (2017 prices) 10,000 8,000 15,000 6,000 10,000 4,000 5,000 2,000 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 manufacturing construction agriculture industry services overall mining & quarrying electricity, gas, water industry overall Source: Authors’ calculations based on PWT10.0, IMF WEO, and World Bank WDI. Although structural transformation accelerated over the past decade, it remains incipient. The contribution of the reallocation of workers across sectors was negative during 1993-2003; increased modestly during 2003-2011, a period in which agriculture productivity increased significantly; and accelerated during the most recent period, attaining a share of nearly 40 percent of the total (Figure 9). Yet structural transformation remains limited: while agriculture continues to lag significantly behind industry and services in terms of labor productivity, it still plays a large role in the economy, employing about two-thirds of total workers compared to just 10 percent of workers employed in the industry sector. Analysis within the industry sector suggests labor productivity improvements have been constrained by workers moving into the ‘electricity, gas, and water supply sector’ (mostly SOEs), which has much lower productivity than manufacturing, which experienced a 19- percentage-point decline in its ‘industry’ employment share over the past decade as construction boomed. Figure 9. The contribution of structural transformation (in orange) to labor productivity has increased over time Change in labor productivity, US$ (2017 prices) -500 -250 0 250 500 750 1,000 1,250 1,500 1993-2003 2003-2011 2011-2020 within agriculture within industry within services between sectors 17 Ethiopia Country Economic Memorandum Source: Authors’ calculations based on PWT10.0, IMF WEO, and World Bank WDI. Note: The bars show changes in labor productivity (in US$ at 2017 prices) over the respective periods that are due to changes within and between sectors. While stagnant TFP might at first glance seem difficult to reconcile with the reported increase in labor productivity, particularly in industry (and specifically construction), the two phenomena are consistent. These two phenomena are consistent because capital accumulation during the most recent period led to a higher capital-to-labor ratio and hence made labor more productive. A potential explanation is growing misallocation across firms and sectors: if capital is put to increasingly less productive use, aggregate productivity may decline. One would expect to see a declining output elasticity of capital in such a setting. Aggregate data does not seem to support this hypothesis; if anything, the aggregate output elasticity of capital has been at historically high levels in the most recent period. But firm-level data for manufacturing does find some evidence of growing misallocation (refer to Chapter 3: Productivity dynamics in Ethiopian manufacturing). 1.4 Subnational economic activity and convergence Given the lack of official figures, subnational GDP estimates at the zone level have been constructed drawing from night-time light data and agriculture surveys. The analysis presented in this section extends previous studies that leveraged big data to build poverty and GDP estimates at the subnational level (Njuguna and McSharry 2016; McSharry et al. 2020). In this analysis, GDP is disaggregated into three major components: (i) industry and services, (ii) agriculture, and (iii) major projects such as industrial parks and hydropower plants. Night-time light data from the NASA Defense Meteorological Satellite Program (DSMP) is used to estimate variations in industry and services activity, as the literature has found a robust correlation (Ebener et al. 2005; Elvidge et al. 2009; Ghosh et al. 2010a; Ghosh et al. 2010b). Estimates for the agriculture sector are based on data on production in quintals and yield (quantity per hectare) from the annual Agriculture Sample Survey and a separate dataset containing producer prices for each zone—both sources are compiled by the Central Statistical Agency. Finally, the start year, valuation, and contribution to GDP for different megaprojects between 2000 and 2019 was estimated by the authors using different data sources. The GDP data is disaggregated to illustrate the spatial distribution of each economic sector at the zone level. Section 1.2 in the Annex provides further detail on the methodological approach used to arrive at subnational GDP estimates. There is a strong (negative) correlation between the regional GDP per capita estimates and the poverty levels. To assess the reliability of the data, the estimates were compared with poverty figures, which were available at the subnational level. A comparison of the poverty headcount and per capita GDP at the regional level indicates a strong relationship between the two measures, with higher incomes associated with decreasing levels of poverty. This relationship can be quantified by the negative correlation coefficient of -60.1 percent and provides reassurance that the GDP estimates behave as expected. Unsurprisingly, the autonomous cities of Addis Ababa and Dire Dawa enjoy the lowest poverty rates and the highest per capita GDP (Figure 10). The latest Ethiopia Poverty Assessment found strong poverty reduction nationally between 2011 and 2016 but highlights the concerning finding that inequality has risen (Bundervoet et al. 2020). 18 Ethiopia Country Economic Memorandum Figure 10. Higher GDP per capita is associated with lower poverty levels (poverty headcount and GDP per capita by region, 2016) Source: Authors’ calculations and Bundervoet et al. (2020). Note: SNNP refers to Southern Nations, Nationalities, and Peoples' Region. On average, the poorest zones have grown faster, but their GDP remains significantly smaller. Over the period 2000 to 2019, the zone-averaged total GDP increased at an annual growth rate of 10.9 percent, in constant terms (Table 1). The bottom 20 percent of the zones in year 2000 in terms of GDP (Q1 in the table), grew at an average of 15.4 percent during this period, as they departed from a lower base. Meanwhile, the top 20 percent of the zones (Q5) grew at 7 percent. In 2019, the average GDP of Q5 zones was 17 times higher than that of Q1 zones, compared to 70 times higher in year 2000. Table 1. Over the past two decades, poorer zones have grown faster than richer zones Zone-averaged GDP levels (2000 and 2019) in constant millions birr and average annual growth (2000-2019) by sector GDP GDP (2000) GDP (2019) Growth (2000-2019) Zones Total Agr Ind&Ser Total Agr Ind&Ser Total Agr Ind&Ser All Zones 4887 2798 2089 24536 8064 16472 10.9 8.5 13.1 Q5 11944 5241 6704 44228 7163 37065 7.0 3.4 12.4 Q4 6598 4672 1927 31065 11254 19811 9.9 5.7 13.2 Q3 3937 2694 1243 27613 10478 17135 10.7 7.5 13.5 Q2 1728 1311 418 17460 10191 7269 12.2 11.3 13.2 Q1 169 67 102 2509 1385 1124 15.4 18.3 13.0 Source: Authors’ calculations based on DMSP, the annual Agriculture Sample Survey, and official GDP estimates. Driven by agriculture, the bottom zones have on average reduced the gap in per capita GDP levels. The average per capita GDP of the bottom zones (Q1) was 580 birr in 2000, then rose to 15,325 birr by 2019, achieving average annual growth of 12.2 percent, and getting close to the average per capita GDP levels of zones in the second and third quintiles (Table 2). Nonetheless, there is still a considerable gap when compared to the per capital levels of Q4 (27,243 birr) and Q5 (34,827 birr). This convergence in per capita GDP is mainly explained by the much faster agriculture growth experienced by the poorer zones: The agriculture GDP of Q1 zones grew at an average of 14.9 percent in per capita terms, compared to just 0.5 percent growth in Q5 zones. In turn, the average per capita growth rates in industry and services were similar across the first four quintiles, and just slightly slower for the top quintile. 19 Ethiopia Country Economic Memorandum Table 2. Largely driven by growth in agriculture, per capita GDP in the bottom quintiles as reduced the gap Zone-averaged per capita GDP levels (2000 and 2019) in constant birr and average annual growth (2000-2019) by sector GDP p.c. Per Capita GDP (2000) Per Capita GDP (2019) Growth (2000-2019) Zones Total Agr Ind&Ser Total Agr Ind&Ser Total Agr Ind&Ser All Zones 5189 2748 2441 22575 9339 13235 7.9 5.6 10.1 Q5 12756 6012 6744 34827 9252 25575 4.2 0.5 9.4 Q4 6357 3217 3140 27243 5371 21871 7.0 2.9 10.2 Q3 3941 2712 1229 18717 8182 10535 7.8 4.7 10.6 Q2 2234 1561 673 16519 10499 6020 9.1 8.2 10.2 Q1 580 237 342 15325 13318 2007 12.2 14.9 10.0 Source: Authors’ calculations based on DMSP, the annual Agriculture Sample Survey, and official GDP estimates. Growth rates in industry and services activity are estimated to be more homogeneous than in agriculture, where there is large dispersion. All zones that experienced growth rates above 20 percent started from a relatively low base (per capita GDP of less than 2,500 birr in 2000). The top three zones in terms of growth are Argoba (36.8 percent), Basketo (27.4 percent), and Derashe (24.0 percent). As shown in Figure 11, disparity in growth stems from agriculture rather than industry and services. The dispersion in performance of zones in the bottom quintile, in particular, is much larger than in the rest, with five zones among the top seven performers and two zones among the worst performers (Figure 11, blue dots in the left-hand-side chart). Figure 11. There is high dispersion in agriculture growth rates, in particular among the poorest zones Source: Authors’ calculations based on DMSP, the annual Agriculture Sample Survey, and official GDP estimates. The maps illustrate that, while most of the zones in the lower quintiles in terms of per capita GDP in 2000 have grown fast, there are some laggards as well. Addis Ababa and zones in the South and East of Oromia, which ranked in the highest quintile in terms of per capita income in 2000, are estimated to have experienced lower growth rates over the past two decades (Figure 12). Meanwhile, some zones in the Southern Nations, Nationalities, and Peoples' (SNNP) Region that were in the bottom quintiles in 2000 experienced high 20 Ethiopia Country Economic Memorandum growth rates over the past two decades. However, growth in some zones in the regions of Somali and Gambela has been sluggish, despite starting from a low base; some of these are areas in which land is not particularly fertile and there is limited concentration of economic activity and population. Figure 12. Generally, zones in the lower per capita GDP quintiles experienced relatively higher growth rates Per capita GDP from all sectors in 2000. Quintile ranking from lowest (dark red) to highest (dark green) Growth in per capita GDP from all sectors (2000 and 2019). Quintile ranking from lowest (dark red) to highest (dark green) Source: Authors’ calculations based on DMSP, the annual Agriculture Sample Survey, and official GDP estimates. 21 Ethiopia Country Economic Memorandum Econometric analysis confirms the overall patter of convergence in economic activity across zones and identifies zone-level drivers of growth. First, the share of population with TVET education in the year 2016 has a strong positive impact on zones’ growth, while other education measures are insignificant. As TVET graduates are likely better fits for industrial jobs enhancing output and perhaps also more entrepreneurial, zones with a higher proportion of TVET graduates would experience higher growth. On the other hand, the average years of schooling variable is always insignificant, possibly due to the low variability it experiences across zones. Second, the presence of hospitals and specialized clinics has a significant positive impact on the growth of Ethiopian zones. Because of their scale, hospitals and clinics are a proxy of fixed investment that may also be capturing the existence of other infrastructure (e.g., storage, education centers, lodgings) that is not explicitly incorporated into the model due to lack of data. Third, the cumulative population reachable within a 30-minute distance has a positive and significant coefficient. Although this result supports the hypothesis that agglomeration growth effects are at play in Ethiopia, the overall impact of proximity dies off quickly as travel time increases. Finally, having a high initial share of industry and services GDP is also a significant driver of growth. Results must be interpreted with caution given data constraints (there is only one year of data available for several variables) and the limited scope of applicable estimation methods. Annex 1.3 provides full detail on the methodology and results from the econometric analysis of the drivers of growth at the subnational level in Ethiopia. Finally, it is worth noting that convergence in per capita GDP does not necessarily imply that was pro- poor. According to household surveys, the consumption of the bottom 40 percent households in rural areas did not increase significantly between 2011 and 2016 (see Bundervoet et al. 2020 and chapter 2 in this report). Simple regressions suggest a negative and significant relationship between growth and non-food consumption at the zone level (meaning that growth has been regressive). The relationship between consumption increases for the bottom 40 percent of the households and growth at the zone level is also negative, although not significant. 1.5 Conclusion and policy implications Over the past decade, economic growth in Ethiopia was driven by capital accumulation, while total factor productivity did not increase, and structural transformation was limited. Contrary to previous studies, the analysis finds that the contribution of capital to growth during 2003-2011 was negative, and that the capital-output ration declined during that period, in which growth was mainly driven by total factor productivity improvements and labor accumulation. On the other hand, growth during 2011-2020 was fundamentally driven by capital accumulation, as TFP declined. In the past decade, Ethiopia experienced an increase in industry value added that was mostly driven by construction, and some labor reallocation took place; nonetheless, two-thirds of the workers remain in agriculture and just 10 percent are employed by industry. It can be argued that, while Ethiopia’s rate of capital accumulation in the 2010s was among the fastest in the world, its impact so far in terms of fostering productivity growth and structural transformation has been modest. This points to potentially weak project selection and execution undermining the effectiveness of public investment, as well as to a series of policies limiting private sector development. There is a need to strengthen public investment management to ensure a higher growth multiplier. New public project proposals would need to be screened under the Public Project Administration and Management System (PPAMS) Proclamation, whose approval in June 2020 has been crucial to ensure projects have a sound economic rationale, there are no maturity mismatches, environmental and social standards are observed, and the public debt remains sustainable. Drawing from the authority to approve investment projects assigned under the PPAMS Proclamation, the National Planning and Development Commission, in collaboration with Ministry of Finance, would need to play a more pro-active gatekeeper role when assessing 22 Ethiopia Country Economic Memorandum public investment projects, including those proposed by SOEs, in order to ensure the efficiency of public spending and its alignment with strategic priorities, while safeguarding fiscal sustainability. Subnational estimates of economic activity suggest that there has been convergence in income at the zone level. Given the lack of official figures, subnational GDP estimates at the zone level have been constructed drawing from night-time light data and official agriculture surveys. On average, poorer zones have been catching up from a lower initial level thanks to their strong performance in terms of agriculture GDP growth. However, no significant catch up is observed in industry and services, and the differences in economic activity between top and bottom zones remains substantial. It is also worth noting that there are some lagging zones (e.g. in the Somali region). However, convergence does not necessarily imply that growth has been pro-poor. Even in those rural areas that had strong growth, growth may not have been necessarily pro-poor: according to household surveys, the consumption of the bottom 40 percent in rural areas did not increase between 2011 and 2016 (Bundervoet et al., 2020). Regressions find a negative and significant relationship between growth at the zone level and non- food consumption in those zones. One potential explanation is that robust agriculture growth may not have sufficiently trickled down, while limitations to labor mobility may have prevented households from participating further on economic transformation. Zone-level infrastructure can play a role in boosting economic growth. The following drivers of growth at the zone level are found to be significant: (i) a high initial share of industry and services GDP; (ii) having vocational training centers and hospitals; (iii) being nearby and well connected to other poles of economic activity. Thus, in addition to investing in some key roads connecting with nearby growth poles, investments in human capital matter. Beyond the basic network of schools and primary health centers, investing in TVET centers and hospital and clinics but also on other infrastructures that are adjacent or linked to those, is beneficial for growth at the zone level. Complementary policies are needed to broaden economic transformation and facilitate the participation of households and firms in the process. In addition to the policy options derived from this chapter (Table 3), other challenges highlighted above are analyzed in more detail in the remainder of this report. Chapter 2 discusses the limitations to labor mobility and household participation in economic growth, while Chapters 3 and 4 discuss constraints to private sector development and productivity improvements, and Chapters 5 and 6 focus on growth sustainability aspects. Table 3. Policy implications towards sustainable capital formation and growth convergence Area Short term Medium term Institution in charge Public Investment Implement the new PPAMS Ministry of Planning Management Proclamation and Development; Ministry of Finance Zone-level growth Ensure key connectivity with Invest in TVET and high-complexity Federal and regional drivers nearby zones and growth poles health centers (hospital, clinics) authorities Inclusion Refer to the recommendations in Chapter 2 in this report Source: authors’ elaboration. 23 Ethiopia Country Economic Memorandum 1.6 References Abreu, M., H. de Groot, L. Henri; and R. Florax. 2004. “Space and Growth.” Tinbergen Institute Discussion Paper, No. 04-129/3. Amsterdam and Rotterdam: Tinbergen Institute. Aiyar, S., and C. Dalgaard. 2009. “Accounting for productivity: Is it ok to assume that the world is Cobb- Douglas?” Journal of Macroeconomics, 31(2): 290-303. Barro, R. and J. Lee. 2013. “A new data set of educational attainment in the world, 1950–2010.” Journal of Development Economics, Vol. 104, September: 184-198. Bundervoet, T., Maiyo, L. and Apurva, S. 2015. Bright lights, big cities: measuring national and subnational economic growth in Africa from outer space, with an application to Kenya and Rwanda. Policy Research Working Paper 7461. World Bank Group. Bundervoet, T., A. Finn, S. Nakamura, B. Beyene, P. Paci, N. Mylenko, and C. Turk. 2020. Ethiopia Poverty Assessment - Harnessing Continued Growth for Accelerated Poverty Reduction. Ethiopia and Washington, DC: World Bank. Caselli, F. 2005. “Accounting for Cross-Country Income Differences.” In P. Aghion and S. Durlauf (eds): Handbook of Economic Growth, Volume 1 (Part A): 679-741. Ebener, S., Murray, C., Tandon, A. and. Elvidge, C. 2005. “From Wealth to Health: Modeling the Distribution of Income per Capita at the Sub-National Level Using Night-Time Light Imagery.” International Journal of Health Geographics 4(5): 1–17. Elvidge, C., Sutton, P., Ghosh, T., Tuttle, B., Baugh, K., Bhaduri, B. and Bright, E. 2009. “A Global Poverty Map Derived from Satellite Data.” Computers and Geosciences 35: 1652-1660. Earth Observation Group (EOG). 2021. Nighttime Lights data available at. https://eogdata.mines.edu/products/dmsp Feenstra, R.., Inklaar, R. and M. Timmer. 2015. "The Next Generation of the Penn World Table." American Economic Review, 105 (10): 3150-82. Ghosh, T., Powell, R., Elvidge, C., Baugh, K., Sutton, P. and S. Anderson. 2010a. “Shedding Light on the Global Distribution of Economic Activity.” The Open Geography Journal (3): 148-161. Ghosh, T., Powell, R.L., Anderson, S. Sutton, P.C. and C.D. Elvdige. 2010b. “Informal Economy and Remittance Estimates of India Using Nighttime Imagery.” International Journal of Ecological Economics & Statistics, 17. Hall, R., and C. Jones. 1999. “Why do Some Countries Produce So Much More Output Per Worker than Others?” Quarterly Journal of Economics 114(1): 83-116. Henderson, J.V., Storeygard, A. and D. Weil. 2012. “Measuring Economic Growth from Outer Space.” American Economic Review 102(2): 994-1028. Hsieh, C., and P. Klenow. 2010. “Development Accounting.” American Economic Journal: Macroeconomics, 2(1): 207-223. Hulten, C. 2010. “Growth Accounting.” In B. Hall and N. Rosenberg (eds.): Handbook of the Economics of Innovation. Vol. 2: 987-1031. Klenow, P., and A. Rodríguez-Clare. 1997. “The Neoclassical Revival in Growth Economics: Has It Gone Too Far?” NBER Macroeconomics Annual 1997, Volume 12. 24 Ethiopia Country Economic Memorandum IMF. 2020. “The Federal Democratic Republic of Ethiopia : 2019 Article IV Consultation and Requests for Three-Year Arrangement under the Extended Credit Facility and an Arrangement under the Extended Fund Facility-Press Release and Staff Report.” Country Report No. 20/29. January 28, 2020. International Monetary Fund, Africa Department. Li, X., Zhou, Y., Zhao, M., and X. Zhao. 2020. “A harmonized global nighttime light dataset 1992–2018.” Scientific Data 7: 168. McMillan, M., D. Rodrik, and I. Verduzco. 2014. “Globalization, structural change and productivity growth with an update on Africa.” World Development 63: 11–32. Mankiw, N. D. Romer, and D. Weil. 1992. “A contribution to the empirics of economic growth.” The Quarterly Journal of Economics 107(2): 407-437. McSharry, P., S. Mahdi, and F. Massarongo Chivulele. 2020. Mozambique spatial economic growth: strengthening economic management for inclusive growth. World Bank Group, forthcoming. Moller, L. and K. Wacker. 2017. “Explaining Ethiopia's Growth Acceleration - The Role of Infrastructure and Macroeconomic Policy.” World Development 96: 198-215. Njuguna, C. and P. McSharry. 2017. “Constructing spatiotemporal poverty indices from big data.” Journal of Business Research 70: 318-327. Trenczek, J. 2020. Essays on Structural Transformation. Doctoral dissertation, University of Mainz. World Bank. 2015. Ethiopia’s Great Run: The Growth Acceleration and How to Pace It.” Washington, DC: World Bank. 25 Ethiopia Country Economic Memorandum 2 Improving access to off-farm jobs and labor mobility in Ethiopia. 2.1 Motivation: what went off-track? Ethiopia's high economic growth over the past decade has not trickled down much to rural areas due to limited labor mobility both into the rural nonfarm sector and from rural to urban areas. The poorest 20 percent in rural areas did not experience consumption growth during the last decade, and growth in average consumption in rural areas between 2011-2016 was only one-fifth of growth in urban areas, which reaped most of the benefits of high economic growth. This is the consequence of labor not shifting out of agriculture even as the share of agriculture in gross domestic product (GDP) declined. Rural labor has also not moved to urban areas rapidly enough: only 12 percent of rural households had a working-age member who moved to urban areas between 2014 and 2016. Analysis shows that structural transformation—as measured by employment sectoral changes, and movement between rural and urban areas—contributed to less than 5 percent of the overall poverty reduction during 2011-2016. This chapter investigates the constraints to labor mobility in Ethiopia and suggests policy directions to help rural Ethiopia benefit more from the country’s economic growth. It first provides an overview of recent trends in poverty reduction and the movement of labor across sectors and between urban and rural areas. It then looks at the challenges associated with rural off-farm employment, with a focus on the potential role that connectivity and growth in small towns can play. It also discusses the benefits of migration in linking the rural population to wider economic growth and the constraints to rural-urban migration. 2.2 Economic growth and the rural-urban divide Ethiopia’s economy has been growing rapidly since the 2000s, bringing significant but uneven growth in household consumption and poverty reduction. Based on the national poverty line, about 12 million people moved out of poverty between 2000 and 2016—a 45 percent decrease. Poverty decreased faster in urban areas than in rural areas, falling from 37 percent in 2005 to 15 percent in 2016 in urban areas, compared to a decrease from 45 percent to 26 percent in rural areas (Figure 13). In earlier periods, poverty reduction in rural areas was stronger than in urban areas, but it has weakened in recent years as robust economic growth has not been accompanied by strong consumption growth in rural areas. Consumption per adult equivalent in rural areas increased by only six percent between 2011 and 2016, while it grew by about one-third in urban areas. Furthermore, rural households at the bottom of the distribution did not experience any consumption growth— a pattern that was also observed between 2005 and 2011. As a result, rural areas experienced weaker poverty reduction, and the rural-urban gap further widened (World Bank 2020). Poverty has thus become more concentrated in rural areas. About 88 percent of the poor lived in rural areas in 2016—this constitutes an increase in the rural share of poverty by more than 2 percentage points during 2011-16 even though the rural population share decreased by 3 percentage points to 80.5 percent in 2016 (Figure 13, right panel). Such trends reflect how poverty reduction has been driven by urban areas in recent periods. Given the concentration of poverty in rural areas, future poverty reduction will need to happen mainly through improvements in rural areas and through increased mobility to urban areas where growth has been high. 26 Ethiopia Country Economic Memorandum Figure 13. Poverty reduction slowed during 2011-2016, particularly in rural areas Trends in national, urban, and rural poverty: Population versus poverty share in 2011 and 2000 to 2016 2016: rural and urban 50.0 100 90 14.1 16.6 12.1 19.5 80 40.0 70 Percentage of poor 60 Percent 30.0 50 40 85.6 83.4 87.9 80.5 30 20.0 20 10 10.0 0 Poverty Population Poverty Population share share share share 0.0 2000 2005 2011 2016 2011 2016 Urban Rural National Rural Urban Source: FDRE (2018); adapted from World Bank (2020). Note: Poverty estimates are based on national poverty line using consumption aggregates from official welfare surveys periodically collected by the Central Statistics Agency (CSA). 4 The structural transformation Ethiopia has achieved has not been accompanied by a significant shift of labor out of agriculture. The share of agriculture in GDP declined from 52 percent in 2005 to 42 percent in 2013 and 33 percent in 2019, but approximately two thirds of the labor force remain employed in agriculture. Recent estimates based on the Ethiopia Socioeconomic Survey (ESS) indicate that 74 percent of the labor force in 2019 worked in agriculture (Figure 14). This suggests that transformation in terms of shifting employment away from agriculture has yet to happen. Figure 14. The decline in agriculture’s share of GDP has not been matched by a decline in its employment share Sectoral GDP share, 2005-2019 Sectoral employment shares, 2014-2019 100% 100 1.1 0.8 3.7 90% 90 17.5 17.6 16.9 80% 80 4.05 5.27 70% 5.3 70 60% Percent 60 Percent 50% 50 40% 40 77.4 76.4 30% 74.2 51.9 30 40.2 37.5 42 20% 33.3 20 10% 10 0% 0 2005 2013 2014 2016 2019 2014 2016 2019 Agriculture Industry Services Agriculture Secondary Tertiary Others/unknown Source: NBE Annual Report 2019-20; Source: Estimates based on Ethiopia Socioeconomic Survey, 2014-2019. 4 While the survey methodology remained largely consistent across survey rounds, an important change was introduced in 2011. Before 2011, consumption information was collected in two month—one immediately after the harvest season and another during the lean period. In 2011 and 2016, consumption data was collected in twelve months. This makes the poverty numbers from earlier periods not fully comparable with those from recent round, but the 2014 Ethiopia Poverty Assessment (World Bank, 2015) found that poverty trends are robust to this methodological change. 27 Ethiopia Country Economic Memorandum In addition, movement of labor from rural to urban areas has been slow. Although Ethiopia’s urban population is rapidly growing, the slow decline in the rural population share suggests that rural-urban migration has been very limited. Over the past decade, the urban share of the total population increased by around 4 percentage points, which is slower than most peer countries except for Rwanda (Figure 15), yet the urban population has grown by nearly 60 percent or 5 percent annually, albeit from a low base. This stems from the fact that a significant part of urbanization has been driven by two factors: the internal expansion of urban areas due to high fertility in small towns and the re-classification of rural areas as urban areas. Even with the fast growth in the urban population, estimates suggest the rural population still expanded by 20 percent during 2010-19,5 adding more people to the country’s population than in urban areas. Figure 15. Urbanization in Ethiopia is accelerating but slower than its comparators Share of urban population Vietnam Cambodia Bangladesh Tanzania Uganda Rwanda Kenya Ethiopia 0 10 20 30 40 50 Percent of total 2020 2015 2010 2005 2000 Source: UNDESA, 2018. The limited transitions to rural off-farm employment and the low rural-urban migration have contributed to low poverty reduction in rural areas, which is consistent with the insignificant poverty reduction role of the population shift, either from rural to urban areas or within sectors . For example, population shifts from rural to urban areas contributed to around two percent of overall poverty reduction during 2011-16 (Figure 16, left panel). Similarly, the lack of structural transformation in the rural economy is also reflected in the small contribution of employment sectoral shifts to overall poverty reduction. Population shift across economic sectors contributed only 5 percent to overall poverty reduction between 2011 and 2016, and this is a higher contribution than in earlier periods (Figure 16, right panel). 5 Based on estimates from the National Bank of Ethiopia, Annual Report 2019-20. 28 Ethiopia Country Economic Memorandum Figure 16. The role of population shift (from rural to urban and inter- sectoral shift) in poverty reduction is negligible Contribution to changes in poverty by Contribution to changes in poverty by residence employment sector 2000-2005 2005-2011 2011-2016 2 0 % change in poverty headcount -1 0 % change in poverty headcount 2000-2005 2005-2011 2011-2016 -2 -2 -3 -4 -4 -5 -6 -6 -7 -8 -8 -9 -10 -10 Agriculture Manufacturing Construction Services Other Population shift Rural Urban Populaton shift Interaction Interaction Source: Adopted from the World Bank (2020). Note: The interaction term is a residual that cannot be explained by other factors. 2.3 Rural off-farm employment and spillover effects of urban areas Off-farm employment generation in rural areas is a crucial channel for rural residents to engage in and benefit more from the country’s fast-economic growth. Off-farm employment and, in particular, non-farm self-employment, is associated with significantly higher consumption when controlling for a host of household and geographic characteristics (World Bank, 2020). This section sheds light on the determinants of access to nonfarm employment opportunities in rural areas. In addition to individual and household factors, the section focuses on the role of local economic factors including proximity to urban areas—particularly small and medium towns—in promoting off-farm employment opportunities, using novel analysis combining household and geospatial data. The analysis is based on the Ethiopia Socioeconomic Survey of 2019 (otherwise referred to as LSMS-ISA 2019), combined with geospatial data from the transport network map of Ethiopia and gridded population datasets to produce measures of proximity and access to markets which are used to quantify rural- urban employment spillovers. The methodological overview is provided in Box 1, while a detailed methodological discussion and results is included in Annex 2. Box 1. Estimating farm and nonfarm labor participation in a rural household farm production model Identifying the determinants of off-farm labor participation in rural Ethiopia is done by estimating a farm household production model using simultaneous probit estimation regressions for joint nonfarm participation of household heads, spouses, and adult children within the household (see Corsi and Findeis 2000; Corsi and Salvioni 2012; Ahearn et al. 2006; and World Bank 2019a). The estimation is restricted to rural households. This analysis jointly estimates the probability of engaging in nonfarm either wage or self-employment activities for household heads and eldest working member pairs (about 80 percent being female spouses). In the conceptual framework for the model, households allocate their labor between farm and nonfarm activities subject to their own individual, household, and farm characteristics as well as local conditions, which include proximity indicators to capture access to markets. Rural-urban spillovers are reflected by the relationship 29 Ethiopia Country Economic Memorandum between rural off-farm employment and urban proximity indicators in our estimation model. Table B.1 below provides the variable description and data sources. Annex 2 provides a full model framework and detailed results. Proximity indicators are based on travel time between origins and destinations pairs (woreda to all other woredas; woreda to all urban centers). This is calculated from GIS data extracted from the 2020 transport network for Ethiopia by determining the shortest possible path for each origin-destination pair and the speed of travel based on the type of roads connecting the pair. Similarly, accessibility in terms of population is estimated by attaching population data to the destination pairs (woredas, urban). Three global population datasets are used to supplement each other: Gridded Population of the World Version 4 (GPWv4) dataset, the LandScan Global Population Distribution (LSG) dataset, and the WorldPop Spatial Distribution of Population (WDP) 1km dataset. Total reachable population within 30 minutes, 60 minutes, 120 minutes, and 180 minutes is calculated from each origin woreda. Lastly, using woreda-to-woreda origin-destination matrices as inputs, accessibility in terms of market access was approximated following earlier work by Donaldson and Hornbeck (2016) and World Bank (2019a). Table B.1. Description of variables and data sources for the household farm production model Variable type Variables Data Source Individual characteristics Age, gender, education, marital status ESS 2018/19 Presence of under-5 children, presence of elderly, land Household characteristics ESS 2018/19 ownership Dummies for crop type (maize, teff, wheat, and barley) and Farm characteristics ESS 2018/19 livestock, farm potential wetness index, farm slope Sector employment shares (agriculture, industry, services), Labor market relative prices (with Addis Ababa as benchmark) for teff, maize, ESS 2018/19 conditions wheat, and barley Population density ESS 2018/19 Local Travel time to nearest urban centers (any type; small towns of Ethiopia Transport Network conditions Geographical 20,000 to 50,000 people) Layer 2020 conditions Gridded Population V4 Market access index - combines travel time with destination (GPWv4), LandScan Global population (LSG), WorldPop (WDP) Access to Presence of banks, presence of markets ESS 2018/19 services Source: Authors’ elaboration. 2.3.1 Rural off-farm jobs in Ethiopia are primarily in the services sector Rural off-farm jobs in Ethiopia are dominated by the services sector, and there is significant gender segregation. Commerce—wholesale and retail trade—is the largest subsector, making up nearly one-quarter of all rural nonfarm jobs. The public sector also plays a considerable role, accounting for 23 percent of rural off-farm jobs, while industry (excluding construction) accounted for less than 15 percent of total rural off-farm jobs in 2019. There is some gender segregation in rural off-farm jobs (Figure 17), with a higher share of women engaged in the hotel and restaurants sector (15 percent) than men (3 percent), who have a higher share engaged in construction (15 percent) and transport (10 percent) instead. Rural women in off-farm jobs are also more likely to be engaged in commerce than any other sector, followed by industry (excluding) construction, while more men are in the public sector. The sectoral composition of rural off-farm jobs and the gender segregation 30 Ethiopia Country Economic Memorandum in occupations is a product of a combination of factors including access to off-farm jobs, individual factors such as education, and local economic factors such as population density and proximity to urban centers. Figure 17. Gender segregation can be seen in rural nonfarm jobs in Ethiopia Females – rural nonfarm jobs composition Males – rural nonfarm jobs composition Other services, Industry excl. Other Industry excl. 11.43 construction, 23 Public services, construction, 8.94 sector, 17.54 16.14 Construction, Construction, 2.35 15.27 Financial services & Public real estate, sector, 27.54 Trade 0.67 Hotels & (wholesale Trade restaurants, & retail), (wholesale 15.02 19.58 & retail), 31.39 Transport & communication, 0 Financial Hotels & services & restaurants, 2.86 real estate, Transport & 1.58 communication, 6.68 Source: Authors’ estimates based on ESS 2018/19. 2.3.2 Education and gender effects on rural off-farm employment Demographic segmentation of the labor market can be seen by age and gender. Older people are significantly less likely to be engaged in nonfarm employment. The prospects of nonfarm employment peak at around 30 years for both male household heads and female spouses (Figure 18, left panel). For male household heads, nonfarm employment prospects peak at 20 percent when they are around 30 years of age but sharply decline to around 12 percent when they are 50 years old. A similar pattern is observed among female spouses, although their participation rates are generally lower. Figure 18. Nonfarm employment prospects appear to be affected by age, gender, and level of education Nonfarm employment and age Nonfarm employment by age and education Source: Authors’ estimates based on ESS 2018/19. 31 Ethiopia Country Economic Memorandum Notes: Based on predicted marginal probabilities of being in nonfarm work for working-age male household heads and female spouses who live in rural areas. The gender gap for nonfarm employment prospects persists across all age groups. The gender gap is largest for the 20-40 years age group but narrows as household members grow old. Female spouses’ prospects of nonfarm employment peak at 12 percent when they are around 30 years of age, compared to 20 percent for male household heads. This gap is still wide among youth: for example, a 20-year-old male has a 70 percent greater chance of being engaged in nonfarm work than a 20-year-old female with similar location, household, and education level. This indicates unequal access to nonfarm opportunities even among new labor market entrants. Education is the most important attribute in determining access to nonfarm employment. Prospects of nonfarm employment for rural people with secondary education are at least double that of people without secondary education across age groups and genders. The demographic segmentation of the labor market is also mostly explained by educational attainment: older workers with secondary education are as likely to engage in nonfarm jobs as people in their 20s at least until their 40s (Figure 18, right panel) so the higher nonfarm job prospects for younger people in general reflects a higher share of secondary-educated people in this age cohort. About 30 percent of people in their 20s have secondary education, compared to 7 percent among people age 40 and above. The gender gap in nonfarm job prospects narrows for people with secondary education. With secondary education attainment, the nonfarm job prospects for female spouses are at least as high as male household heads among people over 30 years of age. In contrast, the gender gap persists for household members without secondary education at all ages. Less than 7 percent of working-age females in rural areas have at least secondary education, which is about half the secondary education attainment rate among working-age men. The gender gap in access to nonfarm opportunities is thus partly explained by women’s lower educational attainment. 2.3.3 Farm and food prices dynamics and rural off-farm employment Households with more land are less likely to engage in nonfarm employment, suggesting that push factors influence nonfarm participation decisions. People in households owning at least two hectares of land are twice as less likely to engage in nonfarm employment than people in households without land (Table 4). This observation is consistent for both male household heads and female spouses. It implies that engagement in nonfarm employment activities might be driven by push factors such as the need to pursue work for acquiring food. Thus, owning more land reduces households’ chances of transitioning out of agriculture. Table 4. Households owning more land are less likely to be in nonfarm jobs Land ownership Household Male Female engages household spouse in nonfarm head No land 0.25 0.2 0.14 Less than 0.2 ha 0.24 0.19 0.12 0.20 - 0.50 ha 0.19 0.14 0.1 0.5 - 1 ha 0.16 0.12 0.09 1 - 2 ha 0.15 0.1 0.08 More than 2 ha 0.13 0.09 0.07 Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted as marginal probablity of nonfarm employment for married male household heads and female spouses in rural areas, then tabulated by land ownership category. The first column is estimated as the joint probability of household head or eldest working household member being engaged in nonfarm work. 32 Ethiopia Country Economic Memorandum Participation in nonfarm activities is lower among households that concentrate on staple food production, a potential indication of self-sufficiency dominating intrahousehold labor allocation decisions. Both male household heads and female spouses in households growing staples such as maize and teff are significantly less likely to be engaged in any form of non-agriculture work, either as paid workers or in self-employment activities (Figure 19). The same applies to households engaged in livestock production, who are twice as less likely to have a person in a nonfarm job compared to those not engaged in livestock production. No such relationship is observed for households producing wheat and barley. Households with land prioritize staple food production to meet their subsistence needs before allocating labor to nonfarm work. Figure 19. Households engaged in staple food production or livestock production are less likely to have nonfarm employment Probability of nonfarm employment 30% 25% 20% Probability (%) 15% 10% 5% 0% Household Male Female Household Male Female Household Male Female Head Spouse Head Spouse Head Spouse Maize Teff Livestock Non-producer Producer Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted as marginal probability of nonfarm employment for married male household heads and females spouses in rural areas, then tabulated by farm type (maize, teff and livestock) and whether a member is a producer or non-producer. The first columns are estimated as one minus the joint probability of household head or eldest working household member not being engaged in nonfarm work. Prioritization of subsistence needs is reflected in food consumers’ labor supply response to price changes. In rural households that are teff consumers, people tend to work less in non-agricultural jobs as teff prices increase (Figure 20, left panel). Similarly, all members in households that are maize consumers tend to work less in nonfarm jobs as maize prices increases (Figure 20, right panel). This suggests that net consumers in rural areas substitute their labor from nonfarm employment to farm-related activities as staple food prices increase to meet their subsistence needs. However, food producers respond differently to price changes depending on the farm production system. As teff prices increase, male household heads in teff-producing households are more likely to engage more in nonfarm activities, while female spouses reallocate their time to agriculture-related activities (Figure 20). In contrast, the prospects of nonfarm engagement by household members in maize-producing households are not responsive to maize price changes. This pattern is likely influenced by how access to markets influences prices of the two crops. 33 Ethiopia Country Economic Memorandum Figure 20. Nonfarm employment for teff and maize consumers tends to fall as teff and maize prices rise Nonfarm employment response to teff price changes Nonfarm employment response to maize price changes Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted marginal probabilities of being in nonfarm work based on crop price changes for male household heads and female spouses. This is estimated separately for producers and consumers of teff and maize. Market access has a strong influence on maize and teff prices, reflecting the balance between demand and supply constraints. Teff fetches higher prices in woredas with greater market access (Figure 21). In contrast, maize prices tend to be higher in woredas with low market access, where nonfarm labor is not responsive to price changes. Teff has a positive income elasticity since it is mostly consumed in richer areas (i.e. economic hubs)—its share in total consumption among the top quintile is double the share of the bottom quintile. Thus, the demand factor dominates the relationship between market access areas and prices for teff. Unlike teff, the share of maize in total expenditure among the bottom quintile is double the consumption share among richest quintile. With a negative income elasticity, supply-side constraints due to low market integration dominates the relationship between market access and maize prices instead, such that prices are high in remote areas with low market access. For this reason, nonfarm labor supply is unresponsive to higher maize prices among maize producers, consistent with the price wedge hypothesis. This postulates that households in areas not integrated with food markets prioritize labor allocation to staple food production because they face higher market prices. Figure 21. Access to markets affects maize and teff prices differently Teff prices response to market access Maize prices response to market access Source: Authors’ estimates based on ESS 2018/19. 34 Ethiopia Country Economic Memorandum Notes: Predicted marginal probabilities of being in nonfarm work based on crop prices changes and access to markets. This is estimated separately for teff and maize. The market access indicator is a summary measure for every origin woreda and all destination woredas of travel-time weighted market (population). The greater the z-score, the more accesible the markets. Overall, higher food prices slow the transition out of agriculture, as households cushion themselves against price shocks by engaging in subsistence farming. Consumers of both teff and maize tend to respond by reallocating their labor out of nonfarm activities when prices are high. Households also tend to keep women in agriculture in response to higher food prices, irrespective of whether they are consumers or producers. This analysis also shows that by lowering local prices through better integration of food markets, improving access to markets can influence household labor allocation decisions in favor of nonfarm employment in poorly connected maize-producing areas, which account for more than 40 percent of the rural population. 2.3.4 Local economic development and non-farm employment prospects Off-farm job prospects are largely determined by the local economic structure, itself driven by population density. The chances of anyone in a rural household being engaged in a nonfarm job declines sharply as the share of agriculture workers in a woreda increases from 50 percent to 70 percent (Figure 22). However, the existence of an agriculture sector can still be accompanied by more off-farm job opportunities. Nonfarm wage job prospects for both males and females are highest when half of the population in the woreda is engaged in the agriculture sector. Under suitable conditions, jobs can be created in the agri-food system, generating off-farm opportunities alongside a vibrant agriculture sector. Figure 22. Prospects for nonfarm jobs decline as the share of agriculture workers in a woreda increases above 50 percent Probability of nonfarm employment and share of agriculture workers 35% Probability of nonfarm employment 30% 25% 20% 15% 10% 5% 0% 25% 40% 50% 55% 65% 70% 75% 80% 85% 90% 95% Share of agriculture workers in woreda Household engages in nonfarm Male household head Female spouse Source: Authors’ estimates based on ESS 2018/19. High population density appears to be the critical condition for availability of off-farm jobs. The evidence from the ESS shows that nonfarm job opportunities in rural areas are mostly created in the services sector such as retail shops trading food and agriculture inputs or those providing agriculture services. The share of workers employed in the services sector gradually increases as population density in woredas increases 35 Ethiopia Country Economic Memorandum (Figure 23). Employment in industry only emerges in very high-density areas with population density of more than 5,000 people per km2 (e.g., the population density of Dire Dawa). Figure 23. Employment in services and industry increases with higher population density Employment by sector and population density 1 0.9 0.8 Share of workers 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 - 100 per 100 - 500 500 - 1000 1000 - 5000 5000+ per km2 per km2 per km2 per km2 km2 Population density Agriculture Industry Services Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted marginal probabilities of being in nonfarm work based on local labor sectoral shares. The share of workers in each sector is tabulated with local population density. However, the majority of the rural workforce is concentrated in areas with low population density . Because services tend to thrive on local demand, nonfarm job opportunities in rural areas require a minimum population threshold to take off sustainably. The availability of such opportunities is thus largely dependent on population density. Yet the rural population is concentrated in low density areas—about 26 percent of people ages 15 years and above reside in areas with less than 100 people per km2, and an additional 53 percent are in areas with 100-500 people per km2. Thus, four in five working-age people in rural Ethiopia live in low-density areas with limited off-farm job creation opportunities. The evidence confirms that local population density is a primary driver of access to nonfarm opportunities in rural areas, after accounting for other factors that influence access to nonfarm jobs . In low population density woredas with less than 100 people per m2, male household heads have a 10 percent chance of being in a nonfarm job. Their off-farm job prospects jump to about 30 percent in woredas with a population density of 1,000-2,000 people per m2 (e.g., similar to population density in regional capitals such as Adama and Mekelle). The same trend is observed among female spouses. However, male household heads’ nonfarm job prospects increase faster than those of female household members as local population density increases (Figure 24). Therefore, the gender gap in involvement in nonfarm employment is wider in highly populated areas. 36 Ethiopia Country Economic Memorandum Figure 24. Nonfarm employment prospects increase with population density for both males and females Nonfarm employment probability by gender and population density Source: Authors’ estimates based on ESS 2018/19 Notes: Predicted marginal probabilities of being in nonfarm work based on local population density for male household heads and female spouses who live in rural areas. The widening gender gap in nonfarm job participation is a result of women having low education, which reduces their ability to take advantage of increased nonfarm opportunities in high population density areas. Job prospects for male household heads and female spouses with secondary education increase by a similar margin as population density increases, but they increase much more for male household heads than for female spouses among people without secondary education (Figure 25). For example, when comparing nonfarm job prospects of people without secondary education in woredas with less than 100 people per m2 to those in woredas with over 5,000 people per m2, the nonfarm job prospects increase by 26 percentage points for male household heads but only by 13 percentage points for female spouses. This shows that low education adds to women’s disadvantages in the labor market, even after accounting for local economic conditions. Figure 25. Lower educational attainment among women widens the gender gap in nonfarm employment prospects, particularly in high density population areas Educational attainment and probability of nonfarm employmeny, by gender 0.5 Probability of nonfarm employment 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 No Secondary Secondary No Secondary Secondary Male household head Female household member 0 - 100 per km2 100 - 500 per km2 500 - 1000 per km2 1000 - 5000 per km2 5000+ per km2 Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted as marginal probability of nonfarm employment for male household heads and female spouses in rural areas, then tabulated by local population density and level of education. 37 Ethiopia Country Economic Memorandum 2.3.5 Small towns, connectivity, and rural-urban employment spillovers Proximity to economic hubs boosts the likelihood of engaging in nonfarm employment, partially mitigating the effects of low local population density. However, the employment spillovers from urban areas to rural areas rapidly decline within 60 minutes of commuting distance between a rural woreda and the nearest urban center (Figure 26). The nonfarm employment spillovers fade away when urban centers are more than 90 minutes from the woreda. Notably, gender gaps in nonfarm participation persist irrespective of proximity to the nearest urban center, with male household heads having a high chance of being engaged in nonfarm activities compared to their female spouses. As rural areas need to be near urban centers to benefit from employment spillovers, small towns could play a significant role in improving access to nonfarm job prospects in rural areas. Figure 26. Proximity to urban centers helps improve nonfarm employment prospects for low population density areas Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted as marginal probablity of nonfarm employment based on travel time to the nearest urban center for male household heads and female spouses. Small towns can play an important role in rural poverty reduction and the creation of off-farm employment. Small towns account for a large share of Ethiopia’s urban population and will continue to do so. In 2015, 51 percent of the urban population lived in small towns (those with a population of below 50,000). Although their relative share will decline, a large proportion of the urban population will continue to reside in small towns in the near future (Figure 27). The role of small towns in poverty reduction between 2011 and 2016 was strong. They can contribute to faster poverty reduction in rural areas by improving agricultural productivity through greater availability of modern and cheaper inputs and increasing demand for agricultural products (World Bank, 2020). Proximity to towns can also create opportunities for off-farm employment through nonfarm employment spillovers from productivity leakages and increased demand for goods and services by rural households, as well as through increased migration to these secondary cities that are closer and have stronger social-cultural ties. 38 Ethiopia Country Economic Memorandum Figure 27. A large portion of the population continues to live in small towns Urban population trends and projections by 55 town/city type, 2007-2035 50 5.8 Less than 50,00 45 50,000 to 100,000 40 100,000 to 500,000 20.2 35 Addis Ababa Millions 30 4.6 25 5.6 9.2 20 15 3.3 3.3 3.5 10 2.7 1.8 20.9 2.3 14 5 1.1 8.9 5.7 0 2007 2015 2025 2035 Source: adapted from the World Bank (2020). Evidence suggests that small towns significantly improve access to nonfarm job prospects in low- density rural areas. People in woredas within an hour commute to the nearest small or medium town have a higher chance of engaging in nonfarm jobs irrespective of local population density. The benefits of proximity to small towns are relatively higher for low population density areas. In woredas with less than 100 people per m2, for example, the chance of being in a nonfarm job is 3 percentage points higher (equivalent to a 51 percent increase) if the woreda is within an hour commute to a small town compared to those farther away (Figure 28, left panel). Nonetheless, the overall effect of local population density dominates over proximity when it comes to prospects for nonfarm employment. The chance of nonfarm employment increases by around 9 percentage points from low to medium population density woredas (of between 500 – 1,000 people per m2), irrespective of their proximity to small towns (Figure 28, right panel). Figure 28. Close proximity to small towns offers relatively higher benefits for areas with low population density Nonfarm employment and distance Nonfarm employment and density 20% 18% Probability of nonfarm employment Probability of nonfarm employment 18% 16% 16% 14% 14% 12% 12% 10% 10% 8% 8% 6% 6% 4% 4% 2% 2% 0% 0% 0 - 100 per km2 100 - 500 per 500 - 1000 per 0 - 100 per km2 100 - 500 per 500 - 1000 per km2 km2 km2 km2 Within 60 minutes Above 60 minutes Less that 20,000 people Above 20,000 people Source: Authors’ estimates based on ESS 2018/19. Notes: Predicted as marginal probablity of nonfarm employment based on travel time to small towns and their local population density. 39 Ethiopia Country Economic Memorandum Overall, access to nonfarm jobs in rural areas is influenced by economic density, predominantly driven by local population density and connectivity to economic hubs. This can be captured using a measure of market access (defined in Box 2) capturing the size of the accessible population to a woreda by factoring the travel time from the woreda to all other woredas and their population. That means a small rural woreda within a short commuting distance to Addis Ababa is still deemed to have a large market access. Because small towns tend to be much closer to most rural areas, they contribute effectively more to market access for rural woredas in the hinterland of Ethiopia. They can be the primary driver for providing the high density required to support take-off of services sector jobs in rural areas. Box 2. Measurement of market access in Ethiopia The market access indicator for a woreda is estimated as the sum of travel time weighted population to destination woredas. With woreda-to-woreda origin-destination matrices, market accessibility is estimated by the following equation (Donaldson and Hornbeck 2016; World Bank 2019a): − = ∑ ≠ Where is market access at woreda “o”, is the trade cost between two woredas “o” and “d”, is the population of woreda “d”, and is the trade elasticity. Trade costs between two woredas, is defined by = exp⁡( ) with = 0.02 and the optimal travel time between woredas using the transport network of 2020. The trade elasticity, has a value of 8.28 (Eaton and Kortum 2002). Better market access increases nonfarm employment prospects for both males and females but benefits women more. At above-average level of market access, nonfarm job prospects for female spouses in the lowest-density areas (less than 100 people per m2) rapidly converges with nonfarm job prospects in woredas with population density between 100-500 people per m2 (Figure 29, right panel). For males, however, the gap in prospects widens. Females thus have more to gain from increasing market access in Ethiopia. Figure 29. Better market access improves nonfarm employment prospects, particularly for women Nonfarm employment prospects and market access in low-density areas Male household heads Female spouses Source: Authors’ estimates based on ESS 2018/19. 40 Ethiopia Country Economic Memorandum Notes: Predicted marginal probabilities of being in nonfarm work based on market access for male household heads and female spouses. This is estimated using local population density. However, the size of the internal market or the immediately accessible population still plays a dominant role in determining nonfarm job prospects. For the lowest-density areas with low market access, the prospects of nonfarm jobs are negligible for both male household heads and female spouses. Job prospects only start to improve noticeably when the level of market access is about 0.25 standard deviations from the average—similar to the market access levels of Adama city (back to Figure 29). The importance of an immediately accessible population is evident in how nonfarm employment is significantly higher in woredas with more than 20,000 people within 30 minutes of the woreda (excluding the woreda’s own population) compared to those surrounded by fewer people. In summary, these findings suggest that rural-urban spillovers are limited by three factors at the individual, household, and community levels. At the individual level, low education is a primary constraint reducing the ability of people—especially women—to access off-farm opportunities. At the household level, the primacy of meeting subsistence needs in household labor allocation decisions in response to food prices inhibits the movement of labor out of agriculture. Rural consumers faced with high food prices tend to engage less in nonfarm activities and more in agriculture instead, and both consumers and producers are biased toward keeping women on farms for this purpose. At the community level, most rural areas have very low population density, so they do not have a population large enough to support service enterprises through which nonfarm employment opportunities are typically introduced in rural areas. This low density is partly linked to low connectivity and low market access, which also affects food prices in a way that slows the transition out of agriculture. 2.4 Rural-urban migration: benefits, determinants, and constraints Rural-urban migration is an important pathway for improving access to better income-generating opportunities in Ethiopia. The above analysis established that low economic density limits access to local economic opportunities and the potential to create these jobs. Thus, increased labor mobility out of low-density areas would help mitigate the disadvantages faced by distant low-density economies. However, rural-urban migration in Ethiopia has been low. This section highlights the potential benefits of rural-urban migration in Ethiopia, analyzes the trends and determinants of rural-urban migration, and discusses its constraints. 2.4.1 Benefits of rural-urban migration in Ethiopia Global evidence indicates that rural-urban migration can improve the welfare of both the migrants themselves and their families in the place of origin (Lall et al. 2006). Migrants can benefit from access to employment as well as improved access to services including education. They can also send remittances to their family members back home, improving the welfare of the recipients not only by directly increasing their consumption but also by improving investment in both human capital and physical capital and encouraging risk-taking behavior. The benefits of migration can be particularly high if those from extremely poor households migrate. Data from Ethiopia also show the benefits from migration. Using the latest two rounds (2005 and 2009) of the Ethiopian Rural Household Survey (ERHS), Brauw et al. (2018) found that migrants from rural areas (particularly those who migrated to urban areas) have higher consumption and nutrition compared to those who stayed in their villages of origin. New analysis based on the three rounds of LSMS data (2012 to 2016) shows that while only a small fraction (five percent) of rural households receive remittances from urban areas, the amount they receive (as a fraction of their total consumption expenditure) is significant, particularly for the 41 Ethiopia Country Economic Memorandum poor (Table 5). The average amount of remittances received by households in 2016 was equivalent to 31 percent of recipient household consumption expenditure nationally and 70 percent among the bottom quintile. Table 5. While a small fraction of households receives remittance, the amount they receive is significant Share of households receiving remittances, by quintile 2012 2014 2016 % of Amount % of con- % of Amount % of % of Amount % of con- households received sumption households received consu household received sumption receiving (birr) receiving (birr) mption s receiving (birr) Poorest quintile 3.6 798 48.6 2.8 1209 74.2 1.6 1099 70.1 Second 5.6 543 19.5 4.6 1026 34.9 4.8 889 31.6 quintile Third quintile 5.3 1124 26.7 6.8 1995 51.1 5.5 1317 31.7 Fourth quintile 5.4 1581 26.7 5.1 1115 19.8 7.0 1958 32.9 Richest quintile 6.7 1338 11.9 7.5 1288 12.8 6.8 1573 15 National 5.2 1059 25.8 4.9 1405 40.8 5.1 1455 31.3 Source: ESS (2012, 2014, 2016); Authors’ calculations. Note: Remittance amount is given in 2012 price and is calculated only for those who receive it. An important aspect of rural-urban migration is how successful migrants are at their destination and which motivates them to migrate. Using the 2013 Labor Force Survey, the 2020 Ethiopia Poverty Assessment (World Bank 2020) documented that migrants’ labor market integration is reasonably good, where the migrants’ employment situation (at least for those who stayed a few years at the destination) is not much different from that of the resident population. However, migrants lag in their educational outcomes and their children’s education vis-à-vis other residents. Nonetheless, a more recent qualitative study (World Bank 2018) shows that most migrants consider their decision to migrate to be right despite the challenges they face. According to the migrants, the major benefits of migration include increased income and better access to education and skills. For females, escaping the traditional gender roles that prevail in rural areas, including early and/or arranged marriage, is another benefit of migrating to urban areas. 2.4.2 Patterns and trends in rural-urban migration in Ethiopia Despite the large income differences between urban and rural areas and the potential benefits of migration, the scale of internal migration in Ethiopia has not changed over time. Only 6 percent of Ethiopians changed their zone of residence between 2008 and 2013. The share of recent migrants (those who migrated in the last five years) in the urban areas has been stagnant, hovering between 17 and 20 percent from 1999 to 2013, while the share of all-time migrants in the urban population has declined. Outside Addis Ababa, high fertility and rezoning of urban areas have contributed more to urban population growth than migration (World Bank 2018). A recent study by Tsegay and Litchfield (2019) using data from a sample of nine woredas in Tigray, Amhara, Oromia and SNNP, selected from zones with relatively higher migration historically, shows that only 10 percent of rural households had an internal migrant, mostly to urban areas, between 2014 and 2018. The evidence on limited rural-urban migration is corroborated by fresh analysis based on the three rounds (2012 to 2016) of ESS panel data. The analysis shows that only between 2-3 percent of rural people migrated to urban areas between 2012 and 2016 (Figure 30, left panel).6 Rural-urban migration was higher among youth but still only around 7 percent this period. There is no clear relationship between welfare status and rural-urban migration (Figure 30, right panel). The overall migrant population figures under ESS seems to 6 The most recent round (2019) starts with new sample households and does not include a module on migration. 42 Ethiopia Country Economic Memorandum be consistent with Labor Force Survey data, which suggest the migrant population increased from 15.1 percent in 2013 to 16.5 percent in 2021. Figure 30. A small fraction of rural people migrate to urban areas across all quintiles Rural-urban migration by age group Distribution of migrants by consumption quintile 8 30 7 6 25 Percent of total 5 20 Percent 4 15 3 10 2 1 5 0 0 All age group Working age (15- Youth (15-29) Poorest Second Third Fourth Richest 64) quintile quintile quintile quintile quintile Between 2012 and 2014 Between 2014 and 2016 Between 2012 and 2014 Between 2014 and 2016 Source: ESS (2014, 2016). Authors’ calculations. Note: Return migration and urban-rural migration are not included, so figures are not net rural-urban migration. Migration of whole household (where all household members migrate) is also not included, although it is expected to be small. The nature of migration shows that small towns facilitate rural-urban migration. Nearly three-quarters of the new rural-urban migrant population between 2012 and 2016 moved to woreda (47-53 percent) or zonal (21-27 percent) towns (Figure 31). These movements were mostly within the same woreda or zone. Also, small towns tend to have a higher share of recent migrants compared to large towns, further demonstrating how small towns are integral to facilitating labor movements. Figure 31. Woreda towns are the most important destination for rural-urban migrants Distribution of rural-urban migrants by destination type 100 10.6 12.6 90 80 15.1 13 70 26.4 21.8 60 Percent 50 40 30 47.8 52.6 20 10 0 Between 2012 and 2014 Between 2014 and 2016 Woreda town Zone town Regional town Other urban Source: Authors’ estimates based on ESS (2014, 2016). 43 Ethiopia Country Economic Memorandum 2.4.3 Profile of migrants and determinants of migration Past evidence shows that both individual and community characteristics determine rural-urban migration (World Bank 2020). Findings from the Labor Force Survey (LFS) 2013 show that migrants tend to be younger and more educated compared to those who do not migrate, while distance to all-weather roads and zonal-level poverty rates are associated with lower likelihood of migration. Analysis of the ESS panel data between 2012 to 2016 shows that the profile of migrants remains largely the same. Migrants tend to be younger and more educated and are more likely to be males. They also tend to come from families with more educated household heads and families that are better connected, i.e. closer to the woreda town (Table 6). Multivariate analysis confirms that age and education are important determinants of migration, all other factors considered (Box 3). Table 6. Migrants tend to be male, younger, and relatively more educated Profile of rural-urban migrants – working age population: 2012 to 2016 Between 2012 and 2014 Between 2014 and 2016 Migrants Non-migrants Migrants Non-migrants Female (%) 39.6 51.17 43 50.8 Age 21 32 21 32 No education (% of total) 8.1 52.9 10.5 50.6 Some primary (%) 55.5 39.2 55.5 39.7 Primary complete (%) 3.7 2 2.3 2.8 Some secondary (%) 10.4 3.1 19.3 3.8 Secondary and above (%) 22.2 2.9 12.4 3 Household size (mean) 6 6 6 6 Number of adults (18-64) 3 2 3 3 Head’s maximum education (years) 8 5 7 5 Distance to nearest woreda (kms) 15.6 20.1 16.3 19.3 Source: ESS (2014, 2016). Authors’ calculations. Education is associated with a higher likelihood of rural-urban migration, even after accounting for other factors. Multivariate regression analysis (using a multinomial logit regression) shows that individuals with secondary education have a 12 percent higher likelihood of migrating compared to those without any education. Even those having some primary education have a two percent higher probability of migrating compared to those who do not have any schooling, underscoring the importance of education in rural-urban linkages. Males are less likely to migrate after controlling for other characteristics, suggesting that higher migration rates among males is driven by more males having characteristics like higher education which are positively correlated with rural-urban migration. Box 3. Analysis of the determinants of rural-urban migration: data and methodology To complement the existing evidence on the determinants of rural-urban migration using findings from more recent and richer data, further analysis was done using the three rounds of the LSMS panel data collected in 2012, 2014, and 2016. The LSMS data is representative of Ethiopia and some of the regions. Although not a specialized migration or labor survey, the LSMS data has detailed information about migrants including their destination and characteristics. More specifically, information about migration of household members was asked in the second and third rounds as part of updating the household roster, providing the opportunity to see who migrated between the rounds (between the first and second rounds and between the second and third rounds). Its panel nature, combined with detailed information on monetary and non-monetary household welfare and geospatial variables, allows better control for initial household characteristics. Accordingly, migration was modeled at two points (2014 and 2016), by observation migration outcomes of individuals in an earlier round, controlling for lagged individual and household characteristics. In other words, for migration observed in 2014, the explanatory variables come from the 2012 round; for migration observed 44 Ethiopia Country Economic Memorandum in 2016, the explanatory variables from 2014 were used. This approach helps minimize the potential endogeneity and control for unobserved heterogeneity. Among other things, the data on migration includes destination of the migrant (other rural areas, urban areas, and abroad) and main reason of migration. Two types of models were run to exploit this information. First, a probit model was run on the pooled data, where the dependent variable is any migration to urban areas (regardless of the reason for migrating. A multinomial logit regression was then run with the possible outcomes—no migration, migration for economic reasons (which includes migration for job and related reasons), migration for education, and migration for family related reasons (e.g., marriage). This helped show if determinants of migration vary by reason of migration. Economic considerations like labor supply and access to local opportunities also influence household migration decisions. Those individuals from households with more working-age members (particularly those with more male working-age members), hence those likely to have more labor supply, have a higher probability of migrating. However, those from households with a higher number of employed members in the initial year are less likely to migrate for economic reasons (Figure 32, left panel). Migration is also less likely among those who participate in Productive Safety Net Programme (PSNP) public works. Moreover, individuals from households with more farmland are also less likely to migrate (Figure 32, right panel). Notably, these factors are only significant determinants for economic migration and not migration for other reasons, indicating that economic migration to urban areas is to some extent driven by push factors. Figure 32. Households with more employed members and more land are less likely to have members who migrate Likelihood of migration for economic Likelihood of migration for economic reasons, by number of employed members reasons, by land ownership 3.5% 1.8% 3.0% 1.7% Probability of migrating Probabilty of migrating 1.6% 2.5% 1.5% 2.0% 1.4% 1.5% 1.3% 1.0% 1.2% 1.1% 0.5% 1.0% 0.0% 0-0.2 ha 0.2 -0.5 ha 0.5-1 ha 1-2 ha Above 2 2 or less 3 or 4 5 or 6 Above 6 ha Number of employed household members Farm land owned by the household Source: Authors’ estimates based on ESS 2012-2016. The push factors for economic migration are stronger among poor households. Migrants do not come from the poorest group. Compared to those who come from the top consumption quartile, only those in the middle (especially those in the third quartile) are more likely to migrate when all types (by reason) are lumped. However, among those migrating for economic reasons, individuals in the bottom three quartiles are significantly more likely to have migrated two years later than individuals from the top quartile, although the marginal effects are small (see Annex 2 for detailed results). Individuals in the bottom quartile have a higher likelihood of migration than the top quartile but have a lower likelihood of migration compared to those in the middle, suggesting that migration at the bottom might be held back by financial constraints. 45 Ethiopia Country Economic Memorandum Indeed, road accessibility (which is a proxy for transportation costs) is the second-most important determinant of economic migration. While not an impediment to those who migrate for family reasons, individuals residing further away from a major road are less likely to migrate for economic reasons. Transport costs make more challenging the migration for remotely placed individuals, adding to already significant costs of housing and job search in the destination areas. Compared to those who do not have any education, those with at least some secondary education are significantly more likely to migration to urban areas for economic reasons. 2.4.4 Barriers to migration The rural-urban wage differentials in Ethiopia are large enough to incentivize migration, at least from a monetary perspective. In search of an explanation for the low rural-urban migration in Ethiopia, some studies have explored whether wages in urban areas are high enough relative to rural wages to provide an incentive to migrate. For example, using data from four African countries including Ethiopia, McCullough (2016) found that differences in wages/productivity between the agricultural and non-agricultural sectors are limited after accounting for differences in hours/days worked. However, analysis using the most recent (2019) round of LSMS shows that while rural-urban wage differences narrow when hours worked are taken into account, they remain significant. The gap is not negligible even after accounting for the difference in the probability of being employed; expected hourly wage is 16 percent higher in urban areas than in rural areas.7 The average monthly wage adjusted for hours worked is three times higher in urban areas (Table 7). The average urban wage is also significantly higher than average agriculture income (including income from crops, livestock, renting agricultural inputs, and wage income in agriculture). It is also worth noting that the wage employed in rural areas are better educated and have higher welfare than the average rural resident which implies that the average wage is an upper bound relative to what a typical rural resident would earn from wage employment.8 The annual average agricultural income (total income divided by household members ages 18 and above) was 10,057 birr in 2019. Table 7. Wages in urban areas are considerably higher than in rural areas Rural-urban wage differentials are significant in Ethiopia (2019) Weekly hours worked Hourly wage Monthly wage Urban 43 36 4,862 Rural 27 27 1,412 National 39 33 3,972 Source: ESS (2019); Authors’ calculations. Note: There are also differences in number of weeks worked in a month (not presented here). The significant rural-wage differential thus points to the existence of other barriers to migration. A recent qualitative study (World Bank 2018) shows that rural migrants face various challenges that make transitioning to urban life and finding jobs difficult. These challenges limit the success of migrants in their destination and, by extension, their potential to support their families back home by sending remittances. Some of the highlighted key challenges include administrative, economic, and socio-cultural barriers as summarized below. 7 Based on the ESS (2019), unemployment rates in urban and rural areas were respectively 19 percent and 6 percent; the probabilities of being employed were 81 percent and 94 percent, respectively. 46 Ethiopia Country Economic Memorandum Administrative barriers Although Ethiopia does not have formal laws or regulations restricting rural-urban migration, a host of administrative barriers limit rural-urban migration. The most important ones include: • Difficulty obtaining an urban ID: Migrants perceive that government services are not as available for migrants as they are for urban residents. Of primary significance is migrants’ inability to access an urban identification (ID) card, due to either their inability to fulfill the requirements or authorities’ reluctance to provide them with IDs. IDs are issued by kebeles (the lowest administrative units below woreda), and for migrants to get an ID from the kebele in which they currently reside, they must produce a release/reference letter along with an ID from their original kebele or prove that they have been living in the current kebele. Most migrants from rural areas lack such documentation. Furthermore, kebele officials require proof of residence before issuing an ID and that is difficult for migrants because they mostly are renters and are required to ask their landlords to accompany them to the kebele offices to vouch for them, something landlords are usually reluctant to do. According to the World Bank’s Identification for Development (ID4D) data, Ethiopia was also among the countries with the highest gender gap where men are about 20 percentage points more likely to own an ID than women (World Bank, 2019b). Thus, the problem is likely to be more serious for women migrants than for men migrants. A kebele ID is required to access a wide range of services including obtaining a SIM card, opening a bank account, and obtaining a passport or driving license as well as being able to travel freely (if stopped by police and unable to produce an ID, an individual could be detained and questioned).9 An ID from the specific kebele is also required to obtain services provided by kebeles/woredas such as low-interest loans, employment support, and subsidized goods and services. As a result, migrants without an ID find it difficult or impossible to access these benefits and services, and local community members treat those without urban ID cards with suspicion. • Unfavorable treatment by urban authorities: Migrants believe that rural authorities view migrants negatively so are reluctant to provide services to them, including IDs (even when they satisfy the requirements). Migrants also mentioned that they sometimes face harassment from police officers. This seems to be consistent with how authorities view migrants: most authorities do not welcome migrants, and some even consider rural- urban migration illegal. They believe that rural youth are not supposed to migrate to urban areas and should stay where they are and rely on agriculture. Some reasons why authorities do not welcome rural migrants include lack of job opportunities in urban areas, fear of more migrants coming, limited budget, and pressure on urban services such as housing and sanitation. While their concerns might be valid, it is worrying that they do not consider rural-urban migration to be a natural process, especially at a time when the country is trying to achieve an ambitious transformation plan which aims to move a significant part of the rural population to urban areas. Economic barriers • Cost of settlement: While transportation costs to urban areas might not be expensive for many potential migrants, the cost of living (e.g., housing, food) at the destination can be expensive, especially in the first few months when migrants are still looking for a job or not making much money. As a result, migrants try to migrate to destinations where they think finding a job is easier and/or where they have a family member, friend, or acquaintance. Migrants without a network find settling in difficult and expensive. Networks also help migrants find jobs more quickly. 9 For details on the nature and function of kebele IDs, see the World Bank’s 2016 ID4D Country Diagnostic for Ethiopia. 47 Ethiopia Country Economic Memorandum • Access to quality jobs: While some rural migrants find jobs easily, the majority have difficulty finding jobs. Some migrants identified nepotism and bribery as a serious challenge, where only those with relatives at the potential workplace or those who can pay bribes to the recruiters or relevant authorities can get jobs. Some migrants mentioned that certain jobs require references, which is difficult if they do not know anyone at their destination and do not have work experience. They also identified lack of transparency and long bureaucratic process as problems associated with obtaining government jobs. Furthermore, as mentioned above, migrants typically do not qualify for government-supported job creation programs (either participating in government projects as casual laborers or getting support to open small businesses) because they do not satisfy the administrative requirements such as having a kebele ID. Even when they are lucky and get a job, the jobs tend to be casual and not stable, require working long hours, and subject them to various abuses. Participants reported facing various forms of abuse by employers, such as denial of salary, false accusation of theft, sexual harassment, untimely payment of salary, and exploitation of labor. Socio-cultural barriers • Transitioning to urban life and facing harassment: Migrants typically find the transition into city life difficult, and females face additional challenges. At the beginning, youth who migrate from rural to urban areas find fitting into social environments and practices common in urban areas difficult. Language is a major barrier for those who migrate to areas where their mother tongue is not spoken. Migrants are also easy targets for theft, and many have experienced harassment including from police officers, landlords, and employers. It is common for new women migrants to face sexual abuse and harassment from their work placement agents and others. Some even end up with unwanted pregnancy and sexually transmitted diseases (STDs). 2.5 Policy options The analysis presented in this chapter points to two pathways for rural households to benefit more from Ethiopia’s economic growth: expanding access to rural nonfarm activities and increasing rural- urban migration to link rural labor to economic growth centers. Creating off-farm opportunities in rural Ethiopia thus requires overcoming the challenge of low economic density. With geographical limitations to rural job creation, promoting rural-urban migration is essential but requires addressing the constraints that slow the integration process for migrants and discourage migration. Potential policy directions for each of these pathways are presented below, and Table 8 at the end of this section provides a summary. 2.5.1 Expanding rural off-farm opportunities Creating off-farm opportunities in rural Ethiopia requires adopting a development strategy that is appropriate for low-density economies. Such a strategy recognizes how the process of growth in remote and sparsely populated areas differs from that of high-density areas in fundamental ways. Unlike in high-density areas, growth in low-density economies (i) is driven by local absolute advantages (mostly in natural resources), (ii) depends on external demand from domestic (i.e. cities) and international markets given their small internal markets, and (iii) depends on micro and small enterprises since it is difficult to achieve scale in sparsely populated areas. Off-farm opportunities in rural Ethiopia could therefore be created by adopting two complimentary strategies that fit low-density economies. One strategy is to develop the secondary economy and expand economic functions, building on the natural resource advantages of rural Ethiopia (basically land and agriculture). The second strategy is to link rural areas to the network economy—urban and international food markets—which is the source of demand for products produced in rural areas. 48 Ethiopia Country Economic Memorandum Developing a secondary economy around the agri-food system Developing the agro-food system offers the best path for rural off-farm job creation in Ethiopia. Since production in low-density economies tends to be based on absolute advantages, off-farm opportunities can be created by developing industries based on these advantages. Ethiopia has abundant land, with high agriculture potential, so growth in the rural off-farm sector can be anchored in the development of the agri-food system. Further growth can be achieved by building a post-production economy in areas that support agricultural production. Such activities include developing producer services such as logistical services (warehousing and post harvesting management, input and output aggregation) and business and accounting services that could also be managed by small and medium-sized enterprises (SMEs) or complement consumer services. Specifically, growth in the rural off-farm sector can be anchored in the development of the agri-food system by: • Developing the skills necessary to provide services to the agriculture sector and post-agriculture production activities. Certain technical skills are needed to expand the economic functions of rural enterprises to support a post-production economy or to link to the network economy. Producer services such as marketing and business planning also require particular skills, and participation in digital platforms requires digital literacy. Yet most labor—especially female—in Ethiopia is low-skilled due to low educational attainment. The analysis also shows that better-educated people have improved access to off- farm jobs and greater chances of migration. Furthermore, women’s disadvantage in access to the labor market is reduced considerably with attainment of secondary education. Interventions needed to support skills development include: o Incentivizing school progression using conditional cash transfers, especially for girls, to broadly improve educational attainment as a foundation for developing a skilled workforce. o Providing vocational training programs for youth and women, focusing on developing skills to promote agriculture services and to provide secondary services to rural enterprises (e.g., business development services). • Promoting agricultural value chains. Developing jobs within agri-food systems requires developing agricultural value chains and promoting household engagement and private sector investment in the postproduction stages of the value chains by providing an adequate infrastructure and regulatory environment. Interventions to support value chain development include: o Investing in logistical services such as warehousing and storage facilities to support promising agricultural value chains. o Establishing food safety and trackability mechanisms to penetrate high-standard, premium food markets. o Promoting farmer market linkages to improve household market participation, production quality, and market response. • Supporting rural SMEs through complimentary investments. Because low economic density limits economies of scale, growth in employment opportunities in Ethiopia’s rural areas must be driven by SMEs. Complimentary investments and policies to improve the business environment for SMEs operating in rural areas are therefore critical. These measures include: o Improving access to finance for rural enterprises and agribusinesses through a combination of modalities such as matching grants for medium enterprises, credit lines, and an expansion of direct financial support to common interest groups. 49 Ethiopia Country Economic Memorandum o Supporting rural entrepreneurship development through business incubation programs, focused on promoting business start-ups on agricultural value chains. Integrating rural areas into the network economy Linking rural areas to the network economy is crucial to make up for low economic density. The lack of density makes growth of nonfarm opportunities in sparsely populated rural areas dependent on external demand to sell more of what they produce or to obtain a higher price for their products. Links to markets, both domestic and global, therefore become vital for expanding access to off-farm opportunities in rural areas. Producers of teff, whose prices increase more with domestic market access, would benefit more from increasing local connectivity, while some products such as maize would benefit from access to global markets, as discussed in Chapter 5 on food prices. Thus, opportunities can be created around regional value chains. To increase rural areas’ access to markets, improvements in both physical and digital connectivity are needed. Good connectivity is essential both for integrating lagging rural areas into the network economy and developing a secondary economy, maximizing rural-urban employment spillovers. By integrating food markets, improved connectivity also makes it easier for households to release labor into non-food production activities and generate incomes to source food from the markets. For rural Ethiopia, improving connectivity would require from: • Expanding the rural road network. More than three-quarters of rural households in 2019 lived in communities only accessible by either a dirt road (19 percent) or dirt tract (58 percent), while 87 percent resided in areas where the main road does not have year-round accessibility by lorries. Thus, investment in rural roads is a necessary last mile connection to increase rural accessibility. • Building secondary roads connecting rural areas to small and medium towns. As discussed earlier, urban-employment spillovers are highest within an hour connection, best achieved with development and connections to small towns. These small and medium towns are also a base for developing the secondary economy. • Adopt incentives to promote the expansion of mobile and broadband infrastructure to underserved areas. While roads make it easier to transport goods to markets or for workers in neighboring localities to work, a smooth exchange of information and connections with intermediaries and/or the ultimate buyers are also needed. Thus, in addition to a good transport network and logistical services, ICT investments are particularly crucial for reducing economic distance and leveraging digital technologies to generate new opportunities in rural areas. While the issuance of new telecom licenses is expected to foster competition and result in an expansion of the ICT network, legal provisions for infrastructure sharing and, possibly, incentives to build towers in less densely populated areas may be necessary to boost the integration of the rural economy. 2.5.2 Facilitating rural-urban migration The above policies to expand off-farm opportunities are also important for supporting rural-urban migration. The same factors inhibiting access to nonfarm jobs in rural areas also constrain rural-urban migration. Poorly educated people are less likely to migrate, and household dynamics that favor labor allocation into agriculture work against migration. In addition, people in remote areas without access to roads are less likely to migrate. The policies presented above to increase off-farm opportunities therefore also help facilitate rural-urban migration. 50 Ethiopia Country Economic Memorandum Additional policies and interventions could focus on addressing constraints that slow the integration process for migrants, potentially discouraging migration. It is recommended that authorities focus on helping migrants better connect to jobs and improving their access to services in their destination areas. Connecting migrants to jobs Improved job intermediation can play an important role in connecting migrants to jobs. Migrants with limited social networks find job search difficult and challenging. This situation could be improved by establishing a well-functioning Labor Market Information Systems (LMIS) or labor exchanges through: • Setting up an Employment Agency and streamlining Public Employment Services (PES). To reduce job search costs and matching frictions, labor market intermediation would need to be strengthened in line with the National Plan for Job Creation 2020-2025. Strengthening labor market intermediation will require addressing the fragmented nature of PES provision to date. The recent creation of the Ministry of Labor and Skills could address this need, at least from a mandate point of view, although the exact institutional set-up to provide employment services is still unclear. • Expanding youth apprenticeships programs. Youth apprenticeship programs like the one under the Urban Productive Safety Net and Jobs Project (UPSNJP) can address the inability of youth to signal their skills and help connect them to better employment opportunities. These programs need to be complemented by greater access to demand-driven training in urban areas through the PES to enhance skills development and connect people to jobs. Improving migrants’ access to services Greater efforts are needed to help migrants access services and employment support from the government. Making the process of acquiring an urban ID and residence easier for new migrants will help in this regard. Important measures include: • Minimizing the burden of requirements needed to be satisfied by the migrant to obtain an urban ID. This can be achieved by reducing the minimum length of time of stay in an area required and/or eliminating the need for a release letter from the origin kebele. The requirements and process for obtaining the IDs could also be made more transparent, and the requirement for a landlord or someone who owns a house to vouch for the applicants should be repealed and replaced with the requirement to show a proof of residence. • Streamlining ID reforms and household registration. The new government National Identity Program (NIDP) establishes a foundational and digital form of identification aimed at providing a unique identity to all residents in Ethiopia by 2027. Individuals will be able to prove their identity to access public and private services for a variety of use cases (e.g. financial sector, social protection). This new system has been introduced to reduce the current identification gap in the country and address the barriers to access faced by citizens, notably those without any existing form of identification. However, there has not yet been a formal decision of replacing the existing Kebele ID with this new system, and local authorities can still opt for not using it and exclude migrants from various services. Thus, focused interventions are still necessary to prevent discrimination. Another priority is to formulate a policy framework conducive for internal labor mobility . Recognizing that rural-urban migration is a natural and necessary process to achieve economic transformation, it is imperative to develop a policy that aims at maximizing the benefits of migration and minimizing the costs, including the burden on urban areas. Such policies could include: 51 Ethiopia Country Economic Memorandum • Conducting awareness creation campaigns. The objectives of such campaigns would be to change the negative attitude of urban authorities toward rural migrants and to help protect female migrants’ rights, given the specific challenges female migrants face. • Accounting for migrant flows in urban planning and budgeting . It is recommended that special attention is given to improving services like sanitation, housing, and schooling in small towns, which are the most important destination of rural migrants but are lagging in infrastructure. This will minimize the pressure on small towns in terms of providing services and is likely to make officials of small towns more accommodating to rural migrants. Table 8. Policy options to expand access to rural nonfarm activities and increasing rural-urban migration Area Short term Medium term Institution in charge Agro-food Provide vocational training Foster school progression using Ministry of Education, systems programs for youth and women, conditional cash transfers, especially Ministry of Labor and focusing on developing skills to for girls Skills Development promote agriculture services Invest in logistical services and Establish food safety and trackability Ministry of Agriculture, storage mechanisms to penetrate high- Agriculture standard, premium food markets Transformation Agency Promote farmer-market linkages Improve access to finance for rural Support rural entrepreneurship Development Bank of enterprises and agribusiness development through business Ethiopia, incubation programs Integrating rural Build secondary roads connecting Expand the rural road network Ministry of Transport, areas into the rural areas to small and medium Ethiopia Roads network economy towns Authority, Ministry of Innovation and Ensure infrastructure sharing and technology, Ethiopia incentives to expand ICT Communications connectivity to rural areas Authority Connecting Streamline Public Employment Expand youth apprenticeships Ministry of Labor and migrants to jobs Services programs Skills Development Improving Minimize the requirements needed Rollout the centralized and digital ID Ministry of Peace, migrants’ access for a migrant to obtain an urban ID, system and facilitate adoption by regional and local to services including by voiding the public and private actors authorities requirement to have landlords vouching for them Conduct awareness campaigns to Take into account migrant flows in Prime Minister Office, address anti-migrant bias and urban planning and budgeting Ministry of Peace, protect female migrant rights Ministry of Urbanization and Infrastructure, regional and local authorities Source: authors’ elaboration. 52 Ethiopia Country Economic Memorandum 2.6 References Ahearn, M., H. El-Osta, and J. Dewbre. 2000. “The Impact of Coupled and Decoupled Government Subsidies on Off-Farm Labor Participation of U.S. Farm Operators.” American Journal of Agricultural Economics. 88 (2). 393-408. Benjamin, C., and A. Kimhi. 2006. "Farm work, off-farm work, and hired farm labour: estimating a discrete- choice model of French farm couples' labour decisions," European Review of Agricultural Economics, Foundation for the European Review of Agricultural Economics, vol. 33(2): 149-171, June. Brauw, A., V. Mueller, and T. Woldehanna. 2018. “Does Internal Migration Improve Overall Well-Being in Ethiopia?” Journal of African Economies. 27(3). 347-365. Corsi, A., and J. Findeis. 2000. “True state dependence and heterogeneity in off-farm labor participation.” European Review of Agricultural Economics. 27 (2). 127-151. Corsi, A. and C. Salvioni. 2012. “Off- and on-farm labour participation in Italian farm households.” Applied Economics 44 (19). 2517-2526. CSA. 2006. “Report on the 2005 National Labour Force Survey.” Addis Ababa: Central Statistical Agency of Ethiopia. CSA. 2014. “Statistical Report on the 2013 National Labour Force Survey.” Addis Ababa: Central Statistical Agency of Ethiopia Donaldson, D. and R. Hornbeck. 2016. “Railroads and American Economic Growth: A ‘Market Access’ Approach.” The Quarterly Journal of Economics 131 (2). 799-858. Eaton, J. and S. Kortum. 2002. “Technology, Geography and Trade.” Econometrica 70 (5). FDRE. 2018. Poverty and Economic Growth in Ethiopia 1995/6 -2015/16. Federal Democratic Republic of Ethiopia: National Planning Commission. Huffman, W., and M. Lange. 1989. Off-farm work decisions of husbands and wives: Joint decision making. The Review of Economics and Statistics 71, 471–480. Lall, S., H. Selod, and Z. Shalizi. 2006. “Rural-Urban Migration in Developing Countries: A Survey of Theoretical Predictions and Empirical Findings.” Policy Research Working Paper. WPS 3915. Washington, DC: World Bank Group. MaCurdy, T. 1981. “An Empirical Model of Labor Supply in a Life-Cycle Setting,” Journal of Political Economy, University of Chicago Press, vol. 89(6): 1059-1085. McCullough, E. 2016. “Labor Productivity and Employment Gaps in Sub-Saharan Africa.” Food Policy. Mishra, A. and B. Goodwin. 1997. “Farm Income Variability and the Supply of Off-Farm Labor.” American Journal of Agricultural Economics, 79, issue 3: 880-887. National Bank of Ethiopia (NBE). 2020. Annual Report 2019/2020. Addis Ababa. Available at: https://nbebank.com/wp-content/uploads/pdf/annualbulletin/Annual%20Report%202019- 2020.pdf Singh, I., L. Squire, and J. Strauss. (Eds.). 1986. Agricultural Household Models: Extensions, Applications, and Policy. Baltimore: Johns Hopkins University Press. Sumner, D. 1982. “The off-farm labor supply of farmers.” American Journal of Agricultural Economics, 64, 499– 509. 53 Ethiopia Country Economic Memorandum Tokle, J., and W. Huffman. 1991. “Local economic conditions and wage labor decisions of farm and rural nonfarm couples.” American Journal of Agricultural Economics 73, 652–670. Tsegay, A. and J. Litchfield. 2019. “Changing Patterns of Migration and Remittances in Ethiopia 2014-2018.” Migration Out of Poverty Research Program Consortium Working Paper 59. Institute of Development Studies. University of Sussex. United Nations, Department of Economic and Social Affairs (UNDESA). 2018. World Urbanization Prospects: The 2018 Revision. UNDESA, Population Division. Data accessed online: https://www.un.org/development/desa/publications/2018-revision-of-world-urbanization- prospects.html World Bank. 2015. “Ethiopia Poverty Assessment 2014.” Washington, DC: World Bank. World Bank. 2016. “ID4D Country Diagnostic: Ethiopia.” Washington, DC: World Bank. World Bank. 2018. “Internal Migration in Ethiopia: Evidence from a Quantitative and Qualitative Research Study.” Washington, DC: World Bank. World Bank. 2019a. “Better Opportunities for All: Vietnam Poverty and Shared Prosperity Update Report.” Washington, DC: World Bank. World Bank. 2019b. “Global ID Coverage, Barriers, and Use by the Numbers: An In-depth Look at the 2017 ID4D – FINDEX Survey.” Washington, DC: World Bank. World Bank. 2020. “Ethiopia Poverty Assessment: Harnessing Continued Growth for Accelerated Poverty Reduction.” Washington, DC: World Bank. 54 Ethiopia Country Economic Memorandum Part II: Competitiveness 55 Ethiopia Country Economic Memorandum 3 Productivity dynamics in Ethiopian manufacturing 3.1 Motivation: what went off-track? Ethiopia has achieved significant success in jumpstarting its manufacturing sector. Following the agriculture-led industrialization strategy in the 2000s, the goal of boosting manufacturing was placed front and center in the Growth and Transformation Plans (GTPs) of the past decade. Today, Ethiopia is regarded as one of the few countries in Africa that has succeeded in kickstarting the manufacturing sector. Since the onset of GTP I, manufacturing experienced an acceleration, growing at an annual average of 17 percent during 2010/11- 2016/17—well above the already extraordinary gross domestic product (GDP) growth rate of 10 percent (Figure 33). Figure 33. During the past decade, the growth rate of manufacturing was well above that of GDP Growth in manufacturing value-added and GDP 30 20 Annual real growth (%) 10 0 -10 Large- and medium-scale All manufacturing Overall GDP Source: Authors’ calculation based on data from the National Planning Commission of Ethiopia. Note: Growth is annual midpoint growth in sectoral GDP at market prices based on constant local currency. GDP deflator is used as a price index, and 2015/16 is used as a base year. However, the manufacturing sector’s growth rate has slowed in recent years, and its contribution to GDP remains modest. In recent years, the manufacturing growth rate has slowed down to single digits, and its share of GDP has remained relatively small at below 7 percent of GDP. Thus, it has been argued that despite the initial success, “the Ethiopian manufacturing sector is still far from being an engine of growth and economic transformation” (Oqubay 2019). This chapter examines firm-level productivity dynamics in manufacturing and finds evidence of productivity declines in recent years. It finds that TFP in manufacturing has been in decline since 2012. This is consistent with the finding in Chapter 0 of this report that economy-wide TFP dropped during the past decade. Meanwhile, capital and labor productivity peaked by 2015, prior to starting to drop. There is evidence of increasing productivity dispersion in manufacturing between 2012-18, which suggests resource misallocation. 56 Ethiopia Country Economic Memorandum The chapter identifies challenges that might explain these trends. Arguably, the anti-export bias of the exchange rate policy under the state-led development model, coupled with a series of episodes of instability and violence, have hindered the performance of the manufacturing sector in recent years. 3.2 Context: Ethiopia’s manufacturing ambition Like many other developing countries, Ethiopia has the aspiration of becoming a manufacturing hub, which is seen as a key to unlocking economic development. For some successful Asian economies, attracting manufacturing was central to their development process, helping boost productivity and create employment for large portions of the population. Manufacturing activity is particularly valued by policymakers as it can benefit from economies of scale and access to international markets, facilitates technological adoption, and seems to have higher potential for spillover effects, although recently it has become increasingly clear that certain services offer some similar opportunities (Nayyar et al. 2021). Manufacturing has been highlighted as a motor of the shift toward private-sector-led development in Ethiopia’s development plans. The most recent Ten-Year Development Plan (2021-2030) projects an increase in the manufacturing share to 17 percent of GDP. To this end, authorities have built industrial parks, provided priority access to foreign exchange, and extended tax incentives to manufacturers. However, a series of international megatrends and domestic constraints may complicate the attainment of such an ambitious increase in manufacturing activity. In terms of megatrends, accelerated technological change and reshoring of value chain activities complicate the prospects for developing countries willing to attract manufacturing. On the other hand, there are also new opportunities derived from trade integration. These trends are discussed in more detail in the next subsection. 3.2.1 Megatrends: challenges and opportunities in attracting manufacturing Technological change is poised to reshape the global manufacturing landscape, and some may pose a challenge for countries with a comparative advantage on abundant unskilled labor. Manufacturing in the era of the fourth industrial revolution is characterized by advanced digital production technologies, including the extensive application of hardware (e.g., industrial robots and cobots, 3D printers), software platforms (e.g., big data analytics, artificial intelligence, cloud computing), and connectivity (internet of things) (UNIDO 2019). These technologies come with opportunities and challenges that can significantly shape the dynamics of manufacturing productivity growth, industrialization, and structural transformation of Ethiopia. Recent evidence shows mostly positive effects of digital technology on productivity in developing countries, including Sub-Saharan African countries.10 However, because these technologies are skill-biased, the scope for substituting for other production inputs is limited. Like other low-income economies, Ethiopia may face erosion or even loss of its comparative advantages (Rodrik 2016 and 2018; Reijnders et al. 2021). In addition, the COVID-19 crisis is expected to accelerate trends such as the reshoring of value chain segments to developed countries. While the concept and existence of the so-called “premature de- industrialization” has been debated at length (Hallward-Driemeier and Nayyar 2017), there is increasing 10 Based on firm-level data for developing countries (of which 26 are in the Sub-Saharan Africa region), Cusolito et al. (2020) examine the effect of digital technologies on factor demand among manufacturing firms. Their findings show that technology adoption is associated with productivity gains, significant positive effects on the number of jobs and demand for capital, and no effect on factor and more importantly labor displacement. 57 Ethiopia Country Economic Memorandum evidence of reshoring of activities by multinationals.11 The reshoring of some production processes toward sources of demand could mean that today’s developing economies run out of industrialization opportunities earlier than the early industrializers did. For instance, processes in the automotive and electronics industries, which are heavily automated, may remain in China or Japan and not be outsourced to Ethiopia. At the same time, there are new opportunities for manufacturing productivity growth through new intra-regional trade. A good example is the AfCFTA, a continental trade agreement that aims to eliminate tariff and non-tariff trade barriers; liberalize trade in services; enhance cooperation in investment, intellectual property rights, competition, and trade-related areas; and establish a dispute settlement mechanism concerning rights and obligations of members as well as build an institutional framework for the implementation and administration of the agreement (African Union 2018). Simulation results on the economic impact of the agreement indicate that manufacturing is likely to be a key driver of income growth and is expected to boost intra-regional trade (Abrego et al. 2019; World Bank 2020). As a signatory member of the agreement, Ethiopia can therefore tap into and take full advantage of the opportunities, especially in terms of market access. Chapter 4 of this report discusses this in more detail and argues that trade policy changes can help boost manufacturing and, importantly, that Ethiopia has an untapped potential in terms of trade in services. 3.2.2 Manufacturing sector performance in Ethiopia The manufacturing sector has been the main target of Ethiopia’s development strategies. It is usually considered a key driver in boosting productivity, creating jobs, raising export earnings, promoting technology transfer and skills development, sustaining economic growth, and accelerating structural change. The Growth and Transformation Plan for the period 2015/16-2019/20 (GTP II) states that the main objective is to make Ethiopia a leading manufacturing hub in Africa, while aiming as well to substitute certain strategic imports. “The strategic directions during GTP II are improving the productivity, quality and competitiveness of both existing and new industries and ensuring structural change, building labour intensive light manufacturing industry that is globally competitive in terms of productivity, quality and price, transforming the medium and large manufacturing industry to become a reliable source of foreign exchange and building industrial engineering and technological capacity. Efforts will also be put to improve the production capacity of existing industries, expand new manufacturing industries and attract new local and foreign direct investment both in quantity and quality. In addition, all the necessary effort will be made to link the development of high tech and light manufacturing industries, expand metal and engineering and chemical and pharmaceuticals industries and substitute strategic imported items by locally produced goods and reduce pressure on foreign exchange demand for imports.” (GTP II, p.136). Several policies have been adopted to pursue a manufacturing takeoff. For example, one policy is the provision of comprehensive support to the sector, which includes establishing industrial parks and developing clusters, strengthening organizational support for private investment, facilitating access to credit and foreign exchange, and providing targeted tax incentives. As a result, several industrial parks have started operation, attracted foreign direct investment, and contributed significantly as sources of employment and export earnings. The ongoing Homegrown Economic Reform Agenda: A Pathway to Prosperity also puts manufacturing at the center and calls for prioritizing industries with strong local content such as agro-processing and leather products; it 11 Banga and te Velde (2018) document evidence of reshoring, in which U.S. companies reshored production from Africa that resulted in job losses. Although the region has not yet experienced reshoring to the degree that it has occurred in Asia, the reshoring trend is likely to continue as automation becomes cheaper over time. 58 Ethiopia Country Economic Memorandum also emphasizes strengthening backward linkages through domestic production of primary and intermediate inputs and promoting the production of import-competing industries. While the share of manufacturing in the economy increased significantly during the first years of the past decade thanks to large- and medium-scale firms, it has remained small since then. Partly thanks to the policy support received, the share of manufacturing in GDP increased from below 5 percent by the end of the 2000s to nearly 7 percent in 2016/17, but it has since remained at the same level. Large- and medium- scale manufacturing are the main sources of value-added in the sector and have been behind the increase in the sectoral share. Meanwhile, the contribution of small-scale and cottage/handicraft establishments, 12 which are predominant in number, has experienced a decline (Figure 34).13 On average, large- and medium-scale establishments registered an annual average growth rate of 17 percent (compare to 7 percent growth among small-scale and cottage establishments) and contributed about two-thirds of the manufacturing value-added during the period 2012/2013 - 2019/2020. Figure 34. Large- and medium-scale firms have driven the increase in manufacturing contribution to GDP in recent years Contributions to GDP, by establishment size 7 6 Contribution to GDP, in % 5 4 3 2 1 Large- and medium-scale Small-scale and cottage All manufacturing Source: Authors’ calculation based on data from the National Planning Commission of Ethiopia. Note: Sectoral shares are obtained by dividing sectoral GDP at constant market prices to overall gross value-added at basic prices. GDP deflator is used as a price index, and 2015/16 is used as a base year. The number of establishments and workforce employed in large- and medium-scale manufacturing have increased steadily. While there were some gains in the early 2000s, it is after 2005 that the number of manufacturing establishment and the sector workforce started to take off, supported first by the Agriculture Development Led Industrialization (2003) and later by the two Growth and Transformation Plans (2010 and 2015). The number of establishments increased almost fourfold between 2005 and 2018, and the size of the 12 According to CSA (2014), manufacturing establishments are divided into three major groups: cottage/handicraft (establishments do not use power-driven machinery); small-scale (establishments have less than 10 workers and use power-driven machinery), and large- and medium-scale (establishments have at least 10 workers and use power-driven machinery). 13 According to a survey conducted in 2013/2014, there were 116,604 small-scale manufacturing establishments hiring more than 1.7 million workers (CSA 2014). 59 Ethiopia Country Economic Memorandum manufacturing workforce tripled in this period. (Figure 35). This manufacturing workforce expansion outpaced what was observed in the manufacturing sectors of Kenya, South Africa, and Tanzania, although it was slower relative to Indonesia and Vietnam.14 It is worth noting that the Figure, based on surveys, does not fully capture the boom of industrial parks, initiated around 2015. Overall, job creation in formal Ethiopian manufacturing has been driven mainly by the entry of new establishments and growth of young establishments, more than by the expansion of the established ones (Abreha et al. 2019). Figure 35. The number of establishments and the number of manufacturing workers have increased steadily Manufacturing: number of establishments and total employment 3,500 300,000 3,000 250,000 2,500 200,000 2,000 150,000 1,500 100,000 1,000 500 50,000 - - 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Number of manufacturing establishments Number of manufacturing workers Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. The data for 2013 on the number of workers looks problematic (about an 88,000 increase in the number of workers compared to 2012), and hence the plot is based on an imputed value by taking the mean of 2012 and 2014. Despite the observed growth, manufacturing is still far from becoming the driver of structural transformation and creating enough formal employment for a rapidly growing population. Compared to an estimate of two million people joining the labor market every year, job creation in manufacturing remains modest. Overall, Ethiopia’s growth acceleration since the early 2000s has not been associated with a substantial increase in the share of formal manufacturing employment, as shown in Diao et al. (2021). The share of manufacturing employment in industry has declined, mostly because that of construction increased sharply.15 This underscores the fact that enhancing growth opportunities for formal firms is critical to the creation of 14 Data from UNIDO’s INDSTAT2 database shows that the number of workers in Ethiopian manufacturing w as about 374,000 in 2015—this constituted 120 percent growth from the size of the workforce in 1998. This pattern resembles the experiences of comparator countries. For instance, the corresponding figures are 5.2 million and 24 percent (Indonesia); 386,000 and 59 percent (Kenya); and 6.1 million and 149 percent (Vietnam) during the same period. Among the benchmark countries, South Africa and Tanzania experienced a reduction in the size of their manufacturing workforces during this period: 1.2 million and -15 percent (South Africa); 122,000 and -8 percent (Tanzania). Note that although the data is comparable across countries, some variations exist in the scope of manufacturing establishments surveyed. See UNIDO’s INDSTAT2 database for more details. 15 Industry comprises ISIC-rev. 4 divisions 10-45 and corresponds to manufacturing, mining, construction, electricity, water, and gas. 60 Ethiopia Country Economic Memorandum more productive employment by facilitating the reallocation of labor from less productive, informal firms to formal firms.16 3.3 Manufacturing productivity dynamics in Ethiopia This section assembles the best available evidence from Ethiopia to analyze the dynamics and drivers of manufacturing productivity growth at the establishment and industry levels. It mainly draws on the annual surveys of large and medium manufacturing and electricity industries. The Central Statistical Agency of Ethiopia conducts an annual survey of all manufacturing establishments with at least 10 employees and that use power-driven machinery. While there is an abundant corpus of literature analyzing this data in Ethiopia, most of the existing evidence depends on previous waves of the surveys (up to 2010).17 This chapter focuses on the most recent years and builds a panel dataset for the period 2012-2018, on which productivity decomposition techniques are applied (Box 4). In addition, World Bank Enterprise Surveys and similar datasets are used for comparing Ethiopia’s manufacturing with other benchmark countries. Box 4. Decomposing productivity growth In this study, an establishment’s productivity is approximated using labor productivity, capital productivity, and total factor productivity (TFP). Aggregate productivity in any given period is a weighted average of establishment-level productivity. As commonly used in the literature, employment and nominal value-added shares are used in computing aggregate labor productivity and TFP, respectively. Labor productivity is value-added per unit of workers (in full-time equivalents). Value-added is defined as sales revenue minus material cost. As detailed in Annex 3.1, TFP is computed after estimating a gross output Cobb-Douglas production function in which labor is approximated by wage bill and capital by beginning-year fixed assets holdings. The three productivity measures are computed using employment shares as weights. Nominal values are converted into real values by using the GDP deflator and 2016 as a base year and are given in constant local currency; the deflator is chosen for being less noisy than other price indexes. Aggregate productivity growth is decomposed into components that reflect productivity growth within incumbent establishments, resource and market reallocation between incumbent establishments, and selection due to establishment entry and exit (Figure B.1). The contribution of incumbents to aggregate productivity growth comes from within productivity changes as well as the reallocation of resources and market shares between them. For example, if the productivity levels of incumbents improve and/or resources and markets move toward more productive incumbents, aggregate productivity increases. On the other hand, the entry channel affects aggregate productivity growth depending on the relative difference in the productivity levels of entrants and incumbents. If entrants are relatively more productive compared with incumbents, their entry raises aggregate productivity, and the opposite holds when entrants are relatively less productive. Similarly, the exit channel compares the productivity levels of firms leaving the market relative to incumbents. If less productive establishments leave, their market departure improves aggregate productivity and vice versa. 16 For instance, findings by McCaig and Pavcnik (2018) on Vietnam show that new export opportunities increased employment generation in the formal sector in Vietnam, inducing reallocation of workers from household businesses toward more productive firms in the registered enterprise sector. 17 The only exceptions are recent studies by Diao et al. (2021), Gebrewolde et al. (2020), and Policy Studies Institute (2020). 61 Ethiopia Country Economic Memorandum Figure B.1. Aggregate productivity growth is decomposed into three components Reallocation Allocation of resources and markets between establishments Selection Within Entry and exit of Productivity changes establishments within establishments Aggregate productivity growth Source: Authors’ illustration. 3.3.1 The disappointing performance of manufacturing productivity in Ethiopia Following a peak in 2015, firm-level labor and capital productivity growth in manufacturing firms turned negative. By 2015, labor and capital were 16.9 and 8.6 percent more productive than in 2012, respectively. While the reallocation component was negative, suggesting that resources and market shares shifted from more productive to less productive establishments, the within, entry, and exit components contributed positively to overall productivity growth. During 2015-18, however, relatively better reallocation of resources among incumbents and exit of less productive establishments were not sufficient to offset declines in the within productivity growth of incumbents and the entry of significantly less productive firms (Figure 36). Figure 36. Firm-level productivity measures in manufacturing were negative across the board during 2015-18 Ethiopia, productivity growth in manufacturing 20 16.9 10 8.6 0 -4.5 Percent change -10 -9.7 -20 -25.1 -30 -40 -46.7 -50 Within Reallocation Entry Exit Overall -60 Labor Capital TFP Labor Capital TFP 2012-15 2015-18 62 Ethiopia Country Economic Memorandum Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. Meanwhile, TFP in manufacturing firms in Ethiopia is estimated to have been in decline since 2012. This finding for the manufacturing sector is consistent with the decline in TFP found at the macro level for the whole Ethiopian economy (see section 1.2 in this report). Declining TFP during 2012-15 was driven by the large negative contribution of the reallocation component (-39.3 percent), suggesting there was a substantial shift of resources toward firms that are less productive and innovate less. Meanwhile, the within, entry, and exit components had a positive (but smaller) contribution to overall TFP growth. In the most recent period of 2015-18, in addition to the negative contribution of the reallocation component, incumbent firms also experienced a decline in TFP growth, and the entry component turned slightly negative (Figure 36). The growth rates of productivity differ significantly across manufacturing industries. For example, between 2015 and 2018, all industries except those of leather and plastic & rubber experienced a decline in labor productivity (Figure 37). Labor productivity shrank by an amount ranging from -147 percent (machinery & equipment) to -8 percent (wood & furniture). In terms of TFP, in addition to leather and plastic & rubber, textiles & apparel and paper & publishing also experienced positive growth, driven mostly by new entrants. The garment and leather industries were also found to have experienced significant increases in capital productivity during this period, driven by new entrants. The opening of several industrial parks and the entry of foreign establishments during this period is likely to have contributed to improved competition and total factor productivity gains in garments. In contrast, the machinery and equipment and non-metallic industries suffered sharp declines across the three productivity measures (labor, capital, and TFP). Figure 37. During 2015-2018, aggregate labor productivity has declined in most manufacturing subsectors, while TFP growth has varied across industries Labor productivity growth by industry, 2015-18 TFP growth by industry, 2015-18 Food & Beverage -17.7% -26.8% Textiles & Apparel -82.9% 20.0% Leather 10.0% 46.2% Wood & Furniture -8.4% -12.9% Paper & Publishing -16.3% 41.0% Chemicals -10.6% -2.0% Rubber & Plastic 7.5% 14.4% Non-Metallic -37.6% -44.5% Metal Products -14.5% -6.4% Machinery & Equipment -147.0% -14.6% Others -2.3% 60.3% -150% -100% -50% 0% 50% -50% 0% 50% 100% Percent change percent change Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Industry and regional characteristics explain only a small fraction of the observed patterns of productivity growth at the establishment level. A multiplicity of factors drive the level and growth rate of 63 Ethiopia Country Economic Memorandum productivity at the establishment level. These factors can be roughly grouped into those that are internal (for which the establishment has a higher degree of control), and those that are external (factors outside the boundaries of the establishment, and for which it exercises limited or no control). The analysis found that sectoral and regional characteristics, as well as time effects, explain a rather small percentage of the variation of the observed patterns in both the level and growth rate of productivity over the sample period. Therefore, productivity growth is determined predominantly by factors that are internal to the establishment. Firm-specific features can explain a substantial part of the productivity of the establishment. Looking at other characteristics of establishments and their relationship with productivity can provide insight into which establishments are responsible for the productivity slowdown. Figure 38 shows that large establishments (with at least 50 workers) have higher labor productivity and hire more workers compared to smaller establishments.18 Exporters and importers have positive productivity premia compared with non-exporters and non-importers, respectively. Establishments with foreign and state ownership also tend to be more productive than domestic private-owned ones. Figure 38. Significant performance differentials can be seen across different types of firms in Ethiopia Productivity differential across different manufacturing establishments Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. By taking the mean number of workers in the first two periods, an establishment is defined as large if it employs at least 50 workers. Similarly, exporters, importers, foreign-owned, and state-owned are establishments that have exported, imported, had foreign ownership, or were publicly owned at least once in the first two periods, respectively. 18 By taking the mean number of workers in the first two periods, an establishment is defined as large if it employs at least 50 workers. Similarly, exporters, importers, foreign-owned, and state-owned are establishments that have exported, imported, have foreign ownership or publicly owned at least once in the first two periods, respectively. 64 Ethiopia Country Economic Memorandum Despite the productivity slowdown over time, there is evidence of industry-level productivity convergence in the manufacturing sector. In terms of growth, labor productivity and TFP have declined over time (“Trend” in Figure 39). At the same time, establishments with initial productivity levels below the median (defined at the industry level) have experienced faster productivity growth over time. Hence, there is some evidence of productivity convergence at the industry level. Similarly, large establishments have become more productive over time, especially during the 2016-2018 period. By contrast, foreign-owned and state- owned establishments have registered lower productivity growth relative to their domestic and privately owned counterparts; expansion in the number of workers in these firm categories has come at the expense of labor productivity. Figure 39. Productivity in below-the-median establishments has grown faster than in the rest Labor productivity and TFP growth differential across different establishments Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. By taking the mean number of workers in the first two periods, an establishment is defined as large if it employs at least 50 workers. Similarly, exporters, importers, foreign-owned, and state-owned are establishments that have exported, imported, had foreign ownership, or were publicly owned at least once in the first two periods, respectively. 3.3.2 What explains the weak manufacturing productivity growth in recent years? Growth in employment has not been accompanied by an equally proportional growth in value-added in recent years. Figure 40 shows that manufacturing establishments experienced significant growth in value- added in 2017 and 2018. At the same time, (and except for 2015, when there was some drop in employment levels), manufacturing establishments saw substantial expansion in their workforces, which outpaced 65 Ethiopia Country Economic Memorandum productivity improvements. For example, in 2016, labor productivity contracted by about 28 percent compared to the previous year. While an average establishment had a comparable level of value-added in 2016 as in 2012, it had expanded in size quite substantially, which resulted in a drop in labor productivity. Figure 40. Employment expansion during 2016-2018 outpaced value-addition, explaining the drop in labor productivity Fixed effects regressions: drivers of labor productivity (base year 2012) Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Based on regressions that control for employment size and for establishment, sector, and region fixed effects. Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. Value-added is defined as sales revenue minus material cost. GDP deflator is used as a price index, and 2016 is used as a base year. 3.3.2.1 A decline in capacity utilization, investment, and scale of operations In recent years, manufacturing establishments saw a significant decline in the utilization of their productive capacity and size of capital (fixed assets) holdings. Productive capacity utilization is one key determinant of the level of production and productivity in manufacturing. It is defined as the ratio of actual production levels to production levels at full capacity. As shown in Figure 41, productive capacity utilization has declined consistently from 2015 onward compared to the 2012 baseline level. Related to productive capacity is the size of capital holdings and investments. In addition, the number of establishments that made investments in fixed assets decreased, although with no difference in the level of investments among those that invested. The decline in sales and productive capacity utilization may be explained by a sudden drop in the availability of imported materials. Given the heavy reliance of manufacturing production on imported inputs, the large fall in the use of imported raw materials indicates that production operations were adversely affected, translating into lower aggregate productivity levels. As discussed in the next section, shortage of raw materials as well as limited access to foreign currency, are among the most mentioned obstacles that severely constrain business operations in Ethiopian manufacturing. 66 Ethiopia Country Economic Memorandum Figure 41. Productive capacity utilization has declined from 2016 onward, possibly driven by lack of availability of imported inputs Fixed effects regressions: productive capacity, capital, and investment (base year 2012) Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Based on regressions that control for employment size and for establishment, sector, and region fixed effects. Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. Value-added is defined as sales revenue minus material cost. GDP deflator is used as a price index, and 2016 is used as a base year. Transportation costs grew over time and likely constrained business operations, while electricity costs have been on decline. After accounting for scale of production, the establishments have faced increasing transportation costs since 2016, which would likely impede their business operations in purchase, production, and distribution. This may have been the result of increased power of cargo and logistics companies (which did not face any foreign competition back then), and potentially, unrest, as discussed below. At the same time, however, Ethiopian manufacturing establishments have incurred declining costs of electricity, which indicates the growing efficiency of the establishments in their electricity usage and/or access to cheaper sources of electricity. It could be argued that one reason for declining electricity costs is that supply has become more reliable and that manufacturing establishments have resorted to backup sources due to frequent power outages; however, the costs of a non-electric sources of energy have not significantly changed over the years. One factor that could possibly account for these trends is the increase in civil unrest and insecurity. The country experienced an increase in civil unrest over land grabs and human rights during 2016 and 2017, especially in the Oromia regional state which hosts the second-largest number of establishments next to Addis 67 Ethiopia Country Economic Memorandum Ababa. Several factories were burned down, and others suffered disruptions—this, together with demand being possibly affected, could be among the factors explaining a decline in capacity utilization.19 A second factor that might account for the trends described above is the introduction of additional controls in foreign exchange management. The National Bank of Ethiopia (NBE), Ethiopia’s central bank, issued its directive Transparency in Foreign Currency Allocation and Foreign Exchange Management (FXD/45/2016) in February 2016.20 This directive instructed banks to develop foreign exchange operations management guidelines and to allocate foreign exchange on a first come, first served basis while observing a priority list of items. The issuance of the directive was driven by the worsening foreign currency shortage in banks when importers were finding it difficult to access foreign exchange, and several instances of malpractice in foreign exchange allocation by the banks were reported. A noticeable result was the significant decline in the imports of raw materials in 2016 and 2017, as available foreign exchange was possibly redirected to priority uses (e.g., imports of fuel, fertilizers, and medicines, plus debt service payment). 3.3.2.2 Constraints to access imported inputs Raw material shortages undermine the productivity of the Ethiopian manufacturing sector. Ethiopian manufacturing entrepreneurs face many business environment constraints that prevent them from growing or being productive. Producers consistently cite a shortage of raw materials—partly caused by constraints in the access to foreign exchange as well as underdevelopment of domestic supplies—as the top constraint affecting their performance (Figure 42). The raw material shortage has also forced Ethiopian manufacturers to operate below installed capacity, with the average manufacturing capacity utilization below 70 percent. In 2018, 33 percent of firms identified shortage of raw materials as major obstacle preventing them from operating at full capacity, up from 29 percent in 2012. The share of firms directly identifying shortage of foreign exchange as the main obstacle more than doubled between 2012 and 2018. Figure 42. Shortage of raw materials has been the top reported constraint affecting the manufacturing sector Obstacles preventing full capacity operation - percentage of establishments responded % of establishments citing as top obstacle Shortage of material inputs 36 Shortage of power and water 17 Others 14 Lack of market of customers 11 Newly established 7 Shortage of foreign exchange 5 Frequent machine breakage 4 Shortage of working capital 3 0 10 20 30 40 Percent 2018 2012 19 See reports by major news outlets such as The Economist, Reuters, and The Washington Post. 20 Directive accessible at: https://nbebank.com/wp-content/uploads/pdf/directives/forex/fxd%2045.pdf 68 Ethiopia Country Economic Memorandum Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Based on regressions that control for employment size, output and establishment, sector and region fixed effects. Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. Material inputs comprise raw materials as well as spare parts. The severity of raw material shortages varies across industries, depending on imported inputs. For the purposes of the analysis, import dependence is defined as the share of imported inputs in total material inputs used in production by a specific manufacturing subsector. Industries that rely heavily on imported inputs due to lack, limited availability, or low quality of local supplies are more affected by the shortage of imported materials (Figure 43). A good example of an import-dependent industry is the machinery & equipment industry, which experienced a substantial decline in labor productivity and TFP, particularly during 2015-18 and among the existing firms. Figure 43. Overall, firms in industries dependent on imported inputs tend to see shortage of materials as a top constraint Severity of material inputs supply shortages and imported input dependence, 2018 60 % establishments for which shortage of inputs is top 55 Rubber & Plastic Chemicals 50 Textiles & Apparel Leather 45 Food & Beverage Paper & Publishing Metal Products Machinery & constraint 40 Equipment 35 30 Wood & Furniture 25 Non-Metallic 20 20 30 40 50 60 70 80 90 100 Import dependence (%) Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Based on regressions that control for employment size, output and establishment, sector and region fixed effects. Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. Material inputs comprise raw materials as well as spare parts. Establishments with an international trade orientation view access to foreign exchange as a severe constraint in their business operations, particularly for those whose international trade participation is exclusively through exporting or importing (Figure 44). This may reflect administrative issues with respect to foreign trade regulation and foreign currency management. However, those establishments that are two-way (simultaneously exporting and importing) seem to be less severely affected by foreign currency access, as they can finance at least part of their imports from their foreign exchange proceeds. It suggests that while import duties are not an issue in priority sectors such as manufacturing, actual production was constrained by a shortage of raw materials, possibly caused by issues they (or their suppliers) are facing when trying to access foreign 69 Ethiopia Country Economic Memorandum exchange. This is consistent with the findings of a recent study that shows how the inability of firms to import key components of production is the binding constraint to the manufacturing sector and most proximate cause of the country’s slowdown (Goldstein 2020). In addition, the reduction in the output of most sectors and the rise in the exit rate of firms are linked with acute shortages of foreign exchange. Figure 44. Establishments engaged in importing or exporting are more likely to view access to foreign exchange as a severe constraint Likelihood of shortage of material input supply as a top constraint Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Based on regressions that control for employment size, output and establishment, sector and region fixed effects. Establishments with unreported or zero sales revenue, employees, industry classification, and regional location are excluded. Material inputs comprise raw materials as well as spare parts. Depending on their trade orientation, import-only and export-only establishments are importers or exporters but not both. Two-way are establishments that engage in both exporting and importing. In contrast, accessing foreign exchange was less of an obstacle for state-owned enterprises (SOEs) in 2016-18. This suggests that SOEs have been benefiting from priority access to foreign exchange at the expense of other firms. Interestingly, as shown in Figure 39 in the previous section, both labor productivity and TFP have been on the decline in the SOEs, suggesting that priority access to resources has not been put to the most productive use. 3.4 Productivity dispersion and resource misallocation Another factor that may be driving the weak productivity performance could be the inefficient allocation of productive resources, with inefficient firms commanding more resources than is warranted by their productivity levels. Resources (labor, capital, technologies) are misallocated if there are deviations from the firms’ efficient level of inputs. Misallocation of productive resources across firms can reduce aggregate productivity by affecting the number and composition of firms operating in an industry. Misallocation results 70 Ethiopia Country Economic Memorandum from policies or regulations that support less competitive firms to survive instead of exiting the market, limit the entry of new firms, and force productive firms to exit. For example, subsidized loans allow the survival of less competitive enterprises—firms that would otherwise exit in a competitive market—fueling misallocation of capital across firms and thereby lowering aggregate productivity. Uneven access to foreign exchange even within the same sectors (as some firms may have political connections) also results in differences in productivity. The analysis suggests that the level of resource misallocation—measured by the dispersion of marginal revenue products of inputs across firms—is considerable within the manufacturing sector and that misallocation has been on the rise. For example, the ratio of 90th to 10th percentiles of revenue productivity (TFPR)21 in 2018 was 13. While many factors could potentially explain such dispersion, firm-level distortions (e.g., due to economic policies or institutions) that prevent resources from moving from the less productive to the more productive ones are identified as the main reason (Hsieh and Klenow 2009). The dispersion of TFPR has risen between 2012 and 2018, with the 2018 distribution showing more mass in both tails (Figure 45). The rise in TFPR dispersion can be a result of a variety of factors that changed during this period, including deterioration in the business environment that reduces the tendency for marginal products to be equalized. The TFPR dispersion using the balanced sub-panel (surviving firms) is much smaller, suggesting that the rise in TFPR dispersion was partly driven by a surge of entry of new firms. Figure 45. TFPR dispersion has increased over time Distribution of revenue total factor productivity (TFPR) in the Ethiopian formal manufacturing sector 21 To calculate firm-level physical productivity, data on physical output is needed. Unfortunately, as in most other firm-level datasets, the firm-level data for Ethiopia doesn’t contain harmonized information on product prices and quantities for most sectors. The datasets, however, report the value of production. One way to obtain the physical output is to deflate the values by the industry price deflator, that is, , where P is the industry price index. However, firm-level prices may vary relative to the industry price index in differentiated product industries. If so, this approach yields a contaminated measure of productivity, since within-industry price differences are lumped into productivity measures. This measure is often referred as revenue total factor productivity ( = × ), where R stands for revenue and Q for quantity, and denotes firm’s product price. 71 Ethiopia Country Economic Memorandum Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Productivity is computed for each firm relative to the four-digit industry average. In the absence of any distortion, the marginal products of each input should be equalized across firms within the four-digit sector. Productivity dispersion levels in Ethiopia appear to be higher than in other countries with manufacturing census data. Although cross-country comparison of the dispersion of TFPR is difficult due to differences in coverage of firm-level datasets, misallocation in Ethiopia is larger than in other countries in the region for which comparable data are available, such as Cote d’Ivoire, Ghana, and Rwanda (Figure 46). Figure 46. Resource misallocation in formal manufacturing appears to be higher in Ethiopia than in other countries for which comparable data is available Degree of dispersion in TFPR in manufacturing Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Resource misallocation is measured by the dispersion of marginal products of inputs across firms. Std. Log is the standar deviation of the log of the TFPR measure; a higher value entails higher productivity dispersion. Large dispersions suggest that frictions in input and output markets prevent the movement of productive resources across firms and that inefficient firms command more resources than warranted by their productivity. This phenomenon of resource misallocation has been observed in most sectors, but the magnitude differs significantly across subsectors. Resource misallocation may be more severe in some industries that are subject to prevalent policy and institutional distortions. Understanding the patterns of dispersion across different subsectors may help improve the focus of policy measures. Figure 47, which presents the dispersion of the marginal products by subsector in 2012 and 2018, shows that productivity dispersion varies significantly across manufacturing subsectors. For example, productivity dispersion is large in textile & apparel manufacturing, implying substantial scope for improving productivity through more efficient resource reallocation. The Figure also shows that most sectors experienced an increase in TFPR dispersion between 2012 and 2018. For example, the metal, machinery & equipment, and textile & apparel subsectors saw a significant increase in productivity dispersion. The increase in TFPR dispersion is more pronounced in sectors that experienced higher firm entry rates, suggesting that the large increase in these sectors can be attributable in part to a surge in the entry of new firms. 72 Ethiopia Country Economic Memorandum Figure 47. Productivity dispersion varies significantly across manufacturing subsectors, and most sectors experienced an increase in dispersion between 2012 and 2018 Productivity dispersion by subsector, 2018 vs. 2012 1.6 1.5 Textiles & Apparel 1.4 Machinery & Equipment Std. Log of TFPR in 2018 1.3 Metal Products 1.2 Rubber & Plastic Chemicals 1.1 Non-Metallic 1 Food & Beverage Wood & Furniture 0.9 0.8 Paper & Publishing 0.7 0.6 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 Std. Log of TFPR in 2012 Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing a nd Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Dispersion is computed for each four-digit sector and then aggregated at broad sector using value-added shares as the weights. Std. Log is the standar deviation of the log of the TFPR measure; a higher value entails higher productivity dispersion. Distortions in output and input markets penalize the more productive firms more heavily. Firm-specific distortions would be even more deleterious if they are positively correlated with a firm's physical productivity. Cirera et al. (2020) provide evidence from Ethiopia that policy distortions penalize more productive firms and subsidize the inefficient ones, significantly undermining the manufacturing sector’s productivity, growth, and employment potential. Firm-level output and input distortions could also have a longer-term dynamic impact on growth by distorting firms’ productivity and employment decisions. For example, firm-level distortions could discourage firms from investing in productivity-enhancing activities such as the use and adoption of better technologies, introduction of new products, and better management practices, further undermining the scope for aggregate productivity growth. A notable difference in the life cycle of firms has been documented between developing and developed countries (Hsieh and Klenow 2014): while firms in advanced countries tend to grow faster after entry, firms in developing countries do not seem to grow as they age, contributing to aggregate productivity differences across countries. Indeed, earlier research shows that Ethiopian manufacturing firms do not grow as they age after entry (Cirera et al. 2020). Hence, boosting aggregate productivity requires not only improving static reallocation but also removing various distortions that limit firms’ growth. There is substantial scope for improving the productivity of the Ethiopian manufacturing sector. Improving manufacturing productivity requires removing or reducing the underlying frictions that prevent the efficient allocation of resources toward more productive producers. The counterfactual experiment suggests that the government can improve manufacturing productivity by at least 100 percent if resources are allocated as efficiently as in the United States. The counterfactual exercise focuses on productivity gains from removing distortions within an industry, abstracting from aggregate TFP gains associated with the reversal of distortions between sectors. Aggregate productivity could be improved further through more efficient allocation of resources across sectors. 73 Ethiopia Country Economic Memorandum The potential productivity gains in manufacturing can be viewed as a reasonable lower bound for the economy at large. The scope for productivity improvements through more efficient resource allocation is likely to be even more prominent in other sectors. Evidence suggests that in other countries, the scope for improving productivity through more efficient resource allocation is larger in services: Cote d’Ivoire (Maemir 2020), Latin America (Busso et al. 2013), and Portugal (Dias et al. 2016). While other factors are at play, the high level of misallocation in the service sector could be because firms in these sectors are much less exposed to international trade and foreign competition (in the case of Ethiopia) compared to the manufacturing sector. Given the service sector’s high share of value-added in the country, economy-wide productivity gains from eliminating misallocation are enormous. 3.4.1 Market power and price markups Misallocation may reflect market imperfections, including market power. Firm-level markups in Ethiopian manufacturing have been increasing over time (Figure 48). This is consistent with a growing body of research that points to a rising markup in other countries (De Loecker et al. 2020). While high markups— the extent to which prices are above marginal costs—could be the result of quality upgrading or change in quality, they could also reflect market power stemming from anticompetitive practices. It has long been recognized that market power has a detrimental impact on productivity through differentiation channels. Market power limits the incentive that firms have to invest in productivity-enhancing activities such as the use and adoption of better technologies, the introduction of new products, and the adoption of better management practices (Aghion et al. 2018; Van Reenen 2011). Moreover, market power discourages business dynamists by limiting the entry of more productive firms and the exit of less productive ones. Market power also has implications for the allocative efficiency of inputs across firms and thereby measured aggregate productivity and welfare (Edmond et al. 2018; Baqaee and Farhi 2020). Rising markups have been associated with a decline in investment rates, firm entry rate, and labor’s share of income in several countries (Syverson 2019). Figure 48. Median markups in the Ethiopian manufacturing sector have increased over time Median Markups 1.2 Price to marginal cost ratio 1.15 1.1 1.05 1 0.95 2012 2013 2014 2015 2016 2017 2018 Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Firm-level markups are estimated using the methodology proposed by De Loecker and Warzynski (2012). The increase in markups is happening in most industries but was especially pronounced in textile & apparel manufacturing. Markup dispersion within industries has also risen in all major manufacturing industries (Figure 49). Previous empirical analysis shows that the dispersion in markups in Ethiopian 74 Ethiopia Country Economic Memorandum manufacturing negatively affect business formation as well as productivity growth and job creation (Damoah et al., 2020). Several factors can contribute to the observed increase in markups in the textile & garment sector, including the growth of global value chain (GVC) participation in these sectors and the entry of new firms setting high markups. GVC participation can lower the costs of inputs or increase the profit rates of these firms (World Bank 2019). Markups are highly heterogeneous across different types of firms. Large, exporter, and state-owned enterprises have higher markups than other firms, which is consistent with the existing empirical evidence. In contrast, importers and foreign-owned establishments tend to have lower markups than the average. State- owned establishments also charge significantly higher markups, reflecting their market power. While SOE presence in manufacturing has been reduced over the years, SOEs still hold a significant share of the market in industries such as paper & publishing, machinery & equipment, and non-metallic products. Figure 49. Markup dispersion within industries has risen in all major manufacturing industries Evolution in markups, by subsector 1.7 1.6 Textiles & Apparel 1.5 Others Price to marginal cost in 2018 1.4 Paper & Publish Wood & Furniture 1.3 Machinery & 1.2 Equipment Metal Products Non-Metallic 1.1 Leather Rubber & Plastic Food & Beverage 1 Chemicals 0.9 0.8 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Price to marginal cost in 2012 Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. Note: Firm-level markups are estimated using the methodology proposed by De Loecker and Warzynski (2012). 3.5 Policy constraints and proposed solutions Declining productivity and increasing TFP dispersion are the result of distortions and market deficiencies that need to be systematically addressed for manufacturing in Ethiopia to live up to its potential. Policies to boost manufacturing productivity cut across many policy areas, and they need to address both external and internal constraints on a firm. Addressing the external constraints will involve enabling markets and reducing distortions by lowering the barriers to firm entry and exit, improving access to finance and to foreign exchange, and fostering competition. To address the internal constraints on a firm and enhance 75 Ethiopia Country Economic Memorandum within-firm productivity, efforts are needed to help upgrade firm capabilities, for example by improving managerial practices and increasing technology adoption. This section discusses challenges and policy solutions in these areas of intervention. 3.5.1 Reforms aimed at enabling markets and reducing distortions 3.5.1.1 Easing business entry and exit Despite some improvements, the regulatory environment in Ethiopia for entry of new firms remains cumbersome. Over the past decade, Ethiopia took some steps to ease the entry of new firms, including removing the need to obtain a certificate of competence for certain types of businesses, removing the requirement to open a bank account for company registration, and eliminating the paid-in minimum capital requirement. Notwithstanding the progress, the regulatory environment for starting and operating a business is still restrictive, complex, and costly. In 2020, Ethiopia ranked 190 out of 213 countries in the time required to start a business (32 days)—well above the average for Sub-Saharan Africa (21 days).22 Foreign investors face additional barriers to starting a business, including challenges accessing land, which is a constraint for manufacturing investments in particular. Thus, considerable scope for reform remains. The automation of the different procedures under the business establishment process could ease the entry of businesses significantly. Full automation of the whole licensing process, together with integration of systems and databases among the several agencies that provide interdependent approvals and services for starting a business, would be crucial to facilitate the entry of new firms. Although a one-stop shop approach has been used in other countries, it does not seem well-suited to the Ethiopian context because of the large number of trade bureaus, which would make it very costly.23 In terms of firm exit, evidence indicates that poorly designed insolvency frameworks, such as the one Ethiopia historically had, can slow down productivity growth. Until 2021, Ethiopia’s regulatory environment for resolving insolvency—the time, cost, outcome, and recovery for a commercial insolvency and the strength of the legal framework for insolvency—was very poor, with the country lagging behind peers. Financially distressed companies in Ethiopia had few to no options other than liquidation, given the lack of an effective regulatory regime that allowed for reorganization and other recovery measures. This severely affected companies that were experiencing financial distress (including due to shocks like the COVID-19 pandemic), limiting the potential for economic recovery. It created impediments to the smooth and timely exit of unviable firms and thereby reduced aggregate productivity. McGowan et al. (2017) show that reforms to insolvency regimes that reduce barriers to business restructuring have improved the productivity of laggard firms and thereby aggregate productivity. Full implementation of the new Commercial Code, coupled with additional measures to simplify exit- related tax audit issues, would help facilitate a healthy churning of firms in Ethiopia. Recognizing the abovementioned challenges, the authorities adopted a new Commercial Code in 2021 that repealed the 1960 Commercial Code in the areas of company formation, operation, and insolvency. The new Code incorporates various good practices, including introduction of reorganization and provision for the concept of the sale of the debtor’s business. It foresees simplified bankruptcy proceedings for SMEs and it allows parties to apply 22 Doing Business Report, 2020. 23 Often, one-stop shops can end up being complicated if agencies simply assign representatives in trade bureaus just to collect applications and forward to their respective headquarters, rather than granting them the ability to approve applications autonomously. In addition, considering that Ethiopia has over one thousand trade bureaus, a one-stop shop would be very costly resource-wise. Processes would need to be streamlined significantly and fully delegated to a single agency for the one-stop shop to be able to provide the needed range of approvals and services efficiently. 76 Ethiopia Country Economic Memorandum various insolvency procedures other than bankruptcy. The provisions of the new Code would give viable companies the possibility of continuing their operations after resolving insolvency, which was not possible under the 1960 Commercial Code. The new Commercial Code also introduces protection of minority investors and upholds best practices on corporate transparency, disclosure, and shareholders rights, which are critical for both domestic and foreign investors operating in Ethiopia. However, the Commercial Code only covers the insolvency-related aspects of business exit. Tax audit-related issues of business exit are outside the Commercial Code and remain a concern, particularly for local businesses; tax audits for business closure purposes currently take years to resolve and often involve unpredictable processes and outcomes. Streamlining these processes is critical to facilitate firm exit. 3.5.1.2 Facilitating access to finance Improving access to finance continues to stand out as one of the most daunting challenges in Ethiopia. A large body of literature has shown that presence of financial frictions at the firm level is a main driver of resource misallocation (Gopinath et al. 2017; Midrigan and Xu 2014). Access to finance is a longstanding challenge in Ethiopia, with the country ranking among the lowest in the world: according to the World Bank Findex, in 2017, just 35 percent of the population ages 15 and over had a bank account (compared to 82 percent in Kenya), and the share of adults who made or received digital payments was just 12 percent (79 percent in Kenya). In the Global Competitiveness Index, the score for financing to SMEs dropped to 3.34 in 2019, the lowest among all the peer countries. Compared to countries like Kenya and Vietnam, Ethiopia lags in areas such as legal rights, capital markets development, and ease of access to loans. While the manufacturing sector appears to have preferential access to finance vis-à-vis other sectors on paper, in practice, firms face a series of challenges due to the nature of their business. As the manufacturing sector is capital-intensive, the standard terms of the loans provided by commercial banks— which usually feature short maturities, stringent collateral requirements, and no grace period—are not favorable to firms in the sector. Available loans focus mostly on the financing of working capital, and there is a lack of instruments aimed at financing the purchase of assets. In addition, until recently, the impossibility of borrowing in U.S. dollars affected domestic firms, particularly those that were not exporters.24 As found by analyzing the manufacturing census survey, small domestic firms are likely to find access to finance more challenging, due in part to lack of sufficient collateral to pledge. This can be a major source of capital misallocation, since relatively more productive firms with limited collateral may not have access to sufficient capital to produce at the optimum level. These obstacles have likely resulted in distortions in the allocation of credit among firms, despite some recent efforts to address them. The Ethiopian government has recently taken positive steps toward revamping the market infrastructure to address imperfections that hinder credit, including by introducing movable collateral, but some complementary measures are needed. Authorities have adopted the building blocks to support the use of movable collateral to secure loans,25 which is likely to improve firms’ access to external finance, 24 A new directive on external loans was adopted in June 2021, allowing manufacturing firms (in addition to foreign investors and exporting firms) to access external loans, although it enforces a cap on the interest rate of the loans (LIBOR+5%). 25 In August 2019, the Parliament approved the Movable Property Security Rights proclamation in line with good international practices. In February 2020, the NBE issued the Directive on Operationalization of Movable Collateral Registry and formally launched phase one of the electronic Collateral Registry as a publicly available registry/database that allows financial institutions to access information about any encumbrances to movable properties (such as vehicles, machinery, livestock, inventory, and accounts receivables) belonging to potential borrowers and to register their security interests over that collateral when giving loans. In September 2020, the NBE issued a directive on Codification, Valuation and Registration of Movable Properties as Collateral, thus fully enabling operations of the collateral registry. 77 Ethiopia Country Economic Memorandum especially for small firms. Studies show that introducing collateral registries for movable assets increases firms’ access to bank finance, lowers interest rates, and increases loans (Love et al. 2013). However, the usage of the movable collateral registry system is far from optimal, and more needs to be done to ensure financial institutions utilize the infrastructure to inform their credit decisions as well as to expand the range of products and services. The existing legal and institutional foundations can be used to develop and roll out asset-backed financing products (e.g., factoring, reverse factoring, channel financing), which are aimed at reducing the high loan collateralization rates. The current credit information system in Ethiopia is operating at suboptimal level and needs to be modernized. This is evidenced by the limited coverage, limited referencing perimeter, and lack of value-added products. By leveraging the existing financial infrastructures, more could be done to strengthen financial institutions’ utilization of credit information to inform credit decisions as well as to support the development of new financial products and spur financing. The informational ecosystem also needs to be developed, including by deepening the presence of accounting and audit firms/finance advisory firms as well as credit ratings agencies for independent third-party validation to strengthen cash flow-based lending and other data- driven lending products. Moreover, leveraging digital technologies to leapfrog credit evaluation using alternate approaches could be considered. Regulations need to allow for this and pilot a sandbox approach that could help bring informal firms into the system more rapidly. It is recommended that the requirement for commercial banks to have a 40 percent share of their portfolio in short-term loans is repealed permanently. As part of the forbearance measure that was enacted in response to the COVID-19 crisis, the NBE temporarily repealed the rule dictating that banks must keep a minimum of 40 percent of their loan portfolios in short-term loans, with long-term loans not to exceed 20 percent of their portfolio. This is a highly distortionary measure that has negatively affected manufacturing firms, which often require long-term loans to finance capital investments. This rule on loan maturity composition of the portfolio should be repealed permanently, since lenders are best placed to manage their asset liability risk profiles and choose loan tenures that best align with their liability profile and institutional risk appetite. The authorities could consider establishing a credit guarantee scheme. A credit guarantee scheme (CGS) provides third-party credit risk mitigation to lenders by sharing a portion of the lender’s losses on the loans made to target borrowers in case of default. As CGSs reduces the banks’ risk exposur es and allows banks to lower the amount of security required in the form of collateral, CGSs can spur growth in financial institutions lending to sectors or segments that are perceived as ‘high risk,’ in this case small manufacturing firms. 3.5.1.3 Removing distortions in access to inputs and foreign exchange While the manufacturing sector is (on paper) a priority to the Ethiopian government in terms of foreign exchange allocation, firms have faced severe constraints in practice, hindering sector potential. Firms often face a lengthy process to get foreign exchange permits (e.g., letter of credit, cash against document, telegraphic transfer). According to the 2015 World Bank Enterprise Survey, it took about 60-65 days on average to get a permit, with firms in industries such as textiles and transport equipment reporting an average wait time of over 100 days. As expected, importers and large establishments were more likely to apply for foreign exchange permits than other firms. Even among firms getting their applications accepted, they usually received foreign exchange for less than the amount requested, thus needing to resort to the parallel market to make up the difference. Foreign exchange scarcity has become increasingly acute in recent years, leading to the widening of the parallel market premium and increasing distortions in firm behavior. For example, some car importers have established parallel coffee export businesses and are willing to operate them at a loss just to access foreign exchange for imports (Tamru et al., 2021). 78 Ethiopia Country Economic Memorandum Full implementation of the planned exchange rate reform will be key to supporting a manufacturing take-off in Ethiopia. Increasing the pace of nominal depreciation is necessary to curb real overvaluation but is unlikely to be sufficient for eliminating the parallel market unless surrendering requirements and controls are phased out. Implementation of the approved Exchange Rate Reform Roadmap would need to be accelerated so that by the time the exchange rate is unified, the following elements are in place: (i) a Code of Conduct for foreign exchange (FX) transactions and platforms to trade FX (including public auctions, IT platforms, etc); (ii) a methodology for calculating reference exchange rate based on market transactions; and (iii) removal of current account restrictions, in particular, Franco Valuta Licensing (Regulation 66/2013) and the queue directive (FXD/62/2019). The boosting of foreign exchange reserve levels and the lifting of restrictions would be preconditions for successful foreign exchange reform, so that hard currency will be available to manufacturers and other private sector actors. 3.5.1.4 Leveling the playing field and limiting the influence of market power Although the regulatory framework for competition in Ethiopia has been improving, institutional weaknesses hamper implementation. The Trade Competition and Consumer Protection Authority (TCCPA), Ethiopia’s competition agency, is accountable to the Ministry of Trade and was established under the Trade Competition and Consumer Protection Proclamation (No. 813/2013). This Proclamation strengthened the competition framework outlined in the 2003 and 2010 laws by provisioning for the control of merges and acquisitions, introducing enforcement measures, and endowing the competition authority with decision-making power. However, enforcement of the Proclamation’s provisions by the TCCPA has been slow, as the institution does not have sufficient capacity to handle the number of cases that have arisen. As of 2019, TCCPA had not initiated any market inquiries, although it has conducted surprise investigations to identify evidence in cases of alleged anti-competitive behavior in sectors including steel and breweries. While private firms have made progress entering sectors that were previously dominated by SOEs, the playing field remains uneven in some subsectors. Private sector actors have made significant inroads in sectors historically dominated by SOEs, such as cement and banking. However, other sectors such as sugar, microfinance, and fertilizer and wheat import and distribution remain heavily concentrated, with state-owned businesses maintaining a dominant role. Abusive, anti-competitive practices by dominant firms that prevent market entry are still an obstacle to a growing private sector. These practices include predatory pricing and captive supply arrangements. In the case of manufacturing, while the revenue share held by private actors has increased across sectors, SOE presence remains significant in paper & publishing, machinery & equipment, non-metallic products, and food & beverages (Figure 50). 79 Ethiopia Country Economic Memorandum Figure 50. SOEs’ revenue share in manufacturing has declined across sectors but remains significant in some areas SOE revenue share in manufacturing subsectors 50% Machinery & 40% Equipment SOEs' revenue share in 2012 Non-Metallic Paper & Publishing 30% Food & Beverage 20% Wood & Furniture Chemicals Textiles & Apparel 10% Rubber & Plastic Metal Products Leather 0% 0% 5% 10% 15% 20% 25% 30% SOEs' revenue share in 2018 Source: Authors’ calculation based on data from Ethiopian Large and Medium Scale Manufacturing and Electricity Industries Survey, Central Statistical Agency of Ethiopia. The adoption and implementation of additional laws can help foster fair and healthy business development. The adoption of the draft Public Enterprise Proclamation would require SOEs to abide by the standards of the new Commercial Code and to compete on a more equal basis with private firms. It will also be important to enact and implement the new Competition Proclamation (as well in draft at the time this report was written), which provides for enterprises with mixed ownership, eliminates remaining provisions that can be used to control prices, and further strengthens the framework to fight against anti-competitive behavior. Finally, it is worth noting that ensuring competitive neutrality would also require eliminating the preferential access to foreign exchange and financing that the SOEs enjoy at the moment vis-à-vis private companies. In addition, the supervision and enforcement capacity of the competition body would need to be reinforced. TCCPA surveillance capacity must be improved to provide timely responses to complaints, initiate investigations of anti-competitive practices, and be able to exact penalties and sanctions. In particular, the competition authority would need to be reinforced to identify anti-competitive practices associated with the abuse of market dominance. To simplify the requirements for launching an investigation, authorities may want to consider adopting market share-based thresholds, over which it is assumed that a firm enjoys a dominant position in that market. 3.5.2 Upgrading firm capabilities Regulatory barriers must be addressed to increase firms’ incentives to invest in productivity- enhancing activities such as the use and adoption of better technologies, introduction of new products, and adoption of better management practices. Firm capabilities are key components of business operations that cannot be acquired through market purchase like other production inputs. Instead, they must be learned, 80 Ethiopia Country Economic Memorandum developed, and accumulated (Sutton 2012, as cited in Cirera and Maloney 2017). They are an integral source of firms’ competitiveness, especially in a dynamic environment that requires constant development and reorganization of resources. 3.5.2.1 Improving managerial practices Ethiopia lags in managerial practices, particularly in terms of instilling and retaining talent. Recent evidence points to managerial practices as one of the determinants of productivity.26 According to the latest data from the World Management Survey in 2015, Ethiopian firms are ranked rather low overall, lagging peers in Kenya, Tanzania, and Vietnam (Figure 51). Taking Vietnam as a reference and controlling for firm size and industry characteristics, Ethiopian firms display poor management performance, especially on human resource management practices related to worker selection, incentive provision, and worker retention. In Ethiopia, the proliferation of survivalist entrepreneurs is indicative of managerial capacity constraints in incumbent firms and the lack of a program dedicated to supporting the entire lifecycle of an SME (Siba and Mekonnen 2019). Limited managerial and marketing capacity may also be limiting local sourcing for foreign direct investment. Furthermore, a survey of the employees of apparel producers in the industrial parks found significant challenges to productivity such as high attrition, absenteeism, lack of sense of urgency for work, and low motivation to work overtime. These features partly reflect weak incentive structures in terms of wage payments and working conditions (Policy Studies Institute 2020). Better-managed firms are found to leverage more incentive-based pay in their compensation packages to motivate workers and spur productivity (Meyer et al. 2021). Other studies have also shown that selection and retention of employees are the most challenging issues Ethiopian firms face (Abebe et al. 2017; Blattman and Dercon 2017). Figure 51. Ethiopian managers feature similar Management score (2015) 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2 Ethiopia Kenya Tanzania Vietnam Source: World Management Survey. In Ethiopia, provision of management practices consulting could help boost productivity. To overcome the challenges above, the authorities could consider options such as: (i) providing group-based management 26 These management practices include operations such as monitoring of performance, introducing targets such as type and time horizon of production goals, and setting up incentives for talent attraction and retention. Bloom et al. (2013), using data from field experiments with Indian textile firms, found that providing management practices consulting resulted in a significant productivity boost as well as operations expansion in the form of plant openings. The study also identified lack of information as a barrier to adoption of better management practices. In a follow-up study on the same set of firms, Bloom et al. (2020) found that about half of the management practices adopted were abandoned, but the gap between adopters and non-adopters remains significant. 81 Ethiopia Country Economic Memorandum consulting, drawing from the Colombian experience;27 (ii) connecting firms with professional business services for direct outsourcing of certain functions (e.g., marketing/sales, finance, accounting), as done in Nigeria (Anderson and McKenzie 2020); and (iii) linking industry with business schools for executive education. In terms of business development support provision in Ethiopia, the Entrepreneurship Development Center (EDC), TVET schools, and private sector firms (DOT Ethiopia, Young Entrepreneur Program) all have a role to play in managerial capacity building training. Table 9. Policy recommendations for improving manufacturing productivity in Ethiopia Area Short term Medium term Implementing agency Firm entry Fully automate the business Simplify and shorten exit-related Ministry of Trade and Industry, and exit establishment process tax audit processes Ministry of Revenue Fully implement the 2021 Commercial Code provisions on company formation, operation, and insolvency Access to Strengthen the credit bureau system Explore the establishment of a National Bank of Ethiopia finance and introduce credit scoring for cash public credit guarantee scheme flow-based lending Leverage digital technologies to Support the development, uptake, leapfrog credit evaluation and implementation of leasing and factoring products Repeal the requirement for banks to have a 40% share of their portfolio in short-term loans Access to fx Set up the platforms and Shift to a market-determined National Bank of Ethiopia methodology to have a market- exchange rate and lift all the determined exchange rate currency allocation mechanisms Lift current account restrictions Competition Enact and implement the new Strengthen the capacity of the Competition Authority (TCCPA) and market Competition Proclamation that Competition Authority for power removes provisions allowing for supervision and enforcement price controls Ensure competitive neutrality Adopt market share-based thresholds between SOEs and private firms to determine market dominant (including in access to forex and position finance) Firm Provide management practices Ethiopian Management Institute, capabilities consulting to improve firm Ethiopian Chamber of Commerce productivity and Sectoral Association, Entrepreneurship Development Centre Source: Elaborated by the authors. 27 A study by Iacovone et al. (2021) on Colombian manufacturing reveals that both individual and group-based consulting resulted in significant gains in labor productivity, revenue, and profitability. Although individual-based consulting has greater and more robust impact, group-based management consulting should be prioritized given cost considerations and limited resource availability to scale up the support. 82 Ethiopia Country Economic Memorandum 3.6 References Abebe, G., S. Caria, and E. Ortiz-Ospina. 2017. “The Selection of Talent: Experimental and Structural Evidence from Ethiopia.” Working paper. Abrego, L., M. Alejandra Amado, T. Gursoy, G. Nicholls, and H. Perez-Saiz. 2019. “The African Continental Free Trade Agreement: Welfare Gain Estimates from a General Equilibrium Model.” IMF Working Paper # 19/124. Abreha, K., P. Jones, E. Lartey, T. Mengistae, and A. Zeufack. 2019. “Manufacturing Job Growth in Africa: What Is Driving It? The Cases of Côte d’Ivoire and Ethiopia.” Washington, DC: World Bank. African Union. 2018. Agreement Establishing the African Continental Free Trade Area. Aghion, P., S. Bechtold, L. Cassar, and H. Herz. 2018. “The causal effects of competition on innovation: Experimental evidence.” Journal of Law, Economics, and Organization 34(2):162-195. Anderson, S. and D. McKenzie. 2020. “Improving Business Practices and the Boundary of the Entrepreneur : A Randomized Experiment Comparing Training, Consulting, Insourcing and Outsourcing.” Policy Research working paper, no. WPS 9502, Impact Evaluation series. Washington, DC: World Bank Group. Ayalew, M., Z. Xianzhi, Y. Dinberu, and D. Hailu. 2020. “The Determinants of Firm’s Innovation in Africa.” Journal of Industry, Competition and Trade 20: 527–567. Banga, K. and D. te Velde. 2018. Digitalisation and the future of African manufacturing. London: ODI. Baqaee, D. and E. Farhi. 2020. “Productivity and misallocation in general equilibrium.” Quarterly Journal of Economics 135(1): 105-163. Blattman, C., and S. Dercon. 2017. “The Impacts of Industrial and Entrepreneurial. Work on Income and Health: Experimental Evidence from Ethiopia.” Working paper. Bloom, N., B. Eifert, A. Mahajan, D. McKenzie, and J. Roberts. 2013. “Does Management Matter? Evidence from India.” Quarterly Journal of Economics 128(1): 1–51. Bloom, N., A. Mahajan, D. McKenzie, and J. Robert. 2020. “Do Management Interventions Last? Evidence from India.” American Economic Journal: Applied Economics 12(2): 198-219. Busso, M., L. Madrigal, and C. Pagés. 2013. “Productivity and resource misallocation in Latin America.” The BE Journal of Macroeconomics 13(1): 903-932. Cirera, X. and W. F. Maloney. 2017. “Managerial Practices as Key Firm Capabilities for Innovation.” In “The Innovation Paradox: Developing-Country Capabilities and the Unrealized Promise of Technological Catch-Up.”, 65-90. Washington, DC: World Bank. Cirera, X., R. Fattal-Jaef, and H. Maemir. 2020. "Taxing the Good? Distortions, Misallocation, and Productivity in Sub-Saharan Africa." World Bank Economic Review 34(1): 75-100. Cornell University, INSEAD, and WIPO. 2020. The Global Innovation Index 2020: Who Will Finance Innovation? Ithaca, Fontainebleau, and Geneva. CSA (Central Statistical Agency). 2014. “Report on Small Scale Manufacturing Industry Survey 2014.” Addis Ababa, Ethiopia, 2014. https://www.statsethiopia.gov.et/wp-content/uploads/2019/06/Small- Scale-Manufacturing-Survey-Report-2014.pdf. 83 Ethiopia Country Economic Memorandum Cusolito, A., D. Lederman, and J. Peña. 2020. “The Effects of Digital-Technology Adoption on Productivity and Factor Demand: Firm-level Evidence from Developing Countries.” Policy Research Working Paper # 9333. Washington, DC: World Bank. Damoah, K., G. Giovannetti, and M. Sanfilippo. 2020. “Markup Dispersion and Firm Entry: Evidence from Ethiopia.” Oxford Bulletin of Economics and Statistics. De Loecker, J., J. Eeckhout, and G. Unger. 2020. “The rise of market power and the macroeconomic implications.” Quarterly Journal of Economics 135(2): 561-644. De Loecker, J., and F. Warzynski. 2012. “Markups and Firm-Level Export Status.” American Economic Review 102 (6 ): 2437-2471. Diao, X., M. Ellis, M. Margaret, and D. Rodrik. 2021. “Africa's Manufacturing Puzzle: Evidence from Tanzanian and Ethiopian Firms.” NBER Working Paper 28344. Dias, D., C. Marques, and C. Richmond. 2016. “Misallocation and productivity in the lead up to the Eurozone crisis.” Journal of Macroeconomics 49: 46-70. Edmond, C., V. Midrigan, and D. Xu. 2018. “How costly are markups?”. NBER Working Paper # 24800. Gebrewolde, T., M. Koelle, P. Krishnan, and A. Mengistu. 2020. “Currency Shocks and Firm Behaviour in Ethiopia and Uganda.” Centre for Economic Policy Research, Discussion Paper DP15524. Goldstein, P. 2020. “Pathways for Productive Diversification in Ethiopia.” http://www.tinyurl.com/y2uyvh9u Gopinath, G., Kalemli-Ozcan, S., Karabarbounis, L. and C. Villegas-Sanchez. 2017. “Capital Allocation and Productivity in South Europe.” Quarterly Journal of Economics 132 (4): 1915-1967. Hallward-Driemeier, M., and G. Nayyar. 2017. “Trouble in the Making?: The Future of Manufacturing-Led Development.” Washington, DC: World Bank. Hsieh, C., and P.J. Klenow. 2009. “Misallocation and Manufacturing TFP in China and India.” Quarterly Journal of Economics 124(4): 1403–1448. Hsieh, C., and P. Klenow. 2014. “The Life Cycle of Plants in India and Mexico.” Quarterly Journal of Economics 129 (3): 1035–84. Iacovone, L., Maloney, D., and D. McKenzie. 2021. “Improving Management with Individual and Group- Based Consulting: Results from a Randomized Experiment in Colombia.” Review of Economic Studies. Volume 89, Issue 1, January 2022: 346–371. Love, I., M. Peria, and S. Singh. 2013. “Collateral Registries for Movable Assets: Does Their Introduction Spur Firms' Access to Bank Finance?” Policy Research Working Paper # 6477. Washington, DC: World Bank. Maemir, H. 2020. Drivers of Manufacturing Productivity in Cote d'Ivoire. Cote d'Ivoire Economic Memorandum. World Bank. McCaig, B., and N. Pavcnik. 2018. “Export markets and labor allocation in a low-income country.” American Economic Review 108(7): 1899-1941. McGowan, M., D. Andrews, and V. Millot. 2017. Insolvency regimes, zombie firms and capital reallocation. OECD Economics Department Working Papers 1399, OECD Publishing. 84 Ethiopia Country Economic Memorandum Melitz, M., and S. Polanec. 2015. “Dynamic Olley-Pakes Productivity Decomposition with Entry and Exit.” Rand Journal of Economics 46 (2): 362–75. Meyer, C., E. Krkoska, and K. Maaskant. 2021. “Wages and Compensation in Ethiopia’s Industrial Parks: Evidence from a Firm Survey.” Washington, DC: World Bank. Midrigan, V. and D. Xu. 2014. “Finance and Misallocation: Evidence from Plant-level Data.” American Economic Review 1014(2): 422-458. National Planning Commission. 2016. “Growth and Transformation Plan II (GTP II): 205/160 - 2019/20.” Addis Ababa. Nayyar, G., M. Hallward-Driemeier, and E. Davies. 2021. At Your Service?: The Promise of Services-Led Development. Washington, DC: World Bank. Olley, G.S., and A. Pakes. 1996. “The Dynamics of Productivity in the Telecommunications Equipment Industry.” Econometrica 64 (6): 1263–97. Oqubay, A. 2019. “The Structure and Performance of the Ethiopian Manufacturing Sector.” In Cheru, F., C. Cramer, and A. Oqubay, Eds., The Oxford Handbook of the Ethiopian Economy. Oxford: Oxford University Press USA. Policy Studies Institute. 2020. Ethiopia Productivity Report. Addis Ababa. Reijnders, L., M. Timmer, and X. Ye. 2021. “Labour demand in global value chains: Is there a bias against unskilled work?” The World Economy. Restuccia, D. and R. Rogerson. 2013. “Misallocation and productivity.” Review of Economic Dynamics 16 (1): 1- 10. Rodrik, D. 2016. "Premature Deindustrialization." Journal of Economic Growth 21 (1): 1–33. Rodrik, D. 2018. “New Technologies, Global Value Chains, and Developing Economies.” NBER Working Paper No. 25164. Cambridge: National Bureau of Economic Research. Siba, E., and B. Mekonnen. 2019. Addressing Challenges of Ethiopian SMEs’ Access to Markets and Production Facilities. Background paper, The World Bank Group. Mimeo. Syverson, C. 2019. “Macroeconomics and market power: Context, implications, and open questions.” Journal of Economic Perspectives 33(3): 23-43. Tamru, S., B. Minten, B. and J. Swinnen. 2021. “Trade, value chains, and rent distribution with foreign exchange controls: Coffee exports in Ethiopia.” Agricultural Economics, 52, issue 1, p. 81-95. UNIDO. 2019. “Industrial Development Report 2020. Industrializing in the Digital Era.” Vienna: UNIDO. Van Reenen, J. 2011. “Does competition raise productivity through improving management quality?” International Journal of Industrial Organization 29(3): 306-316. World Bank. 2019. World Development Report 2020: Trading for Development in the Age of Global Value Chains. Washington, DC: World Bank. World Bank. 2020. Doing Business 2020. Washington, DC: World Bank. World Bank. 2020. The African Continental Free Trade Area: Economic and Distributional Effects. Washington, DC: World Bank. 85 Ethiopia Country Economic Memorandum World Bank and FIRST Initiative. 2015. Principles for Public Credit Guarantee Schemes for SMEs. Washington, DC: World Bank. 86 Ethiopia Country Economic Memorandum 4 Reviving trade 4.1 Motivation: what went off-track? Despite the emphasis on boosting exports in recent government strategies, trade failed to take off and even declined in relative terms during the past decade. While Ethiopia has found some success in establishing footholds in some manufacturing sectors (particularly apparel), and exports of services have experienced healthy growth, propelled by Ethiopian Airlines, the trade performance has fallen short of expectations. Ethiopia’s exports of goods and services as a share of GDP halved over the past decade and, at 7.9 percent in 2019, they are the lowest among peer countries (Figure 52). Exports remain skewed toward agriculture commodities, which have been less dynamic overall, and export growth has not kept pace with GDP. Figure 52. Ethiopia’s exports of goods and services have halved over the past decade and are now the lowest among peer countries Exports of goods and services, % of GDP 120 106.8 100 80 61.1 60 40 21.8 16.7 17.2 15.3 20 12.0 7.9 0 Source: World Development Indicators. Note: Data for 2019, except for Tanzania (2017). This chapter identifies a series of shortcomings in Ethiopia’s export strategy. While merchandise exports are more diversified than in other peer countries in the region, over the past decade there has been a trend towards concentration, and the pace of product discovery has slowed down. An overvalued exchange rate and foreign exchange scarcity has resulted in the stagnation in industrial supplies and capital goods since 2015. Export crops have also faced headwinds during the second half of the past decade. In terms of participation in global value chains, and despite rapid growth in trade integration, there is evidence of stagnation over recent years, backward linkages are limited, and the country has not able to consolidate new comparative advantages. As discussed in this chapter, correcting an adverse exchange rate policy, reducing tariff and non-tariff barriers, and improving trade facilitation are among the avenues for boosting an export revival. In addition, future trade strategies shall complement the current focus on manufacturing with measures to leverage the potential in services trade, where there seem to be untapped opportunities. 87 Ethiopia Country Economic Memorandum 4.2 The quest for export diversification 4.2.1 Merchandise export performance and diversification Exports of goods have stagnated since 2014, although garment exports have kept growing. According to official figures, Ethiopian exports of goods grew at an annualized rate of 16.5 percent between 2002 and 2014—a surge in agriculture commodity exports drove the growth trend, with Ethiopia taking advantage of the commodity price boom in the early 2000s. After 2014, the value of agriculture exports leveled off as prices declined, and exports of goods have stagnated (Figure 53). However, Ethiopia has succeeded in growing exports of textiles, clothing, and footwear from essentially zero early in the sample to around $0.5 billion in 2019. Exports in these categories have continued growing in recent years, although their growth was insufficient to offset the stagnation in other export categories. Among the reasons why Ethiopia has been unable to further develop the light manufacturing industry are an unfavorable exchange rate and business climate. Figure 53. Exports of goods hit their ceiling in value terms in 2014 Ethiopia's exports of goods by product type 3.0 Merchandise exports value in US$ billion 2.5 2.0 1.5 1.0 0.5 0.0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Animal Vegetables Textiles, Clothing Footwear Others Source: UN COMTRADE based on mirror data. Diversification in exports of goods has been limited, with a decline in the number of new products being exported every year. The degree of export diversification varies substantially across sectors, with glass, fuel, mineral, and footwear products being highly concentrated while other categories enjoy greater diversification (Figure 54, left panel). Of concern is the fact that the export product concentration index is generally increasing over time, indicating that exports are becoming more concentrated rather than more diversified. An important exception is textiles and clothing, where the degree of concentration is low and has continued decreasing. Thus, there are parts of the economy where export diversification is taking place, but overall, change is in the opposite direction. Consistent with the stagnation in Ethiopia’s export performance over recent years, the rate of “discovery” (exports of previously un-exported products) has been slowing down (Figure 54, right panel). 88 Ethiopia Country Economic Memorandum Figure 54. Overall, exports have become more concentrated, with a declining number of new product discoveries Product concentration index New product discoveries (HS6) 160 Transportation Mach and Elec Textiles and Clothing 140 Vegetable Miscellaneous 120 Footwear Number of new products Minerals 100 Animal Metals Plastic OR Rubber 80 Chemicals Stone And Glass 60 Hides And Skins Fuels 40 Food Products Wood 20 0 0.2 0.4 0.6 0.8 1 Herfindahl-Hirschman Index 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2016 2008 Source: UN COMTRADE based on mirror data. Note: In the left-hand side chart, a higher Herfindahl-Hirschman Index (HHI) indicates a higher degree of concentration; conversely, a lower score indicates a higher degree of diversification. HS6 (right-hand side) refers to products at the six- digit level according to the Harmonized System international product nomenclature developed by the World Customs Organization. Destination markets for exports of goods are well-diversified, although some trend toward concentration can be seen in recent years. The European Union (about one-third of total exports) and China stand out as the main trading partners for Ethiopian goods, although exports to these destinations stagnated or slightly declined over the second half of the past decade. Meanwhile, the importance of the United States as destination has increased in recent years, reaching 15 percent total Ethiopian exports in 2019 (Figure 55, left panel). As a Least Developed Country in Africa, Ethiopia already enjoys preferential access to the EU market (Everything But Arms) and the U.S. market (African Growth and Opportunity Act28). Other important export destinations include the United Arab Emirates and Saudi Arabia. The role of other African countries is very limited, although some growth has taken place over the past two decades. Overall, in terms of destination markets, Ethiopia is more diversified than other peer countries except Rwanda, although there has been some trend toward concentration in recent years (Figure 55, right panel). Ethiopian goods exports figures show a stable number of export relationships per year. 28 Preferential access for Ethiopian products under the African Growth and Opportunity Act was suspended in January 2022 over concerns regarding international human rights violations during the armed conflict in Northern Ethiopia. 89 Ethiopia Country Economic Memorandum Figure 55. Ethiopia is more diversified that most peers in terms of destination markets, although concentration has increased slightly Ethiopia's exports of goods by destination Concentration in destination markets 3.0 Vietnam 2.5 Cambodia Trade value in US$ billion 2.0 Uganda 1.5 Rwanda 1.0 Tanzania 0.5 Kenya 0.0 Ethiopia 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 0 0.2 0.4 0.6 0.8 Africa ASEAN China Herfindahl-Hirschman Index EU India Japan Middle East Rest of the World UK 2018 2008 USA Source: UN COMTRADE based on mirror data. Note: A higher Herfindahl-Hirschman Index (HHI) indicates a higher degree of concentration; conversely, a lower score indicates a higher degree of diversification. Survival rates for Ethiopian exports seem on par with Kenya and Bangladesh but are well below those of Vietnam. The probability of survival of two traditional agricultural exports (coffee and sesame), one horticultural export (cut flowers), and one garment item (baby clothes) was analyzed. The probability of survival of Ethiopian coffee and sesame exports is lower than that of Vietnam and Kenya. Cut flower survival rates are similar to those observed in Kenya, while the surviving export spells tend to last more years. In garments, export relationships seem to survive longer than those originating in Bangladesh and Kenya (but not Vietnam). Relatively long survival rates in garments may be explained by the fact that investments in the sector are relationship-specific, having been made by foreign investors with well-established suppliers and customers abroad. 4.2.2 Imports of goods Imports of goods surged during the past decade in support of the infrastructure boom but declined in recent years as imbalances deepened and foreign exchange became scanty. In the early 2010s, Ethiopia dramatically increased its imports of machinery, metals, and construction materials, as state-owned enterprises (SOEs) entered a spending spree to build large infrastructure megaprojects and private construction activity boomed. To a lesser extent, investment in machines and acquisition of intermediates for the manufacturing sector contributed to the increase. As external imbalances enlarged and foreign exchange became scanty, imports have been curbed in recent years, in particular those of capital goods and industrial supplies (Figure 56) this has negatively affected export performance in recent years. Consumer goods play a limited role in the Ethiopian import basket, both because the country is relatively populous but still at a low-income level and due to the foreign exchange allocation system that prioritizes oil, fertilizers, medicines, and machinery imports, among others (see Chauffour and Gobezie, 2019). The growing imports of transport equipment largely have to do with purchases of airplanes by Ethiopian Airlines. 90 Ethiopia Country Economic Memorandum Figure 56. Imports of goods peaked in 2017, as access to foreign exchange increasingly became a constraint Ethiopia, imports by end use, Broad Ethiopia, imports by product type Economic Categories (BEC). 12 12 Mechandise import value in US$ billion Merchandise import value in US$ billion 10 10 8 8 6 6 4 4 2 2 0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Machinery/Electricals Metals Goods nec Consumer Goods nec Textiles, Clothing Chemicals Transport Equipment and Parts Capital Goods and Parts Animals Vegetables Fuels and Lubricants Industrial Supplies nec Others Food and Beverages Source: Authors’ calculations using UN COMTRADE based on mirror data. Note:‘nec’ refers to goods not elsewhere classified. Ethiopia relies heavily on a few partner countries to procure its merchandise imports. In 2019, Ethiopia’s three top sources of imports were China (26 percent of total), India (9 percent), and United Arab Emirates (9 percent). European Union countries (including the United Kingdom) held a share of more than one-third of total imports, suggesting the block is important in both directions of trade. Meanwhile, although the United States is a relatively important export market for Ethiopia, it does not provide significant imports. As was the case for exports, Ethiopia’s import sourcing does not rely on Africa, although the relationship has strengthened in recent years. 4.2.3 Searching for a comparative advantage in merchandise exports Ethiopia’s comparative advantage in manufacturing relative to agriculture is greatest in textiles and apparel. Countries are said to enjoy a comparative advantage for a certain good when they can produce it more efficiently (at a lesser cost) than other countries. Figure 57 presents estimates of Revealed Comparative Advantage (RCA) relative to agriculture for the manufacturing sectors included in the Eora database (see Box 5 on the methodology). In manufacturing, Ethiopia’s strongest comparative advantage is in textiles and apparel, followed by food and beverages then other manufacturing. Overall, the sectoral pattern of comparative advantage in Ethiopia is similar to what is seen in the regional comparators, but the Asian comparators have a much stronger comparative advantage relative to agriculture. There is little evidence that Ethiopia’s comparative advantage in manufacturing relative to agriculture is growing substantially stronger over time, as the sector has found limited success in boosting productivity. Ethiopia’s comparative advantage did not increase significantly between 2000 and 2015, which suggests that manufacturing productivity is not growing rapidly relative to agriculture. This is consistent with the findings of the manufacturing census analysis presented in Chapter 3 as well as with the work of Diao et al. (2021), which concur with that overall result but nuance it by using firm-level data to show that large firms are seeing significant productivity gains while small firms are not—yet it is the latter that are generating most of the observed employment growth. 91 Ethiopia Country Economic Memorandum Figure 57. Ethiopia’s comparative advantage in garments is not as pronounced as in Asian comparators Comparative advantage in manufacturing sectors, Ethiopia and comparators, 2015 3 2.5 2 RCA score 1.5 1 0.5 0 Bangladesh Ethiopia 2000 Ethiopia 2015 Cambodia Rwanda Tanzania Uganda Vietnam Electrical & Machinery Food & Beverages Metal Products Other Manufacturing Petroleum, Chemicals etc. Textiles and Apparel Transport equipment Wood and Paper Source: Authors’ calculations based on EORA data. The results are consistent with Ethiopia’s focus in terms of developing a comparative advantage in sectors like textiles and clothing, as well as food and beverage. Other sectors have lower scores, with little evidence of an upward trend. Textile and apparel has the strongest comparative advantage score in Ethiopia, and it is ahead of the scores of Rwanda, Tanzania, and Uganda. However, Bangladesh, Cambodia, and Vietnam, which are among the established garment exporters worldwide, have a large edge over Ethiopia. As the comparative advantage of Ethiopia is not nearly as pronounced as in these Asian comparators, work needs to be done in areas like tariffs, non-tariff measures (NTMs), and trade facilitation, as well as potentially inward investment, to increase sectoral productivity. These issues are discussed in more detail later in this chapter. Box 5. Estimating comparative advantage Costinot et al. (2012) develop a Ricardian model of trade, extending the work of Eaton and Kortum (2002). 29 Their objective is to quantify the importance of productivity differences as a driver of trade. But as a by-product of their investigation, they develop a simple method for analyzing patterns of comparative advantage that is fully consistent with their theoretical setup. Like many models of trade, theirs can be reduced to a gravity-like relation. Specifically, their theory predicts that bilateral trade flows by sector should satisfy the following relation: (1)⁡ = Where is exports from country i to country j in sector k; is a country pair fixed effect capturing structural features of the model, such as trade costs; groups together importer-sector factors in a fixed effect; is a 29 The original source uses a log-linearized model. The presentation here retains the nonlinear form for the reasons set out in Santos Silva and Tenreyro (2006). Estimation is therefore by Poisson Pseudo-Maximum Likelihood (PPML) rather than Ordinary Least Squares (OLS). Similarly, internal trade is included in line with now-standard practice in gravity modeling, and as implied by theory (Yotov et al., 2016). Estimation uses the PPMLHDFE package (Correia et al., 2019). 92 Ethiopia Country Economic Memorandum parameter from the theory capturing intra-industry heterogeneity in productivity; is the fundamental productivity of country i in sector k, taking account of factors like climate, infrastructure, and institutions that affect all producers within a country; and is an error term satisfying standard assumptions. As suggested by the use of a parameter like this, the objective of the exercise is to quantify comparative advantage, not to uncover its sources as in models like Chor (2010), applied to services by van der Marel (2011). Costinot et al. (2012) initially estimate (1) directly, using productivity estimates drawn from available data. However, such an approach is not practical for application to a wide range of countries, particularly developing ones, as such estimates are not readily available on a comparable basis. As the authors note, they are also subject to significant concerns regarding measurement error. An alternative approach is therefore to replace the productivity variable with an exporter-sector fixed effect d: (2)⁡ = The standard PPML estimate will produce consistent estimates of the exporter-sector fixed effects. Once the estimates have been obtained, a value of from the literature can be used to construct revealed productivity ⁄ ) where the numerator is simply the PPML estimate of the measures by exponentiation, i.e. = exp ( exporter-sector fixed effect. There are important advantages to proceeding in this way. First, the only limit on application is the availability of trade data. Second, the revealed productivity measure can be interpreted, as the authors do, in terms of a theoretical revealed comparative advantage measure by scaling it relative to a baseline country (the United States) and a baseline sector in each country (agriculture). The theory-consistent measure of comparative advantage has a number of advantages over the Balassa measure that is more commonly used. First, the Balassa measure is not informative about comparative advantage in a world with varying trade costs (French, 2017), whereas the measure here explicitly controls for the impact of trade costs. Second, the modification used here takes account of domestic production, which arguably is closer to the core idea of comparative advantage than a measure based on trade only, as is the case of the Balassa measure. Third, the measure does not have an artificial cutoff but is instead continuous (Costinot et al., 2012). The data source for estimating comparative advantage is the Eora multi-region input-output table. The reason for using this source is that it includes intra- as well as international trade. While it introduces some amount of smoothing into the results because input-output tables are often interpolated over time, comparison with results from another source (the USITC ITPD-E database) indicate that the interpretation of results is largely unaffected. 4.3 Service exports: growing potential 4.3.1 A lone star: transport exports Ethiopia’s service exports have boomed over the past 15 years, driven by the strong performance of transport and travel. Service exports now account for two-thirds of Ethiopia’s overall export bundle, compared to less than half in 2010. Services exports more than tripled in value terms between 2005 and 2019, following a similar pattern to the comparator countries. At the end of the sample period, Ethiopia’s services exports were between those of Uganda and Kenya (Figure 58, left panel). However, Ethiopia’s services exports remain almost exclusively concentrated in transport and travel, which is a point of distinction among peer countries. The astonishing growth of Ethiopian Airlines, which has become among the top carriers in the African continent in recent years, has been the main (and almost only) driver of service export growth. 93 Ethiopia Country Economic Memorandum Figure 58. Ethiopia’s total services exports fall between Uganda and Kenya but would be negligible if transport and travel is excluded Commercial service exports Commercial service exports, excluding transport and travel 10,000 700 9,000 600 Balanced trade in millions US$ Balanced trade in million US$ 8,000 7,000 500 6,000 400 5,000 300 4,000 3,000 200 2,000 100 1,000 0 0 Ethiopia Kenya Rwanda Uganda Ethiopia Kenya Rwanda Uganda Source: WTO-OECD BATIS Database. Ethiopia’s exports of services remain highly concentrated, as the country has been unable to export beyond transport and travel. Transport exports (mostly air travel), at two-thirds of the total, and travel (including tourism), at one-third of total, are the sole service exports of Ethiopia (Figure 59, left panel). While other comparators such as Cambodia, Kenya, Rwanda, and Uganda also show a high degree of concentration particularly in travel services, they have also been able to ignite exports in other services (albeit still incipient). Meanwhile, Bangladesh has built strong exports of IT-related services and business services (which includes back-office activities), and Vietnam has done so in goods-related services (which add value to final goods produced). Figure 59. Ethiopia’s service imports are more diversified than its service exports Service exports composition Travel Service imports composition 100% Transport 100% 90% 90% Telecommunications 80% and IT 80% 70% Personal, Cultural and 70% Recreational Services 60% 60% Other Business Services 50% 50% 40% Insurance and Pension 40% Services 30% 30% Goods-related Services 20% 20% 10% Financial Services 10% 0% 0% Construction Charges for use of Intellectual Property nie Source: WTO-OECD BATIS Database. 94 Ethiopia Country Economic Memorandum In terms of services imports, Ethiopia purchases transport services as well as business services and goods-related services from abroad. While Ethiopia’s services imports have increased rapidly over the sample period—they were around five times higher in 2019 than in 2005—they remain at a low level relative to some of the comparators, particularly Vietnam and Bangladesh. In terms of composition, Ethiopia imports mostly transport services, business services, and goods-related services (Figure 59, right panel). The last category among the three is indicative of global value chains (GVC)-related activities and suggests in the case of Ethiopia that some traded goods undergo minimal processing in the country, as is typical of simple assembly linkages. 4.3.2 The emergence of new comparative advantages in services exports In contrast to the case of goods, a clear upward trend can be seen in Ethiopia’s comparative advantage in some services sectors relative to agriculture. Transport as well as post and telecommunications stand out as sectors showing some evidence of emerging comparative advantage (Figure 60). The same is true to a lesser extent for hotels and restaurants and wholesale trade. The adoption in 2020 of a new Investment Proclamation, which de facto opens up a number of services sectors to the private sector and to foreign investors (education, health, logistics, some segments within domestic transportation), holds promise in further elevating the service export potential. The exceptions are finance and retail, sectors in which some peer countries have developed a comparative advantage but that remain closed to foreign investors and underdeveloped in Ethiopia. Figure 60. In most service sectors, Ethiopia’s comparative advantage has been improving over time Comparative advantage in services sectors, Ethiopia and comparators, 2015 1 0.9 0.8 0.7 RCA score 0.6 0.5 0.4 0.3 0.2 0.1 0 Bangladesh Ethiopia 2000 Ethiopia 2015 Cambodia Rwanda Tanzania Uganda Vietnam Construction Finance & Business Hotels & Restaurants Maintenance & Repair Post & Telecom Retail Transport Wholesale Source: Authors’ calculations based on EORA data. 4.4 Global value chain participation Participating in GVCs is currently one of the most sought pathways to economic development. The way in which development policy can leverage trade in the early 21st century is fundamentally different from the way in which the “Asian Tigers” did so in the latter part of the 20th century (Baldwin 2011). Those countries (South Korea being the last) tended to develop full supply chains in key manufacturing industries. More recent instances of rapid development, such as Vietnam and in some respects China, have focused on joining existing supply chains rather than developing them domestically from scratch. This is the GVC development model: countries specialize in narrowly defined tasks rather than full sectors, they trade intermediate goods and services extensively, and over time, they move up into higher value-added activities. From this perspective, it is important to gauge the current state of GVC integration in Ethiopia. As trade policies are still more restrictive 95 Ethiopia Country Economic Memorandum than in its comparators, particularly for capital goods and intermediates, this section tries to shed light on whether that translates into decreased GVC integration due to difficulties in accessing inputs. Box 6 provides the details of the approach followed to measure GVC participation. Box 6. Measuring GVC participation The standard method in the literature for analyzing GVC linkages is Wang et al. (2013). Standard trade data are in gross shipments terms, which means they do not accurately track flows of value added, and they count inputs crossing borders multiple times. Wang et al. (2013) develop a fully consistent decomposition of gross value trade data into value added components, using a multi-region input-output table. Their approach makes it possible to identify two types of GVC linkages in trade. Backward linkages refer to a country’s use of imported intermediates to produce its exports. Forward linkages refer to a country’s export of intermediates that are then used by other countries to produce their exports. Summing these two quantities gives the proportion of exports that are believed to take place within GVC structures. World Bank (2020) uses a similar approach to Borin and Mancini (2019) to track GVC integration over time. The report notes that Ethiopia has seen a major increase in GVC integration since the beginning of the period for which data are available (1990). Given that the dynamics have already been investigated, the analysis here focuses on a cross-sectional comparison with the region and extra-regional comparators. The objective of proceeding in this way is to identify any points of similarity or difference in terms of GVC integration, which could in turn be for trade development strategies. The approach taken here applies the Wang et al. (2013) decomposition to Eora data for Ethiopia and the comparators. An important caveat relates to data quality, since Eora is based on estimates and imputations, in particular for low-income countries. Nonetheless, it represents the best source currently available for countries like Ethiopia. 4.4.1 An uneven integration: dominance of forward linkages over backward linkages Ethiopia’s overall level of GVC participation is comparable to what is seen in the comparator countries, although it varies significantly across sectors. Ethiopia’s degree of GVC integration in 2015 was higher than that of Uganda and Rwanda and broadly comparable to what is seen in Tanzania. Looking at extra-regional comparators, Ethiopia’s performance is comparable to that of Vietnam and is generally higher than that of Bangladesh and Cambodia. However, an important nuance to this general picture is that sectoral performance is highly varied: for example, the electrical and machinery sector has a very high degree of integration, while food and beverage activities show little participation in GVCs (Figure 61). While comparators in South Asia and Asia are skewed toward backward linkages, backward linkages are weak in Ethiopia. A key point of contrast between Ethiopia and most of the comparators, particularly in Southeast Asia, is that the latter skew toward backward linkages in their GVC participation, whereas Ethiopia tends to rely more heavily on forward linkages. Backward linkages are strong for electrical & machinery and petroleum products, which is possibly explained by priority purchases by the SOEs and by the government, respectively. In contrast, Ethiopia has the lowest level of backward linkages in textiles and apparel of any of the countries in the sample by a wide margin, which points to challenges in the supply of international inputs. Sourcing high-quality inputs at world market prices is an important element of competitiveness for producers. Difficulties in doing so tend to hinder the development of that competitiveness, which results in a situation where the country risks being locked into a relatively low value addition section of the value chain rather than progressively moving up over time. 96 Ethiopia Country Economic Memorandum Figure 61. Compared to peers, Ethiopia shows a higher degree of variability across sectors in terms of GVC integration in manufacturing GVC integration in manufacturing, Ethiopia and comparators, percent of gross exports, 2015 Source: Authors’ calculations based on EORA data. From a value addition perspective, concentrating on forward linkages may not be optimal for Ethiopia. The importance of forward linkages needs to be understood in terms of the types of products and relationships likely involved. Analytically, the linkage captures goods used by other countries to produce their own exports in the same or other sectors. An example for Ethiopia in the textiles and apparel sector is leather, which accounted for around 4 percent of total exports in 2018. The degree of processing involved in leather exports is relatively low, at least compared to what is seen in other suppliers of textile inputs, which are more capital- abundant and therefore focus on products like artificial fibers and highly processed fabrics. The most labor- intensive parts of the apparel value chain are linked to operations like assembly of finished garments, which imply high levels of inputs sourcing from abroad. The data confirms that Ethiopian manufacturing firms are engaging in less global sourcing of intermediates than are firms in the comparator countries. Service exports also show strong dominance of forward linkages. This dominance is less unusual in services than in manufacturing, however, as it is also the case in some of the other comparators. Still, backward linkages in Ethiopia are much weaker than in any other comparator countries except for Bangladesh (Figure 62). Financial intermediation is the only sector in which backward linkages seem at part with other peers. 97 Ethiopia Country Economic Memorandum Figure 62. Among comparator countries, backward linkages play a relatively more important role for exports in Vietnam and Tanzania GVC integration in commercial services, Ethiopia and comparators, percent of gross exports, 2015 Source: Authors’ calculations based on EORA data. Services are extensively used as intermediate inputs and have an important role to play in driving economic transformation. They are needed to produce other goods and services in the economy. As such, there is evidence of productivity spillovers from local services to manufacturers (Hoekman and Shepherd, 2017). From a value chain standpoint, the use of services inputs to produce goods gives rise to embodied services trade, which is in fact how a substantial proportion of total services trade takes place in the world economy. While data on regulatory policies in services sectors in Africa are generally not available, it is likely that there is scope for productivity-enhancing reforms in key sectors, particularly "backbone" sectors that affect the economy's ability to connect to the rest of the world. Examples of such sectors include transport, logistics, financial services, and telecommunications. Dihel and Grover (2016) show that there is substantial potential for services trade in Africa. However, Ethiopia has had limited success in embedding services into manufactured products to increase their value added. Another way of looking at GVC integration in services is to take the origin perspective of Johnson and Noguera (2012), which assesses the proportion of gross export value in manufacturing that is made up of inputs sourced from the services sectors, distinguishing between domestic and foreign sourcing. On average across the manufacturing sectors considered, Ethiopia has both the lowest proportion of embodied services value added (30.2 percent, compared with 43.1 percent in Rwanda and 42.2 98 Ethiopia Country Economic Memorandum percent in Vietnam) and the widest variability across sectors (Figure 63). While Ethiopia does better than some comparators in embedding international services in exports, the use of domestic services is minimal. Focusing on the apparel sector as the main locus of GVC activity in Ethiopia, embodied services value added is only 18.3 percent, compared with levels of over 45 percent for most of the other comparators and around 37 percent for Tanzania and Uganda. This comparison shows that services sourcing is more limited in Ethiopia than in the comparators and could be one factor holding the country back from further GVC integration. Figure 63. Ethiopia has the lowest proportion of embodied services value added among comparators, with wide variability across sectors Embodied services value added as a percentage of gross exports, by sector, Ethiopia and comparators, 2015 Share of service value in exports Ethiopia's share of service value in exports, by (average) sector Vietnam Oil, Chemicals &… Uganda Wood & Paper Transport… Tanzania Metal products Rwanda Other… Cambodia Electrical &… Ethiopia Food & Beverages Bangladesh Apparel 0.0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 Domestic Foreign Domestic Foreign Source: Authors’ calculations based on EORA data. Legal and trade policy restrictions are likely behind the weak backward linkage performance and the limited services integration in manufacturing. In manufacturing, a priority sector, the limited ability to source inputs from abroad likely has to do with trade policy and trade facilitation challenges, as discussed in the next section. In the case of services, there internationally comparable data to assess the extent of the restrictions in Ethiopia. Nevertheless, it is likely that the observed challenges are due to pervasive restrictions to foreign participation and even private sector activity that have been cramping most service sectors for long. 4.5 Trade policy in Ethiopia and reform prospects 4.5.1 Ethiopia’s current policies for trading goods Ethiopia has traditionally had significantly higher tariffs than comparator countries. Ethiopia’s tariffs were the highest of any of the countries in the sample for consumer goods, food and beverages, and industrial supplies (Figure 64). They are the second highest (after Cambodia) for transport equipment and capital goods. Imposing relatively high tariffs on consumer goods is common for developing countries, as the perception is that it is effectively a consumption tax on relatively well-off consumers, since poorer people tend to consume locally made goods. 99 Ethiopia Country Economic Memorandum Figure 64. Ethiopia stands out as having higher tariffs than comparators in most categories Average effectively applied tariffs, latest available year 40 35 30 Percent ad valorem 25 20 15 10 5 0 Capital goods Consumer Food and Fuels and Goods nes Industrial Transport goods nes beverages lubricants supplies nes equipment and parts Bangladesh Cambodia Ethiopia Rwanda Tanzania Uganda Vietnam Source: World Integrated Trade Solution UNCTAD-TRAINS (Trade Analysis and Information System). High tariffs on capital goods and industrial supplies hinder value addition and technology adoption. These measures make it more difficult for local firms to access foreign technology embodied in machinery and other investment goods as well as to access high-quality intermediate inputs.30 Taking Vietnam as a case in point, there has been a clear policy decision to keep tariffs low on capital goods and industrial inputs exactly because it helps domestic industries be competitive with their output in world markets. High duties and tariffs also have an impact on trade facilitation: they cause more friction in border clearance as there are more discrepancies in valuation of goods as invoice prices are often rejected due to undervaluation, which leads to cascading delays. Goods stay longer at clearance points, and importers take time to settle valuation concerns and arrange additional financing to pay additional duties and taxes. Incentives for the purchase of inputs and capital goods suffer from limitations in their implementation. The authorities have adopted a duty drawback scheme, along with exemptions from import duties for eligible foreign investors.31 More recently, incentives have also been expanded to other firms, although the conditions of access remain relatively restrictive as they lock in a particular input-output relationship for each producer. In principle, the drawback scheme allows exporters to recoup the duty they pay on imported intermediates. The incentive scheme provides an exemption from duties for capital goods and eligible intermediates and may explain the relatively strong performance of the apparel sector, where inward investment is relatively common. However, the drawback system suffers from important limitations. First, it requires the importer to pay the duty up front and later recoup it, which poses a financial burden on smaller firms in particular. Second, analysis based on experiences suggests that exporters find it difficult to use, so its effectiveness in practice is limited (Gebreeyesus and Demile, 2017). This, coupled with foreign exchange scarcity, likely explains the stagnation in both the imports of capital goods and industrial supplies (see Figure 56 earlier) and the exports of transformed goods. 30 Amiti and Konings (2007) show that liberalizing input markets tends to provide a productivity boost to local firms. 31 In broad terms, duty drawback refers to a range of administrative schemes that enable exporters to claim back duty paid on imported intermediates. For details, refer to the Ethiopia Investment Commission website. http://www.investethiopia.gov.et/index.php/investment-process/incentive-package.html. 100 Ethiopia Country Economic Memorandum The 2021 tariff reform by the Ethiopian authorities has increased the number of duty free lines, although the average tariff still remains relatively high by international standards. As of writing in 2022, the government has implemented a new tariff structure that seeks to simplify the previous approach and deal with some of the problems identified here. TRAINS tariffs for the latest year, reported using the Harmonized System classification, have a mean of 16.4 percent and a standard deviation of 11.3 percent, ranging between 0 percent and 35 percent. Of the reported lines, 3.5 percent were tariff free. By contrast, the new tariff, also reported using the Harmonized System classification, has a mean of 14.8 percent and a standard deviation of 12.8 percent, ranging again between 0 percent and 35 percent. Of the reported lines, 21.2 percent are duty free. Thus, the new tariff introduces a significant amount of liberalization, including through the extension of duty- free treatment to a much wider variety of products. The movement in terms of the newly adopted tariff therefore appears to be in the right direction relative to the dynamics set out in this report. A challenge for tariff reforms is that Ethiopia remains reliant on trade taxes as a source of government revenue. In the comparator group, it has the second-highest proportion of trade taxes in total revenue (19 percent) after Bangladesh (25 percent). In contrast, Rwanda has only 4 percent, and the rest of the comparators around 10 percent, in all cases based on data from the World Development Indicators.32 Thus, further reforms, need to be very carefully considered so that revenue loss does not impinge on the achievement of important economic and social goals in other policy areas (e.g., health or education). Over the medium term, an important objective would be to shift to other bases of taxation, such as consumption and income, to replace revenue coming from trade-related taxes. Ethiopia’s many non-tariff measures also likely hamper its trade potential. Ethiopia has the second- highest number of NTMs in the sample, after Vietnam. The number of export measures is particularly significant, as is the number of technical barriers to trade (Figure 65). Notably, NTMs are different from tariffs in that they can have non-protectionist objectives, particularly in the case of Sanitary and Phytosanitary (SPS) measures and Technical Barriers to Trade (TBTs). Thus, the implication of the Figure is not necessarily that Ethiopia is relatively protectionist given its use of NTMs; for such a conclusion to be drawn, the intent behind individual measures as well as their effects would need to be examined. However, it is possible to imply that NTM-related trade costs, regardless of intent, are likely substantial in Ethiopia relative to countries like Bangladesh and Cambodia. 32 The World Development Indicators definition of trade taxes as follows: “T axes on international trade include import duties, export duties, profits of export or import monopolies, exchange profits, and exchange taxes. ” 101 Ethiopia Country Economic Memorandum Figure 65. Ethiopia features a large number of NTMs, in particular technical barriers to trade (TBTs) and export- related measures Number of non-tariff measures, Ethiopia and comparators 900 800 700 600 Number 500 400 300 200 100 0 Ethiopia Bangladesh Cambodia Vietnam Contingent Protection Export Measures Pre-Shipment Inspection Other Measures Price Controls Quantity Controls SPS TBT Source: TRAINS. World Integrated Trade Solution UNCTAD-TRAINS (Trade Analysis and Information System). Note: Base years are as follows: 2015 (Ethiopia), 2017 (Bangladesh), 2018 (Cambodia and Vietnam). ‘SPS’ refers to Sanitary and Phytosanitary Standards; ‘TBT’ to Technical Barriers to Trade. Limitations in trade facilitation in Ethiopia add to the challenges faced by traders. Like NTMs, poor trade facilitation drives a wedge between producer prices in the exporting country and consumer prices in the importing country that is like a trade cost (Anderson and Van Wincoop 2004). Among comparators, Ethiopia has the second-lowest average score in the OECD’s Trade Facilitation Indicators, the lowest being Uganda. Vietnam has the highest score in the group, although Ethiopia’s score is higher than Vietnam’s when it comes to internal coordination (among the various national border agencies). However, Ethiopia continues to lag on information availability, advance rulings, formalities, governance and impartiality, and cooperation with other countries’ border agencies (Figure 66). Some of them are linked to direct increases in trade costs—for instance, additional border formalities or a lack of automation. Others create uncertainty and thereby increase the indirect costs of traders—for instance, limited availability of advance rulings on customs classification and other issues. 102 Ethiopia Country Economic Memorandum Figure 66. Ethiopia lags behind in trade facilitation dimensions pertaining to documentation, automation, procedures, and advance rulings Trade facilitation performance by pillar, 2019 Information availability 2.0 Involvement of Governance trade community 1.5 External 1.0 Advance rulings cooperation 0.5 0.0 Internal Appeal procedures cooperation Procedures Fees and charges Automation Documents Ethiopia Vietnam Source: OECD Trade Facilitation Indicators. Note: the eleven Trade Facilitation Indicators take values from 0 to 2, where 2 designates the best performance that can be achieved. Evidence indicates that improving the trade facilitation environment can help boost both overall trade performance and specifically GVC integration. As Shepherd (2021) shows, poor trade facilitation performance can act as a drag on GVC integration. Combined with Ethiopia’s pattern of tariff protection, the net result is a set of trade policies that are relatively unfavorable to global input sourcing, subject to the exceptions noted above; unlike tariffs, there are few workarounds for companies when it comes to NTMs and trade facilitation. Improving the trade facilitation environment in this context means taking steps to reduce the administrative burden linked to moving goods across borders, in the way set out in the World Trade Organization (WTO) Trade Facilitation Agreement (TFA). Ethiopia is not yet a WTO member, but the country can move in that direction unilaterally. The above analysis of trade policy has focused on goods, and similarly, it would be useful to consider policies that have negative impacts on trade in services. Unfortunately, there is no recent data available for Ethiopia to analyze this dimension. Preferential trade agreements can help Ethiopia reduce trade barriers and improve trade facilitation. According to the WTO’s Deep Trade Agreements (DTAs) database, the Common Market for Eastern and Southern Africa (COMESA)—the only trade agreement Ethiopia had signed at the time of its compilation— has fully legally enforceable provisions in areas such as industrial and agricultural tariffs, customs, export taxes, SPS measures, anti-dumping, and state aid, but not in other WTO disciplines such as TBTs, competition policy, countervailing measures, public procurement, investment, services, and intellectual property. Outside of issues covered by WTO disciplines, only information society, movement of capital, and investment receive strongly legally enforceable treatment in COMESA. All other areas—including environmental laws, health, consumer protection, data protection, and taxation—do not have provisions. Thus, COMESA is a relatively low-ambition agreement in terms of its structure and the areas in which it creates binding rules. In addition to considering WTO accession, implementation of the recently signed African Continental Free Trade Area (AfCFTA) and adhesion to other preferential trade agreements can help Ethiopia deepen rules and institutions to facilitate trade and reduce costs for exporters and importers. 103 Ethiopia Country Economic Memorandum 4.5.2 Trade policy reform scenarios Against the background presented in the previous section, this section discusses four trade policy reform scenarios for Ethiopia. It is useful to think about trade policy reforms that could potentially be beneficial to the country moving forward. “Beneficial” means promoting trade and GVC integration, but not at the price of removing a substantial portion of the government’s revenue base, which is needed for social programs in other areas. This section uses a quantitative trade model to simulate the impacts of two possible reform scenarios.33 The first reform scenario is a shift to an East African Community (EAC)-style tariff band system, as proxied by Kenya’s tariffs. The objective of such a system is to group commodities into broad categories so that intermediate and capital goods can be duty-free or low-rated, while consumer goods and luxuries can be taxed at higher rates. A key component of the first scenario is to change the overall structure of Ethiopia’s tariffs. The starting point is a simple average of 19.8 percent and a standard deviation of 6.8 percent. The counterfactual based on Kenya’s tariffs has a simple average of 14.0 percent and a standard deviation of 6.4 percent. The scenario therefore involves some degree of liberalization but one that is referenced on what is seen in a regional partner. The rationale for considering such a simulation is twofold. First, the band tariff, as in Kenya, is the system used by several neighboring countries and partners. Second, for policy purposes, it emphasizes the potential effects of altering the structure and composition of Ethiopia’s tariff schedule without necessarily lowering rates in an aggressive way, which would be challenging given the country’s revenue reliance on trade taxes. While there is a clear case for lowering tariffs over time, in particular in the context of ongoing talks on WTO accession as well as liberalization through regional agreements like that African Continental Free Trade Agreement, it will be important for the government to move forward in tandem with complementary reforms designed to broaden the revenue base to include income and consumption in more comprehensive ways. The second scenario considered here focuses on NTMs, specifically trade facilitation. Drawing from the OECD Trade Facilitation Index introduced in the previous section, the model looks at the economic impacts of improving Ethiopia’s performance to the level currently seen in Vietnam.34 The economic mechanism generating changes in trade and income is the same in both cases. Changing tariffs or improving trade facilitation changes ad valorem equivalent trade costs. Imports become less expensive relative to domestic goods, so producers are able to source some additional inputs from abroad. As a result, the cost of an input bundle declines, and exports become more competitive. Lowering bilateral trade costs through lower tariffs or improved trade facilitation can therefore be expected to boost both exports and imports (since the trade deficit is exogenous) as well as lead to an increase in GVC participation through changed patterns of input sourcing (see Annex 4.1 for a full description of the model used). The rationale for this scenario is that while improving trade facilitation performance involves upfront costs, it can often result in increased government revenues because it increases trade flows while keeping tariff rates constant. The third reform considers imposing a tariff like the one used by Vietnam. Figure 64 shows that Vietnam typically imposes lower tariffs than Ethiopia, in some cases significantly so. The rationale for using Vietnam as a point of comparison is that it is a country that has seen rapid growth in per capita income, and which has 33 DTC’s Global Trade Model is fully described in Annex 4.1. For policy purposes, the key advantage of this model is that in addition to the standard range of impacts produced by general equilibrium models, it also shows the changes in GVC integration that would be associated with particular policy changes, using the same backward and forward linkage metrics as in section 4.4. The baseline year is 2015, the latest year for which Eora data are available. 34 Shepherd (2020) provides full details of the modeling steps in this process, in particular estimation of sectoral elasticities of trade flows with respect to trade facilitation performance. This section uses those estimates without change. 104 Ethiopia Country Economic Memorandum leveraged its external sector to support that process in part through joining GVCs. Looking at the economics of a Vietnam-style tariff could therefore inform discussions in Ethiopia on the prospects for future liberalization with similar goals in mind. The fourth reform looks at the effects of moving to Ethiopia’s new tariff structure relative to the 2015 baseline. As noted above, the new tariff includes some important moves forward on liberalization, so it is important to assess the impacts of those reforms, and to compare them to other possibilities that could be considered. The exercise is informative in so far as the baseline year for the model is 2015, when the previous tariff was in force. The simulation therefore asks the following question: “what would Ethiopia’s 2015 economy have looked like if it had imposed 2022 tariffs but all other factors had remained constant?” It therefore provides information in a narrow sense on the economic effects of the recent reform, and in a comparative sense on the relative gains and losses involved in moving to some other scenario. All reforms result in some trade gains, albeit modest. Table 10 summarizes simulation results for the two scenarios. The first scenario—a shift to Kenya’s tariffs, as an example of the East African Community (EAC) band system—produces modest trade gains, less than 1 percent changes in exports and imports. Real income falls very slightly because of a loss in tariff revenue, but the change is small. The same dynamic is at play in Scenario 3, as the degree of liberalization involved in moving to a Vietnam style tariff is substantial; however, the trade gains are more significant in Scenario 3 than in Scenario 1. The reason for this result is that the EAC band system has a relatively high top rate, as well as extensive exceptions for sensitive items that are taxed at even higher rates. Scenario 2 produces smaller changes in trade flows but an increase in real income through an increase in tariff revenues: improved trade facilitation reduces the costs of crossing borders, which leads to greater imports and greater exports. Finally, Scenario 4 suggests that the government has skillfully redesigned the tariff in recent times: although the boost to exports is not as large as under the Kenya or Vietnam tariff scenarios, the change in real income is very slightly positive. The reason is that the sectoral pattern of liberalization under the observed shift in tariffs is more in line with maintaining tariff revenue, which falls by less than in the other scenarios. Taking all results together, the simulations are supportive of the government’s new tariff reform relative to other feasible options, but they also highlight the importance of moving forward on non-tariff measures in order to reap larger trade and real income gains. Table 10: Simulation results, key variables, percent change over baseline. Exports Imports Real Income Scenario 1 (Kenya) 0.77 0.43 -0.07 Scenario 2 (Trade facilitation) 0.58 0.32 0.1 Scenario 3 (Vietnam) 1.48 0.83 -0.19 Scenario 4 (Ethiopia 2021 reform) 0.51 0.28 0.01 Source: Authors’ calculations. In addition, all four scenarios result in improvements in GVC participation. The evidence is clear that either shifting to an EAC band tariff, moving towards a Vietnam style tariff, adopting the 2022 tariff reform, or improving trade facilitation to the level currently observed in Vietnam are effective ways of increasing backward GVC linkages in Ethiopia (Figure 67). Impacts are largest in scenario 3, since the degree of effective liberalization involved in moving to a Vietnam style tariff is largest. The analysis is sectoral, focusing on manufacturing only. The changes are modest in an absolute sense, although some sectors stand out: electrical and machinery would see backward linkages increase from 56.0 percent of gross exports in the baseline to 60.9 percent in Scenario 3. Similarly, wood and paper would increase from 18.5 percent to 20.5 percent, and transport equipment would increase from 23.8 percent to 25.6 percent. Interestingly, changes in textiles and 105 Ethiopia Country Economic Memorandum clothing are relatively small relative to the sector’s prominence in discussions regarding industrialization in Ethiopia. Figure 67: All four scenarios would help increase backward GVC linkages in Ethiopia GVC backward integration, all scenarios versus baseline, percent of gross exports Source: Authors’ calculations. These reforms could be a starting point, but more ambitious reforms would be needed to give GVC integration a major boost. The case of apparel stands out as one where the increase in backward linkages in both scenarios is relatively limited, especially since the sector is starting from a low baseline relative to other sectors and relative to comparators. Thus, while the simulations suggest that trade policy changes can help deepen GVC integration, it is important to recognize the limits of relatively minor changes in trade policy, especially considering that Ethiopia still has relatively restrictive policies by international standards. Thus, more ambitious reforms may be necessary to improve GVC integration significantly. A secondary issue is the effective treatment applied to intermediates in the apparel sector under an EAC-style band system: care would need to be taken at the design stage to make sure that inputs are appropriately covered by the zero tariff band in order to increase GVC integration significantly in the sector. 106 Ethiopia Country Economic Memorandum 4.6 Conclusion and policy implications Ethiopia’s trade performance has deteriorated over the past decade, and backward linkages are weak. This chapter has examined Ethiopia’s trade performance and prospects from the perspective of comparative advantage, GVC integration, and the potential for policy reforms affecting tariffs and NTMs to promote greater integration over time. Ethiopia has seen important successes in developing its trade integration over time, but recent years have seen that progress stall. There is little evidence of ongoing export diversification either on the product or market dimension, exports remain concentrated on primary and low-technology products, and policies place significant burdens on acquisitions of foreign capital goods and intermediate inputs through the difficulty of accessing drawback systems. As a result, Ethiopia participates in GVCs mainly through forward linkages, with few backward linkages. Simulation evidence shows that tariff reforms and improvements in trade facilitation could go some way toward improving GVC integration. According to the international experience, developing a competitive manufacturing sector is difficult without sourcing high-quality inputs at world market prices. Simulation evidence shows that trade policy can influence the key outcomes examined here, in particular GVC integration. However, there are no easy answers. Recent reforms go in the right direction in this regard, but the simulations suggest that more ambitious tariff changes would have larger impacts on GVC participation. Similarly, improving trade facilitation performance to the level observed in Vietnam would have similar effects: improvement, but not radical change. On the other hand, large-scale liberalization would have significant implications for the government’s ability to raise revenue for social program so needs to be done in tandem with broader policy reforms designed to broaden the revenue base. Thus, to improve its performance and trade integration, it is important that Ethiopia acts through a combination of reforms in tariffs and trade facilitation (including non-tariff measures) and addresses the issue of the exchange rate (as discussed in the previous chapter). Tariff reductions can help boost exports and GVC integration and would need to be accompanied by alternative revenue-generation measures. The Ethiopian authorities have already taken steps in this direction. But an international comparison shows that further liberalization could be beneficial, based on the experience of a country like Vietnam. While there has been an increase in the number of tariff lines with zero duty rates, broad based reductions in applied tariffs on capital goods and intermediates to zero (or directly exempting them) would mean that producers would not need to go through the cumbersome duty drawback scheme and the application for incentives. In addition, in the context of the AfCFTA, Ethiopia is required to conduct substantial tariff liberalization with other participating countries over the next ten years. Securing alternative revenue bases through taxes on income, consumption, and property can be considered, to compensate for the expected decline in trade tax collection in the medium and long term. Authorities would need to conduct a thorough reassessment and streamlining of NTMs. Undertaking a detailed regulatory review of NTMs contained in the OECD TRAINS database could help identify areas for streamlining or improvement. A particular area for attention is the set of export measures, as the comparator countries typically do not resort to these kinds of measures with the same frequency. There are also numerous TBTs, but many of them likely serve useful social purposes and may need to be maintained in some form. The introduction of Good Regulatory Practice35 in relation to technical regulations and standards (e.g. phytosanitary measures) would help align Ethiopia’s TBTs with international practice, thereby potentially reducing the fixed costs of market entry for local producers and supporting export diversification (Shepherd 2015). 35 Good Regulatory Practice refers to cooperation and exchange of information to improve quality in the formulation, implementation, and impact assessment of regulations. See the APEC-OECD Integrated Checklist on Regulatory Reform. 107 Ethiopia Country Economic Memorandum Aside from streamlining, a National Quality Policy and a series of monitoring arrangements will be needed to transform NTMs from restrictive measures to facilitative ones, to foster productive upgrading over time. Reforming NTMs and particular TBTs and SPS measures is not straightforward. An efficient system of TBTs and SPS measures could be embodied in a robust National Quality Policy which covers all elements of the necessary infrastructure, including metrology and conformity assessment for NTMs, in addition to standards and regulations. Attention will need to be paid to enforcement and monitoring: it is not enough to commit to rationalization of NTMs on paper, and there needs to be a mechanism by which firms can flag measures that impede market access, including the export measures mentioned above. Neighboring countries in East Africa have experience with such a system through the EAC, and developments on this front are also likely in the context of AfCFTA. Ethiopia therefore needs to take full stock of these processes and work to include similar approaches focused on its own firms. Trade facilitation measures can also help boost trade. Ethiopia has substantially reformed its trade facilitation system over the past few years, improving from a low base, but further improvements can help consolidate gains. While Ethiopia is not a member of the WTO, a good starting point would be to review and benchmark against the WTO’s TFA to craft a clear trade facilitation improvement action plan. Some areas for action based on the challenges identified (Figure 66) include the following: • Automation reform. Fully automating Ethiopian agencies participating in trade will be critical to reduce time and cost to trade, improve governance, and reduce revenue losses due to leakages. In particular, it is essential to complete Phase II under the electronic Single Window, covering all trade agencies and reaching an operational service level agreement among regulatory agencies to ensure standardized service delivery and accountability. • Streamlining procedures and implementing risk management. Numerous NTMs have become non-tariff barriers as their implementation is burdensome, costly, and time-consuming, with duplicate and onerous inspections and the need for copious documentation. For example, agencies such as the Ministry of Trade and Industry require close to 100 percent inspection and verification of compliance with National Mandatory Standards through laboratory testing in Ethiopia. This is not only against best practice in the field but is also a source of discontent for traders. While automation will speed up processes, reviewing and removing inefficient and unnecessary procedures and inspecting less by implementing good practice risk management strategies would solidify the gains further. • Improving joint cooperation. Ethiopia shall aim to complete ongoing initiatives that are aimed at enhancing trade and facilitating cooperation with its neighbors (e.g., Djibouti, Kenya) through the establishment of one-stop border posts, use of common documents, sharing of information and data, and cooperation in terms of inspections and regulations. A Trade Information Portal can help improve information availability and also identify areas for further trade facilitation reforms. Authorities can also work with international partners to identify other areas for improvement in trade facilitation, which in any event is a policy area that will need to be considered as part of WTO accession. Significant upfront investments may be required to improve border procedures in this way, hence the need to work collaboratively with donors and international agencies. A useful first step would be to work with development partners to build on the existing Customs website with a view toward developing a full Trade Information Portal, as a way of improving performance in terms of information availability. More generally, the National Trade Facilitation Committee can be positioned to play a lead role in identifying and prioritizing future improvements. Lifting regulatory restrictions to trade in services is expected to help boost GVC integration and value added of exported goods, as well. So far, Ethiopia’s trade in services has been limited to transport (Ethiopian 108 Ethiopia Country Economic Memorandum Airlines) and travel, and the proportion of embodied services value added in merchandise exports is the lowest among comparator countries. Improving the competitiveness of services sectors, including through appropriate regulatory reforms, can help boost productivity and exports throughout the economy. A first key step is to enforce implementation of the new Investment Proclamation by allowing private sector (until now left out) entry into all sectors not included in the negative list. Moreover, authorities shall consider further reducing that negative list in the medium term—in particular, by removing the prohibition for foreign investors to participate in finance, a sector that is underdeveloped compared to other countries and could provide significant value addition. The authorities can also work with international partners and the private sector to assess and reform other pieces of regulation that may be constraining trade in services. In logistics, in particular, there is scope for further reform. Foreign participation in some critical logistics services such as freight forwarding and shipping agency services, warehousing, cargo consolidation, and packaging is limited to 49 percent. This is of concern for some international logistics operators that cannot have greater control on their subsidiaries in Ethiopia. In addition, transport services operations have been fully opened only for trucks of a minimum of 25 metric tons, while they should be open for smaller trucks as well. Authorities need to ensure that red tape does not harm the entry of competitors that have become open de jure, such as the dry ports, which can have a significant impact in reducing logistics costs. In addition to further improving transit through the Addis-Djibouti corridor by interconnecting the Ethiopia and Djibouti National Single Windows, development of new corridors could help improve resilience and diversification. To fully operationalize corridors, there is a need to design functional transit regimes, standardize technical aspects of transport services (especially axle loads), and harmonize rules for market access. Moving forward in areas such as tariff reforms, policies affecting services trade, trade facilitation, and trade of environmental goods requires a coherent national strategy on trade integration and well- defined lines of responsibility for government agencies. Many government departments are involved in administering different parts of trade policy, so it is important for senior leadership to adopt a high-level view of country priorities, including by factoring in AfCFTA implementation and WTO accession. Ethiopia has the opportunity to fast-track WTO accession with the aim of joining before 2025 to enhance the reform program and provide a clear signal to domestic and international investors of a legal commitment to a transparent and predictable trade regime. The WTO is also the premium forum in which global rules on climate change and trade are negotiated and implemented. It can help Ethiopia cope with climate impacts by allowing further export diversification, and it can help influence the policy discussions around trade of environmental goods and services (Box 7). Box 7. The climate change and trade nexus: implications for Ethiopia Ethiopia is one of the most vulnerable countries to climate variability and climate change . This is due to its heavy reliance on rain-fed agriculture and other primary commodities and natural resources and its limited capacity to deal with the potential adverse impacts. Ethiopia’s challenges include ecosystem fragility, underdevelopment of water resources, weak health services, a high population growth rate, low economic development, inadequate road infrastructure in drought-prone areas, weak institutional structure, and lack of climate risk awareness and preparedness. The nexus of climate change and trade is particularly relevant for Ethiopia for several reasons. First, climate change poses significant social, political, and economic risks to Ethiopia since the country is an agricultural economy, with most of its population employed in agriculture. Notably, climate change will change comparative advantages and market dynamics in the agriculture sector to the detriment of low- and middle- income countries and major food-importing countries, increase the costs of production and worsen the country’s terms of trade and balance of payments, increase unemployment, exacerbate the risk of armed conflict and fragility. Second, Ethiopia is one of the fastest-growing economies in Africa and the world, with a nascent 109 Ethiopia Country Economic Memorandum industrial sector experiencing rapid growth. While Ethiopia is already energy self-sufficient and generation is almost entirely coming from renewables, there are opportunities for greening transportation as well as the nascent industry. This is particularly important to reach markets in the European Union, since access is increasingly tied to environmental requirements. The impact of climate change on trade Agriculture productivity will be affected. The World Bank estimates that Ethiopia will lose more than 6 percent of each year’s agricultural output if the current decline in average annual rainfall levels for primary agricultural zones continues to mid-century. Rising temperatures and shifting rainfall patterns may increase soil erosion, make it harder to grow many crops, and shorten the growing season. The livestock sector, a key economic sector for Ethiopia, is also likely to be affected by extreme weather conditions (e.g. cattle suffering from heat stress). Downstream industries such as food processing and leather manufacturing will also be affected, as the availability of primary inputs is reduced and their prices rise. The balance of payments of Ethiopia and the welfare of its people are also likely to be negatively affected, as the country remains a significant food importer and is also dependent on food aid. Remote and marginalized communities are likely to be the most affected by changes in future trade patterns due to climate change. These communities rely predominantly on pastoral and agropastoral livelihoods and are exposed to fragile and conflict affected neighboring states. They face as well limited access to public services and infrastructure, as well as low levels of literacy and formal education. Transport is likely to be affected substantially by extreme weather events. While this sector contributes heavily to greenhouse gas emissions, it also plays a crucial role in supporting Ethiopia’s overall development strategy and supports its export and import networks. Extreme whether events may lead to the destruction of infrastructure and the disruption of transport corridors. In addition, Ethiopia is landlocked, and the port of Djibouti is the lifeline of Ethiopia’s international trade, but it is exposed to the impacts of climate change. Sea- level rise is projected to lead to the loss of a sizable proportion of the northern and eastern coastlines due to a combination of inundation and erosion, with consequential risks for port infrastructure. Trade policy options to mitigating and adapting to the impact of climate change Fast-tracking Ethiopia’s accession to the World Trade Organization to support climate adaptation. The WTO is the premium forum in which global rules on climate change and trade are negotiated and implemented. Additionally, it is the most viable global forum for the enforcement of international environmental law. Ethiopia’s WTO membership is essential for many reasons. First, it would enhance the policy framework for country to diversify the goods and markets of trade, which is imperative for coping with climate shocks on trade (exports particularly), and would provide a forum for reaping the benefits of special and differential treatment to which Ethiopia is entitled as a low-income country. The WTO provides a forum for eliminating trade measures and barriers, such as agricultural subsidies and tariff escalation, in the country’s main export markets. Second, WTO membership also would give Ethiopia a seat at the table where the global rules for environmental goods and services are being negotiated—giving Ethiopia a feasible avenue for securing its global interests in environmental goods and services and for facilitating the green transition domestically. Enhance participation in subregional and regional trade organizations and agreements to which Ethiopia is a party. The AfCFTA provides the greatest prospects for broadening or refocusing the spectrum of export markets as well as for expanding exports into new markets where value added production is in greatest demand. Ethiopia’s main trade partners are currently non-African states, and AfCFTA presents huge opportunities for improving market access, diversification, improving the terms of trade, and industrialization as well as less stringent standards than Ethiopia’s main trade partners (the EU particularly). Even outside of 110 Ethiopia Country Economic Memorandum AfCFTA, the manufacturing content of Africa’s exports to Africa is much higher (46.3 percent) than that of its exports to the world (26.7 percent). Thus, even if primary production in the agriculture sector experiences climate-related declines, transitioning from exports of primary products to higher value-added downstream products can maintain sectoral employment, revenue levels, and the overall value of sectoral exports. For these reasons, enhanced subregional and regional trade offers significant opportunities. Closely monitor developments in the environment-climate requirements of its main trading partners. Monitoring would allow Ethiopia to anchor its own trade policy or trade legal framework strategically in accordance with that of its main trade partners (the European Union and the United States), which are increasingly interested in environmental sustainability. For example, the European Union has stated that the exports of low- and middle-income countries that implement a national carbon tax system or regulation that achieves similar objectives will be exempt from carbon border fees imposed. Ethiopia, with the support of international development agencies, can invest to building the capacity to measure and verify carbon emissions. As countries and corporations implement climate change measures the ability to identify and demonstrate “carbon competitiveness” will become increasingly important for exporting firms in Ethiopia. Source: this Box summarizes the findings of the Ethiopia case study included in the volume “The Trade and Climate Change Nexus” (Brenton and Chemutai 2021). The findings presented in this chapter add urgency to the need to correct the course in trade policy in order to support strong growth and structural transformation going forward. As already noted in World Bank (2014), sustained growth and development in Ethiopia needs a vibrant private sector. There is a clear need to boost competitiveness in manufacturing and commercial services, but this passes through the need to facilitate imports of goods and services, as has been the case in other countries that have successfully pursued outward-oriented development strategies. The comparison with Vietnam is informative: it has an active, development-oriented government but has also made strategic use of liberalization and trade facilitation to increase international integration and GVC participation, which also contributed to boosting manufacturing productivity. The tariff structure and trade facilitation provisions will be key in shaping the future of both domestic and foreign producers, as they affect access to inputs. And as much as Ethiopia is able to improve business climate indicators and investor protection, foreign investors will only settle production platforms in Ethiopia if they can be sure that goods can move in and out of the country in a timely manner and that they can have access to and repatriate foreign exchange. Key policy options discussed in this chapter are summarized in Table 11. 111 Ethiopia Country Economic Memorandum Table 11. Policy recommendations to reviving trade in Ethiopia Area Short term Medium term Implementing agency Tariffs Introduce additional tariff changes, Implement the tariff reductions required Ethiopian Customs possibly reaching the levels of within the AfCFTA agreement Commission Vietnam, to increase GVC participation Zero rate tariffs on capital goods and intermediate inputs, rendering the current duty drawback scheme unnecessary Non-tariff Conduct a comprehensive review Adopt a National Quality Policy and a Ministry of Trade and measures of NTMs, particularly export series of monitoring arrangements to Regional Integration, measures foster productive upgrading of NTMs Ministry of Agriculture, over time Ministry of Health, Ministry Align TBTs (such as phytosanitary of Industry requirements) with international standards, to reduce fixed costs Develop a mechanism by which firms can flag measures that impede market access Trade Complete phase II under the eSW Use the WTO TFA as a guideline for Ministry of Trade and facilitation and reach operational service developing ambitious trade facilitation Regional Integration, agreements reforms Ethiopian Customs Commission Reduce the number of inspections, Cooperate with Djibouti and Kenya to including by implementing a risk- establish one-stop border posts management approach Develop a Trade Information Portal Logistics Complete the reform of the dry Allow whole foreign ownership in Ministry of Transport and ports by reforming regulations and critical logistics subsectors currently Logistics, Ethiopian removing red tape for private capped at 49 percent Maritime Affairs Authority sector entry Trade in Enforce implementation of the Further reduce the negative list, Ethiopia Investment services new Investment Proclamation by including by removing the prohibition Commission allowing private sector entry into for foreign investors to participate in all sectors not included in the finance negative list Vision and Formulate a coherent national Fast track WTO accession with the aim Ministry of Trade and coordination strategy on trade integration and of joining before 2025, to facilitate Regional Integration, Prime well-defined lines of responsibility diversification Minister Office for government agencies Participate in discussions around trade in environmental goods and services at the WTO Source: Elaborated by the authors. 112 Ethiopia Country Economic Memorandum 4.7 References Aichele, R., and I. Heiland. 2018. “Where is the Value Added? Trade Liberalization and Production Networks.” Journal of International Economics, 115(C): 130-144. Amiti, M., and J. Konings. 2007. “Trade Liberalization, Intermediate Inputs, and Productivity: Evidence from Indonesia.” American Economic Review, 97(5): 1611-1638. Anderson, J., and E. Van Wincoop. 2003. “Gravity with Gravitas: A Solution to the Border Puzzle.” American Economic Review, 93(1): 170-192. Anderson, J., and E. Van Wincoop. 2004. “Trade Costs.” Journal of Economic Literature, 42(3): 691-751. Baldwin, R. 2011. “Trade and Industrialization after Globalization’s Second Unbundling: How Building and Joining a Supply Chain are Different and Why it Matter.” NBER Working Paper 17716. Borin, A., and M. Mancini. 2019. “Measuring What Matters in Global Value Chains and Value Added Trade.” Policy Research Working Paper No. 8804, World Bank. Brenton, P. and V. Chemutai. 2021. The Trade and Climate Change Nexus: The Urgency and Opportunities for Developing Countries. Washington, DC: World Bank. Caliendo, L., and F. Parro. 2015. “Estimates of the Trade and Welfare Effects of NAFTA.” Review of Economic Studies, 82(1): 1-44. Chauffour, J.P. and M. Gobezie. 2019. “Exiting Financial Repression: The Case of Ethiopia.” Policy Research Working Paper No. 9082. Washington, DC: World Bank. Chor, D. 2010. “Unpacking Sources of Comparative Advantage: A Quantitative Approach.” Journal of International Economics, 82(2): 152-167. Correia, S., P. Guimaraes, and T. Zylkin. 2019. “ppmlhdfe: Fast Poisson Estimation with High-Dimensional Fixed Effects.” Working Paper, https://arxiv.org/abs/1903.01690. Costinot, A., D. Donaldson, and I. Komunjer. 2012. “What Goods do Countries Trade? A Quantitative Exploration of Ricardo’s Ideas.” Review of Economic Studies, 79: 581-608. Dekle, R., J. Eaton, and S. Kortum. 2007. “Unbalanced Trade.” American Economic Review, 97(2): 351-355. Diao, X., M. Ellis, M. McMillan, and D. Rodrik. 2021. “Africa’s Manufacturing Puzzle: Evidence from Tanzanian and Ethiopian Firms.” Working Paper No. 28344, NBER. Dihel, N., and A. Grover. 2016. “The Unexplored Potential of Trade in Services in Africa: From Hair Stylists and Teachers to Doctors and Accountants.” Washington, D.C.: World Bank. Eaton, J., and S. Kortum. 2002. “Technology, Geography, and Trade.” Econometrica, 70(5): 1741-1779. Egger, P., M. Larch. S. Nigai, and Y. Yotov. 2018. “Trade Costs in the Global Economy: Measurement, Aggregation, and Decomposition.” Working Paper, WTO. French, S. 2017. “Revealed Comparative Advantage: What is it Good For?” Journal of International Economics, 106(C): 83-103. Gebreyesus, M. and A. Demile. 2017. "Why export promotion efforts failed to deliver? Assessment of the export incentives and their implementation in Ethiopia," Working Papers 017, Ethiopian Development Research Institute. 113 Ethiopia Country Economic Memorandum Hoekman, B., and B. Shepherd. 2017. "Services Productivity, Trade Policy, and Manufacturing Exports." World Economy, 40(3): 499-516. Johnson, R., and G. Noguera. 2012. “Accounting for Intermediates: Production Sharing and Trade in Value Added.” Journal of International Economics, 86(2): 224-236. Koopman, R., Z. Wang, and S.-J. Wei. 2014. “Tracing Value-Added and Double Counting in Gross Exports.” American Economic Review, 104(2): 459-494. Santos Silva, J., and S. Tenreyro. 2006. “The Log of Gravity.” Review of Economics and Statistics, 88(4): 641-658. Shepherd, B. 2015. “Product Standards and Export Diversification.” Journal of Economic Integration, 30(2): 300- 333. Shepherd, B. 2020. “Modeling Global Value Chains: From Trade Costs to Policy Impacts.” Working Paper. Shepherd, B. 2021. “Modeling Global Value Chains: From Trade Costs to Policy Impacts.” Working Paper DTC-2021-1, https://developing-trade.com/publications/modeling-global-value-chains-from-trade- costs-to-policy-impacts/ Van der Marel, E. 2011. “Determinants of Comparative Advantage in Services.” Working Paper 38933, LSE. Wang, Z., S.-J. Wei, and K. Zhu. 2013 (Revised 2018). “Quantifying International Production Sharing at the Bilateral and Sector Levels.” Working Paper No. 19677, NBER. World Bank. 2014. 3rd Ethiopia Economic Update: Strengthening Export Performance through Improved Competitiveness. http://documents1.worldbank.org/curated/en/560521468028152818/pdf/895480REVISED00c0U pdate0v0final0web.pdf. World Bank. 2020. World Development Report 2020: Trading For Development in an Age of Global Value Chains Washington, D.C.: World Bank. Yotov, Y., R. Piermartini, J. Monteiro, and M. Larch. 2016. An Advanced Guide to Trade Policy Analysis: The Structural Gravity Model. WTO, Geneva/UN, New York, https://doi.org/10.30875/abc0167e-en. 114 Ethiopia Country Economic Memorandum Part III: Sustainability 115 Ethiopia Country Economic Memorandum 5 Food prices and policy distortions 5.1 Motivation: what went off-track? High growth in Ethiopia has been accompanied by inflationary pressures, and food prices have been volatile. Since 2003, at the beginning of the growth take-off, inflation has only remained below 10 percent during five out of nineteen years. On this metric, the national development strategies have fallen short of the target of keeping the inflation rate in single digits. Food and cereal inflation have been particularly volatile, pushing up inflation in drought periods (Figure 68)36 Since 2014, a progressive buildup on inflation can be seen in both food and non-food items. While already high inflationary pressures did not significantly increase with the COVID-19 outbreak, inflation accelerated in the summer of 2021, as the armed conflict in Northern Ethiopia spread and intensified. At the time this report was being finalized in March 2022, inflation remained over 30 percent, fueled by international oil and wheat price increases resulting from the Ukraine war. This chapter focuses in understanding the short- and long-run drivers of inflation in Ethiopia, which can be helpful to explain recent trends, although it refrains from quantifying the impact of specific shocks on inflation. Figure 68. Ethiopia’s food and cereal inflation has been especially volatile Inflation trends in Ethiopia 180 160 Consumer price index change, year-on-year (%) 140 120 100 80 60 40 20 0 Jun-01 Aug-02 Jun-08 Aug-09 Jun-15 Aug-16 Jul-98 Feb-99 Dec-04 Jul-05 Feb-06 Dec-11 Feb-13 Dec-18 Feb-20 Mar-03 Oct-03 Mar-10 Oct-10 Jul-12 Mar-17 Oct-17 Jul-19 Apr-00 Jan-02 Apr-07 Nov-07 Jan-09 Apr-14 Nov-14 Jan-16 Apr-21 Sep-99 Nov-00 May-04 Sep-06 May-11 Sep-13 May-18 Sep-20 Nov-21 -20 -40 -60 All items CPI Inflation Food CPI Inflation Non-food CPI Inflation Cereals CPI Inflation Source: Central Statistical Agency (CSA). This chapter identifies the determinants of inflation in Ethiopia and the policy distortions that affect the agriculture sector. It first presents results from econometric analysis identifying the drivers of inflation over the past two decades, and that include monetary and fiscal factors, the passthrough of international prices, droughts and shocks to agriculture production, as well as inertia. The chapter then digs deeper into food prices 36 The droughts of 1997/98, 2000, 2002/03, 2008/09, and 2011 affected 5.4, 5.4, 10.5, 13, 5, and 4.5 million people, respectively (NBE’s records). 116 Ethiopia Country Economic Memorandum using a qualitative lens, and with a focus on other potential supply-side constraints not captured by the quantitative analysis. Some of the hypotheses assessed include whether agriculture production improvements have been sufficient to cope with rising demand and urbanization trends (inconclusive, may have become true in recent years), whether domestic markets and value chains have been inefficient and have contributed to inflationary pressures (unlikely), and whether trade policy distortions may be contributing to higher prices and slowing down sector modernization (very likely). 5.2 The long-run determinants of inflation in Ethiopia As Ethiopia moved toward pursuing a developmental state model, it reached new heights not only in growth but also in money creation and inflation. Ethiopia began transitioning toward a market economy with the launch of a structural adjustment program in 1992 with IMF support. During the next decade, base money growth was kept in single digits, and real interest rates were positive; growth rates in this period, however, were modest and relatively unstable. Around 2003 the authorities started to pursue a developmental state model that envisaged an Agriculture Development Led Industrialization (ADLI). Monetary policy became supportive of an expansionary fiscal policy, with money growth and inflation reaching double digits while real interest rates turned negative. While policies were able to boost economic growth over 10 percent, they also resulted in growing macroeconomic imbalances. Over the past decade, financial repression mechanisms and foreign exchange controls were tightened to fuel resources into massive public infrastructure projects, which boosted capital accumulation but also resulted in increasing current account deficits, foreign exchange shortages, and external debt vulnerabilities (Chauffour and Gobezie 2019; IMF 2020). Money growth and inflation seem to have followed a similar upward trend over the years, with peaks and valleys in the Consumer Price Index (CPI) resulting from droughts (2003, 2008, 2011) and other driving factors (Figure 69). Figure 69. Ethiopia’s inflation has been trending upward on par with money growth Ethiopia: trends in headline inflation and money growth 80 60 40 Percent 20 0 Jul-98 Feb-99 Aug-02 Feb-06 Feb-13 Feb-20 Jun-01 Mar-03 Oct-03 Aug-09 Aug-16 Mar-17 Dec-04 Jul-05 Jun-08 Mar-10 Oct-10 Dec-11 Jul-12 Jun-15 Oct-17 Dec-18 Jul-19 Apr-00 Jan-02 Apr-07 Jan-09 Apr-14 Jan-16 Sep-99 Nov-00 May-04 Sep-06 Nov-07 May-11 Sep-13 Nov-14 May-18 Sep-20 Apr-21 Nov-21 -20 -40 CPI Inflation M1 Growth Linear (CPI Inflation ) Linear (M1 Growth ) Source: National Planning Commission of Ethiopia. Studies on the drivers of inflation in Ethiopia have focused on shocks to agricultural production and imported inflation, as well as monetary factors. The high levels of inflation have attracted attention from policy makers as well as the research community due to their deleterious effects on household welfare (Ticci 2011; Alem and Söderbom 2012; Geiger and Goh 2012). Attention to inflation also arises from its perceived distortionary effects on private investment and savings, putting into question the sustainability of the high 117 Ethiopia Country Economic Memorandum growth regime. While the existing evidence overwhelmingly confirms the importance of shocks to agriculture as drivers of inflation, evidence on the effects of monetary factors on inflation remains mixed. In one of the most detailed and methodologically rigorous studies, Durevall et al. (2013) find that disequilibria in the monetary sector have no impact on food inflation in the long run. They do find, however, that money growth has a statistically significant positive impact in the short run, but only on non-food price inflation. As a background paper to this Country Economic Memorandum, an econometric study on the drivers of inflation in Ethiopia over the past two decades was conducted. The study by Ndikumana et al. (2021) explores long-run and short-run effects of both domestic and external factors, using the cointegration and error correction methodology (see Annex 5 for further details). Given the strong dependence of the Ethiopian economy on shock-prone agriculture and its exposure to external shocks, inflation is disaggregated into its various components, as each one may be driven by different factors as illustrated in earlier studies (Durevall et al 2013; Geiger and Goh 2012). The study therefore considers four measures of inflation: cereals, food, non- food, and overall CPI inflation. The study contributes novel insights to the existing literature in various ways. First, by drawing from monthly data for a relatively long period (from 1997/1998 to 2019/2020), the analysis captures the impacts of changes in policy frameworks, shifts in growth regimes, macroeconomic imbalances, and developments in the international economy, including episodes of commodity price shocks. Second, the study examines the role of the fiscal sector in modeling inflation in Ethiopia, which had been relatively overlooked in previous papers. This is motivated by evidence of pervasive financial repression and dominance of public sector borrowing in Ethiopia’s financial system (Chauffour and Gobezie 2019). Third, the study examines the impact of open economy factors on inflation in Ethiopia, notably international prices of grains, non-food items, and petroleum. Finally, the study incorporates foreign exchange scarcity into the analysis by including the evolution of the parallel market premium. The empirical results from this study confirm key findings from the existing literature while also unveiling new drivers of inflation, such as the role of fiscal policy. The results show that disequilibria in the monetary sector, fiscal sector, and cereal sector have long-run effects on cereal, food, and overall inflation. In the short run, inflation is driven by structural factors, notably output gaps and prices of imported goods (e.g., wheat, oil, non-food items). There seems to be some pass-through from the parallel market premium into inflation, as it affects import prices for items outside the priority list. On the demand side, in contrast with Durevall et al. 2013, money growth is found to affect both food and non-food inflation in the short run. Growth of public sector borrowing is also found to exert upward pressure on inflation. Inflation exhibits substantial persistence as illustrated by the statistical significance of its lagged values (see results tables in Annex 5). 5.3 Demand, supply, and market functioning factors potentially affecting food prices To complement the quantitative analysis of long-run inflation determinants, this section uses a more qualitative approach to dig deeper on supply and demand-side constraints in the agriculture sector. While the econometric analysis discussed above captures the effects of swings in domestic cereal production as well as in international prices, other factors that affect food prices could not be included due to insufficient data (e.g., relating to actual marketed shares of production, distribution and supply chain issues, presence of domestic subsidies). The remainder of this chapter does not aim to attribute specific amounts of inflationary pressure to individual causes but rather provides an overview of supply and demand conditions and market performance issues for key agriculture commodities. The chapter takes a long-term view and focuses on structural conditions and policies that give rise to food inflation rather than episodic events such as drought, 118 Ethiopia Country Economic Memorandum flood, or civil conflict that also affect food prices. The analysis draws from the existing literature, interviews with the Ministry of Agriculture and the Agriculture Transformation Agency (ATA), and a detailed assessment of the underlying competitiveness of maize and soybean value chains that involved a broad set of stakeholder interviews undertaken with the IFC in February 2020 before COVID-19 travel restrictions. Overall, the analysis shows that rapid urban growth and limited market participation by smallholder farmers may be contributing to disequilibria in food markets and inflationary pressures. Contrary to widespread perceptions, market inefficiencies relating to intermediation do not seem to be severe. Mark-ups between the farm and consumer levels are relatively modest, and a high share of the final retail price reaches the farm level. Price convergence across geographical areas has also been observed. Meanwhile, urban demand is rapidly growing, yet only around 18-20 percent of all cereals grown in Ethiopia are marketed as surplus, which is low compared to some other African countries. Thus, success in managing food inflation is likely to require efforts to improve the incentives for market participation rather than improve market efficiency and connectivity, which has come a long way already thanks to developments in infrastructure and information sharing. The following subsections discuss the food consumption basket of rural and urban households, the extent to which domestic production has held up with population growth, and whether domestic markets across geographical areas are efficient. 5.3.1 Understanding food consumption patterns in Ethiopian households While most rural households do engage in food production, it does not necessarily make the household food self-sufficient. In interpreting national consumption data, it should be noted that the HICES methodology on which the CPI is based imputes values of crops saved for own consumption as if they were bought in the nearby market. Notably, Worku et al. (2017) report that 41 percent of total food expenditures in rural Ethiopia is of the household’s own agriculture production. This is significant since households that grow their own food are not affected by price changes in the same way as consumers who depend on market purchases. The ways in which food prices affect rural households, however, are quite complex and depend, among other factors, on the size of the household’s food surplus or deficit as well the prices received by surplus producers, who often employ those who have food deficits (Aksoy and Isik-Dikmelik 2008).37 In this regard, price stabilization can ultimately be a more important policy objective from a producer’s perspective than the avoidance of high food price inflation per se. Compared to those in the poorer quintiles, those in the upper quintiles spend less on food in relative terms, while they consume a more varied basket of goods. As shown in Figure 70 (left panel), the share of food in total household consumption decreases as incomes rise, with food accounting for almost 60 percent of total consumption by the poorest quintiles in both rural and urban areas, compared to 52 and 42 percent of total consumption by the richest quintiles in rural and urban areas, respectively. Meanwhile, the composition of the food basket becomes somewhat more diverse as incomes rise (Figure 70, right panel). Nevertheless, dietary diversity in Ethiopia is limited, with staple grains and starchy vegetables and legumes accounting for more than half of total food spending by all income groups except the richest quintile. As incomes rise, consumption of animal products (including dairy) become relatively more important. Cereals and vegetables account for a somewhat greater share of total food consumption in urban areas than in rural locations. 37 It is quite common in places like Ethiopia without a deep financial sector that rural farmers are unable to borrow to finance working capital needs or do not have access to instruments like warehouse receipts. This forces them to sell their harvest to buy inputs for the next production season, and to pay taxes and land use fees, even if that means going back to the market to buy the same commodity (often at higher prices). 119 Ethiopia Country Economic Memorandum Figure 70. The poorest quintiles spend more of their income on food and have a less diverse diet Food consumption as % of total household Composition of Total Food Consumption by consumption, 2016 Rural/Urban Income Quintile, 2016 60% 100% Other foods Other foods 90% 50% 80% Non-alcoholic Non-alcoholic 70% beverages 40% beverages 60% Fats and oils Fats and oils 30% 50% Animal products Animal products 40% 20% 30% Oilseeds and Oilseeds and 20% 10% condiments condiments 10% Vegetables, pulses, 0% Vegetables, pulses, 0% tubers tubers Poorest Poorest Richest Richest Q-2 Q-3 Q-4 Q-2 Q-3 Q-4 Poorest Poorest Q-2 Q-3 Q-4 Richest Q-2 Q-3 Q-4 Richest Cereals Cereals Rural Urban Rural Urban Source: HIECS data (2016). Cereals are the largest single food group across all income quintiles and thus have an important bearing on the CPI. Cereals are particularly important for the poor, accounting for around 40 percent and 35 percent of total food consumption by the two poorest urban and rural quintiles, respectively. The data further reveal significant differences in the consumption of the different cereals, depending on location and income. Maize is consumed much more intensively in rural Ethiopia than in urban parts of the country, and the share of maize in total cereal consumption decreases steadily as incomes rise. The opposite is true for teff, for which consumption increases steadily as incomes rise, especially in urban areas (Figure 71). Teff generally costs more than twice as much per kilo as maize and is widely considered a superior good. When the poor consume teff, it is often mixed with maize to make a less expensive form of injera (World Bank, 2018). Meanwhile, wheat consumption shares seem to be steadier across income quintiles and locations. Figure 71. Among cereals, the rich consume mostly teff, while the poor rely more on maize Cereals as % Total Food Consumption by Composition of Total Cereal Consumption by Rural/Urban Income Quintile, 2016 Rural/Urban Income Quintile, 2016 45% 100% 40% 90% 35% All other 80% All other 30% cereals 70% cereals Sorghum 60% Sorghum 25% 50% 20% Maize Maize 40% 15% 30% Wheat Wheat 10% 20% 5% Teff 10% Teff 0% 0% Poorest Poorest Poorest Poorest Richest Richest Richest Richest Q-2 Q-3 Q-4 Q-2 Q-3 Q-4 Q-2 Q-3 Q-4 Q-2 Q-3 Q-4 Rural Urban Rural Urban Source: HIECS data (2016). Consumption patterns suggest that government spending on wheat subsidies primarily goes to benefit richer quintiles. While the share of wheat in total cereal spending is slightly higher for the poor than the rich, 120 Ethiopia Country Economic Memorandum in absolute terms the rich spend more in cereals than the poor (Figure 72). An individual in the poorest quintile consumed the equivalent of just 187 birr of wheat and wheat products per year compared to an individual in the wealthiest quintile, who consumed 408 birr of wheat and wheat products per year. The government subsidizes imported wheat for all consumers at a rate equal to about 30 percent of the imported cost (Dorosh et al. 2020). While the subsidy is no doubt important to the welfare of poor households, it is clear that richer consumers capture relatively more of the total subsidy, making this budgetary expenditure regressive. Moreover, wheat primarily provides calories, just like maize, which is domestically produced and could be a substitute. Figure 72. The richest quintile spends 2.5 times more on cereals compared to the poorest quintile Annual Spending on Cereals by Quintile, 2016 (birr per adult) 2,500 2,000 Other cereals 1,500 Sorghum Maize 1,000 Wheat 500 Teff 0 Poorest Q-2 Q-3 Q-4 Richest 2016 Source: HIECS data (2016). Among vegetables, dry pulses and legumes account for the greatest share of total consumption across households. Vegetables are the second most important food category in the Ethiopian CPI, accounting for 12.3 percent of the overall CPI and 22.9 percent of the food CPI. Pulses and legumes are an important part of the Ethiopian diet and key sources of protein, especially during more than 200 fasting days per year when millions of Orthodox Ethiopians abstain from consuming animal products. As shown in Figure 73, the importance of dry pulses and legumes in total food consumption gradually declines as incomes rise (left panel) but stays relatively flat as a proportion of total vegetable consumption (right panel). Fresh perishable vegetables are covered in the top three slices of the bar charts and represent a limited share of food consumption by all income quintiles. Kocho is popular staple in rural Ethiopia as a lower cost alternative to injera, especially in the densely populated southern highlands, and its consumption declines as income rises. 121 Ethiopia Country Economic Memorandum Figure 73. The importance of dry pulses and legumes in total food consumption gradually falls as incomes rise Vegetables as % Total Food Consumption by Composition of Vegetable Consumption by Rural/Urban Income Quintile, 2016 Rural/Urban Income Quintile, 2016 18% 100% Tomato Tomato 16% 90% 14% 80% Other fresh veg Other fresh veg 12% 70% ET kale/other 60% ET kale/other 10% brassicas 50% brassicas 8% Kocho 40% Kocho 6% 30% 4% Potato & other root 20% Potato & other crops root crops 2% 10% Onions (dry) Onions (dry) 0% 0% Poorest Q-2 Q-3 Q-4 Richest Poorest Q-2 Q-3 Q-4 Richest Poorest Poorest Q-3 Richest Richest Q-2 Q-4 Q-2 Q-3 Q-4 Dry pulses and Dry pulses and legumes legumes Rural Urban Rural Urban Source: HIECS data (2016). Over the past two years, the prices of cereals, vegetables, and other nutritious foods have been rising faster than overall inflation. Figure 74 shows the evolution of the prices of different food categories in real terms relative to the overall CPI. Until recently, higher-nutrition foods (dairy, meat, fruit, and vegetables including pulses and tubers) were rising faster than cereals, while the prices of low-nutrition foods (oils and fats, sugars) were on the decline. Bachewe and Minten (2019) analyzed real price trends for different food groups from 2007 to 2017 and found similar results: the prices for higher-nutrition foods rose significantly while prices of cereals declined somewhat, and prices of low-nutrition sugars and fats fell sharply. In the more recent period, however, Figure 74 shows that cereal prices started to ramp up in 2019 and were among the main drivers of food inflation in 2020 and the first months of 2021. Figure 74. The prices of cereals and highly nutritious foods have been trending upward Real Price Trends of Major Food Categories, March 2014 to June 2021 (3-mos moving avg, Mar 2014 = 100) Bread & 140 cereals 130 Milk, cheese, eggs 120 Meat 110 100 Fruit 90 Veg incl. 80 pulses & 70 tubers Oils and fats 60 Source: Author’s calculations from CSA data. While real price trends are important in the long term, wages are fixed in the short run, and price spikes can have serious consequences for consumer welfare. In an analysis of the 2007/08 food price 122 Ethiopia Country Economic Memorandum crisis, for example, Heady et al. (2012) gained access to a unique monthly series of casual wage data from 119 locations in Ethiopian cities and towns. The authors constructed a set of “poor person’s price indices” that were used to deflate the daily wage rates and gauge welfare trends among the urban poor. With no descriptive or econometric evidence of wages substantially adjusting to higher food prices except in the long run, this showed that the disposable income of daily labor declined sharply as food prices soared in 2007/08. Similarly, large nominal price swings (both positive and negative) have serious consequences for private investors including farmers, who depend on income from the crops they sell, and for private traders and cooperatives for whom investing in storage, forward contracting, or even spot market transactions can be risky. Thus, rising food prices, either in nominal or real terms, are likely to affect food security as well as the pursuit of a more varied diet, leading to poor nutritional outcomes. Prices of food are critically important in the consumption decisions of low-income households. Tafere et al. (2010) estimated that price elasticities for most food items in Ethiopia are close to -1.0, meaning that a 10 percent increase in price is associated with a 10 percent decrease in consumption. The rising real prices of high-nutrition foods relative to fats, oils, and sugars and even staple grains can therefore be an important driver of poor nutritional outcomes, which in turn are widely regarded as major contributors to chronic and communicable disease in Ethiopia (Melaku et al. 2016; Bachewe and Minten 2019). Despite improvements over time, child stunting remains high, and Ethiopian children consume one of the least diverse diets in all of sub-Saharan Africa (Hirvonen 2016; Kibrewossen, Hirvonen, and Minten 2020). The next two sections of this chapter explore some of the factors that could be driving food prices upward. Section 5.3.2 looks at overall supply and demand conditions. While Ethiopian agriculture has performed extremely well from the output side, increases in production need to be set in the context of population growth and growth in urban demand, especially where consumers depend on marketed surpluses. From this perspective, Ethiopia clearly still has some way to go in developing a modern, market-driven food system. Section 5.3.3 then looks more closely at price transmission and other market performance issues for cereals. 5.3.2 Has domestic production kept pace with demand? Ethiopia has done exceptionally well in increasing agriculture production since the early 2000s. While agriculture is still predominantly subsistence, low-tech, and rainfed, very strong production gains have been made thanks to significant increases in the use of chemical fertilizers and improved seeds, increased labor supply, and investments in rural roads, education, and agriculture extension (World Bank 2016; IFPRI 2020). Across primary food crops (cereals, vegetables, fruits), gross production tripled between 2000 and 2019 against an area increase of just 60 percent (Figure 75). This stands in contrast to most other eastern and southern African countries, where total output of primary foods grew by an average of 46 percent from an area increase of 47 percent over the same period.38 38 Figure calculated from FAOStat data and based on average gross production of 17 eastern and southern Africa comparators (Burundi, Comoros, Djibouti, Eritrea, Kenya, Madagascar, Malawi, Mauritius, Mozambique, Rwanda, Réunion, Somalia, South Sudan, Tanzania, Uganda, Zambia, and Zimbabwe). 123 Ethiopia Country Economic Memorandum Figure 75. Growth in primary food crop production has increased four times faster than the growth in cultivated area Growth in Primary Food Crops, 2000-2019 (2000 = 1.0) 3.5 3 2.5 2 1.5 1 0.5 Gross production Area cultivated Source: FAOStat data (accessed February 12, 2021). A large part of the increase in agriculture production is the result of improved yields, especially for cereals and pulses. Figure 76 compares per hectare yields for major food categories in Ethiopia with peer countries. Because of differences in climate and individual crops that countries grow, yields are difficult to compare directly, yet these data still provide insight into Ethiopia’s growth trajectory compared with other countries. The data show that Ethiopia started out the 2000s with relatively low per-hectare cereal yields compared to other countries but finished the period ahead of all African comparators. With pulses, the data similarly show that Ethiopia experienced stronger and more steady growth in per hectare yields compared to the others. With primary vegetables, the very large jump in Ethiopian yields between 2014 and 2015 is difficult to explain and may be due to a change in how different crops are counted, making the picture less conclusive. Before the jump, per-hectare yields of primary vegetables improved until 2005 before dropping off and leveling out somewhat. Following the 2015 jump, per-hectare vegetable yields in Ethiopia have been trailing off again. Overall, the massive investments in agriculture inputs and extension seem to have paid off for Ethiopia. Figure 76. Ethiopia has achieved much faster cereal and pulse yield growth than comparator countries Cereal Yields, 2000-2019 Pulse Yields, 2000-2019 Primary Vegetable Yields, (MT/ha) (MT/ha) 2000-2019 (MT/ha) 6.0 2.0 18.0 1.8 16.0 5.0 1.6 14.0 4.0 1.4 12.0 1.2 10.0 3.0 1.0 8.0 0.8 2.0 6.0 0.6 0.4 4.0 1.0 0.2 2.0 0.0 0.0 0.0 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Vietnam Bangladesh Uganda Ethiopia Vietnam Kenya Cambodia Ethiopia Tanzania Cambodia Bangladesh Ethiopia Uganda Tanzania Bangladesh Vietnam Tanzania Cambodia Kenya Rwanda Rwanda Kenya Rwanda Uganda Source: FAOStat data (accessed June 2, 2021). Against this backdrop of strong agriculture sector performance and yield improvements, how can lasting food inflation be explained? Persistent rising food prices in the face of strong growth in crop 124 Ethiopia Country Economic Memorandum production has puzzled researchers for years (World Bank 2007; Dorosh and Ahmed 2011; Hilegebrial 2015). Some potential explanations are discussed in the next paragraphs. First, agriculture production has slowed in recent years. Growth in cereals has been robust and reasonably steady over the past two decades, although it has slowed in recent years. Meanwhile, vegetable production overall has been less steady and appears to have stalled and even diminished somewhat since 2012 (Figure 77).39 In volume terms, the Figure shows that more maize is produced than any other cereal, followed by teff, sorghum, and wheat, in that order. For the vegetables, dry pulses dominate in volume terms but are much heavier and bulkier than others in this group, so it is difficult to draw direct comparisons. Figure 77. The total production of cereals and vegetables seems to have stagnated in recent years Gross Production of Cereals, 2000-19 Gross Production of Leading Vegetables, 30000 5000 2000-19 25000 4000 Thousands of tons Thousands of tons 20000 3000 15000 2000 10000 1000 5000 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Dry pulses and legumes Onions (dry) ET kale/other brassicas Other fresh veg Maize Wheat Teff Sorghum Others Tomato Source: FAOStat data (accessed February 12, 2021). Overall, production shares and land area for cereals have been more stable than for vegetables. The total cereal area grew by about 40 percent over the past two decades, and the allocation of land between the cereals has been fairly steady, with only a modest rise in the share of land given to maize and wheat and a small decrease in the share given to teff. On the vegetable side, total area harvested grew by 61 percent from 2000- 04 to 2015-19, with the share of cropland given to dry pulses, onions, and brassicas increasing and the share of area given to tomatoes and other fresh vegetables decreasing. In the group of other fresh vegetables, total area fell by 43 percent, from 65,000 hectares on average in 2000-2004 to 37,000 hectares on average during 2015- 2019. In terms of yields, as shown above, those of cereals and pulses increased markedly, while yields for some fresh vegetables have stagnated or even declined. These trends seem to go hand in hand with the evolution of real prices discussed above, whereby the prices of fresh vitamin A-rich vegetables have risen in relation to other foods, including dry legumes and pulses and cereals, at least until recently (see Figure 74 above). Second, Ethiopia’s production gains need to be set in the context of population growth and rapid urbanization. From 2000 to 2019, Ethiopia’s total population grew by 66 percent, from 66.5 million to 111.2 million, and the urban population grew by 139 percent, from 9.8 million to 23.4 million (UNDESA data). In 2000, just 15 percent of the total population was urban compared with more than 21 percent now—this is equivalent to 4.7 percent annual average growth in the urban population and 2.3 percent annual average growth 39 Kocho, potatoes, and starchy tubers are included with vegetables in the CPI but are excluded from the charts and tables here because of an apparent inconsistency in how volumes are counted over time. 125 Ethiopia Country Economic Memorandum in the rural population. These trends are important for food prices since urban consumers depend on marketed surpluses. While total gross production of cereals and vegetables has increased over the past 20 years, supply has tightened in relation to the urban population. As shown in Figure 78, despite the strong record of increasing total production discussed above, cereal production in per capita terms has not increased significantly in recent years, and production of dry pulses and legumes in per capita terms has declined. When looking at urban per capita terms, these trends are more pronounced. With cereals, total production in urban per capita terms is slightly below the 2014-16 average for each commodity but still shows a tightening of supply overall. In the case of legumes, onions, and tomatoes, the supply constraint is even clearer: urban per capita supply in 2019 was only 75-80 percent of 2014-16 levels. Fresh vegetables and dry pulses are grown on a smaller scale than cereals so by nature are prone to greater volatility in total production. Nevertheless, despite their importance to household spending and nutrition, the government has generally given less emphasis to promoting vegetable crops than staple cereals, which have benefited from substantial public investment in seed and fertilizer supply. Figure 78. In recent years, cereal production in urban per capita terms has not increased significantly, and that of dry pulses and legumes has declined Index of cereals per capita production, Index of Dry Pulses/Legumes and Urban: 2014-16=100 Vegetables per capita production, Urban: 2014-2106=100 140 300 120 250 100 200 80 150 60 100 40 50 20 0 0 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Dry Pulses/Legumes, Urban Wheat, Urban Teff, Urban Maize, Urban Tomato, Urban Dry Onion, urban Source: Authors’ calculations using FAOStat data (accessed February 12, 2021) and UNDESA population figures. Third, only a small part of agriculture production is marketed. Ethiopian agriculture remains largely subsistence-oriented, and the gains in crop production have not been matched by gains in market participation. Figure 79 shows the share of different cereals offered for sale each year after allowing for household consumption and all on-farm uses (e.g., seed, payment of wages in kind, animal requirements). While there has been a general upward trend in market participation, with a fairly substantial increase coming in 2020/21, only around 18-20 percent of total cereal production has been offered for sale in recent years compared to about 16 percent in 2008/09. Over the entire period covered, around 30 percent, 12 percent, and 21 percent of total teff, maize, and wheat production, respectively, have been offered for sale, which seems lower than in other countries in the region.40 40 For comparative purposes, FEWSNET (2017) reports that 50 percent of maize produced in Zambia is marketed as surplus, and Kilimo Trust (2017) reports that 70 percent of maize produced in Uganda is marketed as surplus. In Zambia, maize is the primary staple food, much like teff in Ethiopia. 126 Ethiopia Country Economic Memorandum Figure 79. The share of cereals production that is actually marketed has increased, but remains modest Percent of total cereal production marketed 35 as "surplus" 2008/09 - 2020/21 25 15 5 -5 2008/09 2009/10 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2019/20 2020/21 Teff Maize Wheat All cereals ` Note: Estimates for 2010/11 and 2018/19 not available; data cover the main maher season only. Source: CSA Agriculture Sample Survey Report on Crop and Livestock Product Utilization (various years). 5.3.3 Are domestic markets efficient? Market inefficiencies have long been blamed by government and others as an underlying cause of food price inflation (Yami, Mayer, and Hassan 2019). As the Ministry of Agriculture’s 2020 draft policy document states: Product markets are characterized by inefficiencies emanating from market operations and value chains fraught with brokers and traders that add little or no value. Besides, large market actors who often create artificial scarcity and exacerbate inflationary pressures responsible for the often very high and fluctuating consumer prices operate in such markets. Consequently, producers are forced to receive unfair prices much lower than the prices paid by consumers (MOA, 2020:38). While such views may have gained popular acceptance in the past, they do not appear to be well- supported by recent empirical evidence. One way to measure market performance is to look at price dispersion—the degree to which prices in different markets move together over time. Several studies of prices in paired markets have shown that food and cash crop markets have become much better integrated over time (Minten, Stifel, and Tamru 2014; Anderson et al. 2016). Using a large set of wholesale cereal price data covering the period from 1999 to 2016, for example, Minten et al. (2020) paired Addis Ababa with major regional wholesale markets and found substantial evidence of reduction in price differentials over time. They also found that the time taken for prices to return to their long-run equilibrium after a price shock had been reduced by more than half over the period covered, and marketing margins and transactions costs for maize and wheat had fallen by more than 50 percent over the study period. Similarly, another recent study by Yami, Meyer, and Hassan (2020) that focused on maize prices found no strong evidence of asymmetric price transmission in Ethiopia, with prices in all but 2 of 14 major market pairs having a very close long-run relation with the Addis Ababa price. Novel analysis conducted for this report confirms convergence in cereal prices across the country, suggesting markets operate efficiently. To examine the degree of geographical integration of the cereal markets in Ethiopia, the evolution of prices in ten local markets between July 2001 and March 2021 was assessed against that in Addis Ababa. The capital is considered the reference market since, in the case of wheat for example, regional brokers supply about 70 percent of their stock to the Addis Ababa market. Price differences are calculated relative to the average price between the two markets, to account for overall inflation (Campenhout 2007). The analysis shows how marketplaces that were traditionally “in deficit” (where local production for the cereal is below demand) used to feature a large premium in the cost of cereals vis-à-vis prices 127 Ethiopia Country Economic Memorandum in the capital, but that differential has been reduced over time to nearly zero in most markets. Meanwhile, “surplus” marketplaces (where the cereal was locally produced and abundant) used to charge lower prices than the capital, but that differential has also been reduced over time except in the case of teff, which nowadays is more expensive than in the capital. These findings are in line with previous studies, confirming substantial improvements in market efficiency over the past two decades. Figure 80. Over the past two decades, there has been convergence in cereal prices across the country Average Price Spread of Deficit Markets 50 Teff 40 Maize Wheat 30 20 10 0 -10 -20 Aug-01 Aug-08 Aug-15 Dec-03 Feb-05 Jun-07 Dec-10 Feb-12 Jun-14 Dec-17 Feb-19 Mar-02 Oct-02 Jul-04 Mar-09 Oct-09 Jul-11 Mar-16 Oct-16 Jul-18 Apr-06 Nov-06 Jan-08 Apr-13 Nov-13 Jan-15 Apr-20 May-03 Sep-05 May-10 Sep-12 Nov-20 May-17 Sep-19 Average Price Spread of Surplus Markets 20 Teff 15 Maize Wheat 10 5 0 -5 -10 -15 -20 Feb-02 Aug-05 Feb-09 Feb-16 Jun-04 Mar-06 Oct-06 Aug-12 Dec-07 Jun-11 Mar-13 Oct-13 Aug-19 Dec-14 Jun-18 Mar-20 Oct-20 Jul-01 Jul-08 Jul-15 Apr-03 Jan-05 Apr-10 Jan-12 Apr-17 Jan-19 Sep-02 Nov-03 May-07 Sep-09 Nov-10 Nov-17 May-14 Sep-16 Source: Authors’ estimation based on ETBC data. 128 Ethiopia Country Economic Memorandum Moreover, farmers seem to be reaping a large share of the final price of cereals. A comprehensive study of teff value chains found that 85 percent of the marketed teff moves from the farmgate to the final consumer in two transactions or less, and the producers’ share of final retail price averaged 80 percent.41 In the case of maize, Wondim, Tefera, and Tesfaye (2019) report that farmers received 91 and 81 percent of the rural wholesale and rural retail price, respectively, and 75 percent and 66 percent of the urban wholesale and urban retail price, respectively. As predicted by Minten et al. (2018), fixed costs for transportation and handling lead to the producer receiving a smaller share of the final price for maize than teff. At 66 percent of the final urban price, however, the producer’s share compares well with other countries. In Zambia, for instance, smallholder farmers have been shown to receive only 52-62 percent of the final urban maize price, depending on the time of year when grain is sold (World Bank 2009). In summary, these findings stand in stark contrast to widely held positions in Ethiopia that markets are inefficient and that farmers only receive a small share of the final product value. The progressive opening of space for private sector participants coupled with improvements in roads and telecommunications are likely behind the significant improvement in market efficiency over the past two decades. Certainly, marketing practices have improved greatly compared with the early 1970s and 1980s, when policies such as the ban on the interregional movement of grains and requirement that private traders sell at least half of their purchases to the Agriculture Marketing Corporation were regarded as contributing factors to widespread famine (Dorosh et al. 2020). The Ethiopian Trading Business Corporation (ETBC), the state-owned marketing company formerly known as the Ethiopian Grain Trade Enterprise (EGTE), remains very active in the procurement, marketing, and storage of many commodities —including distribution of imported subsidized wheat and sale of maize and sorghum to aid agencies—but it does not otherwise have a monopoly position in the domestic market and buys and sells freely alongside private traders. In addition, massive investment in national and rural roads as well as improvements in telecommunications helped reduce the cost of travel and improve communication across agents as well as information availability. Another important development in agriculture marketing was the establishment of the Ethiopian Commodity Exchange (ECX) in 2008, which has played a role in reducing domestic price dispersion. The ECX was founded to promote private participation in commodity trade and to help stabilize markets for cereals. It is jointly owned by private sector and government. The ECX operates secure storage sites around the country and trades standard contracts based on a warehouse receipt system with fixed parameters for grades, transaction size, payment, and delivery (Minten et al. 2020). When an ECX transaction is concluded, price information is made available in real time through online systems and other data platforms.42 Anderson et. al (2016) show that ECX warehouses and trading arrangements have substantially reduced the price spread between market pairs around the country. Ayalew and Belay (2020) similarly use a difference-in-difference approach to compare the spatial price dispersion of wheat and maize that are traded on the ECX with teff that is not traded on the ECX and show that the markets are much better integrated for wheat and maize than for teff. Even when wheat and maize are not traded over the ECX platform, they find evidence of farmers and traders using ECX price information for their own price determination.43 Teff is more commercialized than 41 The study involved fielding a detailed sample survey that covered all stages from farm production to rural assembly, wholesale marketing, storage, processing, and retail sale. Data were collected from zones representing 38 percent of national teff area, 42 percent of the commercial surplus, and more than 90 percent of supply to Addis Ababa. 42 There are presently 11 commodities traded on the ECX. These include maize and wheat and six different pulses (red kidney beans, white pea beans, green mung beans, chickpeas, pinto beans, and white pigeon peas) that are important food crops plus coffee, sesame, and soybeans that are mainly for export. 43 A 2012 interview published in The Guardian with the head of the ECX reported that the Exchange was receiving over 1.2 million calls per month for price data, with 70 percent of calls coming from rural areas (https://www.theguardian.com/global-development/2012/dec/13/africa-commodity-exchange-ethiopia-economy). 129 Ethiopia Country Economic Memorandum wheat and maize because farmers can earn larger margins producing it, and its market should be better integrated, except for the fact that is not listed at the ECX. Nonetheless, it can be argued that the ECX has been much less effective in attracting interest in domestically traded food crops compared with exports. Dorosh et al. (2020) observe that limited access to foreign exchange and other constrains have largely prevented private traders from importing grains or from competing with the government’s subsidized wheat distribution system. For these reasons, they argue that it has been difficult for the ECX to attract much interest in cereals or other domestic crops that are banned from export. In contrast, all exports of coffee, sesame, and soybeans must pass through the ECX platform, leading to these crops being the primary focus of ECX operations. Dried beans, chickpeas, and other pulses that are allowed for export must also be traded over the ECX (USDA/FAS 2018). Trade barriers are discussed in more detail in the next section. 5.4 Policy distortions hindering agriculture potential in Ethiopia 5.4.1 Bottlenecks affecting agriculture inputs and intensification To meet the increasing demand for food and to alleviate foreign exchange requirements for imports, the Ministry of Agriculture (2020) places considerable emphasis on agriculture commercialization and intensification. Cluster and shareholding farming are identified as two modalities through which small and scattered farm plots could be consolidated to create larger, more productive, and market-oriented units. Agriculture commercialization clusters (ACCs) are meant to serve as models for learning as Ethiopia intensifies production and scales up best practices across the country. Within the ACCs, farm scale is achieved through farmer production clusters (FPCs), in which 30-200 individual farmers are grouped together on adjacent land to farm as a single group using the latest farm recommendations including improved seeds, recommended fertilizers, mechanization, and other farming best practices. Over time, FPC farmers are expected to move toward becoming established commercial entities. Currently, the Ethiopian Agricultural Transformation Agency reports that around 1.32 million farmers using the clustering approach are spread across 31,000 FPCs in Amhara, Oromia, SNNP, and Tigray.44 An analysis of the indicative costs and returns to maize and soybeans under different management levels has been conducted. The study was designed to look at the competitiveness of local stockfeed ingredients and involved wide-ranging consultations with large and small farmers, as well as with other stakeholders with detailed knowledge of the costs of agriculture production, including the ATA. The analysis was structured around three levels of progressively intensive family farm (FAM) production, a cluster farm model (CLU), and a large-commercial farm model (COM). These crop budget models are believed to reflect the broad trade-offs between costs and returns of different production systems but are not based on a sample survey so should be thought of as indicative only. Per-hectare yields for maize and soy improve along the continuum because of progressively more intensive input use, while cluster and commercial farmers also receive a higher price for their output because of improved market linkages.45 44 The cluster model focuses on promotion of nine priority crops including wheat, maize, sesame, malt barley, tomato, onion, banana, mango, and avocado. See: http://www.ata.gov.et/our-approach/agricultural-commercialization- clusters-2/ and http://www.ata.gov.et/farmer-production-clusters-fpc/ 45 Specifically, three family farm models (FAM) are based on cultivation with a traditional ox plow and range in intensity from very basic, low-input management using self-saved inputs and family labor (that do not require cash expenditure) to an intensive, high-input model using improved seed and fertilizer supplied by the Ethiopian 130 Ethiopia Country Economic Memorandum Results suggest that intensification may increase production costs for certain agriculture commodities. The data shows that maize and soybeans both become significantly more expensive to grow as on-farm management improves (Table 12). This is particularly true in cash terms, where production at the high- input family farm and cluster levels requires spending of several thousand birr on purchased seed and fertilizer. With the cluster system, additional cash is required for machinery operation. Whether or not and how these higher costs contribute to food price inflation are complex questions that depend on the overall volumes being supplied and downstream market linkages, among other things. FAM-low and FAM-medium farmers have much lower costs both per hectare and per unit of output in total and in cash terms but may be less likely to produce a marketable surplus, thereby restricting supply to urban consumers. Table 12. The additional costs associated with market-oriented cluster farming and intensification generally are much greater than incremental income from selling surpluses for cash Selected Financial Indicators for Different Levels of Maize and Soybean Production Yield Farmer's Gross Cash Non- Total Cash Cash Total Net Profit Net Price Revenue costs cash costs costs costs per costs per (gross rev - total rate of costs as % Quintal Quintal costs) return (Qt/ha) (Birr/Qt) (Birr/ha) (Birr/ha) (Birr/ha) (Birr/ha) total (Birr/Qt) (Birr/Qt) (Birr/ha) (US$/ha) Maize FAM-low 7% 15.0 650 9,750 200 2,835 3,035 13 202 6,715 210 2.21 FAM-med 46% 25.0 650 16,250 1,743 2,085 3,827 70 153 12,423 389 3.25 FAM-high 75% 30.0 650 19,500 7,266 2,390 9,656 242 322 9,844 308 1.02 CLU 80% 40.0 750 30,000 14,020 3,546 17,566 351 439 12,434 389 0.71 66% COM 60.0 900 54,000 19,439 10,117 29,556 324 493 24,444 766 0.83 Soy FAM-low 6% 7.0 1,000 7,000 200 3,395 3,595 29 514 3,405 107 0.95 FAM-med 13% 12.0 1,000 12,000 420 2,845 3,265 35 272 8,735 274 2.68 FAM-high 53% 15.0 1,000 15,000 3,983 3,590 7,573 266 505 7,427 233 0.98 CLU 64% 23.0 1,100 25,300 10,495 5,996 16,491 456 717 8,809 276 0.53 60% COM 26.0 1,200 31,200 16,374 10,717 27,091 630 1,042 4,109 129 0.15 Source: Summary of results from 2020 IFC Ethiopia Maize/Soybeans competitiveness assessment. Notes: Non-cash costs include family labor, saved seed, and capital recovery on fixed implements; net rate of return = net profit/total costs. Calculations based on official exchange rate at time of data collection (Birr 39.1 = US$1.00 in February 2020). Finding ways to make higher-input production both more affordable and more remunerative are therefore key challenges that Ethiopia will need to address. According to the models above, FAM-medium generates nearly as much total profit as CLU and more profit than FAM-high, meaning that many farmers may very rationally decide to produce just at a medium level, rather than to adopt recommended higher-level Agricultural Businesses Corporation (EABC) and hired labor for certain tasks. The cluster farm system (CLU) is indicative of the practices recommended by ATA. It is based on mechanical cultivation and includes purchased seeds, fertilizers, and pesticides supplied through the EABC and a combination of family and hired labor for weeding and harvest. The cluster approach does not yet include soybeans, and the model in this case was purely hypothetical based on cluster-type recommendations for consolidated production using mechanization and improved input use. Finally, the commercial model (COM) is for a very large-scale, fully mechanized farm that uses professional management and hired labor. 131 Ethiopia Country Economic Memorandum commercial management practices. These findings are fundamental to the incentives for farmers to produce surplus crops for market sale and, in turn, to the prospects for managing food inflation overall. As a further disincentive to agriculture intensification, at least some rural households likely have the potential to produce all the food they need for their own consumption from low-input management only. According to HICES and CSA data, the average individual consumed 51 kg of maize per year in 2011 (cited by Hassen et al. 2018). While rural maize consumption is higher than the national average, a family of five that consumes twice this amount could produce all the maize it requires by farming just one-third of a hectare using low-input management or one-fifth of a hectare with medium-input management.46 Rural households do of course have other consumption needs than maize and require at least some cash income, so it is not possible to predict a household’s optimal production decision from these models alone. Nevertheless, it seems apparent that significantly lower net rates of return from moving from one management level to the next is a clear challenge to increasing crop production and market participation. Thus, while consolidation of small farms into larger, more market-oriented units can simplify market logistics and have other important advantages, many additional bottlenecks and policy distortions need to be addressed for this type of farming to take off. As in other areas of agriculture, Ethiopia has recorded substantial gains in the use of improved inputs, but this has relied almost entirely on the cooperative system and state institutions with little to no room for private investment and competition so arguably holds production back. The Ministry of Agriculture (2020) draft policy document identifies several areas for improvement, including improved supply of seed and fertilizer. Barriers to private participation in seed supply have significant implications for farm productivity that undermine the incentives for market-oriented surpluses production. While the current seed policy does allow private seed companies to engage in multiplication of cereal seed, prices and quantities to be supplied are determined centrally, and distribution is done through the cooperative system (Ministry of Agriculture 2020). The Ethiopian Seed Association is comprised of around 23 local firms that multiply varieties maintained by the Ethiopian Agricultural Businesses Corporation (EABC). Otherwise, there is presently only one international seed company operating in Ethiopia which multiplies and sells four international varieties of maize seed, all for one fixed price set by the government. The private varieties are nearly double the price of EABC seed but result in more than twice as much yield so provide excellent value to farmers who can afford the cost. In the models above, for instance, total cash costs of cluster farming which is assumed to use EABC seed would rise by around 28 percent per hectare from using more expensive private seed, but net profits would jump by 244 percent because of the higher yield. In terms of incentivizing farmers to produce marketed surplus, wider uptake of private varieties could therefore be of critical importance, not only to the success of the cluster model but also to the commercialization of individual family farms. Unfortunately, in the current policy environment, private varieties routinely sell out every year. According to seed experts met in Ethiopia, one of the main constraints to international participation in seed supply is access to the foreign exchange needed to import parent seed for local multiplication. Restricted access to foreign exchange also limits the ability to repatriate profits and pay royalties on intellectual property. Pan-territorial fixed prices for seed are another barrier to private sector investment. The one international company operating in Ethiopia reported that it has many other varieties of seed for maize and 46 The analysis for IFC did not consider the costs and profitability of teff under different management conditions. With this limitation in mind, however, Hassen et al. (2018) report that average per capita consumption of all forms of teff is 34 kg per capita per year (ranging from just 3 kg/capita in Somali to 101 kg/capita in Addis Ababa) and around 35 kg per capita in Amhara and Oromia, two major production areas. With average national teff yields now almost 3 mt per hectare, this means that a family of five could potentially produce all the teff it requires from barely 0.06 hectares or perhaps 0.1ha using low-input, unimproved management. 132 Ethiopia Country Economic Memorandum other crops in its catalog that it believes would outperform EABC varieties but said it is not currently feasible or profitable to bring these into the country for the reasons mentioned above. In a more liberal trading environment, many other international seed companies would likely have strong commercial interest in Ethiopia. There are also important opportunities to improve fertilizer supply. By law, all fertilizer sold to family farmers and cluster farmers is done through the cooperatives. Imports are managed centrally by EABC, which distributes fertilizer to the co-ops which in turn sell to farmers at centrally determined break-even prices, with only minor variations by region according to differences in transport costs. Officials say that the primary advantage of this arrangement is that it helps keep fertilizer prices down, thereby improving overall access. Large commercial farms do not have access to this fertilizer and must import on their own. Figure 81 shows the overall growth and composition of public fertilizer supply over the past several years. While the total supply of fertilizer has grown modestly over the period covered, the types of fertilizer supplied consist almost entirely of nitrogen (N) and phosphorus (P), without any potassium (K) which is also needed for optimal crop growth.47 Figure 81. While the public supply of fertilizer has increased in recent years, it does not include potassium Supply of Fertilizer by Type, 2012-2020 ('000 metric tons) 1600 1400 1200 NPS+ZnB 1000 NPS+Zn 800 NPS+B NPS 600 DAP 400 UREA 200 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 Source: MoA data. According to the fertilizer experts consulted, it appears that a key reason for not supplying potassic fertilizer or agriculture lime is that Ethiopia has its own natural deposits of potassium and limestone. They said the logic has been that Ethiopia should not spend scarce foreign exchange on imports when there is potential for domestic supply. Until now, however, these deposits have not been commercially exploited, and the minerals remain in their natural state rather than being used for agriculture. While the full set of reasons for not bringing potassium and lime into commercial production are unclear, these items could certainly be imported, meaning that the reluctance to spend foreign exchange in the short run almost certainly comes at very high cost for food supply and growth in the long run. In a recent field experiment in Jeldu woreda in Oromia, for instance, agriculture workers report that the application of lime led to yields of barley improving from 1.0 tons/ha to 5.3 tons/ha (Assefa 2020). 47 Soil tests in Ethiopia between 2013 and 2016 found significant deficiencies of potassium, sulfur, boron, zinc, and iron (EthioSIS soil survey results 2013, 2015, 2016, cited by ATA 2019). Additionally, Ethiopia faces significant challenges with acidic soils covering over 40 percent of the country (IFPRI 2020, Assefa 2020). Agriculture lime can be used to improve the soil pH but, like potassic fertilizer, is not widely available to farmers. 133 Ethiopia Country Economic Memorandum 5.4.2 Export bans affecting marketed amounts One major barrier to agriculture trade is that all major cereals are banned from export. Except for a few temporary pauses, Ethiopia has had long-standing bans on maize, wheat, and teff, at least since the 2008 global food price crisis (Yami, Meyer, and Hassan 2020). According to the ATA (2019), the authorities’ primary aim in banning food exports is to secure domestic supply and limit price increases that may result from food shortages. Many other countries in Africa and elsewhere have used similar reasoning to restrict food trade, but there is strong evidence that export bans are rarely successful in achieving these objectives and instead lead to greater price volatility and reduced investment by large and small farmers and by commercial traders alike. In Tanzania, for instance, Dao and Kennedy (2016) show that banning cross-border maize trade has had little effect on the food price index and that the few price benefits that did arise were captured by urban households at the expense of rural producers, thereby increasing poverty in the country. In a study of export bans in Ethiopia, Kenya, Tanzania, Zambia, and Malawi, Porteous (2017) shows that they result in significantly greater price volatility in all implementing countries during the ban period compared to model simulations without a ban. Smuggling usually also arises when trade bans are in place, leading to higher rents and greater price differentials than in a normal trading environment. Despite the prohibition on exports, moderate volumes of maize are known to leak across Ethiopia’s borders to neighboring countries each year. The Joint Cross-Border Market and Trade Monitoring Initiative, which monitors informal cross-border trade in East Africa including at 12 Ethiopian border- crossings, finds that several hundred to a few thousand tons of maize are exported to Kenya and Somalia every month through informal channels (FSNWG, various 2019-2020). There is also evidence of maize grain, maize flour, and maize bran being exported to deficit markets in Sudan and Djibouti (Yami, Meyer, and Hassan 2020). While current exports are generally small, the prohibition on maize exports can have major consequences for producer incentives and production of tradable surpluses. To the extent that higher prices are available in neighboring markets, any ban on selling to those markets will have negative consequences on the prices paid to farmers and incentives to adopt higher-yielding technologies needed for sector growth and commercialization, including the incentives to participate in cluster farming. Export bans on food can also encourage farmers to turn away from these crops to non-food commodities that are not subject to trade bans. In a similar way, export bans undermine the incentives for private traders and cooperatives to invest in crop storage and other marketing services. Partial equilibrium modeling of Ethiopian maize prices under an export ban by Yami, Meyer, and Hassan (2020) helps bring these costs into focus: according to their model, a ban during a bumper harvest that results in a 20 percent yield rise would reduce domestic prices by 81 percent.48 With rapid growth in maize production, Ethiopia is now marginally surplus in maize and has potential to become a large regional exporter. Table 13 presents a set of simplified export parity calculations for potential maize exports to Marsabit, a chronically deficit market town in northern Kenya. Based on 2019 average market prices and transport and handling costs reported by the Ethiopian Shipping and Logistics Services Enterprise (ESLSE), these calculations show that higher marginal profits could have been earned from exporting to Kenya than from selling domestically, resulting in more total value available to be passed up the chain to farmers. The calculations also show how distortions in the exchange rate regime affect the incentives 48 While the actual profitability of exporting from different locations will vary depending on transport costs, the authors report that in a normal year, domestic prices overall have been hovering near the border price, thereby making exports unprofitable and the ban unnecessary. In the event of a bumper harvest that leads to a large fall in prices, however, regional exports would not only be profitable but would serve to mitigate the negative effects of collapsing domestic prices. In the case of a reduced harvest and rising domestic prices, exports would not be profitable, and there would be no reason for an export ban unless prices in the neighboring market were to rise by significantly more than in Ethiopia. 134 Ethiopia Country Economic Memorandum to export, whereby the available profits are about one-third higher when using the parallel exchange rate (as one would if the transaction is illegal) compared with the official rate. In this way, the Table further illustrates how normalization of the exchange rate regime would reduce the incentives to export and contribute to greater price stability. Table 13. Exporting at the overvalued official exchange rate significantly diminish trader earnings Simplified Export Parity Calculations for Maize to Northern Kenya (2019 prices) US/ton Birr/Qt Official Birr/Qt Parallel White maize, cif Marsabit (2019 average) 450.00 1,308 1,743 Less road from AA (1,000km @ 0.12/mt/km) 120.00 349 465 Less documentation, fumigation, handling 10.00 29 39 Equals AA export parity 320.00 930 1,239 AA wholesale price (2019 average) 307.93 895 1,193 Available profit from export trade 12.07 35 47 Source: Authors’ calculations from GIEWS/FPMA data (accessed April 12, 2021) for Addis Ababa wholesale price, FEWSNET (2020) for Marsabit average price, interviews with ESLSE for local transport and handling costs, and National Bank of Ethiopia for 2019 average official and parallel market exchange rate information. In its recent study on Ethiopia’s export bans, the ATA (2019) concludes that export bans are inconsistent with objectives for commercial agriculture development and sets out a spectrum of policy recommendations, ranging from modest and specific to radical and broad. These recommendations are not framed in the specific context of strategies to mitigate inflation but as key parts to a broader strategy for mitigating the risks of cereal shortages and price volatility needed for long-term growth and commercialization. The first set of recommendations focuses on improving the transparency and predictability of the trade environment, the second on replacing export bans with variable export taxes, and the third on doing away with export bans entirely. The latter (removing the bans) seems the most desirable option, given that introducing export taxes would be distortionary in its own way as shown by the international experience. Yami, Meyer, and Hassan (2020) offer a similar range of options, observing that market integration with regional maize deficit markets would reduce maize price instability in times of bumper harvests in Ethiopia. 5.4.3 Centralized wheat procurement resulting in prices above import parity Ethiopia is both a major producer and importer of wheat, although it is trying to make progress toward self-sufficiency. Most domestic production is saved by farmers for household use and, with growing demand in urban and peri-urban areas, Ethiopia is also a major wheat importer. Between 2012 and 2019, imports accounted for 24 percent of total domestic supply on average, including production retained for home use (Table 14) With only about one-fifth of total domestic production sold as surplus, however, imports accounted for 61 percent of market supply on average. Together with large production increases, the share of domestic wheat available in the market increased from 27 percent in 2011/12 to 43 percent in 2019/20, and the imported share of market supply decreased from 73 percent to 57 percent. Ethiopia is therefore making progress toward self-sufficiency but still needs further increases not only in domestic production but also in the share of domestic production marketed if it is to displace the need for imports. The share of domestic production marketed increased just from 18.1 percent to 21.9 percent over the past decade. 135 Ethiopia Country Economic Memorandum Table 14. While Ethiopia has increased the share of domestic production that is marketed, it remains dependent on wheat imports Wheat Production, Marketing, and Imports in Ethiopia 2012-2019 Total Share of Domestic Official Total Imports as Market Supply production production "surplus" imports supply % total Total Domestic Imported ('000 tons) marketed ('000 tons) ('000 tons) ('000 tons) supply ('000 tons) share share (a) (b) (a * b) (c) (a + c) (c /(a + c)) (a * b) + (c) 2011/12 18.1% 32% 27% 73% 3,435 622 1,639 5,074 2,261 2012/13 19.4% 29% 32% 68% 3,925 761 1,618 5,544 2,380 2013/14 18.4% 20% 43% 57% 4,232 779 1,041 5,273 1,820 2014/15 21.5% 21% 44% 56% 4,651 1,000 1,259 5,910 2,259 2015/16 20.8% 35% 28% 72% 4,538 944 2,483 7,021 3,427 2016/17 21.7% 19% 48% 52% 4,643 1,008 1,077 5,720 2,084 2017/18 20.6% 19% 47% 53% 4,838 997 1,141 5,979 2,138 2018/19 21.2% 20% 46% 54% 5,315 1,128 1,306 6,621 2,434 2019/20 21.9% 22% 43% 57% 5,100 1,117 1,460 6,560 2,577 Average 20.4% 24% 39% 61% 4,520 928 1,447 5,893 2,350 Source: Total production and import data to 2018/2019 from FAOStat (accessed February 12, 2021); data for 2019/20 are from USDA/FAS (2021); share of production marketed from CSA Agriculture Sample Survey Report on Crop and Livestock Product Utilization (various years). Notes: Domestic "surplus" = total production * share marketed. Official imports include international food/development assistance. Total supply = domestic production + official imports. Market supply = domestic surplus + official imports. CSA did not report share of total production marketed in 2018/19 so used average of 2017/18 and 2019/20 instead. Apart from the share of wheat that arrives as food aid and development assistance (around 15 percent), all imports are managed centrally by the government. Since 2008, all commercial imports have been centrally managed by the government for sale to designated mills and bakeries at fixed prices about 30 percent below import parity (Dorosh et al. 2020). Until recently, all wheat imports by the government have been done by tender through the Public Procurement and Property Disposal Service (PPPDS), which is under the Ministry of Finance. Upon milling, the flour is sold for a fixed price to designated bakeries, who in turn sell bread products for centrally determined subsidized prices to consumers. Within the past year, at least two wheat contracts have been canceled due to default or contractual non-performance by the tender winner (Dadhi, 2021). This suggests issues with the centralized public procurement of wheat, including due to the demanding requirements imposed on participants, such as providing several months of supplier credit prior to receiving a payment. Wheat subsidy spending is regressive and import restrictions have resulted in a dual market. Domestic wheat markets are segregated from the subsidized wheat imports because of the limited access of the imported product to designated mills and bakeries only. As discussed above, wealthy Ethiopians consume more than twice as much wheat as the poor, making the universal subsidy on wheat regardless of income regressive. Moreover, subsidized wheat has often been in short supply due to incidents and delays in the procurement process, leading some mills and bakeries to opt out of the program in favor of domestic grain for which price controls do not apply and significantly higher prices prevail. The subsidy undoubtedly helps hold prices down 136 Ethiopia Country Economic Memorandum for consumers who are able to access the limited subsidized supply, yet with more and more of the total market supply coming from higher-price domestic wheat, prices overall are on the rise. As a result of import controls and market duality, wheat prices in domestic markets have been well above import parity prices and also more volatile than in other countries. Import parity serves as a natural cap on the domestic price, and even with market segregation, competition with subsidized imports would drive local prices down. This, however, has not been the case in Ethiopia, where domestic prices have stayed considerably above import parity for much of the past 15 years, implying that under-supply and other inefficiencies with the state procurement and distribution system have permitted these large price gaps to emerge. Since January 2020, wholesale prices of local wheat in domestic markets have been 67 percent higher than import parity for free market imports from Ukraine (Figure 82).49 If import restrictions were lifted and if provided foreign exchange was available, private importers would naturally seek to fill this price gap. Domestic prices in Ethiopia are also found to have been substantially more volatile than in other wheat-producing countries. Figure 82. Wheat prices in Ethiopia have been considerably higher than import parity prices Wheat, AA wholesale compared to import parity (US$/ton, 3-month moving average) 800 700 600 Ethiopia (AA) 500 400 Ex 300 Ukraine (0% 200 duty) 100 0 Aug-12 Aug-06 Aug-07 Aug-08 Aug-09 Aug-10 Aug-11 Aug-13 Aug-14 Aug-15 Aug-16 Aug-17 Aug-18 Aug-19 Aug-20 Feb-07 Feb-08 Feb-09 Feb-10 Feb-11 Feb-12 Feb-13 Feb-14 Feb-15 Feb-16 Feb-17 Feb-18 Feb-19 Feb-20 Feb-21 Source: Authors’ calculations from GIEWS/FPMA price data (accessed March 15, 2021), Ethiopian Shipping Lines (ESL) and worldfreightrates.com for sea and land freight costs, and interviews with clearing agents for local handling costs. Note: All local prices converted from birr at official exchange rate. Keeping the price of subsidized imports well below the domestic price also implies huge excess profits or rents for millers and bakeries able to access the imports. This problem has been around for some while. Dorosh and Ahmed (2011) suggested that auctioning of government imports in the domestic market would be a way to eliminate these rents, improve government revenues, and bring prices closer to true import parity. Lower domestic wholesale prices would be a disincentive to the producer in the short run but would be an incentive to the consumer, leading to increased domestic demand and higher and more stable prices in the long run that are needed to incentivize on-farm improvements and market participation by smallholder, cluster, and large commercial wheat growers alike. According to USDA/FAS (2021), the government recently signaled its intention to partially liberalize wheat imports, but it is not clear what this will mean in practice. 49 Authors’ calculations from data in Figure 82. 137 Ethiopia Country Economic Memorandum 5.5 Conclusion and policy implications 5.5.1 Fiscal and monetary policy options to curb inflation Fiscal policy is among the inflation drivers in Ethiopia and can be shaped to minimize inflationary pressures. Ethiopia has multiple drivers of inflation. In the short run, inflation is driven by structural factors, notably output gaps and prices of imported goods (e.g., wheat, oil, non-food items). In addition, disequilibria in the monetary and fiscal sector are found to affect cereal, food, and overall inflation (Ndikumana et al. 2021). Containing government borrowing would need to be part of the general toolkit for controlling inflation. In this regard, treasury bills and bonds are preferrable to direct advances from the National Bank of Ethiopia (NBE), as the latter imply money printing and are thus expected to be more inflationary. Full implementation of the planned monetary and exchange rate policy reforms would also likely bring substantial benefits in terms of controlling inflation. The empirical results suggest that keeping money growth in check is desirable to avoid exacerbating other pressures on inflation from the demand and supply sides. So far, the NBE has been targeting base money growth in relation to the targeted nominal GDP growth (which has been around 20 percent a year) in the development plans. The NBE has committed to shifting toward a modern monetary policy framework, including by developing liquidity management tools and reviving the use of the interest rate as a tool for monetary policy. In addition, the move toward a market clearing exchange rate, coupled with the progressive phasing out of foreign exchange restrictions as envisaged in the Exchange Rate Reform Roadmap endorsed by the NBE, is expected to gradually eliminate the parallel market premium and mitigate its effects on inflation. 5.5.2 Boosting agriculture production through improved access to inputs Despite success to date, rapid urban population growth requires Ethiopia to further boost agriculture production. Since the early 2000s, total cereal and vegetable production has nearly tripled. With cereals, these gains can be attributed to significant yield improvements thanks to heavy state spending on centralized seed and fertilizer supply and farmer extension. In comparison, vegetable crops have received less attention from the state system, and although total production has also grown substantially, overall performance in this part of the food sector has been less with per-hectare yields of some major vegetable crops even falling over time. Moreover, for both cereals and vegetables, growth in total production has slowed and even leveled off in recent years and has generally failed to keep pace with rapidly growing urban demand. Ethiopia has considerable scope to increase private sector participation in seed supply. Improved access to foreign exchange is a key requirement and would enable private seed firms to import additional new varieties and larger amounts of breeding material for local multiplication. Allowing seed firms to set their own prices rather than charge centrally determined pan-territorial prices would be another way to support private sector growth. The EABC still has an important role to play in ensuring a steady seed supply, but measures to open the market to increased competition would not only give farmers greater choice but could also be a route to improved incentives for the commercial market participation needed to meet the growing demand for food. Such a move could also free the EABC, as a public institution, to focus more attention on the development and marketing of open pollinated varieties of maize seed and improved varieties for closed pollinated crops, including wheat and teff, which are of less interest to private sector. An adequate and streamlined process for the quality control of private inputs and the approval of new seed varieties will need to be put in place. Similarly, although Ethiopia has made great progress in increasing fertilizer use, the state system has so far failed to provide fertilizer with potassium or agriculture lime. While the government system has 138 Ethiopia Country Economic Memorandum been very effective at meeting most fertilizer needs and has important advantages in terms of price and dependability, competitive measures that improve farmer choice also have an important role to play in agriculture commercialization and in meeting the growing urban demand for food. The absence of potassium in the fertilizers being supplied at the moment could mean that farmers in Ethiopia are not sufficiently addressing nutrient deficiencies in the soils. Investments in fertilizer blending is recommended so that farmers can purchase the exact amounts of nutrients required for their land (based on soil mapping work). As prior attempts to invest in blending by the public institutions failed, this is an area in which the private sector may succeed if operating space is granted. Allowing the private sector space to participate in the facilitation of inputs is needed to support agriculture modernization and development. Experience in other countries shows that private seed and fertilizer companies often invest in demonstration plots and farmer field days to increase their sales and build loyalty to their brand. These can complement the state system and help free public resources to focus more attention on neglected crops, including vegetables. While access to private seed and fertilizer is unlikely to result in lower costs per hectare as input prices may be higher than under the state system, because of those inputs being more productive (improved seed varieties, lime, and/or fertilizer with potassium as the case may be), they are expected to result in greater yield and lower costs per unit of output. 5.5.3 Promoting market participation and value-addition Beyond improving gross production, the share of commercialized production remains low, due not to inefficient markets but to lack of incentives for farmers. While the growth in production of most crops has been little short of spectacular, only a small share of this production is marketed as surplus. Conventional wisdom in Ethiopia places much of the blame for poor market participation on unscrupulous traders and intermediaries who connive to drive prices up, yet the empirical evidence does not bear out these perceptions. Rather, in the context of smallholder agriculture, the markets in Ethiopia appear to perform very well overall, with a large share of the final retail price of teff and maize reaching up to the farm gate and with most mark- ups on the way to the final consumer relating to transport. Indeed, Ethiopia has already done a lot to improve domestic market performance through massive investments in rural roads and telecommunications infrastructure and the creation of market information systems. State-sponsored cluster farming seems to have lower rates of return than traditional farming so far. One of the main ways the Ethiopian government has tried to encourage market participation is by clustering smallholder farmers into consolidated units to be managed under high-yield technology. Buying from a few large producers is certainly easier and more cost-effective than buying from many individual farmers, and clustering makes marketing easier and more efficient through improved economies of scale. However, it may not be enough to take agriculture to the next level: evidence discussed here suggests that the cluster model provides a much lower rate of return to a farmer’s expenditure than unimproved family farming and results in much higher costs per quintile. Facilitating access to rural credit as well as physical investments (e.g., in storage) can help support activity modernization and the uptake of contract farming. Given that commercial lending sector likely considers smallholder agriculture too risky and faces high transaction costs in processing a large number of relatively small and frequent loans to smallholder farmers, leveraging the outreach capacity and the products of credit and savings institutions of the different regions and the Rural Saving and Credit Cooperatives can help better support farmers in their commercialization efforts. This can be done through arrangements with credit associations to provide input credit that helps farmers resolve their liquidity constraints during farming seasons as well as through marketing channels to help farmers get their fair share of the price margins. 139 Ethiopia Country Economic Memorandum In addition, matching grants and uptake of new technologies can help increase market participation and unlock agriculture potential. Matching grants to farmers and input-distributing firms can contribute to establishing new linkages and help increase the amounts supplied to the market. In addition, demonstrations of how to leverage new technologies as well as continued support for extension will remain key. While the Ethiopian Agricultural Research Institute and its regional branches are tasked with identifying new technologies and new improved seeds, and with showcasing them to farmers, they face significant capacity constraints. Those constraints could be addressed through the provision of training to institutions under the Ministry of Agriculture, and by linking them with technical and vocational schools and the Ministry of Innovation and Technology to develop smart farming solutions. 5.5.4 Removing barrier to trade As discussed earlier, export bans are holding back the agriculture sector from attaining its potential and their removal needs to be considered. In the case of surplus production, a ban on exports can lead to a price collapse that disincentivizes farm production, especially market-oriented surplus production. Whereas export bans have been justified in Ethiopia on grounds of minimizing food price spikes, extensive research from around the world (including Ethiopia) shows that trade bans usually have the opposite effect and usually result in much greater price volatility and high net costs to social welfare than if borders were left open. Use of export taxes has been proposed as a more transparent way to control the flow of food supplies but would risk further encouraging smuggling and is generally less ideal than leaving the borders open and allowing markets to work. Thus, it is recommended that authorities consider removing altogether all the existing exports bans on cereals and other agriculture products. Export bans are inconsistent with objectives for commercial agriculture development and generally unnecessary in the case of a domestic shortage since prices would naturally rise, leading to stocks remaining within the country. Similarly, opening of the wheat market to private imports could go a long way in holding consumer prices down and improving price stability. Recent tenders have attracted little interest from the private sector due to restrictive bid conditions, thereby leading the government to pay well above world market prices. Especially in recent months, very restrictive tender conditions have led to an undersupply of imported wheat and corresponding domestic price hikes. The desire to protect consumers with subsidies is understandable, yet current arrangements are regressive and likely to drive prices up overall. Moreover, the war in Ukraine is significantly adding to the public import bill. As a first step, auctioning of tendered grain to local buyers could reduce the rents associated with government imports. Authorities could also consider to partly substitute subsidized imported wheat with maize or with locally produced cassava flour, in their supply to mills and bakeries. In the medium term, improved access to foreign exchange that allows private importers to participate in the global market would be a more lasting way to fill the cereal gap, leading to greater stability and lower consumer prices overall. Subsidies could be then phased out. During the transition, a more equitable way to support poor consumers would be through social cash transfers or wheat vouchers. 140 Ethiopia Country Economic Memorandum Table 15. Policy options for mitigating food inflation and making the most of Ethiopia’s agriculture potential Area Short term Medium term Implementing agency Fiscal and Phase out direct advances to the Move toward a modern monetary National Bank of Ethiopia monetary government, replacing them with policy framework featuring real policy treasury bill auctions to mitigate positive interest rates and inflation increased central bank independence Fully implement the Exchange Rate Reform Roadmap Access to Put in place a process for the quality Ministry of Agriculture agriculture control of private inputs and the inputs approval of new seed varieties Agriculture Transformation Agency Allow the private sector to provide Invest in fertilizer blending, fertilizer with potassium or agriculture including by opening space for Ethiopian Agricultural Research lime to further boost production the private sector Institute Extension Facilitate access to rural credit for Invest in inputs and extension Credit and savings institutions of support inputs as well as physical investments services for fruits, vegetables, the different regions and (e.g., in storage) and other nutritious foods for incentives which demand is increasing Rural Saving and Credit Provide matching grants to farmers and Cooperatives input-distributing firms Support the demonstration and adoption of new technologies by Ministry of Agriculture building capacity at the Ethiopian Agriculture Transformation Agricultural Research Institute Agency Trade Remove export bans on cereals to Ministry of Trade and Regional barriers maximize potential, in particular for Integration maize exports Partly substitute imported wheat with Phase out wheat subsidies, which Ministry of Finance maize or cassava flour are regressive and distortive (and the poor could be compensated Ministry of Regional Integration Allow private firms to import wheat, through vouchers) while facilitating access to foreign exchange Source: Elaborated by the authors. 141 Ethiopia Country Economic Memorandum 5.6 References Aksoy, M., and A. Isik-Dikmelik. 2008. Are Low Food Prices Pro-Poor? Net Food Buyers and Sellers in Low-Income Countries, Policy Research Working Paper 4642, Development Research Group. Washington DC: World Bank. Alem, Y., and Måns Söderbom. 2012. “Household-Level Consumption in Urban Ethiopia: The Effects of a Large Food Price Shock.” World Development 40 (1): 146–62. Anderson, C., M. Bezabih, and A. Mannberg. 2016. The Ethiopian Commodity Exchange and Spatial Price Dispersion. Environment for Development Discussion Paper Series No 16-02. Addis Ababa: The Environment for Development Centre. Assefa, S. 2020. Visit to Jeldu practice of Integrated Soil Fertility Management for Acid Soil Reclamation, AgriProFocus Ethiopia, 28 September. ATA 2019. Project Transition Report: Fertilizer Blending. January 2019. The Agriculture Transformation Agency (ATA), Addis Ababa. Ayalew, H., and D. Belay. 2020. “The Ethiopian Commodity Exchange and Spatial Price Dispersion: Disentangling Warehouse and Price Information effects.” Department of Food and Resource Economics, Faculty of Science, University of Copenhagen. IFRO Working Paper, No. 2020/01. Bachewe, F., and B. Minten. 2019. The Rising Cost of Nutritious Foods: The Case of Ethiopia, ESSP working papers 134, International Food Policy Research Institute (IFPRI). Campenhout, V. 2007. “Modelling trends in food market integration: Method and an application to Tanzanian maize market.” Food Policy, 32, 112-127. Chauffour, J., and M. Gobezie. 2019. Exiting Financial Repression: The Case of Ethiopia. Policy Research Working Paper; No. 9082. Washington, DC: World Bank. Dadhi, H. 2021. “International Wheat Procurement Falters, Again.” Fortune News (24 April). AddisFortune, Addis Ababa. https://addisfortune.news/international-wheat-procurement-falters-again/ Dao, X., and A. Kennedy. 2016. “Economywide Impact of Maize Export Bans on Agricultural Growth and Household Welfare in Tanzania: A Dynamic Computable General Equilibrium Model Analysis” in Development Policy Review, 34:1, pp. 101-134. Dorosh, P., and H. Ahmed. 2011. “Foreign Exchange Rationing and Wheat Markets in Ethiopia.” Ethiopian Journal of Economics, Vol. XX, No. 2. (October), pp. 83-104. Dorosh, P., J. Smart, B. Minten, and D. Stifel. 2020. “Droughts, Cereal Prices, and Price Stabilization Options” (Chapter 9) in Ethiopia’s Agrifood System: Past Trends, Present Challenges, and Future Scenarios. P. Dorosh and B. Minten eds. Washington, DC: International Food Policy Research Institute (IFPRI). Durevall, D., J. Loening, and Y. Ayalew Birru. 2013. “Inflation dynamics and food prices in Ethiopia.” Journal of Development Economics, 104, issue C, p. 89-106. FEWSNET (Famine Early Warning Systems Network). 2017. Zambia Maize Market Fundamentals. Lusaka. FEWSNET (Famine Early Warning Systems Network). 2020. Kenya Price Bulletin. Nairobi. FSNWG (Food Security and Nutrition Working Group). 1999-2021. East Africa Cross-Border Trade Bulletin, FSNWG, Market Analysis Subgroup. FEWS/NET, FAO, WFP Joint Cross-Border Market and Trade Monitoring Initiative, Nairobi. 142 Ethiopia Country Economic Memorandum Geiger, M., and C. Goh. 2012. Ethiopia Economic Update: Overcoming Inflation, Raising Competitiveness. Washington, DC: World Bank. Hassen, I., M. Regassa, G. Berhane, B. Minten, and A. Taffesse. 2018. “Teff and its Role in the Agricultural and Food Economy” (Chapter 2) in The Economics of Teff: Exploring Ethiopia’s Biggest Cash Crop, B. Minten, A. Teffesse, and B. Brown, eds. Washington, DC: International Food Policy Research Institute (IFPRI). Heady, D., F. Nisrane, I. Worku, M. Dereje, and A. Taffesse. 2012. Urban Wage Behavior and Food Price Inflation: The Case of Ethiopia. Ethiopia Strategy Support Program II (ESSP II), Working Paper 41. Washington, DC: International Food Policy Research Institute (IFPRI); Addis Ababa: Ethiopian Development Research Institute (EDRI). Hilegebrial, Z. 2015. “Determinants of Food Price Inflation in Ethiopia.” Food Science and Quality Management, Vol. 41, pp. 30-44. Hirvonen, K. 2016. "Rural-urban differences in children's dietary diversity in Ethiopia: a Poisson decomposition analysis." Ethiopia Strategy Support Program II (ESSP II), Working Paper 89. Washington, DC: International Food Policy Research Institute (IFPRI); Addis Ababa: Ethiopian Development Research Institute (EDRI). IFPRI (International Food Policy Research Institute). 2020. Ethiopia’s Agrifood System: Past Trends, Present Challenges, and Future Scenarios. P. Dorosh and B. Minten, eds. Washington, DC: IFPRI. IMF (International Monetary Fund). 2020. “Federal Democratic Republic of Ethiopia. 2019 Article IV Consultation.” International Monetary Fund, Staff Report No. 20/29. Kibrewossen, A., Hirvonen, K., and B. Minten. 2020. “Farm Size, Food Security, and Welfare.” (Chapter 6) in Ethiopia’s Agrifood System: Past Trends, Present Challenges, and Future Scenarios. P. Dorosh and B. Minten eds., International Food Policy Research Institute (IFPRI), Washington DC. Kilimo Trust. 2017. Characterization of Maize Markets in the EAC: Regional East African Community Trade in Staples (REACTS). Kampala: Kilimo Trust. Melaku, Y., A. Temesgen, A. Deribew, G. Tessema, K. Deribe, B. Shale, S. Abera, T. Bekele, F. Lemma, A. Amare, O. Seid, K. Endris, A. Hiruye, A. Worku, R. Adams, A. Taylor, T. Gill, Z. Shi, A. Afshin, and M. Forouzanfar. 2016. "The impact of dietary risk factors on the burden of non-communicable diseases in Ethiopia: findings from the Global Burden of Disease study 2013." International Journal of Behavioral Nutrition and Physical Activity, 13 (1):122. Ministry of Agriculture. 2020. Draft Revised Agriculture Development Policy of Ethiopia. Addis Ababa: Ministry Of Agriculture. Minten, B., D. Stifel, and S. Tamru. 2014. “Structural Transformation of Cereal Markets in Ethiopia.” Journal of Development Studies 50, no. 5:611–29. Minten, B., M. Dereje, F. Bachewe, and S. Tamru. 2020. “Evolving Food Value Chains”, Chapter 7, in Ethiopia’s Agrifood System: Past Trends, Present Challenges, and Future Scenarios. P. Dorosh and B. Minten eds. Washington, DC: International Food Policy Research Institute (IFPRI). Minten, B., S. Tamru, E. Engida, and T. Kuma. 2018. “Supply Chain From Production Areas to Addis Ababa” (Chapter 11) in The Economics of Teff: Exploring Ethiopia’s Biggest Cash Crop, Bart Minten, Alemayehu Seyoum Teffesse, and Betra Brown, eds. Washington, DC: International Food Policy Research Institute (IFPRI). 143 Ethiopia Country Economic Memorandum Ndikumana, L., J. Nkurunziza, M. Sánchez-Martín, S. Mulugeta, and Z. Kelborne. 2021. “Monetary, Fiscal and Structural Drivers of Inflation in Ethiopia: New Empirical Evidence from Time Series Analysis.” Policy Research Working Paper; No. 9881. Washington DC: World Bank. Porteous, O. 2017. “Empirical effects of short-term export bans: The case of African maize.” Food Policy, 71, pp. 17-26. Tafere, K., A. Taffesse, S. Tamiru, N. Tefera, Z. Paulos. 2010. Food demand elasticities in Ethiopia: Estimates using household income consumption expenditure (HICE) survey data. Ethiopia Strategy Support Program II (ESSP II), Working Paper 11. Washington, DC: International Food Policy Research Institute (IFPRI); Addis Ababa: Ethiopian Development Research Institute (EDRI). Ticci, E. 2011. “Can Inflation Be a Good Thing for the Poor? Evidence from Ethiopia.” Development in Practice 21 (4–5): 629–41. USDA/FAS. 2018. Ethiopia, Pulse Crops Market Update 2018, Report no. ET1819, United States Department of Agriculture, Foreign Agriculture Service (USDA/FAS), Global Agriculture Information Network (GAIN). Washington, DC. USDA/FAS. 2021. Ethiopia Grain and Feed Annual Report, 2021. Report no. ET2021-0008, United States Department of Agriculture, Foreign Agriculture Service (USDA/FAS), Global Agriculture Information Network (GAIN). Washington DC. Wondim, D., T. Tefera, and Y. Tesfaye. 2019. “Value Chain Analysis of Maize: The Case of Dembecha District, West Gojjam Zone, Ethiopia.” International Journal of Arts Humanities and Social Sciences Studies, Vol. 4, No. 8, pp. 63-74. Worku, I., M. Dereje, B. Minten, and K. Hirvonen. 2017. “Diet Transformation in Africa: The Case of Ethiopia.” Agricultural Economics, Vol. 48 (S1), pp. 73–86. World Bank. 2007. Ethiopia: Explaining Food Price Inflation, Policy Note. J. Klugman, P. Dorosh, and J. Loening. Washington, DC: World Bank. World Bank. 2009. Awakening Africa’s Sleeping Giant: Prospects for Commercial Agriculture in the Guinea Savannah Zone and Beyond (Directions in Development). M. Morris, H. Binswanger-Mkhize, and D. Byerlee. Washington, DC: World Bank. World Bank. 2016. Ethiopia’s Great Run: The Growth Acceleration and How to Pace It. Washington, DC: World Bank. World Bank. 2018. Cereal Market Performance in Ethiopia: Policy Implications for Improving Investments in Maize and Wheat Value Chains. Washington, DC: World Bank. Yami, M., F. Meyer, and R. Hassan. 2019. “Should traders be blamed for soaring food prices in Ethiopia? Evidence from wholesale maize markets.” International Food and Agribusiness Management Review, Vol. 23:1, pp. 19-33. Yami, M., F. Meyer, and R. Hassan. 2020. “The impact of production shocks on maize markets in Ethiopia: implications for regional trade and food security.” Agricultural and Food Economics, Vol 8:8, Springer Open Access. 144 Ethiopia Country Economic Memorandum 6 Making the most of natural capital 6.1 Motivation: what went off-track? In recent years, while massive physical capital accumulation was taking place, Ethiopia lost natural capital. Ethiopia is not a mineral-rich country, with renewable natural resources comprising 99 percent of its total natural capital. From 2015 to 2018, Ethiopia’s renewable natural capital declined as the value of pastureland, forest timber, and Protected Areas (PAs) and Other Effective Area-Based Conservation Measures (OECMs)50 dropped (Figure 83).51 Generally, such declines have been brought by land degradation from factors such as free grazing and tree removal from farmland, ploughing on steep slopes, and continuous tillage. Other causes include high population pressure and associated insecure land access and land concessions, deforestation, and forest degradation, as well as droughts and floods that are amplified by climate variability and change. All these drivers are mutually reinforcing. This chapter assesses the current state of natural capital in Ethiopia and identifies areas for action. It begins by providing background information on Ethiopia’s natural capital and on its vulnerability to climate change. It then looks in more detail at the current state of natural and productive landscapes and the cost of environmental degradation in Ethiopia. Other challenges such as solid waste and air pollution are discussed in detail in a companion technical paper to this report. Figure 83. From 2015 to 2018, the value of Ethiopia’s natural capital declined Value of Ethiopia's natural capital 250,000 200,000 constant 2018 US$, millions 150,000 100,000 50,000 0 1995 2000 2005 2010 2015 2018 Forests, timber Forests, non-timber Mangroves Fisheries Protected areas Cropland Pastureland Fossil fuel energy Minerals Source: World Bank (2021a). 50 Such as church forest under religious sites and in situ conservation sites. 51 Natural capital is the stock of renewable and non-renewable resources (e.g., plants, animals, air, water, soils, fossil fuels, minerals) that combine to yield a flow of benefits to people. The concept of natural capital extends beyond nature as a source of raw materials for production (e.g., timber) to include the role of the environment in supporting human well-being through the supply of such important goods and services as clean water, fertile soils and valuable genetic resources (United Nations et al. 2014). 145 Ethiopia Country Economic Memorandum 6.2 The importance of natural capital Climate inaction is likely to expose the global economy to significant losses in the coming decade. About half of global GDP is estimated to depend directly on nature (Dasgupta 2021). The sectors that rely most on nature are forestry, agriculture, fisheries and aquaculture, food and related sectors, and construction, as well as heat, electricity, and related activities. Given this high dependency on nature and the ecosystem services it renders, the world economy is extremely vulnerable to changes in environmental factors. Extreme weather, climate change inaction, human environmental damage, and biodiversity loss are all in the top five most likely risks to the global economy, together with infectious diseases (WEF 2021). The World Bank estimates that global GDP is likely to contract by between 0 percent and 4 percent from 2020 until 2030 if the market is rigid to environmental changes, meaning economic actors do not respond adequately to the expected environmental changes (Johnson et al. 2021). The loss in ecosystem services also increases people’s vulnerability, since biodiversity underpins both current and future generations’ health, welfare, and economic prosperity (OECD 2020). Ethiopia is no exception, as nature loss endangers the pursuit of strong economic growth and poverty reduction. The costs of inaction on biodiversity loss are high. Under a business-as-usual scenario (BAU), the loss of ecosystem services is expected to reduce Ethiopia’s real GDP per capita by 0.5 percent by 2030. Under a partial ecosystem collapse scenario in which ecological tipping points are reached, Ethiopia’s GDP in 2030 would be 14.7 percent lower than under the BAU, and GDP growth would be reduced by half (Johnson et al. 2021). The forestry sector would be particularly affected under a partial ecosystem collapse scenario, with productivity decreasing by 16 precent relative to the baseline and with gross output shrinking by 20 percent. Manufacturing and services outputs would also be severely affected (-42 and -26 percent, respectively). To compound the challenges brought about by the decline in the natural capital, human pressure on biodiversity increases the risk of infectious diseases. Land-use change caused by factors such as agricultural expansion, unsustainable logging, and infrastructure development is the most common driver of infectious disease emergence (World Bank 2020b; OECD 2020). This leads to a downward spiral of more degradation, higher resource use, and overexploitation of wildlife, with more human-animal contact leading to greater risk of zoonotic diseases, thus heightening the possibility of another pandemic. The COVID-19 pandemic has highlighted the strong yet delicate and intricate links between conservation, human health, and the economy (Hockings et al. 2020). Due to the deterioration of economic activity during the global lockdown, income levels declined, reducing the ability to acquire nutritious food and thus compromising human immune systems further. The importance of natural capital and the reliance of economic actors on ecosystems services highlight the need for action to pursue sustainable and resilient development. In this context, three concurrently running strategic themes for action and decision-making have emerged globally: (i) investing in both the protection and expansion of natural capital as a safeguard and a catalyst for human health, economic development, and climate resilience (Deutz et al. 2020; Palahi et al. 2020; World Bank 2020a); (ii) restoring degraded ecosystems to secure the supply of services in perpetuity (Dasgupta 2021; Palahi et al. 2020; World Bank 2020a); and (iii) achieving a low emissions development trajectory and reducing air pollution (Kaza et al. 2018; World Bank and IHME 2016). 6.2.1 Ethiopia’s natural capital The composition of Ethiopia’s total wealth has changed substantially over the past 25 years. Natural capital accounting provides a measure of the size of and change in role of natural assets in the wealth of nations, which is critical given the strong dependence of a country’s development trajectory on them (Box 8). In 146 Ethiopia Country Economic Memorandum Ethiopia, human capital increased its share from 54 percent in 1995 to 69 percent in 2018, while produced capital increased its share in total wealth from 8 percent to 17 percent. These increases have resulted in a decline in the share of renewable natural capital from 44 percent to 17 percent. These structural changes are to be expected, given the transition toward a more industrialized country. Among peer countries, Ethiopia’s share of natural capital in total wealth in 2018 was like that of Kenya and Vietnam and below Cambodia, Rwanda, and Tanzania Table 16. In per capita terms, however, only Uganda and Bangladesh have lower natural capital. Box 8. The wealth of a nation Although a macroeconomic indicator like GDP provides an important measure of economic progress, it measures only income and production and does not reflect changes in the underlying asset base. Used alone, GDP may provide misleading signals about the health of an economy. It does not reflect the depreciation and depletion of assets or whether the growth comes at the expense of natural capital. More inclusive measures, such as the adjusted net savings rate (or genuine savings rate) and comprehensive wealth accounts, are required to fully capture whether the growth of an economy is sustainable or if the country is effectively living off depleting its capital. Comprehensive wealth accounts include produced capital, natural capital, human capital, and net foreign assets. To fully account for the contribution of natural capital to economic growth and wealth requires natural capital accounting (World Bank 2021). Table 16. Ethiopia’s renewable natural capital per capita is below most peer countries Natural capital relative to total wealth in a selection of countries (2018) Total wealth per Renewable natural Natural capital as % capita capital per capita total wealth Ethiopia 10,792 1,793 16.6% Bangladesh 19,272 1,167 6.1% Cambodia 18,396 4,696 25.5% Kenya 22,047 3,617 16.4% Uganda 10,403 1,387 13.3% Rwanda 11,318 2,886 25.5% Tanzania 15,375 3,118 20.3% Vietnam 34,082 5,829 17.1% Source: World Bank (2021). Note: All values expressed in 2018 US$ terms. It is concerning that, in addition to the shifts in relative shares, Ethiopia has experienced an absolute drop in the value of natural capital. As shown in Figure 83 above, Ethiopia’s natural capital was valued at about US$196 billion (or US$1,793 per capita) in 2018, down from about US$220 billion (or US$2,182 per capita) in 2015. Ethiopia experienced declines of 28 percent in the value of protected areas (US$531 to US$383 per capita), 28 percent in pastureland (US$576 to US$414 per capita), and 17 percent in forest timber (from US$178 to US$148 per capita). As discussed later in this chapter, the reasons for this degradation are multi- fold, induced by policy, population growth, land fragmentation, and land-use practices. Ethiopia relies heavily on renewable resources, which cannot be replaced by produced or human capital. Ethiopia is not a mineral-rich country, with renewable natural resources comprising 99 percent of its total natural capital. Forests and land can be destroyed and degraded, which would affect both their productivity as well as the livelihoods of those dependent on them. At the same time, having a broad base of renewable resources can offer an important advantage: just a small improvement in the rate of change has the potential to 147 Ethiopia Country Economic Memorandum have a significant impact on the welfare and well-being of all Ethiopians for generations to come. Once managed sustainably, renewable natural capital can produce benefits in perpetuity. 6.2.2 Climate impacts and the importance of adequate land use In the absence of proactive and anticipatory measures to reduce vulnerability, climate change will have serious adverse effects on the agricultural sector, exacerbating the cost of degradation. Due to the terrain and a long history of cultivation as well as other factors, land degradation —especially soil erosion and nutrient depletion—is a key environmental problem that has curtailed agricultural productivity in Ethiopia. A large body of evidence indicates that the adverse impacts of climate change are going to worsen, and they will affect poor countries disproportionately. Ethiopia is particularly vulnerable because it is heavily dependent on rain-fed agriculture. In the absence of major efforts to mitigate those risks, increases in spells and intensity of drought due to climate change would have severe impacts on the economy. At the same time, an increase in the incidence of floods and landslides due to climate change threatens to damage crops and exacerbate soil erosion, in addition to causing loss of arable land. If the current decline in average annual rainfall continues without major adaptation, Ethiopia could lose more than 6 percent of each year’s agricultural output. If no adaptation measures are taken, losses of area suitable for cultivation could amount to over 18 percent for maize, 11 percent for teff, and 37 percent for barley (Ethiopian Panel on Climate Change 2015). The livestock sector is also threatened by climate change; problems of feed shortage, water shortage, reduced productivity, and decreased mature weight will arise or be exacerbated. In terms of climate change mitigation, Ethiopia’s success in achieving its emission reduction targets depends on improvements in land use practices. Under the updated baseline currently under consideration that revises the nationally determined contributions (NDCs) under the Paris agreement, Ethiopia would need to reduce its emissions by 14 percent in 2030, compared to the level in 2010 (GoE and World Bank, 2021). The conditional target, which comprises the implementation of additional measures, would be even more ambitious, requiring a 68.8 percent net reduction in emissions.52 Since Ethiopia generates more than 95 percent of its energy from clean sources and is still at its initial stages of industrialization, the focus to achieve the targets would need to be on land use, including the management of forests and other biodiversity and the adoption of sustainable agriculture and livestock practices through employing ecosystem-based adaptation53 (Figure 84). 52 The level of ambition that can be achieved in the NDCs is clearly demarcated by unconditional and conditional targets. The unconditional target refers to the mitigation and adaptation interventions and contributions that can be achieved by the country itself with domestic funding. The conditional target comprises additional mitigation and adaptation interventions that can be achieved with international / donor support and financing. 53 Ecosystem-based adaptation is the use of biodiversity and ecosystem services as part of an overall adaptation strategy to help people to adapt to the adverse effects of climate change. 148 Ethiopia Country Economic Memorandum Figure 84. Ethiopia´s unconditional and conditional emission pathways, with contributions by sector Ethiopia greenhouse gas emissions and targets 450 400 350 300 MtCO2e 250 200 150 100 50 0 Historical emissions in BAU emissions in 2030 2030 Conditional 2030 Unconditional 2010 target target Land use change and forestry Energy Industry Waste Managed soils Livestock Targets Needed reduction Source: GoE and World Bank (2021). Note: The “NDC Conditional” bar represents the combined impact of unconditional and conditional elements. Climate change adaptation measures will also be crucial to mitigate the impacts. Sustainable land management and climate-smart agriculture practices are critical for improving climate resilience, and they are also tightly linked to both biodiversity conservation and landscape restoration. In addition, climate considerations must be built into public investments across sectors, as weather events have negative impacts on road transport and energy infrastructure (e.g., river flow disruptions affecting hydroelectricity production). It is recommended that interventions are based on a holistic perspective on land-use policy and planning that encompasses the mobility of people and goods, cities, and living conditions, as well as the integrated design of infrastructure. 6.2.3 The authorities’ response to the challenge Authorities have set ambitious goals for protecting and developing renewable natural capital. In its 2011 Climate Resilient Green Economy (CRGE) Strategy, Ethiopia seeks to mitigate the adverse effects of climate change while building a green economy that will help realize its ambition of reaching middle-income status by 2025. The latest Ten-Year Plan (PDC 2021) includes “Resilient Green Economy” among its ten strategic pillars and lists a series of goals that are consistent with the CRGE and with the challenges listed above, including fighting climate change, protecting and restoring marine and terrestrial ecosystems, and reversing land degradation. However, implementation of the plans has proven to be challenging. The Government commenced a process to integrate the CRGE vision into the National Integrated Land-Use Plan and Policy (NILUPP). However, a series of legacy challenges persist, including the lack of autonomous institution at national level to coordinate and lead the NILUPP initiative. While efforts have been made to formulate land-use plans at different government levels, progress has been hindered by several factors: the lack of a national land use policy to guide plan development; weak implementation and insufficient monitoring and evaluation mechanisms; the absence of a standard and comprehensive land use planning approach; and the lack of attention to existing land rights and administrative boundaries in land use planning, enforcement and regulatory mechanisms at national and regional level. Factors undermining the success of past land-use planning initiatives included insufficient awareness and sensitization among decision makers, lack of involvement of the major stakeholders 149 Ethiopia Country Economic Memorandum (beneficiaries), poor coordination among different government agencies, lack of a legal framework, and limited implementation capacity. Above all, the lack of a national institution to coordinate, lead, and follow up on implementation of the land-use plans was a major gap (Haddis et al. 2017). Since sustainably managing competing land uses is crucial for Ethiopia to achieve the goals set under its plans and strategies (Danyo et al. 2017), both from a natural capital perspective and from a climate change mitigation point of view, challenges in this area are analyzed further in the coming sections. Weak land tenure and land use planning linkages encourage the spread of informal/illegal settlements and transactions in informal land markets that hinder combating climate risks. In urban areas, 30 percent of residential holdings are informal, and their residents often lack durable houses, adequate living space, and access to safe water and sanitation, making them more vulnerable to natural disasters. Many informal settlements are concentrated on the peri-urban periphery, where properties are acquired informally from nearby farmers and built illegally or rented informally to avoid the complex formal land market procedures. Informal settlements are also often located in the high-risk areas of a city—such as wetlands, floodplains, landfills, garbage dumps, and rocky areas (World Bank 2017). In rural areas, regulatory restrictions, information asymmetry, cultural and social pressure, and transaction costs preclude the efficient transfer of land use rights, leading to insecurity of tenure and parcel fragmentation. This, in turn, undermines landholders’ incentives to adopt sustainable land management practices with long-term payoffs and complicates land use planning and enforcement due to the large number of stakeholders. Where land use planning is implemented, it often lacks sufficient attention to landholders’ tenure security and thus their incentives to maximize land use efficiency, productivity gains, environmental sustainability, and climate adaptation and resilience. 6.3 The costs and drivers of land degradation The lack of proper land-use planning in Ethiopia results in degradation of natural capital. Land in Ethiopia is used for numerous sectors: agricultural production (smallholder and some commercial food production, livestock grazing and fodder, woodlots for household energy and construction material, and feedstocks for biofuels), forestry, protected areas for conservation and tourism, urban settlements, rural settlements, and, increasingly, mining (Danyo et al. 2017). Lack of integrated land-use planning leads to unregulated land use conversions, increasing land degradation, and vulnerability to disasters. If land-use planning was properly coordinated, potential inconsistencies and conflict could be reduced. Land degradation hampers Ethiopia’s achievement of its stated objectives, and the costs have been much higher than in comparator African countries.54 The cost of land degradation during 2001-09, estimated at US$34.8 billion (about 22.5 percent of GDP), or US$38.5/ha/year, was significantly higher than in Kenya, Malawi, and Tanzania (Table 17). The cost of land degradation would have been even higher if the off-site environmental effects of social erosion and deforestation were considered. The off-site effects of degradation include impacts on the biodiversity of the country and on many ecosystem services (e.g., nutrient cycling and soil formation), regulating services (e.g., flood regulation and water purification), and cultural, spiritual, and recreational services for present and future generations. 54 Degradation is calculated as the difference between the total economic value (TEV) before and after the land cover change. 150 Ethiopia Country Economic Memorandum Table 17. The cost of land degradation has been higher in Ethiopia than in comparator African countries Cost of land degradation, Ethiopia and comparators (2001-2009) Cost of land Annual cost of Cost of land Annual cost of land degradation: land degradation degradation: degradation 2001-2009 % of GDP US$, billions % US$/ha/year Ethiopia 34.825 4.353 22.5 38.49 Kenya 10.645 1.331 4.9 22.88 Malawi 1.980 0.248 6.8 21.01 Tanzania 18.474 2.309 13.7 24.53 Source: Kirui and Mirzabaev (2015). Ethiopia has experienced large losses in forest and grassland cover. Kirui and Mirzabaev (2015) found that from 2001 to 2009, Ethiopia lost 25.8 percent of its forest cover (approximately 1.4 million ha) and 10.6 percent of its grasslands (more than 3 million ha) due to an increase of 5.4 million ha for croplands and shrublands combined (Table 18). Losses in forests and grasslands are the largest among comparator countries in the table. Such land transformation contributes to the loss of topsoil and a loss in ecosystem goods and services, reducing system resilience and increasing the vulnerability of people. Table 18. In the 2000s, Ethiopia lost forest cover and grasslands due to an increase in croplands and shrublands Change in land cover, Ethiopia and comparators (2001-2009) Unit Forest Cropland Grassland Woodland Shrubland Bareland Ethiopia Ha -1,412,899 2,783,381 -3,035,811 -333,918 2,753,523 -696,317 % -25.8 32.7 -10.6 -1.5 6.6 -12.3 Kenya Ha 371,322 9,074 7036319 125,040 -6,069,477 -667,004 % -22.7 0.4 32.1 3.3 -24.0 -32.3 Malawi Ha 30,597 -52,749 1042056 -959,338 -65,021 6,341 % 7.7 -33.6 18.3 -31.0 81.4 56.9 Tanzania Ha -684,551 -1,273,497 5591728 -2,916,689 -582,400 232,928 % -17.1 -41.3 10.9 -10.1 -59.4 28.6 Source: Kirui and Mirzabaev (2015). Little of the forest loss has led to permanent agriculture, and much of the deforested land remains degraded. Deforestation is a process in which trees are first removed for fuel or (more rarely) for industrial use, then the land area that has lost its tree cover is converted for agriculture, pasture, or left idle with little or no economic use. While the actual driver for deforestation is always site-specific, some the major drivers of deforestation and forest degradation include agricultural expansion, overexploitation, urbanization, fire incidence, and habitat expansion, which are driven by human population growth (Zegeye 2017). It is worth noting that only about a quarter of the deforested forestland has been converted to permanent agricultural land that would contribute to increasing natural capital (Figure 85). 151 Ethiopia Country Economic Memorandum Figure 85. Only a small percentage of deforested forest land is converted to permanent agricultural land Land-uses replacing forest over the period 2000-2013, as percentage of the total forest loss over the period 4 7 27 17 23 24 Open woodland Agriculture Grassland Shrubland Bareland Other Source: Modified from Federal Democratic Republic of Ethiopia (2017). Several factors have driven land degradation in Ethiopia: • The advance of croplands and commercial agriculture. Since the 1970s, large-scale investments in commercial agriculture have led to the displacement of communities, which increases the degradation of the land onto which the communities have been displaced. The shift in property rights from common land used by pastoralists to private land in large‐scale plantations has thus aggravated degradation in semiarid drylands (Bekele et al. 2020; Nalepa et al. 2017). Soil erosion rates vary greatly across the country, reaching ranges of 61.3 to 180.4 ton/ha/year in the northwest, with the highlands of the Amhara region being the worst affected (Hurni et al. 2015). • Unsustainable land-use practices. Free grazing and tree removal from farmland, ploughing on steep land, and continuous tillage accelerate land degradation. • Population pressures. Except in urban centers where master plans and zoning are in effect, people in Ethiopia have been using land in an unplanned and uncontrolled manner. The rate of expansion of urban centers, industries, agriculture, agro-industries, and hotels has been fast,55 and this rapid change is not always happening with due regard to optimal land use and preservation considerations. Consequently, important wetland ecosystems, high-potential arable lands, grasslands, and forest areas 55 For example, in the Chemoga basin, Finchaa catchment, Gojeb river catchment, and Muga watershed, the expansion of built-up areas and agricultural lands is the major driver of land fragmentation and degradation and the intensification of natural resource use for firewood, charcoal, timber, and housing materials (Damtea et al. 2020; Didaba et al. 2020; Dagnachew et al. 2020; Belay and Mengistu 2019). 152 Ethiopia Country Economic Memorandum have been converted to urban centers and industrial sites, with undesirable environmental and social consequences. Going forward, population growth is likely to exert enormous pressure on the country’s natural resources and institutions, as Ethiopia may reach a population of nearly 200 million by 2050 (UNDESA 2017). Moreover, environmental factors will affect mobility patterns in the future,56 and in ways that are hard to predict. • Property rights. Despite the ongoing efforts to issue land certificates, lack of land tenure and common property rights remains a challenge (see section 6.5.1). • Governance challenges. The widening power gap between the customary and statutory governance systems reinforces unsustainable land use. While on paper there is regulation to protect forests and nature, it was adopted without sufficient consultation and sensitization about the use of natural resources among communities.57 In practice there is corruption and weak law enforcement, which is made worse by lack of effective property rights. The combination has resulted in deforestation and a loss of biodiversity, due to an ineffective local forest governance structure (Girma and Beyene 2015). 6.4 The potential of Ethiopia’s natural and productive landscapes This section focuses on the distinguishing features and challenges of different landscapes where investment in conservation and restoration would contribute to a resilient and sustainable economy. Investments in Ethiopia’s forests (both natural and planted commercial) and protected areas, in particular, are expected to help counter some of the costs highlighted in the previous section. This report does not focus on the sustainability of croplands and pasturelands, since they have been examined extensively in other studies (ELD Initiative & UNEP 2015; Hurni et al. 2015; Molla and Woldeyes 2020). 6.4.1 Natural forests The vast majority (80 percent) of Ethiopia’s natural forests are tropical dry forests (FAO 2020), which are very valuable in several respects. Tropical dry forests are precious for their biodiversity, carbon sequestration, and other environmental externalities; they are also valuable to local communities for the timber, gums, and resins, some of which are exported. While much of the use is informal and not properly recorded, it has been estimated that around 75 million Ethiopians use forest resources for their livelihood at least part-time (Johanson and Mengistu 2013). The land area covered by natural forest has declined steadily since the 1990s, despite the policies adopted. Even optimistic estimates put forest cover at only about 15 percent (FDRE 2017). The extension of Ethiopia’s naturally regenerating forests declined from about 19 million ha in 1990 to almost 16 million ha in 2020 (FAO 2020) (Table 19).58 This decline implies a loss of about 0.6 percent of forest cover per year. 56 Fransen and Kuschminder (2009) confirmed and elaborated on drought-induced mobility over the past 50 years. Major droughts occurred in 1964-65, 1972-73, 1982-83, 1991-93, 1997-98, and as recently as 2015-16, leading to significant population mobility (see also Comenetz and Caviedes 2002). 57 One major driver of degradation is the institutional challenges that have arisen in the process of decentralizing forest management. Such decentralizations were a gradual process commencing in the 1970s with the establishment of state- owned farms, later to be replaced by foreign investments. The repurposing and demarcation of regional forest priority areas was done without adequate public involvement and participation in the process. 58 The national definition of forest in Ethiopia: “Land spanning at least 0.5 ha covered by trees and bamboo, attaining a height of at least 2m and a canopy cover of at least 20 percent or trees with the potential to reach these thresholds in situ in due course” (Minutes of forest sector management, MEFCC, Feb. 2015 as quoted in MEFCC 2016). 153 Ethiopia Country Economic Memorandum Hundreds of thousands of hectares of bamboo forest have also been lost, negatively affecting livelihoods and biodiversity.59 The degradation and loss of forests are the main contributing factor to the frequent and devastating flash floods in regions such as Afar and Somali. Over the years, these floods have caused significant loss of life and property (Atmadja et al. 2019). Table 19. The spatial extent of naturally regenerating forests in Ethiopia has declined significantly Forested area: 1,000 ha 1990 2000 2010 2015 2020 Naturally regenerating forest 18,919 18,189 17,058 16,462 15,984 % change per annum -0.4% -0.6% -0.7% -0.6% Source: FAO (2020). Unsustainable land use practices and poor law enforcement are causing deforestation and degradation. The direct drivers of forest change are often free livestock grazing, wood-based biofuels (fuelwood collection and charcoal production), farmland expansion, fires, and wood collection for construction. Over 90 percent of wood-based bioenergy comes from natural forests (Box 9). Households and communities also collect non-commercial non-timber forest products (NTFPs) such as wild coffee, herbs, honey, spices, liana, and fodder.60 Because these are informally collected and since forests are not significantly contributing to the official economy with goods or services (e.g., tourism), they are not actively managed, and they have become de facto an open access resource. Problems relating to unsecured land tenure and poor law enforcement compound this challenge. Since 2018, communities have been allowed to have forest ownership rights in state forests. While the legislation is conducive to participatory and inclusive management of natural forests, implementation of these regulations is largely insufficient, exacerbating the open access situation. Improved management of the forest resources would require clarification of community rights to adjacent forests, leading to improved monitoring of their use (McLain et al. 2019; MEFCC 2018). Box 9. The use of biomass for fuel in Ethiopia Only 5 percent of Ethiopia’s population has access to clean cooking fuels and technologies.61 Traditional biomass (wood, charcoal, dung) accounts for more than 90 percent of total primary energy use in Ethiopian households—about 84 percent in urban areas and 99 percent in rural areas (Johanson and Mengistu 2013). According to the multi-tier framework (MTF) household survey data (2017), the vast majority (77 percent) of households use three-stone or other self-built stoves, while about 22 percent of households use a manufactured stove for their cooking needs. In rural areas, 67 percent of households use a three-stone stove as their primary stove, while 55 percent of urban households use a manufactured stove. For most Ethiopian households, firewood is the main fuel (71 percent), followed by charcoal (21 percent). About 4 percent of households use electricity, and a smaller percentage of households use biogas. Without meeting the clean cooking target under Sustainable Development Goal 7 (SDG 7.1.2), the cost of inaction—driven by negative externalities for health, gender, and climate—would total US$35.44 billion per year in Ethiopia. Most of that cost (US$ 21.9 billion annually) is due to lost productivity from extended time spent on cooking-related tasks, including fuel collection, cooking, and stove cleaning. The health impact cost is estimated at US$7.76 billion per year, calculated by quantifying the deaths and disability-adjusted life years (DALYs) linked to household air pollution (HAP) produced by stoves and fuels. It is estimated that around 59 The change in spatial extent does not imply that there is no woody biomass left on the land parcel, only that the land cover has changed to something else that can no longer be defined as forest 60 The value of NTFPs can be significant. In a case from southwestern Ethiopia, the share of NTFPs was 26 percent of household income for the poorest 20 percent of the population, with wood fuels providing another 26 percent (Beyene et al. 2020). 61 Household Energy Database, WHO, January 2020. 154 Ethiopia Country Economic Memorandum 32,771 premature deaths and 1,644,047 DALYs are attributable to HAP in Ethiopia (HEI 2020). Finally, the climate impact cost is estimated at US$5.78 billion per year, resulting from 110.98 million tons of CO2 emissions per year in rural areas and 14.82 million tons of CO2 emissions in urban areas. Unsurprisingly, the updated NDC states that the largest driver of LUCF emissions is biomass energy use for cooking and baking (GoE 2021). 6.4.2 Commercial forests While production forestry plantations are still a relatively small land-use category in Ethiopia, the sector could offer large economic and environmental value if efficiently managed. According to Indufor (2016), production plantation areas in Ethiopia cover approximately 190,000 hectares, or about 1 percent of the forest area. Plantations are managed by several state forest enterprises (SFEs). This section discusses the insights from an assessment of SFEs in Ethiopia prepared for this Country Economic Memorandum. Overall, the management of these plantations is often inadequate, leading to the underutilization of their economic and environmental potential. Even if the commercial production forestry sector were to grow, energy can be expected to remain the main wood use in Ethiopia. Currently, around 97 percent of timber removals are for energy,62 which is not a high-value wood product. Most of the feedstock for energy comes from natural forests (Figure 86), often with adverse environmental impacts as discussed earlier. Formalization and commercialization of the wood energy subsector could provide opportunities for increasing efficiency and sustainability both in plantation forestry and in natural forest management. The SFE plantations generate revenue for the regional governments but have low productivity. For example, the production plantation profits of the Oromia Forest and Wildlife Enterprise (OFWE) have been in the range of 50-60 million birr (US$1.2-1.5 million) annually. With a plantation area of 56,000-58,000 ha, profit per hectare of plantation is just around US$23. The establishment cost of the plantations is estimated at US$2,200/ha, meaning that the return on investment is only about 1 percent, much lower than international reference levels for plantation investments.63 One reason for the low returns is the low productivity of the current plantations. The current mean annual increment (MAI)64 of the plantations is estimated to be 7m3/ha (pines) and 12 m3/ha (eucalyptus). With improved management and genetic material, this could be doubled or tripled. 62 Data for 2019. FAOSTAT (http://www.fao.org/faostat/en/#data/FO). 63 This is a simplification, and additional financial data from the SFEs in not readily available. As a comparison, a case study from Vietnam found internal rates of return (IRRs) for smallholder plantations for Acacia magnium and Eucalyptus to be 34.8 percent and 32.3 percent, respectively. In South Africa, an IRR of 24.8 percent was recorded for Eucalyptus spp. but only 3.4 percent for pine (Pinus spp.). See Cuong et al (2020) and Cubbage et al. (2020). 64 MAI measures the average annual volume of gross increment over stand rotation period less that of natural losses for the growing stock. In practice, this means the average annual yield of harvestable wood from the plantations. 155 Ethiopia Country Economic Memorandum Figure 86. Most of the feedstock for energy comes from natural forests Wood energy – sources of supply and use, 2019 120 7 5 Million m3 (roundwood equivalent) 100 80 60 109 111 40 20 - Fuelwood supply Fuelwood use Natural forest, woodlands, shrublands Private and municipal woodlots SFE plantations Industry residues Fuelwood Household charcoal Industry charcoal Source: Data collected by Unique Consulting for the CEM. While forest production does provide some income-generating opportunities, there is significant underemployment in the sector. Most income-generating opportunities in forestry production occur in the informal sphere, especially in activities such as woodfuel and charcoal production (around 276,000 persons) and small-scale woodlot planting (around 800,000 hectares) (MoEFC 2017). Formal employment (part- and full-time) in forest production is roughly 15,000 full time equivalents and mainly generated by the SFEs, either by direct employment or outsourcing to local communities. However, the quality of employment is a concern, as the sector is characterized by low wages and reliance on casual jobs without formal contracts, all in the context of a weak rural labor market with few employment opportunities (Negedea et al. 2015). The downstream wood-based industries generated 770,000 employment opportunities during the 2017/18 fiscal year. Of these, around 95 percent were in microenterprises such as furniture production (MoEFC 2017). The SFEs dominate roundwood supply to large wood industries (sawmilling) and also participate in the wood products market through own processing industries, which crowds out the private sector. The private industry actors surveyed consider this a monopoly, and the resulting supply chain and market structures are perceived as a key entry barrier. Private firms are not willing to enter a market in which they would compete in the product market with their main—or even only—raw material supplier. The SFEs are not able to lease out their lands to private operators, as these are officially concessions from the regional government and sub-leases are not allowed. In addition, investment-ready land for greenfield plantations is not available, and the establishment of large-scale outgrower schemes with small woodlot owners is beyond the capacity of the private forest industry. In the processing industries, the SFEs prefer joint ventures in which the private partner is a minority shareholder, which is seen as too risky for commercial investors. These limitations have led to a situation in which private investments in commercial forest-based value chains can only be found in the board industry, not in plantations or sawmilling (Table 20). 156 Ethiopia Country Economic Memorandum Table 20. The private sector is not involved in industrial forest plantations or sawmills and has only limited involvement in boards Forest SFE vs. private sector capacity in Amhara and Oromia Forest SFEs Private firms Industrial forest plantations 83,000 ha 0 ha Sawmills* 23 locations 0 locations Plywood, particle board, & MDF 2 operational 3 operational 3 advanced planning 3 advanced planning or 3 early planning under construction** Source: Data collected by Unique Consulting for the CEM. Note: * Sawmills with an installed capacity of >5,000 m³ roundwood intake per year. **Partly in joint venture with the Amhara Forest Enterprise (AFE). Production forestry and wood processing could play a significant role in the Ethiopian economy but are hampered by numerous technical, institutional, and regulatory impediments. Ethiopia is a net importer of wood products, and the overall trends are largely driven by the construction industry (Figure 87).65 Increased wood production and processing could mitigate the current trade deficit in forest products and provide additional employment; currently, however, the wood industry sees little new investments, and existing production units operate below their installed capacity largely due to lack of access to raw material. The few private wood processing units procure their raw material mainly from private woodlots, not from the SFEs. Globally, wood industries prefer to obtain their raw material from their own plantations or through long-term supply contracts from large formal sector wood producers, while additional marginal volumes may be procured from smallholders and through outgrower schemes.66 Private and community woodlots are a notable source of (semi) industrial roundwood, but their significant potential has not yet been tapped. Currently, the sector is highly disorganized, and woodlot owners are not well-linked to processing industries. There is little information on the exact locations, species composition, and sustainable yield of these important resources. 65 The drop in imports in 2018–2019 was largely due to a lack of foreign exchange allocation to the sector, not increased competition from domestic producers. 66 This applies particularly in developing and emerging economies. The situation is different in countries with well- established roundwood markets (e.g., Nordic countries), where firms are more willing to rely on wood supply from non-industrial private forest owners. 157 Ethiopia Country Economic Memorandum Figure 87. Import volumes in the forest sector largely follow trends in the construction industry (2000-2019) Imports of wood and construction sector activity in Ethiopia 180,000 80 160,000 60 140,000 GDP change p.a. (%) 120,000 40 Product units 100,000 20 80,000 60,000 0 40,000 -20 20,000 - -40 Imports wood-based panels (m³) Imports sawnwood (m³) Imports industrial roundwood (m³) Construction sector GDP change p.a. (%) (Right-hand axis) Source: Data collected by Unique Consulting for the CEM. The Ethiopian wood industry has been growing rapidly but with low levels of modern investment and value added, likely due to unpredictable raw material supply. While Ethiopia’s furniture and wood industries have been growing fast in terms of value added, this is partly explained by a low starting level. These industries continue to lag those in Indonesia and Peru in terms of value addition. Particularly in the furniture sector, growth has taken place in the micro industries segment and not in large formal enterprises, and the industry still has low levels of value addition (Figure 88). The gross value added in any industry depends on numerous factors, but global experience indicates that industries are reluctant to invest in processing facilities if raw material supply is not reliable and predictable. Particularly in more value-added segments, these aspects of raw material supply may be more important than the nominal price level. Figure 88. Gross value added per employee in wood industries in Ethiopia remains low Growth in gross value added (GVA) in wood-based industries (2011-2017) 30,000 Peru furniture 25,000 US$ GVA/employee (2017) Peru wood industry 20,000 Indonesia wood industry 15,000 Ethiopia furniture 10,000 5,000 Indonesia furniture Ethiopia wood industry - - 50 100 150 200 250 Growth in GVA: 2011-2017 (%) Source: Data collected by Unique Consulting for the CEM. 158 Ethiopia Country Economic Memorandum A series of Public-Private Dialogues (PPDs) in 2017 identified three subsectors as being most promising for wood sector development. In 2017, the World Bank and the Ethiopian Chamber of Commerce & Sectoral Association (ECCSA) organized a series of PPDs to identify the key impediments to wood sector development and the areas of potential. This was supported by technical analysis and a formal position paper by ECCSA summarizing the main issues identified in the PPD (GoE 2017; ECSSA 2016). In the process, three subsectors were identified as being most promising in terms of generating value addition and employment: construction industry, furniture industry, and production of utility poles. To reach their full potential, considerable investments and an increase in raw material supply would be needed (Table 21). A significant amount of new employment would be generated as a result.67 Table 21. A significant number of jobs can be created in commercial forestry under reasonable investments Commercial forestry: Investment needs and employment potential by 2033 Construction Furniture Utility Poles Total Roundwood gap (thousand m³) 3,400 500 500 4,400 Area required in 1,000 ha 245 35 30 310 Plantation investment (US$, millions) 196 28 24 248 Investment processing (US$, millions) 68 19 20 107 Total investment (US$, millions) 265 47 44 356 Additional employment (FTE) 64,150 8,703 5,034 77,887 Source: ECSSA (2016). The PPD and technical analysis identified specific challenges and areas in which public investments and policy actions are necessary. • Forest policy and the regulatory environment are fragmented, and their implementation is lacking. Formal private sector operators face a costly and difficult-to-deal-with administration. • Lack of access to suitable land for commercial forest development prevents private investments in production forestry. Land tenure is unclear, leading to risks of community conflict. In theory, land is offered on favorable terms (e.g., tax holidays), but actual demarcated plantation sites are not made available to private operators. • Access to finance is a major constraint, and high (perceived) risk and short financing periods make debt capital unavailable. This is a common issue in forest financing and not specific to Ethiopia. • The State plays a major role in the Ethiopian economy, and the absence of a PPP framework specific to the forest sector limits collaboration, partnerships, and joint ventures between the SFEs and private firms. • Forest sector education and research and development (R&D) need strengthening to improve value addition and productivity, as demonstrated in part by the low levels of tree growth. This is caused not only by inadequate management of the stands but also by lack of knowledge on the genetic material. A lack of skills and scientific infrastructure also leads to poor outcomes. 67 Indufor (2016) presents another estimate for long-term employment potential in the forest and wood industry sector, concluding that the employment potential is considerable. Direct employment potential is estimated at 7,000 people in plantations and 3,000 in industry, leading to 10,000 jobs directly with notable employment (during construction of the new processing units) and another 60,000 jobs indirectly in value chains by the 2030s. Total job creation is estimated to be 80,000 in 2031–2050. All these employment estimates are for the formal sector and do not include wood energy. 159 Ethiopia Country Economic Memorandum • Overall, the forest sector is seen as poorly organized, with weak or non-existent industry associations, poorly developed markets, lack of public information and awareness, and generally having low productivity at all stages of the value chains. 6.4.3 Protected areas In providing a wide range of ecosystem services, protected areas play an important role in society from a socio-economic development perspective, and significant investments are needed to maintain them (Van Zyl 2015). Protected areas are also one of the most effective ways to mitigate climate change and safeguard ecosystem services that support livelihoods as well as support ecosystem-based adaptation (Francesca et al. 2018; IUCN 2012). Given the pressures faced by protected areas both locally and globally, there is a need to unlock their value much more effectively, which costs money. It is estimated that more than US$1 billion per annum is required to maintain the effectiveness of protected areas in Africa (Lindsey et al. 2018). Thus, it is essential to estimate the economic value of protected areas ecosystem services to facilitate the sustainable development of the economy, as well as its transition onto a nature-smart trajectory (Soe Zin et al. 2019). Ethiopia has been expanding its protected areas and OECMs. Currently, about 14 percent of the total land area in Ethiopia is officially demarcated as protected areas (Habtamu et al. 2019). These protected areas include 3 wildlife reserves, 3 sanctuaries, 6 open hunting areas, 6 community conservation areas, 20 controlled hunting areas, 21 national parks, and 58 national forest protected areas, of which 37 are protected forests (Young 2012). However, weak institutional capacity to enforce protection and low investment result in numerous challenges. Ethiopia’s protected areas (and OECMs) are surrounded by and utilized extensively by agrarian and pastoralist communities in search of arable land, pastureland, fuel wood as energy sources, etc. (Amare 2015; Wassie 2020). Despite being protected, in practice the conversion of natural habitat to agricultural land, overgrazing, weak institutional capacity, human-wildlife conflict, poaching, and expansion of invasive species are major threats and compromise the ability of natural areas to provide benefits to the economy and environment (EWCA 2009). In addition, the habitats and wildlife in the national parks and protected areas are frequently exposed to fires that affect wildlife migration, which is significant as Ethiopia boasts the world’s second-largest wildlife migration (Amare 2015; Naidoo et al. 2016). To make matters worse, government- banked investment land that is allocated and leased to investors has overlaps with protected areas, contributing to national parks degradation (Nalepa et al. 2017). Limited financial resources have also hindered the management of protected areas and OECMs, affecting the ability to implement effective operational activities, infrastructure development, staff capacity development training, community outreach, and tourism and research development (Namaga et al. 2020). Much of this degradation is the result of many of the protected areas existing only in paper; de facto there is open access and no oversight, which opens the door for excessive grazing and overexploitation of resources, compromising the value the parks offer to society and the local communities (Van Zyl 2015; Mengist 2020). As a result, the value of protected areas in Ethiopia has been on decline and could decline further. The value of the stored above- and below-ground carbon in the protected areas managed by the Ethiopian Wildlife Conservation Authority (EWCA) is estimated to be about US$938 million, but around US$19 million/year is lost due to deforestation—this is one of the drivers of the decline in the natural capital wealth of the country. Similarly, the economic value of biodiversity in protected areas has been estimated at 6.5 billion birr/year 160 Ethiopia Country Economic Memorandum (US$325 million/year) (Figure 89),68 but this figure could decline to 3.7 billion birr within 20 years due to the destruction of natural ecosystems (Van Zyl 2015; see also Namaga et al. 2020). Figure 89. Watershed protections services and harvesting of natural products comprise two thirds of protected areas value Current total protected areas system value = ~6.5 billion birr/year Pollination and pest control 3% Tourism and recreation 8% Existence and cultural values 4% Watershed protection services 42% Grazing 18% Harvesting of natural products 21% Medicinal plant harvesting 4% Source: Van Zyl (2015). An assessment of two national parks conducted for this report suggests that despite the degradation experienced over the past decade, investing in the management of protected areas still renders significant benefits. An assessment of ecosystem services in the Nechisar National Park and the Gambella National Park found that their value had declined over the past decade, with a net loss of about US$500/ha/year in both cases. This was mainly driven by degradation in high forests and wetlands. Despite this loss, the return on investment in these parks was still significant: US$62.10 per dollar invested in Nechisar and US$194.2069 per dollar invested in Gambella (Namaga et al. 2020). This finding highlights the need for improved management as well as the potential returns that investments in natural capital could have. 6.5 Policy options for rebuilding and leveraging natural capital to achieve sustainable and green growth Based on the findings presented above, this section provides policy options that could help Ethiopia restore and invest in natural capital, which is expected to have positive social, economic, and 68 This figure was calculated by estimating, using market values, the grazing value of the parks by animals and the value of harvesting natural products, including medicinal products. Using a benefit transfer method from other eastern African studies, the value of the watershed protection and water provisioning services, carbon sequestration, and pollination and pest control were calculated. 69 Calculated as the total economic value of the respective protected areas divided by the management cost. 161 Ethiopia Country Economic Memorandum biodiversity outcomes. Interventions will be needed to preserve and revamp those assets that have experienced an erosion in their value (e.g., protected areas, forests). It will also be important to adopt more integrated approaches for land-use planning in the country. Table 22 at the end of this section summarizes the policy options, and discussion on the financial instruments that can be leveraged to implement some of these recommendations can be found in a companion background note to this report. 6.5.1 National, integrated land-use planning Land policy and land administration in Ethiopia are fragmented, and property rights registration and enforcement are weak. The sale or exchange of lands is not allowed per Article 40 of the 1995 Constitution, and land administration is fragmented, with different legal and institutional frameworks for urban and rural lands and poor coordination across local and federal institutions.70 In rural Ethiopia, of the 50 million estimated parcels of landholdings, about 22 million parcels (44 percent) have been adjudicated and mapped for certification, of which 18 million parcels (36 percent) have been issued with second level landholding certificates (Abab et al 2021)., By contrast, out of an estimated 5.6 million urban parcels, only 10.4 percent have been adjudicated and only 6.1 percent registered by the end of 2021, according to the Ministry of Urban Development and Infrastructure. Registration and certification of rural land holdings are conducted without formally enacted registration and cadastral laws, leaving pastoral and communal lands and their land rights— which are often not demarcated, registered, or mapped—vulnerable to infringement by land use changes, including commercial agriculture investment. In rural areas, restrictions on the transfer of land use rights lead to insecurity of tenure and parcel fragmentation. These challenges are compounded by insufficient land supply for urban development and a reliance on compulsory acquisition of rural land to meet the growing demand for urban land. In addition, the weak land administration system and the lack of integrated land-use planning enable the unmanaged spread of informal/illegal settlements into high-risk areas and former farmlands. To address the challenges discussed above, it will be critical to expedite the approval of the draft Rural Land Administration and Use Proclamation, harmonize the laws governing urban and rural lands, complete cadastral mapping and registration, and integrate existing information systems. Adoption of the Proclamation will need to be accompanied by the harmonization of existing laws governing urban and rural lands through a participatory process involving affected communities, including to identify alternatives to compulsory land acquisition to facilitate inclusive and sustainable urban growth. This will require from developing and adopting a rural-urban land use conversion strategy that ensures recognition of social, administrative and legal tenures and resources rights, as well as from harmonizing rural and urban land adjudication and registration procedures. For the reforms to be effective, it will also be important to complete the cadastral mapping and registration of land use rights in both urban and rural areas and to integrate the existing land information and geospatial information systems to increase transparency, facilitate inter- institutional coordination, inform land-use planning, and secure individuals’ and communities’ existing land use rights. It would also be useful to adopt the proposed national land-use policy and establish an independent agency to develop a national land-use plan and land-use classification system. A land-use policy is a 70 Rural land is defined by federal Proclamation 456/2005, while urban land is defined and administered by federal Proclamation 721/2011 and other proclamations. Responsibility for land administration in Ethiopia is divided among different institutions in the rural and urban spaces. At the federal level, rural land is handled by the Ministry of Agriculture, while urban land is handled by the Ministry of Urban Development and Construction. The Federal Integrated Infrastructure Development Agency is responsible for compulsory land expropriation, valuation, compensation, and resettlement for federal infrastructure development projects. The Land Bank and Development Corporation is mandated to manage public lands. All regional states also have their own regional land policies and laws. 162 Ethiopia Country Economic Memorandum useful tool for balancing the trade-offs among economic growth, sustainable land development, conservation, and land resource utilization (GoE 2017). The proposed national land-use policy foresees the creation of an independent agency to regulate competing land uses and mitigate potential conflicts from multiple demands on finite land resources. Its adoption and implementation would provide a framework for informed decision- making and planning at multiple levels, including through the development and implementation of a national integrated land-use plan, guidelines for harmonizing urban and rural land management, and the development of a land-use classification system through an assisting spatial planning process approach.71 6.5.2 Natural forests Sustainable forest management and landscape restoration are critical to building the productivity and resilience of rural landscapes and expanding economic opportunities. Reversing deforestation through expanded re/afforestation and restoration of degraded forest areas therefore has important economic and environmental benefits for Ethiopia. It could contribute not only to the recuperation of forest resources but also toward enhanced ecosystem services and climate resilience. At stake is also minimizing the loss in carbon stock and meeting NDC goals.72 Policies and procedures are needed for the development of incentive measures such as Payment for Ecosystem Services (PES) and results-based financing. PES provide incentives to landowners and managers to undertake certain activities to improve their land or watershed management (Box 10). If adequately defined, these PES would contribute to increased use and application of agro-forestry and silvopastural practices as well as the re/afforestation and restoration of an estimated 8 million ha by 2030, as per the updated NDC target under the conditional pathway. The measures would help achieve the NDC target, which is divided into the reforestation of 3 million ha and the restoration of 5 million ha and a further 9 million ha by 2050. Such restoration would also help protect the country’s watersheds. Box 10. Payments for environmental services Payments for environmental services (also known as payments for ecosystem services or PES) are payments to farmers or landowners who have agreed to take certain actions to manage their land or watersheds to provide an ecological service. PES is a market-based mechanism, similar to subsidies and taxes, to encourage the conservation of natural resources. A classic example is compensation payments to upstream land managers by downstream water utilities, hydroelectric power companies, breweries, etc. to ensure that they have steady supply of high-quality surface water, saving water purification costs and reducing sedimentation. Results-based payments involve a mechanism through which a funder is willing to make payments to an agent who assumes responsibility for achieving pre-defined results. Funding is only released upon the achievement of these results, which are verified independently. The rationale behind this approach is to link financing more directly with outputs and outcomes rather than inputs and processes. The objective is to increase 71 The approach will assist Ethiopia to move towards e-economies, e-service and e-commerce to improve services to citizens, build capacity for using geospatial technology, enhance informed government decision-making processes, facilitate private sector development, take practical actions to achieve a digital transformation, and to bridge the geospatial digital divide in the implementation of national strategic priorities and the 2030 Agenda for Sustainable Development. Currently, Ethiopia is preparing the country action plan of the framework. 72 Due to ongoing degradation, average above- and below-ground carbon declined. The carbon stored in dead wood also declined from 322.8 tCO2/ha (88.03tC/ha) in 1990 to 314.4 tCO2/ha (85.75tC/ha) in 2020 (FAO 2020). This implies both a spatial loss in forest cover as well as lower capacity to store carbon over time. The total carbon stored thus declined from 6,107 MtCO2 to 5,026 MtCO2. In terms of value, assuming a conservative price of US$ 30/tCO2, the value of stored carbon in Ethiopia’s forests declined from US$183 .2 billion to US$150.8 billion at a rate of 0.6 percent per year between 1990 and 2020. 163 Ethiopia Country Economic Memorandum accountability and create incentives to improve program effectiveness. Results-based payments are expected to be the long-term solution in REDD+ schemes73 in which forest countries are compensated ex post for their increased carbon sequestration. In Ethiopia, the Climate Action Through Landscape Management Program for Results (CALM) is an example of a results-based payment scheme. Source: https://www.iied.org/markets-payments-for-environmental-services, and https://www.oecd.org/dac/peer-reviews/Results-based-financing-key-take-aways-Final.pdf In terms of promoting clean cooking, a fuel-switching program could help shift from unsustainable biomass energy demand (charcoal and firewood) to electric stoves, renewable biofuels (e.g., residues), and improved cookstoves. The use of biomass for cooking and baking is the largest driver of emissions in the land-use change and forestry sector (GOE 2021). Fuel switching can be encouraged through the establishment of a supportive policy environment and standards for improved cook stove program integration, harmonization, and alignment with other initiatives. It also requires a market-related program management approach and implementation that is linked to the carbon market and financing. Clean cooking design standards for efficient cookstoves (for example, Tier 4 and above74) as well as processes to enhance local production capacity are also needed. Policies and processes to strengthen the existing community-based management systems will also be essential. Natural forests are under pressure in part due to a weak and ineffective regulatory framework that is also tied to an incomplete information base and lackluster implementation. The existing community-based management systems can be strengthened by a review of the open access policy and ensuing legislation as well as the enforcement of policies controlling access to and use of the resources.75 Policies for the development of a geospatial-technology-oriented data management system will also be needed. This system would contribute to landscape-level planning and link natural forest management with protected areas management and plantation development, allowing for improved resource management and ensuring that knowledge on the resources with respect to biodiversity conservation is accurate and the inventory data regularly updated. 6.5.3 Commercial forests Improving the management and expansion of planted forests, which could include measures such as certification of sustainable forest management (SFM), would need to play a role in Ethiopia’s green recovery agenda. Unlike extractive industries and minerals, timber plantations are a resource that can be used and managed in perpetuity to produce both economic and environmental services (including carbon sequestration). Plantation management would need to be based on sustainable resource management and can provide for biodiversity protection and improved land and water resource management, particularly when combined with forest certification. While there is no evidence of price premia for certified timber, it has better acceptance and market access in (international) markets. Independent third-party certification would also contribute to greater transparency and, ultimately, improved governance of the sector.76 In addition, independent forest certification could provide confidence for private investors that forests are well-managed. 73 Reducing emissions from deforestation and forest degradation (REED+) is a framework created by the UNFCCC Conference of the Parties to guide activities in the forest sector that reduces emissions from deforestation and forest degradation, as well as the sustainable management of forests and the conservation and enhancement of forest carbon stocks in developing countries. 74 https://www.cleancookingalliance.org/technology-and-fuels/standards/iwa-tiers-of-performance.html 75 For instance, “Access to genetic resources and community knowledge, and community rights proclamation (Federal Democratic Republic of Ethiopia Council of Ministers Regulation No. 169/2009)”. 76 Forest certification is a voluntary process whereby an independent third party (the “certifier”) assesses the quality of forest management and production against a set of requirements (“ standards”) predetermined by a public or private 164 Ethiopia Country Economic Memorandum It will also be important for Ethiopia to develop regulatory procedures for establishing public-private partnerships (PPPs) for commercial forestry and wood-based industries. Given the significant role played by production forestry and wood processing in the Ethiopian economy, institutional and regulatory impediments to attracting private investment could be removed. Doing so will not only facilitate the development of PPPs but also allow for the development of markets for institutional investors (such as the insurance sector and other players in the financial market) to make equity investments in partnerships with the SFEs to expand planted forests and establish sustainable production forests.77 This process would unlock the entire value chain, from wood production to wood product end-users (e.g., consumers, construction, furniture and other wood product industry, exports). Private sector investment in plantations also requires that land tenure and rights be clear and that investors (including smallholder farmers) face a predictable and stable policy environment. In addition to removing institutional barriers to private sector involvement, better management of the resources is needed to increase supply. Such management interventions include improved silvicultural practices as well as investment in better genetic material. To ensure that these improved management practices are performed, standards and norms with respect to industry best practices must be developed and applied. This will further increase profitability and attract private sector investment, reinforcing the process discussed above. It will also stimulate market expansion since Ethiopia has significant potential to enlarge its plantation estate,78 and it will go a long way in helping Ethiopia fulfill its NDC commitments while increasing forest-based carbon sequestration. Facilitating private sector investment in plantations also requires mapping the potential areas for commercial plantation development in alignment with Ethiopia’s new NDC targets, as well as developing information and R&D systems. Guidelines and procedures for both the resource mapping and the R&D system are required to facilitate a reliable and increasing wood supply for industries and end-users. These upstream investments create an enabling environment and functional sector infrastructure that are a necessary—although alone not entirely sufficient—condition for sustainable private sector job growth. Public funding and support structures are also needed when private plantations and woodlots are established to reduce the inherent high risks largely stemming from the long incubation periods required before planted forests generate revenue. Investment in the forestry sector can be accelerated to embark on a PES system, in conjunction with the natural forest areas as described above. Through PES, land managers can monetize the environmental certification organization. Forest certification, and associated labelling, is a way of informing consumers about the sustainability of the forests from which wood and other forest products were produced (source: FAO). The best-known SFM certification schemes are the Programme for the Endorsement of Forest Certification (PEFC) and the Forest Stewardship Council (FSC). Currently (available data as of March 2022), Ethiopia has no certified forests. The market access benefits are most evident in environmentally sensitive markets in Europe and North America. However, anecdotal evidence (e.g., from World Bank operations in Uganda) indicates that clients in Africa are also increasingly interested in SFM certification and the due diligence it provides. 77 In many countries, the investor base for plantations has expanded, and sustainably managed production plantations have become (still a relatively small) part of asset portfolios of large institutional investors. This has happened mainly in developed industrial countries, although emerging economies like Brazil and Uruguay have seen pension funds investing in plantations (Binkley et al. 2020). In Ethiopia, the current regulatory situation does not allow this, as all pension funds are mandated to invest only in Government bonds yielding negative returns (World Bank 2020b). While plantations can be only a small part of any responsibly managed institutional investor’s portfolio, allowing pension funds to invest in plantation development—possibly in partnership with the SFEs—would strengthen the financing for plantations and likely yield positive returns to these financial institutions. 78 New plantation establishment should take place on degraded land, abandoned and low productivity agricultural and pastureland, and other areas with low conservation value. This would still be only a small part of the national reforestation target, which is to double the country’s forest cover by 2030 (PDC 2020). 165 Ethiopia Country Economic Memorandum externalities to generate additional income (or compensate foregone other revenues or increased management costs). Generating new investment streams in commercial forestry would be beneficial if positive environmental externalities, such as soil stabilization and increased carbon sequestration, were internalized, e.g., through payments for ecosystem services around watershed protection. This is more likely to work in conditions where a water or energy utility benefits from investments in natural capital and is willing to provide compensation payments. Private and community woodlots can play a key role to support inclusive development of commercial forestry. The development of woodlots would serve many purposes: they provide direct income livelihood support to local communities; when well-managed, they sequester carbon; site-specific management practices can be designed in a way that provides local environmental services (e.g., erosion control); and they can be linked to agriculture and animal husbandry to develop agroforestry and silvo-pastoral systems. Development of these woodlots requires support, including to clarify tenure rights, provide extension services, and ensure that social equity issues (e.g., gender, the youth) are addressed. It would also require that woodlot owners have access to high-quality genetic material, both fast-growing exotic species as well as indigenous species. 6.5.4 Protected areas Ethiopia needs to develop feasible conservation business tools to strategically tackle financial constraints for biodiversity conservation. The return on investment in protected areas in Ethiopia is significant, with economic and social benefits ranging from US$6.80 to US$583 per US$ invested. Despite this, the current funding gap (excluding foreign aid) is between US$10.3-21.3 million, or between US$344/km2 and US$710/km2 annually (Namaga et al. 2020). Unlocking such investments will require the development of policy mechanisms that will continuously monitor, report, and verify operational and management efficiency and recommend changes in such PA management systems when and if needed to ensure that protected areas are managed according to international best practices. Increasing the value of Ethiopia’s protected areas and OECMs will require private sector involvement through PPPs, among other measures. Policies and measures could be developed by EWCA to, on an ongoing basis, seek and operationalize external sources of funding that could act as investments in the protected area system through concessional resources, results-based financing, or recapitalization of the tourism sector. They could also contribute to the expansion of protected area surface through innovative financing schemes to enlarge the protected footprint. The commercialization of protected areas can also be brought about through income diversification, which will require policy measures to develop and promote schemes such as tourism revenue retention schemes, PES, trust funds, concession fees,79 and traversing rights.80 In addition to private sector finance, public sector finances can be raised for the support of protected areas. This can be done at zero cost to the government budget through the removal of all explicit or implicit subsidies that incentivize land-use change in and near the protected areas—for example, access to government development loans, tax holidays (both reduction and exemption), land lease with minimum rate, and longer credit grace period. The reduction and repurposing of agricultural subsidies in favor of protected areas would also be supportive. 79 Concessions could be awarded to private operators in the hospitality industry as well as in nature-based protection and conservation to co/manage protected areas, or parts thereof, in conjunction with the communities to reduce human- wildlife conflict and reduce resource use. 80 Traversing rights, extensively used in South Africa, involve payment for the right of way. Operators pay for the right to use certain roads, a cost the operator recoups from the visitors. This is a form of the PES system—the beneficiaries (tourists) contribute to upkeep and management of the park through the private enterprises. It also incentivizes private sector investment and targeted marketing of the resources. 166 Ethiopia Country Economic Memorandum Finally, improving land use planning and management in the protected area buffer zones, in collaboration with communities, would be important. A market-based approach, like PES, could be piloted to support communities and local landowners to co-manage and maintain the protected areas and OECMs through collective action. Based on the proceeds received, the protected areas could also consider a revenue sharing model, such as the one in Rwanda,81 to share the returns from their activity with the local communities in various forms. Revenue sharing is likely to improve the trust among the protected area managers and the local communities while offering opportunities for mutual learning and co-management. Table 22. Policy recommendations to make the most of natural capital in Ethiopia Area Short term Medium term Lead institution(s) Integrated Approve the National Land-Use Policy Develop and update regional land-use Environmental land-use plans for a multisector institutional Protection Agency planning Establish an independent, high-level response to key challenges (EPA) agency focused on land administration Ministry of and use Agriculture (MoA) Adopt guidelines toward the Develop land use/spatial decision support EPA development of a system for land-use tools that inform PES schemes and Benefit MoA classification, sharing mechanisms Adopt the new Rural Land Complete the cadaster and integrate land MoA Administration and Use proclamation information and geospatial information systems Develop and adopt a rural-urban land Harmonize rural and urban land MoA use conversion strategy that ensures adjudication and registration procedures Ministry of Urban recognition of social, administrative Development and and legal tenures and resources rights Infrastructure (MUDI) Ministry of Planning and Development (MoPD) Natural forests Develop guiding principles and Continue to pursue results-based Ministry of Water, procedures for implementing a payments for forestry under REDD+ and Irrigation and nationwide (or regional) PES scheme future carbon market mechanisms under Electricity (MoWIE) for watershed protection the Paris Agreement Reduce the demand for wood from Develop minimum standards for efficient MoWIE natural forests through active sourcing cookstoves and phase out unsustainable and roll-out of energy-efficient stoves firewood and charcoal Adopt incentives to move away from unsustainable firewood and charcoal production Strengthen existing community-based Develop a data management system and Ethiopian management systems, including by ensure that knowledge on natural Biodiversity Institute reviewing the open access policy resources and biodiversity is accurate, and (EBI) that inventory is updated regularly 81 The National Parks in Rwanda are surrounded by areas with a high population density, which is the highest in Africa at 525 per km2. Currently most of the residents in the proximity of the Parks, up to 80 percent, depends on agriculture (NISR 2017). The important role played by communities surrounding the Parks in Rwanda’s conservation efforts have been recognised by the Rwandan government. Since 2005 the government has therefore adopted a revenue sharing policy. Initially this entailed 5 percent of the park revenues being given back to the communities by funding community-based projects. This amount was recently doubled to 10 percent of all park revenue. The tourism revenue sharing scheme constitutes the main source of income for communities around the parks, who were getting approximately $54.7 (RwF50,000/month) per household (RDB 2020). This is affecting families, education for children, sanitation, food security and the security of the park. Between 2005 and 2019, the total amount for revenue sharing stood at $5.3 million and has funded about 690 projects (RDB 2019). 167 Ethiopia Country Economic Memorandum Ensure the enforcement of legislation Ethiopian Wildlife and policy controlling access to and use Conservation of the resources Authority (EWCA) Commercial Initiate independent certification of EPA forests and sustainable forest management (PEFC MoA plantations or FSC) of SFE-planted forests. Develop regulatory procedures for Disinvestment/divestment of SFEs assets EPA establishing PPPs for commercial where public ownership is not needed MoA forestry and wood-based industries, Ethiopian Chamber and remove impediments for private of Commerce & sector participation Sectoral Association (ECSSA) Improve the management of existing Develop information and R&D systems to EPA plantations to increase their facilitate reliable wood supply for MoA productivity industries and end-users ECSSA Conduct mapping of areas for potential Develop private and community woodlots, EPA commercial plantation development in including by clarifying tenure rights and MoA alignment with Ethiopia’s new NDC providing extension services targets Develop a PES system for the EPA commercial forestry sector that aligns MoA with that of the natural forest sector Protected areas Develop policy mechanisms for EWCA, EBI and OECMs continuous monitoring and reporting of operational and management efficiency in protected areas Invest in the of protected areas through Expand protected areas through EPA concessional resources, results-based innovative financing schemes EWCA financing, and recapitalization of the EBI tourism sector Pursue commercialization of the Remove all explicit or implicit subsidies protected areas by reviewing and that incentivize land-use change in and adjusting the entry fees, concession near the protected areas, including fees, traversing rights, hunting quotas, repurposing agricultural subsidies etc. Pilot PES or revenue sharing schemes EPA with communities and local EWCA landowners for co-management and EBI maintenance of protected areas Source: authors’ elaboration. 168 Ethiopia Country Economic Memorandum 6.6 References Abab, S. A., F. Wakjira, and T. Negash. 2021. “Determinants of the Land Registration Information System Operational Success: Empirical Evidence from Ethiopia.” Land, 2021, 10 (12). Amare. A. 2015. “Wildlife Resources of Ethiopia: Opportunities, Challenges and Future Directions: From Ecotourism Perspective, A Review Paper.” Natural Resources, Vol 6: 405-422. Atmadja, S., A. Eshete, and M. Boissière. 2019. Guidelines on sustainable forest management in drylands of Ethiopia. Rome: FAO. Bekele, A. E., D. Drabik, L. Dries, and W. Heijman. 2020. Large‐scale land investments, household displacement, and the effect on land degradation in semiarid agro‐pastoral areas of Ethiopia. Land degradation and development, 32(2):777-791. https://doi-org.ez.sun.ac.za/10.1002/ldr.3756 Belay, T., and D.A. Mengistu. 2019. Land use and land cover dynamics and drivers in the Muga watershed, Upper Blue Nile basin, Ethiopia. Remote sensing applications: society and environment, 15(100249). https://doi.org/10.1016/j.rsase.2019.100249 Beyene, A., A. Mekonnen, M. Hirons, E. Robinson, T. Gonfa, T. Gole, and S. Demissie. 2020. Contribution of non-timber forest products to the livelihood of farmers in coffee growing areas: evidence from Yayu Coffee Forest Biosphere Reserve. Journal of Environmental Planning and Management, 63:9, 1633- 1654, DOI: 10.1080/09640568.2019.1679615 Comenetz, J., and C. Caviedes. 2002. Climate variability, political crises, and historical population displacements in Ethiopia. Global Environmental Change Part B: Environmental Hazards, 4(4):113-127. Cuong, T., Chinh, T.T.Q., Zhang, Y. and Y. Xie. 2020. “Economic Performance of Forest Plantations in Vietnam: Eucalyptus, Acacia mangium, and Manglietia conifera.” Forests 2020, 11, 284; Cubbage, Frederick et als. 2020. “Global timber investments, 2005 to 2017.” Forest Policy and Economics 112. Dagnachew, M., A. Kebede, A. Moges, and A. Abebe. 2020. Land Use Land Cover Changes and Its drivers in Gojeb River Catchment, Omo Gibe Basin, Ethiopia. Journal of agriculture and environment for international development, 114(1). https://doi.org/10.12895/jaeid.20201.842 Damtea, W., D. Kim, and S. Im. 2020. Spatiotemporal Analysis of Land Cover Changes in the Chemoga Basin, Ethiopia, Using Landsat and Google Earth Images. Sustainability, 12(9):3607. https://doi- org.ez.sun.ac.za/10.3390/su12093607 Danyo, S., A. Abate, M. Bekhechi, G. Köhlin, H. Medhin, A. Mekonnen, A. Fentie, T. Ginbo, B. Negede, H. Tesfaye, and A. Wikman. 2017. Realizing Ethiopia’s Green Transformation: Country Environmental Analysis, Environment and Natural Resources Global Practice. Washington, DC: World Bank. Dasgupta, P. 2021. The Economics of Biodiversity: The Dasgupta Review. London: HM Treasury. Deutz, A., G.M. Heal, R. Niu, E. Swanson, T. Townshend, L. Zhu, A. Delmar, A. Meghji, S.A. Sethi, and J. Tobin-de la Puente. 2020. Financing Nature: Closing the global biodiversity financing gap. The Paulson Institute, the Nature Conservancy, and the Cornell Atkinson Center for Sustainability. URL https://www.paulsoninstitute.org/key-initiatives/financing-nature-report/ (Accessed February 2021). Dibaba. T., T.A. Demissie, and K. Miegel. 2020. Drivers and Implications of Land Use/Land Cover Dynamics in Finchaa Catchment, Northwestern Ethiopia. Land, 9(4):113. 169 Ethiopia Country Economic Memorandum ECSSA (Ethiopian Chamber of Commerce and Sectorial Association). 2016. “Position Paper on the Barriers and recommended measures for the Commercial Forestry and Wood Processing Industry Development in Ethiopia.” ELD Initiative & UNEP (United Nations Environment Programme. 2015. The Economics of Land Degradation in Africa: Benefits of Action Outweigh the Costs. Available from www.eld-initiative.org. Ethiopian Panel on Climate Change. 2015. First assessment report: Working group II – climate change impact, vulnerability, adaptation and mitigation. https://eas-et.org/sites/default/files/2017- 06/Agricalture%20and%20Food%20Security.pdf. EWCA (Ethiopian Wildlife Conservation Authority). 2009. “The Value of the Ethiopian Protected Area System: Message to Policy Makers, Key Policy Message.” EWCA, Austrian Federal Forests. Addis Ababa FAO (Food and Agriculture Organization of the United Nations). 2020. Global Forest Resource Assessment 2020: Ethiopia. Rome. FDRE (Federal Democratic Republic of Ethiopia). 2017. Ethiopia’s forest reference level submission to the UNFCCC. https://redd.unfccc.int/files/ethiopia_frel_3.2_final_modified_submission.pdf (accessed March 2021). Fransen, S., and K. Kuschminder. 2009. Migration in Ethiopia: History, Current Trends and Future Prospects. Paper Series: Migration and Development Country Profiles, Maastricht Graduate School of Governance. Girma, W., and F. Beyene. 2015. Institutional Challenges in Sustainable Forest Management: Evidence from the Gambella Regional State of Western Ethiopia. Journal of Sustainable Forestry, 34:233–258. DOI: 10.1080/10549811.2014.1003245 GoE (Government of Ethiopia). 2017. Ethiopia Forest Sector Review: focus on commercial forestry and industrialization. Technical report. GoE (Government of Ethiopia) and World Bank. 2021.Ethiopia's Nationally Determined Contribution (NDC) Update, Technical Report, July 2021 Habtamu, M., Mohammd, A., and A. Mezgebu. 2019. “Review on Economic valuation and conservation of national park of Ethiopia.” The International Journal of Business Management and Technology, 3 (1): 79-91. Haddis, Z., S. Bekure, A. Dagnew, T. Gebremeskel, and B. Tafere. 2017. Ethiopia’s move to a national integrated land use policy and land use plan. Paper presented at the responsible land governance conference, Washington. Hailu, Z. 2016. Land Governance Assessment Framework Implementation in Ethiopia. Washington, DC: World Bank. HEI (The Health Effects Institute). 2020. State of Global Air: A Special Report on Global Exposure to Air Pollution and its Health Impacts. The Health Effects Institute/The Institute for Health Metrics and Evaluation Available at https://www.stateofglobalair.org/data/estimate-exposure. Hockings, M., N. Dudley, W. Elliott, M. Ferreira, K. Mackinnon, M. Pasha, A. Phillips, S. Stolton, S. Woodley, M. Appleton, O. Chassot, J. Fitzsimons, C. Galliers, R. Kroner, J. Goodrich, J. Hopkins, W. Jackson, H. Jonas, B. Long, M. Mumba, J. Parrish, M. Paxton, C. Phua, R. Plowright, M. Rao, K. Redford, J. Robinson, C. Rodríguez, T. Sandwith, A. Spenceley, C. Stevens, G. Tabor, S. Troëng, S. Willmore, and A. Yang. 2020. Covid‐19 and protected and conserved areas. Parks, 26(1):7-24, 10.2305/IUCN.CH.2020.PARKS‐26‐1MH.en. 170 Ethiopia Country Economic Memorandum Hurni, K., G. Zeleke, M. Kassie, B. Tegegne, T. Kassawmar, E. Teferi, A. Moges, D. Tadesse, M. Ahmed, Y. Degu, Z. Kebebew, E. Hodel, A. Amdihun, A. Mekuriaw, B. Debele, G. Deichert, and H. Hurni. 2015. Economics of land degradation (ELD) Ethiopia case study. Soil degradation and sustainable land management in the rainfed agricultural areas of Ethiopia: An assessment of the economic implications. Report for the Economics of Land Degradation Initiative. Indufor. 2016. Ethiopia Commercial Forest Plantation Investment Plan. Report to the International Finance Corporation/World Bank Group. Helsinki. IHME (Institute for Health Metrics and Evaluation). 2020. Global Burden of Disease Project. Johanson, F.X., and A.T. Mengistu. 2013. Alternative future pathways for household biomass use in Ethiopia. Discussion Brief Series, Stockholm Environment Institute, Stockholm.Johnson, J. A., G. Ruta, U. Baldos, R. Cervigni, S. Chonabayashi, E. Corong, O. Gavryliuk, J. Gerber, T. Hertel, C. Nootenboom, and S. Polasky. 2021. The Economic Case for Nature: A Global Earth-Economy Model to Assess Development Policy Pathways. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/35882 License: CC BY 3.0 IGO. Kaza, S., L. Yao, P. Bhada-Tata, and F. Van Woerden. 2018. What a Waste 2.0: A Global Snapshot of Solid Waste Management to 2050. Urban Development Series. Washington, DC: World Bank. doi:10.1596/978-1- 4648 -1329-0. License: Creative Commons Attribution CC BY 3.0 IGO. Kirui, O.K., and A. Mirzabaev. 2015. Cost of land degradation in Eastern Africa. Poster presented at the 29th ICAE conference in Milan, Italy. https://www.zef.de/fileadmin/downloads/Oliver_Kirui_- _ICAE_2015.pdf Lindsey, P., Miller, J., Petracca, L., and L. Hunter. 2018. More than $1 billion needed annually to secure Africa’s protected areas with lions, Proceedings of the National Academy of Sciences, 115 (45), E10788- E10796, 10.1073/pnas.1805048115 McLain, R., H. Kassa, S. Lawry and B. Yazew. 2019. Fostering tenure security for forest landscape restoration in Ethiopia Creating enabling conditions for the 2018 Forest Proclamation. CIFOR Infobrief No. 267, September 2019. https://doi.org/10.17528/cifor/007410 MEFCC (Ministry of Environment, Forest and Climate Change). 2016. Ethiopia’s Forest Reference Level Submission to the UNFCCC. January 2016. MEFCC (Ministry of Environment, Forest and Climate Change). 2018. National Forest Sector Development Program, Ethiopia Volume I: Situation Analysis. Mengist, W. 2020. “Challenges of Protected Area Management and Conservation Strategies in Ethiopia: A Review Paper.” Advances in Environmental Studies. 4(1):277-285. MoEFC (Ministry of Environment, Forest and Climate Change). 2017. Ethiopia Forest Sector Review. Focus on commercial forestry and industrialization. Federal Democratic Republic of Ethiopia. Ministry of Environment, Forest and Climate Change. Molla, K.G., and F.B. Woldeyes. 2020. Pathways to Sustainable Land-Use and Food Systems in Ethiopia by 2050 In: FABLE. 2020. Pathways to Sustainable Land-Use and Food Systems, 2020 Report of the FABLE Consortium. Laxenburg and Paris: International Institute for Applied Systems Analysis (IIASA) and Sustainable Development Solutions Network (SDSN), pp. 262-291. https://doi.org/10.22022/ESM/12-2020.16896 Naidoo, R., Weaver, L.C., Diggle, R.W., Matongo, G., Stuart-Hill, G. and Thouless, C. 2016, Complementary benefits of tourism and hunting to communal conservancies in Namibia. Conservation Biology, 30: 628-638. https://doi.org/10.1111/cobi.12643 171 Ethiopia Country Economic Memorandum Nalepa, R., A. Short Gianotti, and D. Bauer. 2017. Marginal land and the global land rush: A spatial exploration of contested lands and state-directed development in contemporary Ethiopia. Geoforum, 82:237-251. Namaga, S.E., H. Tadele, M. Tessema, and Z. Tefera. 2020. Underfunding, the challenge of federally managed protected areas of Ethiopia. Advances in environmental studies, 4(2):307-317. DOI: 10.36959/742/227 Negedea, B., Pirarda, R., and H. Kassa. 2015. Employment in industrial timber plantations – An Ethiopian case supported by a global review. CIFOR Infobrief no. 122. May 2015. NISR (National Institute of Statistics of Rwanda). 2017. Rwanda Agriculture Households Survey. Kigali, Rwanda. OECD (Organisation for Economic Co-operation and Development). 2020. Biodiversity and the economic response to COVID-19: Ensuring a green and resilient recovery. Brussels. Palahí, M., M. Pantsar, R. Costanza, I. Kubiszewski, J. Potočnik, M. Stuchtey, R. Nasi, H. Lovins, E. Giovannini, L. Fioramonti, S. Dixson-Declève, J. McGlade, K. Pickett, R. Wilkinson, J. Holmgren, K. Trebeck, S. Wallis, M. Ramage, G. Berndes, F. Akinnifesi, K. Ragnarsdóttir, B. Muys, G. Safonov, A. Nobre, C. Nobre, D. Ibañez, A. Wijkman, J. Snape, and L. Bas. 2020. Investing in Nature as the true engine of our economy: A 10-point Action Plan for a Circular Bioeconomy of Wellbeing. Knowledge to Action 02, European Forest Institute. URL https://efi.int/publications-bank/investing-nature-true-engine-our-economy- 10-point-action-plan-circular-bioeconomy. https://doi.org/10.36333/k2a02 (Accessed February 2021). PDC (Planning and Development Commission). 2021. Ten Years Development Plan. A Pathway to Prosperity. 2021- 2030. Federal Democratic Republic of Ethiopia. RDB (Rwanda Development Board). 2019. Tourism annual report: 2018. Kigali, Rwanda. RDB (Rwanda Development Board). 2020. Rwanda Tourism Performance Statistics. Kigali, Rwanda. UNDESA (United Nations Department of Economic and Social Affairs). 2017. World Population Prospects: The 2017 Revision. New York: United Nations. https://population.un.org/wpp/DataQuery/ United Nations, European Union, Food and Agriculture Organization of the United Nations, International Monetary Fund, Organisation for Economic Co-operation and Development and the World Bank. 2014. System of Environmental-Economic Accounting 2012--Central Framework. New York. ST/ESA/STAT/Ser.F/109 Van Zyl, H. 2015. The Economic Value and Potential of Protected Areas in Ethiopia. Report prepared for The Sustainable Development of the Protected Areas System of Ethiopia (SDPASE) project and the Ethiopian Wildlife Conservation Authority (EWCA) Wassie. B. S. 2020. “Natural resource degradation tendencies in Ethiopia: a review.” Environmental Systems Research, 9(1), 33, 10.1186/s40068-020-00194-1 WEF (World Economic Forum). 2021. The global risk report 2021. Geneva. World Bank. 2017. Safe and Resilient Cities in Ethiopia: CityStrength Diagnostics in Nine Regional Capitals and Dire Dawa City Administration. World Bank, Washington, DC. World Bank. 2020a. Mobilizing private finance for nature. Washington, DC.: World Bank. World Bank. 2020b. The Next Generation Africa Climate Business Plan: Ramping Up Development-Centered Climate Action. Washington, DC: World Bank. World Bank. 2021. The change in the wealth of nations. World Bank. In press. 172 Ethiopia Country Economic Memorandum World Bank and IHME (Institute for Health Metrics and Evaluation). 2016. The Cost of Air Pollution: Strengthening the Economic Case for Action. http://documents1.worldbank.org/curated/en/781521473177013155/pdf/108141-REVISED-Cost- of-PollutionWebCORRECTEDfile.pdf (Accessed February 2021). Young, J. 2012. Ethiopian Protected Areas A ‘snapshot’. A reference guide for future strategic planning and project funding. Ethiopian Wildlife Authority. Zegeye, H. 2017. Major Drivers and Consequences of Deforestation in Ethiopia: Implications for Forest Conservation. Asian Journal of Science and Technology Vol. 08, Issue, 08, pp.5166-5175, August, 2017. Zin, W.S., Suzuki, A., Peh, K., and A. Gasparatos. 2019. “Economic Value of Cultural Ecosystem Services from Recreation in Popa Mountain National Park, Myanmar: A Comparison of Two Rapid Valuation Techniques.” Land, 8 (12): 194, doi:10.3390/land8120194. 173