Boosting Productivity in Sub-Saharan Africa Policies and Institutions to Promote Efficiency César Calderón Boosting Productivity in Sub-Saharan Africa Boosting Productivity in Sub-Saharan Africa policies and institutions to promote efficiency César Calderón © 2021 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 24 23 22 21 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 Execu- tive 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 immuni- ties of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://­ creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution —Please cite the work as follows: Calderón, César. 2021. Boosting Productivity in Sub-Saharan Africa: Policies and Institutions to Promote Efficiency. Washington, DC: World Bank. doi:10.1596/978-1-4648-1550-8. License: Creative Commons Attribution CC BY 3.0 IGO Translations —If you create a translation of this work, please add the following disclaimer along with the attribu- tion: 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 attribu- tion: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adap- tation 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 responsibil- ity to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. 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. ISBN (paper): 978-1-4648-1550-8 ISBN (electronic): 978-1-4648-1551-5 DOI: 10.1596/978-1-4648-1550-8 Cover art: © Sergio Andrés Moreno Téllez / World Bank. Further permission required for reuse. Cover design: Bill Pragluski, Critical Stages, LLC. Library of Congress Control Number: 2020914241 Contents Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix About the Author. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii 1 Boosting Productivity in Sub-Saharan Africa. . . . . . . . . . . . . . . . . . . . 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Sub-Saharan Africa’s Long-Term Performance: Still Far from the Frontier. . . . . . . . . . . . . . 3 Sources of Productivity Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Dimensions of the Productivity Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Data and Measurement Issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Plan of the Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2 Needed: Boosting the Contribution of Total Factor Productivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 The Divergent Paths of Malaysia and Senegal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Development Accounting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Dismal Growth Performance: The Negligible Contribution of TFP Growth. . . . . . . . . . . . 29 Lagging Structural Transformation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3 Resource Misallocation in Sub-Saharan Africa: Firm-Level Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Resource Misallocation in Agriculture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 v vi   C o n t e n t s Resource Misallocation in Manufacturing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4 Policies and Institutions that Distort Resource ­ ub-Saharan Africa. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Allocation in S Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Land Market Imperfections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Agricultural Subsidies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Taxation and Informality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Trade Policy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Financial Market Imperfections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5 Agenda for Future Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Appendix A:  Output per Worker, Factor Accumulation, and Total Productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Appendix B:  Country Productivity Analysis in Sub-Saharan Africa. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Boxes 1.1 Building upon the World Bank’s Productivity Research Agenda. . . . . . . . . . . . . . . . . . 11 2.1 The Contribution of Natural Capital to Growth per Worker. . . . . . . . . . . . . . . . . . . . . 33 3.1 Resource Misallocation: Theoretical Underpinnings. . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.1 Land Institutions in Selected Sub-Saharan African Countries. . . . . . . . . . . . . . . . . . . . 53 4.2 Trade Liberalization and Within-Firm Changes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 4.3 The Role of Transportation Infrastructure in Agriculture. . . . . . . . . . . . . . . . . . . . . . . 70 Figures 1.1 Output per Worker in Sub-Saharan Africa and EAP5 Countries, Relative to the United States, 1960–2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2 Output per Worker in Sub-Saharan Africa versus Selected Country Groups, 1960–2016. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3 Sources of Resource Misallocation That Reduce Total Factor Productivity. . . . . . . . . . . 7 2.1 Outputs, Inputs, and Productivity Gaps between Malaysia and Senegal, 1960 and 2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.2 Relative Labor Productivity of Sub-Saharan African Countries, 1980 versus 2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.3 Sources of the Labor Productivity Gap between Sub-Saharan Africa and the United States, 1960–2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 C o n t e n t s   vii 2.4 Share of Labor Productivity Differences due to TFP in Sub-Saharan African Countries, 1980–89 versus 2010–17. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2.5 Traditional Solow Decomposition of Labor Productivity Growth, Selected Regions and Country Groups, 1960–2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 2.6 Traditional Solow Decomposition of Labor Productivity Growth in Sub-Saharan Africa, by Country Group and Period, 1961–2017. . . . . . . . . . . . . . . . . . 32 B2.1.1 Decomposition of Labor Productivity Growth, Including Natural Capital, in Sub-Saharan Africa, 1996–2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.7 Decomposition of Labor Productivity Growth, including Role of Public Capital, in Selected Regions and Country Groups, 1961–2014. . . . . . . . . . . . . . . . . . . 34 2.8 Sectoral Employment Shares, Sub-Saharan Africa versus Low- and Middle-Income Countries in Other Regions, 1990–2016. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 2.9 Sectoral Labor Productivity Relative to Agriculture: Sub-Saharan Africa and Low- and Middle-Income Countries in Other Regions, 1990–2016. . . . . . . . . . . . 36 3.1 Farmers’ Productivity, by Input Use and Yields, in Uganda. . . . . . . . . . . . . . . . . . . . . . 45 3.2 Quantity versus Revenue Productivity across Selected Sub-Saharan African Countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 4.1 Government Spending on Agricultural Input Subsidies, by Type, in Sub-Saharan African Countries with the 10 Largest ISPs, 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.2 Distribution of ISPC Taxpayers in Mozambique: 2010 versus 2015 . . . . . . . . . . . . . . . 62 4.3 Size Distribution of Formal Firms versus All Firms and US Benchmark, Selected Sub-Saharan African Countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 4.4 Simulated Impact of Business Registration Reform on Occupational Choice and Income in Cameroon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.5 Changes in Price Dispersion before and after Mobile-Phone Coverage in Niger’s Grain Markets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.6 Modeling the Impact of Financial Frictions on Sector-Level TFP . . . . . . . . . . . . . . . . . 79 Maps 2.1 Labor Productivity, by Country, Relative to the United States, 2017. . . . . . . . . . . . . . . 25 2.2 Capital-Labor Ratio, by Country, Relative to the United States, 2017. . . . . . . . . . . . . . 25 2.3 Human Capital Index, by Country, 2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.4 Efficiency of Production, by Country, Relative to the United States, 2017. . . . . . . . . . . 27 Tables 3.1 Gap between Actual and Potential Agricultural Yields, Selected Sub-Saharan African Countries, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 3.2 Optimal Crop Choice and Aggregate Yield Gains, Selected Sub-Saharan African Countries, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.3 Dispersion of Revenue and Quantity Productivity across Manufacturing Firms, Selected Sub-Saharan African Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.1 Policy-Related Sources of Potential Resource Misallocation Affecting Farm and Firm Productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.2 Impact of Land Rental on Resource Misallocation among Farmers in Ethiopia, 2013/14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.3 Effects of Actual and Efficient Distribution of Land, Capital, MPL, and MPK among Farms in Malawi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 viii   C o n t e n t s 4.4 Impact of Weather Shocks on Input Use and Output on Farmers in Uganda, 2009–14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 A.1 Development Accounting in Sub-Saharan Africa and in Non-African Developing Countries, Relative to the United States, 1960–2017 . . . . . . . . . . . . . . . . 102 A.2 Estimated Output Elasticities to Private and Public Capital, by Country Income Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 A.3 Classification of Sectors of Economic Activity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 A.4 Classification of Sub-Saharan African Countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Acknowledgments T his report was prepared by a team led • Barrot-Araya, Luis Diego, César Calderón, by César Calderón (Lead Economist), and Luis S er ven. 2019. “S ec toral under the guidance of Albert G. Zeu- Productivity Shifts in Sub-Saharan fack (Chief Economist of the Africa Region) A frica.” Unpublished manuscript , and Hafez Ghanem (Vice President). Core World Bank, Washington, DC. members of the team included Fernando • Chen, Chaoran, and Diego Restuccia. Aragón, Diego Barrot, Catalina Cantú, 2018. “Agricultural Productivity Growth Chaoran Chen, Margarida Duarte, Robert in Africa.” Unpublished manuscript, Fattal-Jaef, Nicolas Pierre P. Gonne, Apara- Department of Economics, University of jita Goyal, Megumi Kubota, Emmanuel K. Toronto. K. Lartey, Taye Alemu Mengistae, Diego • Cirera, Xavier, Robert N. Fattal Jaef, Restuccia, Juan Pablo Rud, Luis Servén, and Hibret B. Maemir. 2000. “Taxing Rishab Sinha, and Måns Soderbom. Ken- the Good? Distortions, Misallocation, neth Omondi provided able administrative and Productivity in Sub-Saharan Africa.” support to the report team. The report was World Bank Economic Review 34 (1): edited by Mary Anderson. 75–100. The following background papers served • Duarte, Margarida, and Diego Restuccia. as important inputs: 2018. “Structural Transformation and Productivity in Sub-Saharan Africa.” • Aragón, Fernando M., and Juan Pablo Unpublished manuscript, Department of Rud. 2018. “Weather, Productivity and Economics, University of Toronto. Factor Misallocation: Evidence from • Goyal, Aparajita, Keith Fuglie, and Felipe Ugandan Farmers.” Unpublished manu- Dizon. 2018. “Agriculture Productivity script, Department of Economics, Simon and Economic Transformation in Sub- Frasier University. Saharan Africa.” Unpublished manuscript, • Barrot-Araya, Luis Diego, César Calderón, World Bank, Washington, DC. and Luis Serven. 2019. “Growth in Sub- • Jones, Patricia, Emmanuel Lartey, Taye Saharan Africa: A TFP Boost is Needed.” Mengistae, and Albert Zeufack. 2019. Unpublished manuscript, World Bank, “Market Size, Sunk Costs of Entry, and Washington, DC. ix x   A C K N O W L E D G M E N T S Transport Costs: An Empirical Evaluation E conomics and Law, University of of the Impact of Demand-Side Factors ver- Gothenburg. sus Supply-Side Factors on Manufacturing We are grateful to the peer reviewers who Productivit y.” World Bank Group supported the preparation of this report from Policy Research Working Paper 8875, the concept stage, through the authors’ work- Washington, DC. shop, to the decision stage: Luc Christiaensen Sinha, Rishabh, and Xican Xi. 2018. •  (World Bank), Daniel Lederman (World “A g r o n o m i c E n d o w m e n t , C r o p Bank), and Richard Rogerson (Princeton Choice and Agricultural Productivity.” University). Several World Bank staff pro- Unpublished manuscript, World Bank, vided valuable comments and engaged in pro- Washington, DC. ductive discussions at various stages of the Söderbom, Måns. 2018. “Productivity •  development of this report. Our thanks also Dispersion and Firm Dynamics in go to the production team, including Mary Ethiopia’s Manufacturing Sector.” Fisk, production editor; Jewel McFadden, Unpublished manuscript, Department acquisitions editor; and Orlando Mota, print of E conomics, School of Business, and electronic conversion coordinator. About the Author César Calderón, a Peruvian national, is a Since 2014, he has been a core team mem- Lead Economist in the Office of the Chief ber of the “Africa’s Pulse” regional flagship Economist of the Africa Region (AFRCE). on recent macroeconomic developments in He joined the World Bank in 2005. Before Sub-Saharan Africa. He also has been a task joining the A FRCE , he worked in the team leader of AFRCE regional research Latin America and the Caribbean Regional projects, such as Africa’s Macroeconomic Chief Economist Office and the Finance Vulnerabilities. He has worked on issues of and Private Sector Development (FPD) open economy macroeconomics, growth, and Chief Economist Office, as well as on the development—especially the growth impact World Development Report . He was a of infrastructure development and outward- core team member of the Global Financial oriented strategies. Development Report 2013: Rethinking He is currently working on issues of trade the Role of the State in Finance and the diversification and growth, digital infrastruc- World Development Report 2014: Risk ture and development, and policy determi- and Opportunity: M anaging Risk for nants of macroeconomic resilience. He holds Development, where he authored the chap- a master’s degree and a PhD in economics ter on macroeconomic risk management. from the University of Rochester, New York. xi Abbreviations AfCFTA African Continental Free Trade Area AFRCE Office of the Chief Economist of the Africa Region (World Bank) CBA Commercial Bank of Africa Ltd. CPIA Country Policy and Institutional Assessment CES constant elasticity of substitution COVID-19 coronavirus disease 2019 DT digital technology EAP5 East Asian “dragons” (Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand) EWS early warning systems FAO Food and Agriculture Organization FCS fragile and conflict-affected states GAEZ Global Agronomic Ecological Zones GDP gross domestic product GIS geographic information system GPS Global Positioning System GSS Ghana Statistical Services HCI Human Capital Index ICT information and communication technology ICP International Comparison Program ILO International Labour Organization ISA Integrated Survey of Agriculture ISIC International Standard Industrial Classification ISP input subsidy program ISPC Simplified Tax for Small Taxpayers LLMCs low- and lower-middle-income countries LSMS-ISA Living Standards Measurement Studies–Integrated Surveys on Agriculture MPK marginal product of capital NGO nongovernmental organization NIPA national income and product accounts xiii xiv   A b b r e v i a t i o n s NIS National Innovation System OECD Organisation for Economic Co-operation and Development PIM perpetual inventory method PPP purchasing power parity PWT Penn World Table R&D research and development RFID radio frequency identification SIC Standard Industrial Classification SMS short message service SSA Sub-Saharan Africa TFP total factor productivity TFPQ total factor productivity quantity TFPR total factor productivity revenue UMICs upper-middle-income countries UN-NAC United Nations National Accounts WBES World Bank Enterprise Surveys WDI World Development Indicators Boosting Productivity in Sub-Saharan Africa 1 Introduction There has been an intense debate as to whether the observed variation in output per A stylized fact of development economics worker across countries is attributable to dif- is the wide disparity in output per worker ferences in factor accumulation or to differ- across the world’s countries. On the one ences in total factor productivity (TFP). The hand, there are large disparities in labor pro- evidence shows that TFP accounts for most of ductivity across countries at any point in the differences in income per worker across time. The richest countries (top decile) were countries (Caselli 2005; Hall and Jones 1999; 23 times as productive as the poorest nations Hsieh and Klenow 2010; Jones 2016; Klenow (bottom decile) in 1960 (Feenstra, Inklaar, and Rodríguez-Clare 1997). This implies and Timmer 2015). That gap increased to 37 that some countries, sectors, and firms pro- times by 2017. The productivity gap between duce more than others with the same amount the richest countries and those in the middle of inputs (labor, human and physical capital, of the distribution has fluctuated by about land, and intermediate inputs, among others). fourfold between 1960 and 2017. In fact, TFP overwhelmingly explains the On the other hand, a country’s own out- cross-country differences in income per cap- put per worker also tends to move signifi- ita (Caselli 2005; Jones 2016). An important cantly over time—such that the country’s lesson emerges from these aggregate account- growth successes and failures may be unre- ing exercises: productivity improvement is lated to the initial level of development. For essential to sustained economic growth (Kim instance, Botswana, a very poor country at and Loayza 2019). 3.8 ­percent of United States gross domestic Theoretical and empirical efforts to go product (GDP) per worker in 1960, rose to beyond the country-level analysis and to 30.8 ­percent of US GDP per worker by 2017. understand the microeconomic foundations In contrast, Malawi, at a similar stage of of aggregate behavior have also provided development in 1960 (at about 4.4 ­ p ercent valuable insights on productivity. Two dif- of the US GDP per worker), dropped to ferent but complementary explanations may 2.2 ­percent in 2017.1 account for the productivity differences 1 2  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca between richer and poorer countries: The sectors of economic activity and across pro- first emphasizes the slow diffusion and adap- duction units (farms or firms). This is par- tation of technology as well as (production ticularly important in agriculture, where (a) organization and management) best practices sectoral productivity of African countries is to poorer countries (Bloom and Van Reenen extremely low relative to that of high-income 2007; Bloom et al. 2013; Parente and Prescott countries, and (b) the sector employs most of 2000, 2005). The second focuses on the dif- the population. That so much of the popula- ferences across sectors and firms in the allo- tion works in a sector of very low economic cation of resources in the production process. activity explains why Sub-Saharan Africa T he literature on misallocation of lags the rest of the world’s regions in struc- resources, which is the focus of this volume, tural transformation (Duarte and Restuccia argues that poorer countries are less effective 2010, 2018). than wealthier countries in allocating their At the production unit level, the low factors of production to their most efficient productivity of African countries can be uses. Conversely, the efficient allocation of explained by policies and institutions that resources across firms and sectors boosts foster a systematic redistribution of resources TFP by enabling productive firms to grow, from the more-productive establishments to low-productivity firms to exit the market, the less-productive ones. This allocative inef- and new firms to emerge (Foster, Haltiwan- ficiency across African production units can ger, and Krizan 2001; Hsieh and Klenow be attributed to market imperfections (for 2009; Restuccia and Rogerson 2008, 2013, example, regarding credit and land), pref- 2017). erential trade policies, size-dependent taxa- tion policies, and informality, among other causes. History and Context Economic growth in the Sub-Saharan Africa Effects of COVID-19 region historically has been plagued by a series of shocks: wars, political instability, The COVID-19 (Coronavirus) pandemic natural disasters, epidemics, terms-of-trade has significantly hit output and productiv- deterioration, and sudden stops in capital ity across African countries, sectors, and inflows, among others. These shocks have firms. This asymmetric supply shock has had lingering effects on the factors of produc- led to a disproportionate fall in demand as tion (physical capital and human capital) as production shuts down in some sectors, well as on TFP because of structural charac- and the lower demand is transmitted to teristics that exacerbate the impact of those less-contact-intensive sectors. In other words, shocks. These characteristics include, among the COVID-19 shock has caused a reduction others, the lack of diversification of economic in aggregate demand larger than the original activity, reliance on volatile commodity reduction in labor supply. This type of shock exports, weak governance, inadequate regu- has been labeled as a “Keynesian supply lation in labor and output markets, shallow shock” (Guerrieri et al. 2020). land markets, and underdeveloped financial Uncertainty increased dramatically at the markets with low access to financial prod- onset of this dramatic health shock, and this ucts. Some of these features are the outcome higher uncertainty has caused firms to tem- of policies and institutions that distort the porarily suspend their hiring and investment allocation of resources from their most effi- spending. Amid the COVID-19 pandemic, cient use. aggregate productivity is expected to fall The aggregate productivity gap of Afri- sharply as the decline in hiring and invest- can countries relative to high-income econ- ing slows the reallocation of resources from omies (notably, the United States) reflects low- to high-productivity firms (Baker et al. substantial productivity differences across 2020; Bloom 2014). This productivity impact B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   3 underlies the theories of uncertainty-driven two decades of uninterrupted growth, the business cycles, which emphasize how uncer- “Afro-pessimism” of the 1980s was being tainty shocks reduce investment, hiring, and replaced by “Afro-optimism.” In 13 of the 19 productivity (Bloom et al. 2018). years from 1996 through 2014, Sub-­ Saharan The policy response to COVID-19 cur- Africa grew faster than East Asia. In the rently focuses on (a) emergency relief mea- aftermath of the Global Financial Crisis and sures to protect lives (such as strengthening amid a growth slowdown in high-income the health sector; securing the food supply; economies, the region still grew at an average and enhancing access to water, sanitation, annual rate of 4.8 percent from 2011 to 2014. and handwashing stations); and (b) policies However, the region’s growth perfor- to protect livelihoods (for example, provid- mance is less stellar when accounting for pop- ing income support to the most vulnerable ulation growth. Real GDP per capita grew workers and extending credit to still-viable at an annual average rate of 1.95  ­ p ercent firms). However, the policy response should from 1996 through 2014. Amid the “Africa also consider measures to protect the future Rising” euphoria after 2000 —given the of the Africa region. Such a response requires persistence of the region’s rapid economic a comprehensive productivity policy agenda growth—the poverty rate in Sub-Saharan that addresses the human capital crisis, lever- Africa decreased (from 54.3 percent in 1990 ages digital technologies for trade and govern- to 40.1 percent in 2018), albeit more slowly ment effectiveness, and fosters intra-African than in East Asia (from 61.6 percent in 1990 value chains under the umbrella of the Afri- to 2.3 percent in 2018) and South Asia (from can Continental Free Trade Area (AfCFTA). 47.3 percent in 1990 to 12.4 percent in 2018) Policy makers in the region need to engage (World Bank 2020b). However, the num- with development partners to think ahead ber of poor people in Sub-Saharan Africa and design policies that build greater resil- increased from 278 million in 1990 to 416.4 ience and boost productivity so that African million in 2015, as the region’s population economies can recover faster and thrive in continued to expand rapidly. Most of the the post–COVID-19 era. A robust produc- world’s poor live in Sub-Saharan Africa, and tivity policy agenda would not only shorten without drastic policy actions, that number the region’s recovery time but also put it on a will only continue to grow.3 path of economic transformation with more, better, and more-inclusive jobs (World Bank Performance against Benchmarks 2020a). The sluggish growth of output per worker in Sub-Saharan Africa widened the region’s Sub-Saharan Africa’s Long-Term labor productivity gap relative to two familiar Performance: Still Far from the benchmarks: a global efficiency benchmark Frontier and an aspirational development benchmark. Overall Economic Growth and Poverty The former is proxied by the United States,4 Reduction while the latter refers to the “East Asian dragons” (or “EAP5,” comprising Indonesia, From 1996 to 2014, growth of real economic the Republic of Korea, Malaysia, Singapore, activity in Sub-Saharan Africa sharply accel- and Thailand). erated to an annual average rate of 4.8 per- cent, up from 1.4 percent in 1978–95.2 Most Global Efficiency Benchmark countries in the region experienced the ris- According to Penn World Table 9.1 data ing hopes and expectations that came along (updated from Feenstra, Inklaar, and Tim- with robust growth. Six of the world’s 10 mer 2015), the (population-weighted) aver- fastest-growing countries were in Sub-Saha- age output per worker in Sub-Saharan Africa ran Africa (Economist 2011). After nearly relative to that of the United States declined 4  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca from 11.9 percent in 1960 to 7.7 percent those in Sub-Saharan Africa. The contrasting in 2017 (figure 1.1). 5 This reveals not only evolution of labor productivity in Sub-Saharan stagnant but also very low labor productiv- Africa and the EAP5 indicates not only dif- ity in Sub-Saharan Africa over the past half ferences in the pace of (human and physical) century. capital accumulation but also a growing diver- The same cannot be said about the (pop- gence in TFP between the two regions. ulation-weighted) average output per worker Zooming in on the evolution of output in the EAP5 countries, which climbed from per worker in Sub-Saharan Africa relative to 8.5 percent in 1960 to 29.8 percent in 2017 the EAP5 shows that, in 1960, the region as (figure 1.1). Unlike East Asia, Sub-Saharan a whole had a head start in terms of labor Africa failed to make headway against the of productivity relative to Korea (by 20 per- global efficiency benchmark. cent), Indonesia (by 30 percent), and Thai- land (more than double), while its labor Aspirational Development Benchmark productivity was almost equal to Singapore’s The labor productivity trends show that (figure 1.2, panel a). In the 1980s, however, Sub-Saharan Africa lost its productivity labor productivity contracted in Sub-Saharan edge over the EAP5 countries over the past Africa while it monotonically increased in six decades. Workers in Sub-Saharan Africa all the EAP5 countries—albeit at varying during the 1960s were, on average, about speeds. By 2017, workers in Indonesia and 40–45 percent more productive than those in Thailand were more than twice as productive the EAP5. By the 2010s, workers in the EAP5 as those in Sub-Saharan Africa, and those in were more than three times as productive as Korea and Singapore were more than 6 and FIGURE 1.1  Output per Worker in Sub-Saharan Africa and EAP5 Countries, Relative to the United States, 1960–2017 0.30 0.25 Share of US output per worker 0.20 0.15 0.10 0.05 0 60 63 66 69 72 75 78 81 84 87 90 93 96 99 02 05 08 11 14 17 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 Sub-Saharan Africa East Asian 5 Sources: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: “EAP5” (or “East Asian 5”) refers to five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand. The aggre- gate real output per worker for Sub-Saharan Africa and the EAP5 is a population-weighted average. These calculations use the output-side real GDP per capita at chained purchasing power parity (PPP) rates (in US$, millions, at 2011 prices). The figure presents the Hodrick-Prescott permanent component of the ratio of output per worker of each region relative to the United States (US output per worker = 1.0). B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   5 FIGURE 1.2  Output per Worker in Sub-Saharan Africa versus Selected Country Groups, 1960–2016 a. Sub-Saharan Africa versus EAP5 countriesa 12.0 11.5 11.0 ln output per worker 10.5 10.0 9.5 9.0 8.5 8.0 7.5 19 0 20 4 20 6 20 8 20 0 20 2 20 4 16 19 0 19 2 19 4 66 19 8 72 19 4 76 19 8 80 19 2 19 4 19 8 19 0 92 19 4 96 20 0 20 2 86 20 8 7 0 1 1 1 6 6 6 6 7 7 8 8 8 9 9 0 0 0 9 0 19 19 19 19 19 19 19 19 Sub-Saharan Africa (weighted average) Indonesia Korea, Rep. Malaysia Singapore Thailand b. Sub-Saharan Africa versus Brazil, China, and India 12.0 11.5 11.0 ln output per worker 10.5 10.0 9.5 9.0 8.5 8.0 7.5 19 0 19 2 19 4 66 19 8 19 0 72 19 4 19 6 19 8 80 19 2 19 4 19 6 19 8 19 0 19 2 19 4 19 6 20 8 20 0 20 2 20 4 20 6 20 8 20 0 20 2 20 4 16 6 6 6 6 7 7 7 7 8 8 8 8 9 9 9 9 9 0 0 0 0 0 1 1 1 19 19 19 19 Sub-Saharan Africa (weighted average) Brazil China India Sources: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: The figure plots the natural logarithm (ln) of the output per worker of each country or group. These calculations use the output-side real GDP per cap- ita at chained purchasing power parity (PPP) rates (in US$, millions, at 2011 prices). The figure presents the Hodrick-Prescott permanent component of the output per worker (in logs). ln = natural log. a. “EAP5” refers to five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand. 10 times as productive, respectively. Overall, Finally, labor productivity in Sub-Saharan there is a clear divergence in labor produc- Africa also lost ground relative to three large tivity between Sub-Saharan Africa and its and dynamic emerging market economies: aspirational development benchmark. ­ Brazil, China, and India. Labor productivity 6  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca in Sub-Saharan Africa was double that of industry average) and the exit of low-pro- China and India in 1960, while it was slightly ductivity firms (relative to the industry lower than Brazil’s (figure 1.2, panel b). The average). It captures the aggregate effect of region’s boom-bust cycles in labor produc- firm churning (or turnover) on productiv- tivity led it to gradually diverge from these ity growth. other countries. By 2017, labor in India was A growing strand of the literature inves- nearly twice as productive, China more than tigates aggregate productivity as the result 2.5-fold, and Brazil more than triple that of of firm-level decision-making processes, Sub-Saharan Africa. whereby firms are assumed to have different levels of productivity even within narrowly Sources of Productivity Growth defined economic activities. (See, for instance, Bartelsman, Haltiwanger, and Scarpetta Drawing on firm-level census data, this report 2013; Foster, Haltiwanger, and Krizan 2001; evaluates the sources of firm productivity and Syverson 2011.) In this context, the sem- growth. Productivity gains within each sector inal work by Restuccia and Rogerson (2008) of economic activity are primarily the out- and Hsieh and Klenow (2009) has argued come of increased dynamism within produc- that the microstructure of production estab- tion units. Resource reallocation from less- to lishments in different economic sectors can more-productive firms and activities also con- help explain the development gap between tributes to industry-level productivity growth rich and poor countries. In their framework, in any market economy—especially in low-­ the production units exhibit different levels of income economies with greater distortions. productivity and hence size. Aggregate TFP Broadly speaking, the sources of produc- is, in turn, influenced by the distribution of tivity growth at the firm level (for countries productivity across production units, those either pushing the production possibility fron- units’ corresponding allocation of resources, tier or catching up to the productivity leaders) and the number of firms per capita.6 are as follows (Cusolito and Maloney 2018): • T  he within component, which accounts Role of Resource Misallocation for the productivity growth within firms. This report will focus on resource misallo- It depends on changes in the efficiency and cation as a potential explanation of low pro- intensity with which inputs are used in pro- ductivity (levels and growth) in Sub-Saharan duction (that is, to upgrade firms) owing Africa.7 Resource misallocation refers to to increased firm capabilities (including distortions in the allocation of inputs (such improved managerial skills, labor skills, as capital, land, and labor) across produc- innovation, and technology adoption tion units of varying sizes. In other words, it capacity). occurs when different production establish- The between component, which reflects •  ments are taxed at different rates. This focus the role of factor reallocation across firms on misallocation is grounded in the following in aggregate productivity growth. Increases dimensions: in the “between” component imply that the most-productive firms would com-  irst, the increasing role of TFP differences • F mand the most resources—thus rendering in explaining the labor productivity gap the largest output and productivity gains. between African countries and both the However, multiple distortions may limit global efficiency and aspirational develop- the productivity gains arising from this ment benchmarks. component. Second, the limited availability of firm-level •  The selection component, which accounts •  census data that would primarily permit for the gains arising from the entry of the testing of the static effects of misalloca- high-productivity firms (relative to the tion on aggregate productivity. In the few B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   7 countries with longitudinal data from firm- there is a unique allocation of labor and capi- level censuses (for example, Côte d’Ivoire tal across producers that maximizes total out- and Ethiopia), there will be an exploration put net of fixed operating costs. of the static and dynamic implications of Theoretically, inefficiencies in the allo- misallocation (which includes not only cation of labor and capital across heteroge- reallocation among incumbents but also neous producers will affect aggregate output reallocation by churning). and productivity through three different Third, the prevalence of policies and •  channels: institutions (including social norms) in  he technology channel reflects the level of • T Sub-Saharan African countries that drive productivity of each producer. If techno- production units away from efficiency logical changes increase the productivity of benchmarks. all producers, output will be greater. The selection channel reflects the choices •  Framework of Resource Allocation–or of producers that would operate in a given Misallocation industry, given the costs of entry and their levels of productivity. This strand of the literature on resource mis- The misallocation channel reflects the •  allocation assumes that aggregate output is allocation of capital and labor among the produced by several producers (N) that have operating producers. different (individual) levels of productivity (Ai). Firm i’s technology is summarized by a These three channels are not independent: production function (f ) that is strictly increas- any policy or institution that misallocates ing and strictly concave. There is a fixed cost resources across producers will potentially of operation (c) for any producer. Given an generate additional effects through both the aggregate demand of labor (H) and capital (K), selection and technology channels (figure 1.3). FIGURE 1.3  Sources of Resource Misallocation That Reduce Total Factor Productivity Productivity levels Factor utilization Firm capabilities Operating environment Market failures Statutory provisions Discretionary provisions Technology Selection Misallocation channel channel channel Total factor productivity losses Source: Original figure for this publication. 8  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca According to this framework, lower values cross-country differences in aggregate pro- of Ai reflect either slow adoption or inefficient ductivity levels, it is then crucial to investi- use of technology. The efficient allocation in gate the sources of misallocation. Resource this economy maximizes final output and is misallocation across different production characterized by two decisions: (a) the num- units might reflect the following (Restuccia ber of operating establishments (that is, estab- and Rogerson 2017): lishments that can pay the fixed cost, c); and • S tatutory provisions, including some fea- (b) the allocation of capital and labor across tures of the tax code and regulations—for the operating establishments. If either of instance, tax code provisions that vary with these decisions is distorted, the economy will firm characteristics (say, age or size); tariffs have lower output and hence lower aggregate targeting certain groups of goods; employ- TFP—as aggregate factor inputs (K and H) in ment protection measures; and land regu- the industry are constant. lations, among others An allocation of inputs that maximizes • Discretionary government (or bank) provi- output across production units (say, either sions that favor or penalize specific firms— firms or farms) takes place when, condi- for instance, subsidies, tax breaks, or tional upon their operation, the marginal low-interest loans granted to specific firms; (and average) products are equal across all preferential market access; and unfair bid- production units. In this equilibrium, no ding practices for government contracts, output gains would be obtained by reallo- among others cating inputs of production (such as capi- • Market imperfections such as monopoly tal, land, and labor) from production units power; market frictions (for example, in with low marginal products to those with credit and land markets); and enforcement high marginal products. In the efficient of property rights. allocation, the most productive operating establishments will demand more inputs. In other words, a production unit’s productiv- ity and size are positively associated in the Dimensions of the Productivity efficient allocation. In addition, production Assessment units with similar productivity levels com- The main objective of this report is to charac- mand the same amount of inputs and are of terize the evolution of output and productiv- identical size. ity in Sub-Saharan Africa. To accomplish this Deviations from the efficient allocation task, the report documents the region’s (labor of resources across firms may have implica- and multifactor) productivity trends on an tions for aggregate output and productivity. international, regional, and country basis. It Input choices that differ from the efficiency benchmarks productivity levels and growth model, even if they allocate more factors to in Sub-Saharan Africa in relation to countries the more-productive production units, will in other regions as well as in various African generate lower aggregate output. Given the country groups, classified by their degree of constant aggregate amount of inputs (such as natural-resource abundance and condition of capital, land, and labor), the output loss asso- fragility.8 Overall, the analysis of productiv- ciated with an inefficient allocation is also an ity trends is conducted for three different lev- aggregate TFP loss. In this context, misallo- els of data aggregation: aggregate, sectoral, cation refers to situations where resources are and establishment. not allocated efficiently across production units, and the cost of misallocation is typi- Aggregate Level cally measured in terms of aggregate output or TFP losses. First, the report estimates the level and If the misallocation of resources across growth of labor and multifactor productiv- these different producers helps explain ity in Sub-Saharan Africa (for the region as B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   9 a whole as well as across countries) and the Database, and International Labour Orga- extent and nature of productivity gaps in nization statistics to unbundle the industry relation to international benchmarks at the and services sectors. Within the industry aggregate level. Labor productivity is mea- sector, it distinguishes manufacturing sured by the ratio of real GDP to the number from nonmanufacturing activities (such as of persons employed. construction; mining and quarrying; and The report not only illustrates the region’s electricity, water, and gas). In the services labor productivity trends but also identifies sector, it classifies the different activities the sources of the persistent differences in as either market or nonmarket services. labor productivity between Sub-Saharan (Market services include wholesale and Africa and benchmark countries or regions. retail trade; hotels and restaurants; trans- To that end, the development accounting portation, storage, and communications; framework is used to decompose the differ- financial intermediation; and real estate. ences in the level of labor productivity into Non ma rket ser v ice s compr ise publ ic (a) differences in input intensity (such as administration and defense; education; capital-use intensity and land-use intensity); health and social work; and other com- and (b) differences in production efficiency munity, social, and personal service activ- (Hsieh and Klenow 2010). ities.) Using data on labor productivity In addition, the growth accounting and labor shares, this report examines the framework is used to examine the sources shifts of resources across sectors over the of growth of African economies. In other recent decades. words, it quantifies the proportion of growth attributed to factor accumulation and TFP Establishment Level growth (Solow 1957). The analysis of the sources of variation of labor productivity Third, the report presents evidence on using these two frameworks is fully presented (labor and multifactor) productivity at the in chapter 2. establishment level. Using the World Bank’s Living Standards Measurement Studies– Integrated Surveys on Agriculture (LSMS– Sectoral Level ISA) and manufacturing firm-level censuses Second, the report depicts labor productivity of select Sub-Saharan African countries, the trends at the sectoral level in Sub-Saharan report calculates quantity and revenue pro- Africa. Current research typically classifies ductivity (TFPQ and TFPR, respectively) at economic activity into three broad sectors: the farm level in agriculture and at the firm agriculture, industry, and services (see, for level in manufacturing. The coverage of instance, Duarte and Restuccia 2010; Her- countries in the region as well as time peri- rendorf, Rogerson, and Valentinyi 2014). This ods depends on the availability of microeco- classification has been broadly used to ana- nomic data. lyze the role of structural change—captured The core analysis of this report will by the reallocation of labor from low- to rest upon the assessment of the implica- high-productivity sectors—in explaining the tions of aggregate productivity of produc- differences in labor productivity in low- and tion decisions across agricultural farms middle-income countries (Diao, McMil- and m­ anufacturing firms in Sub-Saharan lan, and Rodrik 2017; Gollin, Lagakos, and Africa. Using farm- and firm-level data, it Waugh 2014; McMillan, Rodrik, and Ver- will assess the performance of production duzco-Gallo 2014) and particularly in Afri- units in terms of their productivity levels can countries (McMillan and Harttgen 2014; across African establishments relative to McMillan, Rodrik, and Sepulveda 2017). an efficiency benchmark by computing the T he repor t uses input-output data, extent of resource misallocation. This cal- the United Nations National Accounts culation will provide information on the 10  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca role of allocative inefficiencies in explaining Data and Measurement Issues productivity differences between establish- ments across Sub-Saharan African coun- One of the main challenges of empirical tries relative to those of other benchmark work in low- and middle-income countries, countries or regions. notably in Sub-Saharan Africa, is the issue The establishment-level analysis identifies of data quality. The poor quality of the data and discusses the different policies and insti- on macroeconomic, financial, and structural tutions that affect productivity and drive the indicators for less-developed countries and misallocation of resources across farms and for economies with large informal sectors— firms in Sub-Saharan Africa. Specifically, it particularly in Sub-Saharan Africa—has discusses a comprehensive but not exhaus- been well documented (Jerven 2010, 2013a, tive set of policies and institutions that are 2013b, 2013c). categorized by these potential sources of Empirical work on productivity in Sub-­ misallocation (Restuccia and Rogerson Saharan Africa is plagued by problems con- 2017): cerning data availability, comparability, and quality. At the national level, these problems • M arket imperfections. The analysis dis- are often tied to issues of capacity: The pro- cusses credit market imperfections (that duction of high-quality data for national is, lack of access to finance due to the lack income and product accounts (NIPA), con- of collateral); lack of land titling, affect- sumption surveys, and firm-level censuses ing the allocation of land; and informa- is technically complex. It involves the large- tion frictions, affecting producers that are scale mobilization of sizable financial and not connected to markets or farmers who human resources as well as the setup of have inadequate information on weather robust quality control mechanisms. Addition- forecasts. ally, the failure of statistical offices to adhere • Statutory provisions. Also discussed are to methodological and operational standards size-dependent policies—more specifically, leads to data comparability and quality issues tax provisions and regulations that depend (Beegle et al. 2016). on features of the different production At the aggregate level, problems with units (such as size and age) as well as trade NIPA quality in Sub-Saharan African coun- policies that protect specific categories of tries have been extensively reported (Jerven goods. 2010). Inaccuracies in the output and pro- • Discretionary provisions. In addition, the ductivity data reported by national statistical report captures government provisions that systems have led to (potentially) misleading favor or penalize certain types of produc- country productivity rankings. Output and tion units—for instance, subsidies to farm- productivity estimates in international cur- ers, low-interest lending to specific firms, rency showed significant variation across and preferential market access for specific countries because of the varying reliability of groups of producers, among others. the data sources or differences in the meth- Finally, this repor t— launched and ods chosen to express the data in interna- financed by the World Bank’s Office of tional currency.9 the Chief Economist of the Africa Region Output and productivity estimates across (AFRCE)—is part of the Bank’s program- African countries can also be volatile, not matic agenda on the drivers of productiv- only because of the low quality of statistical ity worldwide, emphasizing the factors that services but also partly because of the large explain the productivity gap of emerging weight of sectors (such as agriculture) that markets (and, notably, Sub-Saharan African are prone to volatile domestic shocks and vul- countries) relative to the high-income world. nerable to fluctuating international commod- Box 1.1 succinctly describes the goals of some ity prices. This report highlights some of the of these research projects. data production problems facing the region’s B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   11 BOX 1.1  Building upon the World Bank’s Productivity Research Agenda This report is part of the World Bank Group’s program- in B­ razil, Côte d’Ivoire, Ethiopia, Hungary, India, matic agenda on productivity and part of the regional Indonesia, Mexico, South Africa, Thailand, Tunisia, studies program of the Bank’s Office of the Chief Econ- and Turkey. Its findings reveal that high-growth firms omist of the Africa Region (AFRCE). It complements are powerful engines of job and output growth. They other research projects conducted or already published also create positive spillovers for other businesses along in the region as well as the output from the Productivity the value chain. Project, an initiative of the Vice Presidency for Equita- H ar ve s t ing Pr os pe r it y: Tec h n ol og y a n d ble Growth, Finance and Institutions. Productivity Growth in Agriculture (Fuglie et al. The Productivity Project seeks to bring frontier 2019) uses recent impact evaluations to examine the thinking on the measurement and determinants of constraints to farmers’ adoption of technology and productivity, grounded in the developing-country dissemination of productivity-enhancing technologies. context, to global policy makers. Among its reports It also discusses recent developments in agriculture are the six described below. value chains and the emergence of new institutional The Innovation Paradox: Developing-Country arrangements to include smallholder farms in these Capabilities and the Unrealized Promise of value chains. Technological Catch-Up (Cirera and Maloney Industrializing for Jobs in Africa? (Abreha et al., 2017) documents the small investments in inno- forthcoming) addresses (a) the lack of industrializa- vation undertaken by low- and middle-income tion in postindependence Sub-Saharan Africa and country firms and governments even though the some countries’ patterns of deindustrialization; (b) the returns from these investments are potentially high. prospects for the region’s countries to undergo indus- Underlying this “innovation paradox,” the evidence trialization through participation in regional or global suggests, is the lack of complementary physical manufacturing value chains over the next two decades; and human capital—in particular, firm managerial and (c) industrial policy tools that might foster country capabilities—needed to reap the returns to inno- participation in the right regional or global manufac- vation investments. Countries need to build firms’ turing value chains. The authors’ analysis is conducted capabilities and embrace an expanded concept of the at three levels of aggregation: country level, country National Innovation System (NIS), incorporating a groups defined by resource abundance, and income broader range of market and systemic failures.a groups and natural trade groupings. Productivity Revisited: Shifting Paradigms in Inclusive Digital Africa (Begazo-Gomez, Blimpo, Analysis and Policy (Cusolito and Maloney 2018) and Dutz (forthcoming) addresses why digital tech- presents a “second wave” of thinking in productivity nology (DT) is important for Africa’s development. analysis and its implications for productivity poli- It focuses on understanding the current drivers of cies. It tests these hypotheses across select middle-in- DT adoption by individuals, households, and enter- come countries (for instance, Chile, Colombia, and prises, as well as the linkages between DT adoption Malaysia). It provides a more accurate calculation of and business-driven productivity growth, output, and distortions and examines more rigorously their impor- aggregate jobs expansion as a contribution to poverty tance as the primary barrier to productivity growth. reduction and inclusion outcomes in Africa today. It recommends a more comprehensive analysis of firm More specifically, it addresses the extent to which performance that includes efficiency, quality upgrad- (a) barriers impede DT adoption in Africa; (b) DT ing, and demand expansion. The authors advocate adoption by existing and new enterprises (firms and ­ an integrated approach to productivity analysis that farms, both formal and informal), as well as by people accounts for the need to (a) reduce distortions, (b) cre- who are or could be working in these enterprises, can ate human capital capable of identifying opportunities generate productivity gains and aggregate output and offered to follower countries, and (c) upgrade firm jobs expansion; and (c) these gains can have greater capabilities. impact on poverty reduction and inclusion outcomes. High-Growth Firms: Facts, Fiction, and Policy a. The NIS refers to the institutions, human capital, and interactions between Options for Emerging Economies (Grover Goswami, them that facilitate the creation and diffusion of knowledge. They focus on Medvedev, and Olafsen 2019) examines whether tar- policies that not only foster research and development (R&D) investments but geting high-potential firms can enable more economic also upgrade firm capabilities. In addition to addressing barriers to knowledge capital accumulation, an NIS should also consider barriers to the accumulation dynamism. It presents evidence on the occurrence, of all types of capital—barriers such as business climate, bankruptcy laws, poor features, and determinants of high-growth firms ­ product and factor regulation, and so on (Maloney 2017). 12  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca national statistical systems and some of the systems are trapped in a vicious circle where data usage problems facing researchers. inadequate funding undermines the produc- tion of high-quality data, in turn reducing demand for the data, which further reduces Outdated Output and Productivity resources as well (Jerven 2013c). Estimates Many statistical offices in the region The recent slew of national account rebasing still use outdated methods and data, and exercises (updating the base year of constant they lack the capacity to handle an efficient price estimates) in Sub-Saharan Africa (for and transparent revision of their national example, in Ghana and Zambia in 2010, and accounts. Even if more of the region’s statis- in Kenya, Nigeria, Tanzania, and Uganda tical offices were to apply more recent stan- in 2013) have called attention to the use of dards of national accounts (say, base year outdated economic structures, lack of adher- 2008), they are unequally adapted at the ence to international standards of national national level. Moreover, there is no agree- accounts measurement, and more broadly, ment on methods to deal with the growth the unreliability of income and product effects of these revisions. The best practice in estimates. the buildup of NIPA data should focus not For instance, the Ghana Statistical Ser- only on international standardization (such vices (GSS) revised its 2010 GDP to Ȼ44.8 as the International Comparison Program billion, 60.3 percent higher than its previous [ICP] or the Penn World Tables) but also on estimate of Ȼ25.6 billion (Jerven 2013a). The fostering local conditions across statistical upward revision was attributed to the inclu- offices to timely and reliably produce and dis- sion of new data on unmeasured parts of the seminate surveys. economy as the GSS changed the base year from 1993 to 2006. Bias in Human Capital Assessment: Nigeria revised its 2013 GDP upward to Overstating School Enrollment Data US$509 billion after changing the base year for calculation from 1990 to 2010. It now There is evidence of systematic biases in includes previously uncounted industries like administrative data systems in the reporting telecoms, information technology, music, of primary schooling enrollment data. These online sales, airlines, and film production. biases do not necessarily reflect the lack of Thanks to this new calculation, Nigeria over- analytical capacity. In some cases, they are took South Africa as the largest economy in the outcome of incentives to overestimate the region and the 26th largest in the world progress in the sector. Overestimation of (Blas and Wallis 2014). school enrollment data results, at least partly, The rebasing of output and productivity from incentives provided by the governance estimates in Sub-Saharan Africa reveals that and funding structures of the Ministries of important segments of economic activity had Education, especially in low-income, highly gone missing for decades—say, air transpor- aid-dependent countries (Sandefur and Glass- tation and information and communication man 2015). technology (ICT) services—and that the lack As national government and line min- of backward estimates thwarts the accurate istries seek to allocate resources between accounting of economic history in these school districts and evaluate teacher perfor- countries based on official statistics (Jerven mance, the decision process is fed by infor- 2013b). These revisions also reconfigure the mation from administrative systems based map of income, productivity, and growth in on teacher self-reporting. The administrative Sub-Saharan Africa. And they raise concerns data show significant evidence of overreport- about the status of income and product sta- ing of enrollment growth: the average change tistics in other Sub-Saharan African coun- in enrollment is nearly one-third higher tries—many of whose national statistical (3.1 percentage points) in administrative data B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   13 than in survey data across 21 African coun- for the livelihood of large segments of the tries, and this optimistic bias is completely population. Consequently, their govern- absent in data outside Africa (Sandefur and ments cannot assign their limited funding Glassman 2015).10 Overall, the resulting to improvement of statistical quality (FAO systemic misreporting undermines the gov- 2008; World Bank 2004). ernment’s ability to manage public services, The global strategy spearheaded by the especially in remote rural areas. Food and Agriculture Organization (FAO) of the United Nations to improve agricultural and rural statistics, and the corresponding Unreliable Employment and Wage regional action plans, represents efforts to Labor Estimates improve agricultural standards and practices. The importance of nonmonetary, subsistence, The advent of new (and relatively affordable) informal, or unrecorded economic activities technologies and rigorous research is help- in Sub-Saharan Africa may place into ques- ing foster the adoption or improvement of tion the reliability of the reported data on cost-effective standards in agricultural sta- labor and income. Informality alone makes tistics (Carletto, Jolliffe, and Banerjee 2015). it difficult to draw the production bound- Finally, statistical systems need to promote ary of Sub-Saharan African economies (Jer- enhanced integration of agricultural data and ven 2010). For instance, recent employment other types of data sources (for example, on surveys in Tanzania suggest that self-em- poverty and nutritional, socioeconomic, and ployment is by far the most prevalent type environmental conditions) to better inform of employment relationship in the informal sectoral policies. economy. Wage labor in Sub-Saharan African coun- Limited Availability of Firm-Level tries might also be underestimated for two Census Data other reasons (Rizzo, Kilama, and Wuyts 2015): (a) labor force surveys that lump it The scope and breadth of the microeconomic together with self-employment, and (b) poor analysis of productivity in this report is lim- understanding of the trends toward subcon- ited by the sparse availability of firm-level tracting of informal labor services (instead census data across Sub-Saharan African of direct production of goods). In the case of countries. Fewer than a handful of countries Tanzania, labor surveys failed to capture the in the region conduct surveys at the establish- heterogeneity of employment relations found ment level—and even fewer provide longitu- in the informal economy and the heterogene- dinal firm-level census data. This statistical ity of relationships between capital and labor deficit clearly hurts African countries’ ability that mediate poor people’s participation in to formulate good policy decisions. To over- the (informal) economy (Rizzo, Kilama, and come the lack of firm-level census data across Wuyts 2015). Remedial measures include the region’s countries, researchers have used abandoning the misplaced aggregation in the alternative sources of data such as the World classification of labor regimes, which results Bank Enterprise Surveys (WBES). from conflating into a single catchall category The evidence suggests that mismeasure- various forms of production and employment ment in the distribution of firms at the four- that are essentially different. digit Standard Industrial Classification (SIC) level across African manufacturing indus- tries—which might overrepresent large firms Inadequate Agricultural Statistics relative to firm-level census data—leads to There are severe weaknesses in the mea- biases in the computed extent, using WBES surement of agricultural outcomes in Sub-­ data, of resource misallocation. Inaccuracies Saharan Africa—especially in the poorest in the measurement of value-added shares of countries that depend critically on this sector industries in narrower industry groups would 14  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca likely overestimate or underestimate the as well as the methodology and periodic- degree of misallocation because they would ity of firm-level censuses. Such new and reflect both the true share and a sampling increased data impose other challenges: (a) error. The WBES-based measure of misal- wider availability of output price data rather location would tend to be overestimated if than input price data at the establishment sectors with higher misallocation are over- level; (b) reported output prices that are, in represented relative to their shares in the most cases, unit values; and (c) the need to census. Evidence for African countries shows undertake surveys at the product level if most that most industries would have smaller mis- manufacturing establishments in a specific ­ allocation in the WBES than their dispersion sector are multiproduct. in the census data. Hence, the WBES might Having greater data availability on output underestimate the true misallocation of each and input prices does not prevent the need sector and therefore underestimate manu- to impose more structure to identify the role facturing productivity dispersion (Cirera, played by demand shocks in the measured Fattal-Jaef, and Maemir 2018). ­ TFPR. Recent research using firm-level cen- sus with price data shows that there is still a larger dispersion of TFPR across manufac- Limited Interpretation of turing firms in Ethiopia, and this is mirrored Microeconomic Evidence by large differences in physical productivity. One of the most widely used measures of Prices tend to vary significantly less than pro- firm-level productivity in the literature is ductivity levels and do not constitute a major total factor productivity revenue (TFPR)— driving factor of TFPR differences (Söder- typically defined as the ratio of firms’ sales bom 2018). (or revenues) to input costs (appropriately weighted by their production elasticities). It has been argued that TFPR is a mea- Plan of the Volume sure of profitability (or firm performance) This volume documents the productivity rather than productivity. Hence, differ- trends in Sub-Saharan Africa in three dif- ences in TFPR across firms may capture ferent dimensions, assessing productivity at not only differences in physical efficiency the aggregate level, the sectoral level, and the but also differences in prices, which reflect establishment level. It characterizes the evo- product differentiation and markups in lution of productivity in the region relative to addition to costs (De Loecker and Goldberg other countries and regions as well as coun- 2014). The emergence of (output and less try groups in Africa classified by their degree often input) price data and new techniques of natural-resource abundance and condition applied to databases with firm-level prices of fragility. has enabled researchers to compute more The core of this volume rests upon the accurate measures of physical efficiency. assessment of the implications for aggregate Evidence on the use of these techniques for productivity of production decisions across emerging markets is presented in Cusolito agricultural farms and manufacturing firms and Maloney (2018) and references therein. in Sub-Saharan Africa. The next three chap- Future work in Africa needs to distinguish ters will present evidence on aggregate pro- productivity shocks (or technical efficiency) ductivity from the perspective of production from demand shocks in the measures of units, using recent household surveys for TFPR among Sub-Saharan African produc- farmers and firm-level surveys for select Afri- tion establishments. This requires the timely can countries as well as frontier estimation availability and recurrent production of techniques. The empirical work presented in high-quality data on output and input prices this volume can provide further guidance for at the establishment level—a task that does productivity analysis and the design of a pol- not preclude improving the country ­ coverage icy agenda for the region. B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   15 Chapter 2, “Needed: Boosting the Con- medium-size firms as well as large formal tribution of Total Factor Productivity,” sector firms. On the other hand, the unavail- documents the growth performance of ability of firm-level data for the services sec- Sub-Saharan Africa over the past half cen- tor prevents us from extending the services tury both across countries and across sec- sector analysis to African countries. tors of economic activity. Despite an uptick The evidence shows that agriculture in labor productivity since 1996, the region and manufacturing in Sub-Saharan Africa has failed to catch up to either high-income are plagued by severe misallocation of countries (notably, the United States) or to resources. The region’s low agricultural groups of middle- to high-income countries productivity is not attributed to the qual- such as the EAP5 (Indonesia, Korea, Malay- ity of its soil or the amount of rainfall. It sia, Singapore, and Thailand). The sizable is overwhelmingly explained by inefficien- gap in output per worker between Sub-­ cies in the allocation of resources. In man- Saharan African countries and those two ufacturing, the misallocation is captured by benchmark groups is primarily attributed to TFPR dispersion—which is larger than that a lower relative stock of physical and human of other low- and middle-income countries capital (from the 1960s to the 1980s). During (China and India) and the efficiency bench- 2000–17, inefficiencies in the region’s factor mark (United States). production use have played an increasing role Both agricultural and manufacturing pro- in explaining this gap. duction units tend to face higher distortions At the sectoral level, the analysis in in Sub-Saharan Africa than in other regions. this volume unpacks the various indus- In turn, these distortions decelerate the try and ­ s ervices sectors into a five-sector growth of the production units, disincentiv- classification: agriculture, manufacturing, ize their adoption of productivity-enhancing nonmanufacturing, market services, and technologies, and reduce the ability of their nonmarket services. Sectoral labor produc- peers to learn new techniques. tivity in Sub-Saharan Africa exhibits long Chapter 4, “Policies and Institutions that swings in the medium term over the past Distort Resource Allocation in Sub-Saharan quarter century, and it is lower than in the Africa,” explains how policies and institu- United States, especially in agriculture. tions have distorted the allocation of inputs Broadly speaking, the structural transforma- (capital, land, and labor) across heteroge- tion of Sub-Saharan Africa tends to lag that neous production units. These policies and of other world regions. Agricultural employ- institutions can be classified into potential ment shares have declined more slowly and sources of misallocation: (a) market imper- remain higher than in other regions. fections (restricted access to finance, lack Chapter 3, “Resource Misallocation in of land titling or rental markets, and infor- Sub-Saharan Africa: Firm-Level Evidence,” mation frictions affecting market connectiv- documents the extent of resource misal- ity); (b) statutory provisions (size-dependent location across agricultural and manufac- taxes and regulations); and (c) discretionary turing production units in Sub-Saharan provisions (targeted subsidies and preferen- Africa. The agriculture sector analysis uses tial trade policies). household-level panel data from the World ­ Allocative inefficiencies affect output and Bank’s LSMS-ISA initiative for selected coun- productivity levels through three channels: tries in the region as well as geographically technology, selection (occupational choices), gridded data on actual and potential crops, and misallocation. These three channels can crop choices, and land endowments from the be interdependent. For instance, policies or FAO’s Global Agronomic Ecological Zones institutions that lead to resource misallo- (GAEZ) database. The manufacturing sector cation can potentially generate additional analysis uses firm-level manufacturing census effects through both the selection and tech- data that adequately accounts for small and nology channels. 16   Boosting Productivity in Sub-Saharan Africa The pervasive misallocation of land in is still at the world’s technological frontier Sub-Saharan Africa can be influenced by (Duarte and Restuccia 2006; Restuccia 2011). the lack of land titling and the underdevel-   5. These figures roughly implied that the average opment of land rental markets. This volume US worker produced in 28 days what the aver- age worker in Sub-Saharan Africa produced shows evidence that rental activity can help throughout a year in 2017, down from an reallocate land from less- to more-productive average of 43 days in 1960. farms. Still, land markets are subject to other   6. See Hopenhayn (2014) for a unifying theoret- frictions, and farms that rent land operate ical framework and review of the literature on far from the efficiency benchmark. Credit misallocation. market imperfections introduce distortions   7. The report will not focus on the drivers of to entry and technology adoption decisions. within-firm productivity. In other words, pro- By distorting entrepreneurs’ entry decisions, ductivity improvements due to better mana- credit market imperfections can lead to pov- gerial practices, greater input quality, product erty traps. Asset grant programs to the poor innovation and research and development can help them escape from poverty by iden- (R&D) investments, and firm structure deci- sions, among others, are beyond the scope of tifying potential high-growth entrepreneurs this report. For a comprehensive review of and facilitating their growth. these issues, see Cirera and Maloney (2017), This chapter also highlights the adoption Syverson (2011), and the references therein. of digital technologies to reduce some of  8. Resource-rich countries, in this report, these market frictions. For instance, mobile are those nations with rents from natural money has raised financial inclusion in sev- resources (excluding forests) that exceed 10 eral African countries. The insertion of percent of GDP; that is, the sum of oil rents, digital technologies in finance has granted natural gas rents, coal rents (hard and soft), individuals access to savings instruments and and mineral rents should exceed 10 percent loan products. of GDP over the past decade. Estimates of Chapter 5, “Agenda for Future Research,” natural-resource rents are based on Lange, Wodon, and Carey (2018). On the other discusses further avenues of research that may hand, fragile and conflict-affected situations provide further insights on the productivity are defined as economies having either (a) a dynamics across countries in the region—for harmonized Country Policy and Institutional instance, distinguishing demand from supply Assessment (CPIA) rating of 3.2 or less, or forces—and identify the different channels of (b) the presence of a United Nations and/or policy transmission to enhance productivity. regional peacekeeping or peace-building mis- sion during the past three years.   9. In this context, the extent of inaccuracies in Notes the data cannot be easily evaluated because   1. Percentages calculated by the author from it also reflects the underdevelopment of the GDP per worker data in Feenstra, Inklaar, region’s different countries. and Timmer (2015). 10. Discrepancies between administrative and   2. Regional economic growth data for Sub-­ survey data series in Kenya and Rwanda were Saharan Africa are calculated by the author. concomitant with the shift from bottom-up   3. The rising concentration of extreme poverty financing of education (through user fees) to in Sub-Saharan Africa over the past quarter top-down finance (through per pupil central century can be attributed to two primary fac- government grants). This highlights the inter- tors: (a) economic growth that has been nei- dependence of public finance systems and the ther as fast as population growth nor inclusive integrity of administrative data systems. enough to put a big dent in poverty, and (b) the persistently low contribution of TFP to economic growth. References   4. The development accounting literature uses Abreha, Kaleb G., Woubet Kassa, Emmanuel the United States as a benchmark given that K. K. Lartey, Taye A. Mengistae, Solomon it is a large, stable, and diverse country that Owusu, and Albert G. Zeufack. Forthcoming. B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   17 The Prospects of ­ I ndustrialization in Sub-­ Sub-Saharan Africa.” World Bank Economic Saharan Africa: Seizing O ­ pportunities in Review 34 (1): 75–100. Global Value Chains. Washington, DC: Cirera, Xavier, and William F. Maloney. 2017. World Bank. The Innovation Paradox: Developing-Country Baker, Scott R., Nicholas Bloom, Steven J. Davis, Capabilities and the Unrealized Promise of and Stephen J. Terry. 2020. “COVID-­ Induced Technological Catch-Up. Washington, DC: E conomic Uncertainty.” Working Paper World Bank. No. 26983, National Bureau of Economic Cusolito, Ana Paula, and William F. Maloney. Research, Cambridge, MA. 2018. Productivit y Re visited: S hif ting Bartelsman, Eric, John Haltiwanger, and Stefano Paradigms in Analysis and Policy. Washington, Scarpetta. 2013. “Cross-Country Differences DC: World Bank. in Productivity: The Role of Allocation and De Loecker, Jan, and Pinelopi Koujianou Goldberg. Selection.” American Economic Review 103 2014. “Firm Performance in a Global Market.” (1): 305–34. Annual Review of Economics 6 (1): 201–27. Beegle, Kathleen, Luc Christiaensen, Andrew Diao, Xinshen, Margaret McMillan, and Dani Dabalen, and Isis Gaddis. 2016. Poverty in a Rodrik. 2017. “The Recent Growth Boom in Rising Africa. Washington, DC: World Bank. Developing Economies: A Structural Change Begazo-Gomez, Tania, Moussa P. Blimpo, and Perspective.” Working Paper 23132, National Mark A. Dutz. Forthcoming. Digital Africa: Bureau of Economic Research, Cambridge, MA. Building the Infrastructure Foundations and Duarte, Margarida, and Diego Restuccia. 2006. Facilitating the Adoption of Technologies for “The Productivity of Nations.” Federal Jobs. Washington, DC: World Bank. Rese r ve Bank of R ichmond Economic Blas, Javier, and William Wallis. 2014. “Nigeria Quarterly 92 (3): 195–223. Almost Doubles GDP in Recalculation.” Duarte, Margarida, and Diego Restuccia. 2010. Financial Times, April 7. “The Role of the Structural Transformation in Bloom, Nicholas. 2014. “Fluct uations in Aggregate Productivity.” Quarterly Journal of U n c e r t a i n t y.” J o u r n a l o f E c o n o m i c Economics 125 (1): 129–73. Perspectives 28 (2): 153–76. Duarte, Margarida, and Diego Restuccia. 2018. Bloom, Nicholas, Benn Eifert, Aprajit Mahajan, “Structural Transformation and Productivity David McKenzie, and John Roberts. 2013. in Sub-Saharan Africa.” Unpublished manu- “Does Management Matter? Evidence from script, University of Toronto. India.” Quarterly Journal of Economics Economist. 2011. “Africa Rising.” December 3. 128 (1): 1–51. FAO (Food and Agriculture Organization of the Bloom, Nicholas, Max Floetotto, Nir Jaimovich, United Nations). 2008. “Agricultural B ­ ulletin Itay Saporta-Eksten, and Stephen J. Terry. Board on Data Collection, Dissemination 2018. “Really Uncertain Business Cycles.” and Quality of Statistics (ABCDQ) project” Econometrica 86 (3): 1031–65. d eveloped by the FAO Statistics Division. ­ Bloom, Nicholas, and John Van Reenen. 2007. http://faostat.fao.org/abcdq/. “Measuring and Explaining Management Feenstra, Robert C., Robert Inklaar, and Marcel Practices across Firms and Countries.” P. Timmer. 2015. “The Next Generation of Quarterly Journal of Economics 122 (4): the Penn World Table.”  American Economic 1351–1408. Review 105 (10): 3150–82. C a rl e t to , C a lo g e ro , D e a n J ol l i f f e , a nd Foster, Lucia, John C . Haltiwanger, and Raka Banerjee. 2015. “From Tragedy to C. J. Krizan. 2001. “Aggregate Productivity Renaissance: Improving Agricultural Data Grow th: L essons from M icroeconomic for Better Policies.” Journal of Development Ev ide nc e .” I n N e w D e ve l o p m e n t s i n Studies 51 (2): 133–48. Productivity Analysis , edited by Charles Caselli, Francesco. 2005. “Accounting for Cross- R. Hulten, Edwin R. Dean, and Michael Country Income Differences.” In Handbook of J. Harper, 303–72. Chicago: University of Economic Growth, Vol. 1, Part A, edited by Chicago Press. Philippe Aghion and Steven Durlauf, 679–741. Fug l ie , Keit h , M ad hu r G oya l, Apa rajit a Amsterdam: Elsevier. Gautam, and William F. Maloney. 2019. Ci rera, X avier, Rober to Fat tal-Jaef, a nd Harvesting Prosperity: Technolog y and Hibret Maemir. 2018. “Taxing the Good? Pro d u c t i v i t y G ro w t h i n A g r i c u l t u re . Distortions, Misallocation, and Productivity in Washington, DC: World Bank. 18  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Gollin, Douglas, David Lagakos, and Michael a Diagnosis of Africa’s Statistical Tragedy E. Waugh. 2014. “Agricultural Productivity towards a Statistical Renaissance.” African Differences across Countries.” American Arguments (online platform of the Royal Economic Review 104 (5): 165–70. African Society), September 26. Grover Goswami, Arti, Denis Medvedev, and Jones, Charles I. 2016. “The Facts of Economic Ellen Olafsen. 2019. High-Growth Firms: Growth.” In Handbook of Macroeconomics, Facts, Fiction, and Policy Options for Vol. 2A, edited by John B. Taylor, and Harald Emerging Economies . Washington, DC: Uhlig, 3–69. Amsterdam: Elsevier. World Bank. Kim, Young Eun, and Norman V. Loayza. G u e r r i e r i , Ve ro n i c a , G u id o L o r e n z o n i , 2019. “Productivity Growth: Patterns and Ludwig Straub, and Iván Werning. 2020. Determinants across the World.” Policy “Macroeconomic Implications of COVID-19: Research Working Paper 8852, World Bank, Can Negative Supply Shocks Cause Demand Washington, DC. Shortages?” Working Paper No. 26918, Klenow, Peter J., and Andrés Rodríguez-Clare. National Bureau of Economic Research, 1997. “The Neoclassical Revival in Growth Cambridge, MA. Economics: Has It Gone Too Far?” In NBER Hall, Robert E., and Charles I. Jones. 1999. “Why Macroeconomics Annual 1997, edited by Ben Do Some Countries Produce So Much More S. Bernanke and Julio J. Rotemberg, 73–103. Output per Worker than Others?” Quarterly Cambridge, MA: MIT Press. Journal of Economics 114 (1): 83–116. Lange, Glenn-Marie, Quentin Wodon, and Kevin Herrendorf, B erthold, R ichard Rogerson, Carey, eds. 2018. The Changing Wealth of and Ákos Valentinyi. 2014. “Growth and Nations 2018: Building a Sustainable Future. Structural Transformation.” In Handbook of Washington, DC: World Bank. Economic Growth Vol. 2, edited by Philippe Maloney, William F. 2017. “Revisiting the Ag hion and Steven Du rlauf, 855 –941. National Innovation System in Developing Amsterdam: Elsevier. Countries.” Policy Research Working Paper H o p e n h a y n , H u g o A . 2 0 14 . “ F i r m s , 8219, World Bank, Washington, DC. Misallocation, and Aggregate Productivity: A McMillan, Margaret, and Kenneth Harttgen. Review.” Annual Review of Economics 6 (1): 2014. “What Is Driving the ‘African Growth 735–70. Miracle’?” Working Paper 20077, National Hsieh, Chang-Tai, and Peter J. Klenow. 2009. Bureau of Economic Research, Cambridge, “Misallocation and Manufacturing TFP MA. in China and India.” Quarterly Journal of McMillan, Margaret, Dani Rodrik, and Claudia Economics 124 (4): 1403–48. S epu lveda. 2017. “St r uc t u ral Change, Hsieh, Chang-Tai, and Peter J. Klenow. 2010. Fundamentals and Growth: A Framework and “Development Accou nting.” A m e ric an Case Studies.” Working Paper 23378, National Economic Journal: Macroeconomics 2 (1): Bureau of Economic Research, Cambridge, MA. 207–23. McMillan, Margaret, Dani Rodrik, and Íñigo Jer ven , Mor ten. 2010. “A f ric a n Grow t h Verduzco- Gallo. 2014. “Globalization, Recurring: An Economic History Perspective S t r u c t u ra l C h a n ge , a nd P ro du c t iv it y on Africa Growth Episodes, 1690 –2010.” Growth, with an Update on Africa.” World Economic History of Developing Regions 25 Development 63: 11–32. (2): 127–54. Parente, Stephen L., and Edward C. Prescott. Jerven, Morten. 2013a. “Comparability of GDP 2000. Barriers to Riches. Cambridge, MA: Estimates in Sub-Saharan Africa: The Effect MIT Press. of Revisions in Sources and Methods Since Parente, Stephen L., and Edward C. Prescott. Structural Adjustment.” Review of Income 2005. “A Unified Theory of the Evolution of and Wealth 59 (S1): S16–S36. International Income Levels.” In Handbook Jerven, Morten. 2013b. Poor Numbers: How We of Economic Growth, Vol. 1, Part B, edited Are Misled by African Development Statistics by Philippe Aghion and Steven Durlauf, and What to Do about It. Ithaca, NY: Cornell 1371–1416. Amsterdam: Elsevier. University Press. Restuccia, Diego. 2011. “Recent Developments in Jerven, Morten. 2013c. “Why We Need to Invest Economic Growth.” Federal Reserve Bank of in African Development Statistics: From Richmond Economic Quarterly 97 (3): 329–57. B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca   19 Restuccia, Diego, and Richard Rogerson. 2008. Manufacturing Sector.” Unpublished manu- “Policy Distortions and Aggregate Productivity script, University of Gothenburg, Sweden. with Heterogeneous Establishments.” Review Solow, Robert M. 1957. “Technical Change of Economic Dynamics 11 (4): 707–20. and the Aggregate Production Function.” Restuccia, Diego, and Richard Rogerson. 2013. Review of Economics and Statistics 39 (3): “Misallocation and Productivity.” Review of 312–20. Economic Dynamics 16 (1): 1–10. Syverson, Chad. 2011. “W hat Determines Restuccia, Diego, and Richard Rogerson. 2017. Productivity?” Journal of Economic Literature “The Causes and Costs of Misallocation.” 49 (2): 326–65. Journal of Economic Perspectives 31 (3): 151–74. World Bank. 2004. “The Marrakech Action Plan R i zzo, M at teo, Blandina K ilama, and for Statistics. Better Data for Better Results: Marc Wuyts. 2015. “The Invisibility of An Action Plan for Improving Development Wage Employment in Statistics on the Statistics.” Paper presented at the Second Informal Economy in Africa: Causes and International Roundtable on Managing for Consequences.” Journal of Development Development Results, Marrakech, Morocco, Studies 51 (2): 149–61. February 4–5. Sandefur, Justin, and Amanda Glassman. 2015. World Bank. 2020a. Africa’s Pulse: Assessing the “The Political Economy of Bad Data: Evidence Impact of COVID-19 and Policy Responses from African Survey and Administrative in Sub- Saharan Africa, vol. 21 (April). Statistics.” Journal of Development Studies Washington, DC: World Bank. 51 (2): 116–32. World Bank. 2020b. Pover t y and Shared S ö d e r b o m , M å n s . 2 018 . “ P r o d u c t i v i t y Prosperity 2020: Reversals of Fortune. Dispersion and Firm Dynamics in Ethiopia’s Washington, DC: World Bank. Needed: Boosting the Contribution of Total Factor 2 Productivity The Divergent Paths of Malaysia two countries had similar initial conditions in and Senegal terms of labor productivity and similar factor endowments (figure 2.1). In 1960, the productivity stories of Malaysia By 2017, labor productivity in Malaysia and Senegal were roughly similar. Despite had already navigated a different path from their geographical distance—one in East that of Senegal. Output per worker in Malay- Asia, the other in West Africa—the two sia was 6.6 times larger than in Senegal countries were quite close in terms of labor (US$49,630 and US$7,532, respectively, in productivity and its corresponding deep fun- 2011 dollars at current PPP prices). Although damentals: factor endowments and total fac- the capital-output ratios of Malaysia and tor productivity (TFP). Senegal have remained almost invariant over For instance, the 1960 output per worker the past six decades, the amount of physical in Malaysia and Senegal was US$7,261 and capital per worker in the East Asian country US$7,899, respectively (in 2011 dollars at cur- is now more than six times that of the West rent purchasing power parity [PPP] prices).1 African country—specifically, 6.4 times as That is, labor productivity in Malaysia was large in 2017. Human capital continues to be about 92 percent that of Senegal. The cap- higher in Malaysia than in Senegal, although ital per worker and capital-output ratios in the gap has increased, from about 29 percent Malaysia were also close to those in Sene- in 1960 to 92 percent in 2017. In addition, gal. 2 Physical capital per worker in Malay- Malaysian workers became more productive sia and Senegal in 1960 was US$22,874 and than Senegalese workers—by 2017, 3.5 times US$23,175, respectively (in 2011 dollars at as productive (figure 2.1). current PPP prices), while their corresponding In sum, the greater gap in labor produc- capital-output ratios were 3.15 and 2.94. In tivity between Malaysia and Senegal over addition, Malaysia’s human capital index was the past six decades could be attributed not 29 percent higher than Senegal’s in 1960. 3 only to greater differences in factor endow- Finally, workers in the East Asian country ments but also to differences in TFP. In other were about 70 percent as productive as those words, Senegal lost ground to Malaysia not in the West African country. Arguably, these only because of lower investment (in physical 21 22  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 2.1  Outputs, Inputs, and Productivity Gaps between Malaysia and Senegal, 1960 and 2017 y (1960) 0.92 y (2017) 6.59 k (1960) 0.99 k (2017) 6.36 k/y (1960) 1.07 k/y (2017) 0.97 h (1960) 1.29 h (2017) 1.92 A (1960) 0.69 A (2017) 3.5 0 1 2 3 4 5 6 7 Ratio between Malaysia and Senegal 1960 data 2017 data Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: The bars display ratios of Malaysian to Senegalese indicators in their corresponding years. y = labor productivity; k = capital stock per worker; k/y = capital-output ratio; h = PWT human capital index; A = relative total factor productivity (TFP). The year of measurement is shown in parentheses. and human capital) but also because of lower This report proposes several explanations: efficiency in combining the different factors First, physical capital is scarce (as manifested of production. by low capital per worker), and economic activities in Sub-Saharan Africa are less capital-intensive than in other regions where Development Accounting growth took off (say, among the EAP5 coun- The divergent (labor and multifactor) pro- tries). The gap in physical capital per worker ductivity paths of Malaysia and Senegal is persists despite increased public and private only one example of Sub-Saharan Africa’s investment spending in the region. failure to converge economically with the five Second, the region exhibits relatively East Asian “dragons” (referred to here as the poorer levels of human capital and declining “EAP5,” comprising Indonesia, the Republic educational quality as a result of insufficient of Korea, Malaysia, Singapore, and Thailand) investment and poor learning outcomes. and the United States, as further discussed in In fact, 23 of the 25 countries with the chapter 1. In fact, labor productivity in all the lowest Human Capital Index (HCI) are in region’s countries declined sharply relative to Sub-Saharan Africa (World Bank 2019). the EAP5 between 1960 and 2017—with half Finally, the region’s poor economic perfor- of the Sub-Saharan African countries ­ having mance is attributed not only to scarce (and 2017 labor productivity levels that were low-quality) resources but also to inefficien- less than or equal to only a quarter of their cies in the operation of production technolo- corresponding levels in 1960. From these ­ gies. These inefficiencies reflect the prevalence facts, a question emerges: What explains of policies and institutions among the region’s Sub-­Saharan Africa’s dismal performance on countries that impede the more-productive labor productivity compared with the rest of establishments from demanding more factors the low- and middle-income world?4 of production, thus limiting the growth of N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t al Fac t o r P r o d u c t i v i t y    23 their respective firms or farms. For instance, countries with the lowest labor productivity nonmarket mechanisms of land allocation, relative to this benchmark. In about 80 per- differential access to bank credit, tax eva- cent of the region’s countries (37 out of 44), sion, and informality may help explain why the output per worker is less than one-fifth ­ different factors of production in the econ- that of the United States. Within this group, omy are not necessarily r ­ eallocated from the relative GDP per worker of five countries the least- to the most-productive units. In was below 2.5 percent of the US benchmark in other words, scarce resources, compounded 2017 (Burundi, the Central African Republic, by inefficient allocation across the different Liberia, Malawi, and Niger). In contrast, the productive units, translate into low aggregate output per worker in eight countries exceeded labor productivity. 20 percent of the US benchmark—and within Sub-Saharan Africa needs policies to this group, it exceeded 40 percent of the boost productivity across all sectors of eco- benchmark in three countries (Gabon, Mau- nomic activity, especially in those sectors ritius, and the Seychelles).5 where most poor people make their living. The region needs policies that improve pro- Capital-Labor Ratio ductivity in the agriculture sector, foster The substantial gap in output per worker rural development, and create jobs for the between Sub-Saharan Africa and the United youth bulge that is joining the labor force. States is attributable to the region’s scarce availability of inputs of production (phys- ical and human capital) as well as its less Impacts of Low Resource Endowments efficient combination of these inputs. Many and Production Inefficiency countries in the region (31 out of 45) had Relative Labor Productivity 2017 capital-labor ratios below US$50,000 The aggregate growth performance of the at 2011 prices (map 2.2). The stock of capi- region masks the very different growth tal per worker in this undercapitalized group experiences across Sub-Saharan A fri- of countries in 2017 ranged from US$2,500 can countries—where surges, expansions, to US$41,400—with a median capital-labor recessions, and collapses have taken place ratio of US$12,093—and 12 of them had throughout the economic history of the Afri- capital-labor ratios below US$10,000. can subcontinent. From 1980 to 2017, about Only 21 countries outside the region two-thirds of the region’s countries (30 of had capital-labor ratios below US$50,000, 44) experienced a decline in the relative gross but the median ratio for this group (about domestic product (GDP) per worker (figure US$33,319) was significantly higher than 2.2). Relative labor productivity in 2017 was in Sub-Saharan Africa. At the other end of less than half that of 1980 for nine countries the spectrum, only 3 countries in the region in the region: the Central African Repub- had 2017 capital-labor ratios exceeding lic, the Comoros, the Democratic Republic US$200,000 (Equatorial Guinea, Gabon, of Congo, Guinea, Liberia, Niger, Nigeria, and the Seychelles), while that was the case Togo, and Zimbabwe. In contrast, two coun- for 52 countries outside Sub-Saharan Africa. tries (Botswana and Equatorial Guinea) saw their GDP per worker more than double Human Capital Index during the 1980–2017 period. The issue of resource scarcity is not limited to physical capital in Sub-Saharan Africa. Labor Productivity Human capital is also scarce and of low There is great dispersion of labor productiv- quality, as measured by the World Bank’s ity across countries in Sub-Saharan African 2017 Human Capital Index (HCI).6 The HCI countries relative to the United States, which assigns values between 0 and 1 that reflect is the global frontier benchmark (map 2.1). worker productivity relative to a benchmark The region houses the largest number of of complete education and full health (World 24  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 2.2  Relative Labor Productivity of Sub-Saharan African Countries, 1980 versus 2017 6 5 Log real output per worker, 2017 (US = 100) 4 GAB GNQ BWA MUS NAM ZAF SWZ 3 CPV AGO SDN MRT ZMB STP LSO CIV NGA KEN COG 2 SEN MLI CMR GMB BFA TCD TZA COM BEN UGA GNB GIN ZWE ETH RWG SLE COD 1 TGO MDG MOZ NER MWI LBR BDI CAF 0 1 2 3 4 5 Log real output per worker, 1980 (US = 100) Sub-Saharan Africa Non-SSA low- and middle-income countries Industrial countries Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: The figure depicts the output per worker of all countries relative to that of the United States (US = 100). The ratio is then expressed in logs. “Industrial” countries refers to high-income Organisation for Economic Co-operation and Development (OECD) economies. Selected countries and economies are labeled using International Organization for Standardization (ISO) alpha-3 codes. GNB = Guinea Bissau; BFA = Burkina Faso; TCD = Chad; MRT = Mauritania; STP = São Tomé and Principe. Bank 2019). On this measure, 24 out of had HCI scores below 0.4, varying from 40 countries with HCI data in Sub-Saharan 0.369 to 0.398. Among the Sub-Saharan Africa registered low HCI scores (below 0.4) African countries with HCI scores above in 2017, varying within a narrow band from 0.4 (16 out of 40 countries), the median was 0.361 to 0.396 (map 2.3). The median HCI 0.42, varying from 0.476 to 0.678. score for this group of 24 countries (0.37) implies that the future productivity of a child Efficiency of Production born in 2017 is 63 percent below what the The low relative output per worker of sev- child could have achieved with complete edu- eral African countries can be attributed not cation and full health. only to low stocks of capital per worker but Only three countries outside the region also to poor human capital. However, the (Iraq, Pakistan, and the Republic of Yemen) region’s large and persistent gap in output N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t a l F a c t o r P r o d u c t i v i t y    25 MAP 2.1  Labor Productivity, by Country, Relative to the United States, 2017 Relative output per worker (US=1) (y/y*) < 0.2 0.2 – 0.4 0.4 – 0.6 0.6 – 0.8 > 0.8 No information Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: Relative labor productivity of a given country, y/y*, is the ratio of the output per worker in that country to output per worker in the United States (US = 1.0). MAP 2.2  Capital-Labor Ratio, by Country, Relative to the United States, 2017 Relative capital per worker (US=1) (k/k*) < 0.1 0.1 – 0.25 0.25 – 0.5 0.5 – 0.9 > 0.9 No information Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: The relative capital-labor ratio of a given country, k/k *, is the stock of physical capital per worker (US = 1.0). per worker relative to comparator country in the combination of the scarce factors of groups (say, the EAP5 countries or the United production. States) is not only a story of scarce (physical On this score, the TFP of the global effi- and human) capital but also of low efficiency ciency benchmark (the United States) is at 26  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca MAP 2.3  Human Capital Index, by Country, 2017 Human Capital Index 0.2 – 0.4 0.4 – 0.6 0.6 – 0.8 > 0.8 No information Sources: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: The figure plots the values of the World Bank’s 2017 Human Capital Index (HCI) for all countries. The HCI measures three components: probability of survival, expected ­learning- adjusted years of school, and health. The HCI values (from 0 to 1.0) reflect prospective worker productivity relative to a benchmark (1.0) of complete education and full health. least five times that of 31 countries (out of 37) output per worker between Sub-Saharan in Sub-Saharan Africa (map 2.4). ­ Specifically, Africa and the United States were driven US TFP is 5 times that of Botswana, Côte by differences in the relative endowment d’Ivoire, and Kenya; 10 times that of Ghana of (physical and human) capital. Since the and Zambia; and more than 20 times that of 1990s, differences in TFP became the main Nigeria and Tanzania.7 driver explaining the output-per-worker gaps (figure 2.3, panel b).8 Overall, two findings emerge from this Drivers of Labor Productivity Gaps analysis of the widening gap in aggregate between Sub-Saharan Africa and the labor productivity between the United States United States and Sub-Saharan Africa: Development Accounting Analysis • Differences in output per worker were Labor productivity in Sub-Saharan Africa, mainly driven by undercapitalization relative to the global efficiency benchmark in Sub-Saharan Africa from the 1960s (the United States), exhibits long swings to the mid-1980s. The region’s lower (from 5 percent to 15 percent) between 1960 relative accumulation of (physical and and 2017 (figure 2.3, panel a). This rela- human) capital became the main cul- tive productivity declines from an average prit of the labor productivity gap. of 12 percent in the 1970s to a trough of 6 • Gaps in factor accumulation between percent in the 1990s, and then it recovers to Sub-Saharan Africa and the United 8 percent from 2010 to 2017. States still play a role in explaining dif- The development accounting analysis ferences in relative output per worker. shows that, from the 1960s to the mid- However, the gap in the region’s 1980s, more than half of the differences in efficiency in combining its factors of N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t al Fac t o r P r o d u c t i v i t y    27 MAP 2.4  Efficiency of Production, by Country, Relative to the United States, 2017 Relative TFP (US=1) (A/A*) < 0.2 0.2 – 0.4 0.4 – 0.6 0.6 – 0.8 > 0.8 No information Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: The relative total factor productivity (TFP) of each country, A/A*, was computed using data on output per worker, capital-output ratios, human capital, and the share of labor in output relative to the US (= 1.0). production—as captured by the share Drivers of the Labor Productivity Gap due to TFP—has become increasingly The extent and persistence of the labor pro- relevant to explanations of produc- ductivity gap between Sub-Saharan Africa tivity gaps from 2000 to 2014. The and the United States differ markedly across decrease in TFP in Sub-Saharan Africa countries in the region. However, country evi- relative to the United States could be dence supports the aggregate story of changes attributed, among other things, to in the main drivers of these persistent gaps in resource misallocation. output per worker. First, output-per-worker differences There is a shift in the narrative of what between Sub-Saharan African countries and explains the persistent gap in labor produc- the United States from 1980 to 1989 were tivity between Sub-Saharan Africa and the primarily driven by differences in the stocks United States. It has shifted from an under- of physical and human capital.9 Lower capi- capitalization story (reflected by the substan- tal-output ratios and human capital relative tially lower relative capital-output ratios from to the United States explain more than half of the 1960s to the mid-1980s) to a production the labor productivity gap during that period inefficiency story (captured by the region’s in 22 out of 37 Sub-Saharan African coun- lower relative TFP). In turn, Sub-Saharan tries. The median share of labor productivity Africa’s lower TFP levels could be attributed, differences attributed to factors of produc- among other things, to resource misalloca- tion was about 67 percent. tion. In sum, the region’s scarce physical and Second, disparities in labor productivity human capital, compounded by the misallo- between Sub-Saharan African countries and cation of these resources, translates into an the United States were larger in 2010–17 even lower level of (labor and total factor) than in 1980–89. Furthermore, factor accu- productivity. mulation and TFP played increasing roles in 28  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 2.3  Sources of the Labor Productivity Gap between Sub-Saharan Africa and the United States, 1960–2017 a. Real output per worker (US = 100)a 16 14 Share of US real output per worker (%) 12 10 8 6 4 2 0 19 0 19 2 64 19 6 19 8 19 2 19 4 19 6 19 8 19 0 19 2 84 19 6 19 8 90 19 2 19 4 19 6 20 8 20 0 20 2 14 16 70 98 20 0 20 2 20 4 20 6 6 6 7 7 7 7 8 8 6 6 8 8 9 9 9 0 0 0 0 1 1 0 19 19 19 19 19 20 20 b. Share of gap explained by factor accumulation and TFPb 100 90 80 Share of labor productivity gap (%) 70 60 50 40 30 20 10 0 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Disparities in factor accumulation Disparities in TFP Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: TFP = total factor productivity. a. Panel a shows the employment-weighted average of relative real output per worker among countries in Sub-Saharan Africa, as a percentage of that in the United States (US = 100). b. Panel b shows the proportion of the differences in output per worker that are attributed to either factor accumulation or TFP differences. N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t al Fac t o r P r o d u c t i v i t y    29 driving these differences. Out of 37 countries Main Source of Productivity Growth: in the region, the undercapitalization narra- Factor Accumulation tive (that is, factor accumulation explaining Overall Trends more than 50 percent of the labor produc- Growth per worker in Sub-Saharan Africa tivity differences) holds for 3 countries (with has been overwhelmingly driven by physi- a median share due to factor accumulation cal capital accumulation from 1960 to 2017. of 59 percent). On the other hand, the inef- Almost three-quarters of the region’s labor ficiency narrative (that is, TFP differences productivity growth from 1960 to 2017 is explaining more than 50 percent of the explained by growth of physical capital per labor productivity differences) holds for the worker. The contribution of TFP, on the other remaining 34 countries in the region. For hand, is negligible. The narrative on the eco- these 34 countries, about 75 percent of the nomic performance of Sub-Saharan Africa is output-per-worker differences are attributed one of growth at the extensive margin rather to differences in the efficiency with which than at the intensive margin—a typical fea- workers combine the factors of production. ture of low- and lower-middle-income econo- Third, TFP differences have played a mies. Growth per worker and the role played larger role in explaining the gap in relative by factor accumulation (relative to TFP output per worker across Sub-Saharan Afri- growth) in the region is comparable to that of can countries over time. The share attributed Latin America and the Caribbean. to TFP has increased for all 37 countries in Labor productivity growth in the EAP5 the region between 1980–89 and 2010–17 countries was more than triple that of (figure 2.4). The median share due to TFP Sub-Saharan Africa (average annual growth increased from 44 percent in 1980–89 to rates of 3.5 percent and 1.0 percent, respec- 76 percent in 2010–17. This finding implies tively) over the 1960–2017 period. The con- that the narrative of “inefficient use of cur- tribution of TFP growth has also been far rent technologies”—attributed partly to more significant: more than 20 percent of the resource misallocation—is getting more mile- growth per worker in the EAP5 countries was age when explaining output-per-worker dif- driven by greater efficiency in combining the ferences in Sub-Saharan Africa. factors of production. Growth per worker in India (3.3 percent per year) is comparable Dismal Growth Performance: to that of the EAP5, and the contribution of The Negligible Contribution of TFP growth is significantly higher (about TFP Growth 50 percent). Sub-Saharan Africa has failed to catch up Intraregional Trends with both the aspirational development and Resource-rich versus non-resource-rich coun- global efficiency benchmarks—the EAP5 and tries. Within Sub-Saharan Africa from 1960 the United States, respectively—over the past to 2017, non-resource-rich countries out- six decades. From 1960 to 2017, Sub-­ Saharan performed resource-rich countries in terms Africa registered the lowest annual average of growth per worker (with annual average growth per worker of any region in the world rates of 1.2 percent and 0.7 percent per year, (figure 2.5): its average annual rate of growth respectively). The engines that supported the per worker over the 57-year period was 1 per- growth records of these country groups were cent—smaller than that of either ­ industrial also different. Capital accumulation was the (high-income) countries (2.1 ­ percent) or the main engine of growth for both resource-rich EAP5 countries (3.5 percent).10 30  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 2.4  Share of Labor Productivity Differences due to TFP in Sub-Saharan African Countries, 1980–89 versus 2010–17 CAF 100 BWA ZWE LBR NER TZA AGO SEN MWI UGA 80 NGA ETH MDG ZMB GHA BEN TGO MOZ SLE GAB CMR COG KEN BDI GMB COD RWA Share due to TFP (%), 2010–17 CIV LSO 60 MLI BFA MUS ZAF MRT SWZ NAM 40 SDN 20 0 20 40 60 80 100 Share due to TFP (%), 1980–89 Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: TFP = total factor productivity. Selected countries and economies are labeled using International Organization for Standardization (ISO) alpha-3 codes. and non-resource-rich countries; but in basis during 1960–77, bolstered by favorable resource-rich countries, declines in TFP oil prices, was followed by a 1978–95 con- dragged down the growth in overall labor traction characterized by adverse external output (figure 2.5). This suggests that, in shocks and macroeconomic instability. From this group of countries, either (a) part of the 1996 onward, growth per worker recovered capital expenditure may not have translated and continued to expand amid a favorable into a greater amount of physical capital, or external environment (commodity price (b) the combination of factors of production boom and ample capital inflows), improved may have been largely inefficient. Growth per macroeconomic frameworks, and adequate worker in the non-resource-rich countries, on (policy and liquidity) buffers. These buffers, the other hand, was primarily explained by built during the years of expansion, allowed factor accumulation, but TFP growth had a some African countries to formulate policies positive and economically important contri- to withstand the unprecedented 2008–09 bution (about 20 percent). external shock of the Global Financial Crisis. Swings in productivity growth over time. Changes in sources of growth. These Growth per worker in Sub-Saharan Africa swings in labor productivit y g row th exhibited long swings from 1960 to 2017: were accompanied by changes in the rel- The expansion of real GDP on a per worker ative importance of the different sources N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t al Fac t o r P r o d u c t i v i t y    31 FIGURE 2.5  Traditional Solow Decomposition of Labor Productivity Growth, Selected Regions and Country Groups, 1960–2017 4.0 3.5 Average annual growth per worker (%) 3.0 2.5 2.0 1.5 1.0 0.5 0 –0.5 –1.0 s SS e P5 a an A) t t le an an rie di cl. m gi SS be EA A In ex co nd nd nt fra a( rib es -in ou bu bu A ric Ca tri le lc SS -a -a Af d ria ce ce e un id n th ur ur st co m ra du d so so ha nd an re re In Sa -a ica n- A b- w SS no er Su Lo Am A SS tin La Physical capital Human capital TFP Output Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: Group averages are employment-weighted averages. “Industrial” countries refers to high-income Organisation for Economic Co-operation and ­Development (OECD) countries. “Fragile” refers to fragile and conflict-affected states, as defined by the World Bank. “EAP5” refers to five East Asian ­economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand. SSA = Sub-Saharan Africa; TFP = total factor productivity. For the detailed definitions and methodology, see appendix A. of growth. Despite the swings in economic Within Sub-Saharan Africa, the rebound performance over time, the fortunes of of labor productivity growth between 1996 growth in Sub-Saharan Africa are still and 2017 was experienced by resource-abun- tightly connected to factor accumulation dant, non-resource-abundant, and fragile (figure 2.6). For instance, the region’s accel- countries alike. This recovery came along eration of growth per worker in 1996–2017 with an acceleration of TFP growth. For percent per year, up from 0.7 percent (2.3 ­ instance, average annual growth per worker per year in 1978–95) is mostly accounted for of non-resource-rich countries jumped from by the accumulation of physical and human –0.5 percent in 1978–95 to 2.4 percent in capital (about 65 percent). The relative con- 1996–2017 (with TFP growth increasing tribution of TFP growth (about 35 percent from -0.7 percent to 0.8 percent, respec- in 1996–2017) is comparable to that of other tively). In fact, TFP contributed positively to low- and middle-income countries (about growth per worker in all regional country 30 percent). However, these findings should groups: its relative contribution amounted be taken with caution because the contribu- to 33 percent of growth per worker in non-­ tion of TFP growth might be overstated by resource-abundant countries, 40 percent in the omission of factors such as the accumula- resource-abundant countries, and 43 percent tion of natural capital (box 2.1). in fragile countries.11 32  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 2.6  Traditional Solow Decomposition of Labor Productivity Growth in Sub-Saharan Africa, by Country Group and Period, 1961–2017 4 3 Average annual growth per worker (%) 2 1 0 –1 –2 –3 –4 7 5 4 7 5 4 7 5 4 7 5 4 97 99 01 97 99 01 97 99 01 97 99 01 –1 –1 –2 –1 –1 –2 –1 –1 –2 –1 –1 –2 61 78 96 61 78 96 61 78 96 61 78 96 19 19 19 19 19 19 19 19 19 19 19 19 Sub-Saharan Africa (SSA) SSA resource-abundant SSA non-resource-abundant SSA fragile Physical capital Human capital TFP Output Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: Group averages are employment-weighted averages. “Fragile” refers to fragile and conflict-affected states, as defined by the World Bank. TFP = total factor productivity. For the detailed definitions and methodology, see appendix A. Role of Public Capital in Sub-Saharan Public capital has been an important driver Africa’s Economic Growth of economic growth in Sub-Saharan Africa from 1960 to 2017: about half of the region’s Public investment can be an important cat- growth per worker is attributed to the accu- alyst of economic growth by delivering mulation of public capital (figure 2.7). The important public services as well as connect- slump of TFP growth in this period might be ing citizens, farms, and firms to economic partly associated with inefficiencies in public opportunities. After the Global Financial spending.12 This pattern of growth and cap- Crisis, public investment played (and still ital accumulation is even more pronounced plays) a role in supporting long-term growth among the region’s resource-abundant coun- by deploying (own and borrowed) resources tries, where physical capital accumulation to finance infrastructure projects—especially (and especially the dynamics of public invest- among Sub-Saharan African countries (IMF ment) explains growth per worker over the 2015). From 1960 to 2017, the stock of pub- past six decades. However, the extent of inef- lic capital grew faster than the stock of pri- ficient public investment spending is trans- vate capital only in Sub-Saharan Africa (by lated into greater misallocation of resources 1.9 percent and 1.4 percent per year, respec- and a negative contribution of TFP growth.13 tively), although public and private capital Higher public investment may not auto- both grew more slowly than in industrial matically translate into commensurate countries and in other low- and middle-in- increases in the capital stock or in higher come countries. growth benefits, because of a low-quality N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t al Fac t o r P r o d u c t i v i t y    33 BOX 2.1  The Contribution of Natural Capital to Growth per Worker The contribution of total factor productivity (TFP) to tion of natural capital in sectors such as energy growth per worker across Sub-Saharan Africa, espe- (as in Chad, the Republic of Congo, Gabon, and cially among the resource-abundant countries, tends Nigeria) and extractives (as in Botswana, the to decline when the production technology accounts Democratic Republic of Congo, and Zambia). for the use of natural capital as an additional factor of Accounting for the accumulation of natural capi- production. Natural capital—the stock of all extract- tal reduces the contribution of TFP to growth per able resources such as geology, soils, air, water, and worker by almost 1 percentage point per year. living organisms—accounted for more than half of This decline is even larger (more than 150 basis the region’s growth per worker from 1996 to 2017. points per year) for Sub-Saharan African coun- The increased share of growth due to TFP in tries that are abundant in minerals and metals the region might be attributed to the contribu- (figure B2.1.1). FIGURE B2.1.1  Decomposition of Labor Productivity Growth, Including Natural Capital, in Sub-Saharan Africa, 1996–2017 3.0 2.5 Average annual growth per worker (%) 2.0 1.5 1.0 0.5 0 –0.5 –1.0 –1.5 Conventional Natural Conventional Natural Conventional Natural Conventional Natural resource resource resource resource Sub-Saharan Africa (SSA) SSA resource-rich SSA oil-rich SSA metals-rich Capital (composite) Human capital TFP Output Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: Natural capital refers to the stock of all extractable resources such as geology, soils, air, water, and living organisms. The “natural resource” decomposition treats natural capital as a factor of production; the “conventional” decomposition does not. Group averages are population-weighted averages. The methodology to compute the output elasticity of physical capital and natural capital as well as the specification of technology and the computation of total factor productivity (TFP) growth is described in appendix B, “Country Productivity Analysis in Sub-Saharan Africa.” SSA = Sub-Saharan Africa. pipeline of investment projects or inefficien- countries (Keefer and Knack 2007). Closing cies and waste in the selection and imple- efficiency gaps in public investments could mentation of these projects. This disconnect significantly increase the public investment is particularly acute when governance is multiplier. For instance, closing the gap weak—as it is in Sub-Saharan African between the top and bottom quartiles of 34  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 2.7  Decomposition of Labor Productivity Growth, including Role of Public Capital, in Selected Regions and Country Groups, 1961–2014 4.0 3.5 Average annual growth per worker (%) 3.0 2.5 2.0 1.5 1.0 0.5 0 –0.5 –1.0 –1.5 s SS e P5 a an A) t t le an an rie di cl. m gi SS be EA A In ex nco nd nd nt fra a( rib ou bu bu A ric es e-i Ca SS lc -a -a Af tri dl ria ce ce e un id n th ur ur st ra co m d du so so ha nd an re re In Sa -a ica n- A b- w SS no er Su Lo Am A SS tin La Public capital Private capital Human capital TFP Output Sources: Penn World Table (PWT) 9.0 data from Feenstra, Inklaar, and Timmer 2015; PWT 9.1 updates by World Bank. Note: Group averages are population-weighted averages. “Industrial” countries refers to high-income countries such as Organisation for Economic Co-operation and Development (OECD) member countries. “Fragile” refers to fragile and conflict-affected states, as defined by the World Bank. “EAP5” refers to five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand. SSA = Sub-Saharan Africa; TFP = total factor productivity. For the detailed definitions and methodology, see appendix A. public investment efficiency could double the Sectoral Employment impact of such investment on growth (IMF In 1990, the region’s share of agricultural 2015). employment was about 40 percent—higher than in either high-income economies or the Lagging Structural low- and middle-income countries of other Transformation regions (figure 2.8). By 2016, this share had declined to only 31 percent, which is still One of the main features of sectoral structure substantially greater than in high-income and long-term growth in Sub-Saharan Africa economies (2 percent) and other low- and is the region’s substantial lag in structural middle-income countries (18 percent) (Bar- transformation for two reasons: A large share rot, Calderón, and Servén 2018b). of people still work and make a living from Although the average share of agricultural agriculture across countries in the region. employment in the region still exceeded 30 And the region’s employment share in agri- percent in 2016, countries varied greatly in culture has been declining more slowly than the proportion of people engaged in agricul- has historically been the case in other world tural activities. It remains above 60 percent regions (figure 2.8). N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t al Fac t o r P r o d u c t i v i t y    35 FIGURE 2.8  Sectoral Employment Shares, Sub-Saharan Africa versus Low- and Middle-Income Countries in Other Regions, 1990–2016 a. Sub-Saharan Africa 50 40 Share of employment (%) 30 20 10 0 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Agriculture Manufacturing Nonmanufacturing Market services Nonmarket services b. Low- and middle-income countries outside Sub-Saharan Africa 50 40 Share of employment (%) 30 20 10 0 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Agriculture Manufacturing Nonmanufacturing Market services Nonmarket services Source: Barrot, Calderón, and Servén 2018b. Note: Regional sectoral labor shares are GDP-weighted averages of country sectoral labor shares. For detailed definitions and methodology, see appendix A. in 13 (out of 28) countries: Burundi, Camer- employment tend to exhibit low levels of oon, the Central African Republic, Eswatini, ­ agricultural productivity (Duarte and Restuc- Madagascar, Malawi, Mali, Mauritania, cia 2010, 2018; Herrendorf, Rogerson, and Mozambique, Niger, Nigeria, Rwanda, and ­Valentinyi 2014). Uganda. This finding reflects the fact that The region’s share of manufacturing countries with higher shares of agricultural employment remains low, declining from 10.3 36  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca percent in 1990 to 8.4 percent in 2016. Within The aggregate employment share in market ­ anufacturing the region, 10 countries have a m services increased from 23 percent in 1990 to employment share below 5 ­ percent: Angola, 33 percent in 2016 (figure 2.8, panel a). This Botswana, Burundi, Gabon, Mali, Mozam- trend holds for countries across the region bique, Rwanda, Sierra Leone, Uganda, and regardless of income level and the extent of Zambia. On the other hand, Sub-Saharan resource abundance, albeit at different speeds. Africa is experiencing a rapid shift of work- Only three countries have an employment ers from agriculture to market services. share in market services below 10 percent FIGURE 2.9  Sectoral Labor Productivity Relative to Agriculture: Sub-Saharan Africa and Low- and Middle-Income Countries in Other Regions, 1990–2016 a. Sub-Saharan Africa 16.8 14.8 Ratio of value added per worker 12.8 (agriculture = 1) 10.8 8.8 6.8 4.8 2.8 0.8 19 0 91 19 2 20 3 93 19 4 19 5 19 6 97 19 8 99 20 0 20 1 20 2 04 20 5 06 20 7 08 20 9 10 20 1 12 20 3 14 20 5 16 9 9 9 0 9 9 9 0 0 0 0 0 0 1 1 1 19 19 19 19 20 20 20 20 20 20 20 Manufacturing Nonmanufacturing Market services Nonmarket services b. Low- and middle-income countries outside Sub-Saharan Africa 6.8 Ratio of value added per worker 5.8 4.8 (agriculture = 1) 3.8 2.8 1.8 0.8 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Manufacturing Nonmanufacturing Market services Nonmarket services Source: Barrot, Calderón, and Servén 2018b. Note: Regional sectoral labor productivity figures are GDP-weighted averages of country sectoral labor productivity relative to agriculture (=1.0). For detailed definitions and methodology, see appendix A. N e e d e d : B o o s t i n g t h e C o n t r i b u t i o n o f T o t a l F a c t o r P r o d u c t i v i t y    37 (Burundi, the Central African Republic, and over time may reflect either frictions in labor Malawi), whereas the share exceeds 40 per- allocation or issues in measurement and spec- cent in another three countries (the Gambia, ification (Duarte and Restuccia 2018). Mauritius, and South Africa). Notes Sectoral Labor Productivity   1. Productivity data and ratios for Malaysia and Sectoral labor productivity exhibits large Senegal are the author’s calculations using swings over time in most Sub-Saharan Afri- Penn World Table (PWT) data. can countries. However, it has improved in   2. The capital-output ratio is the amount of capital needed to produce each extra unit most of the region’s countries since the mid- of output. As such, it is an indicator of how 1990s (Duarte and Restuccia 2018). Labor efficiently new investment contributes to eco- productivity experienced sharp upswings in nomic growth. agriculture (averaging 4.5 percent per year)   3. The human capital index is calculated from and manufacturing (averaging 3 percent per the average years of schooling and an assumed year) from 1990 to 2016. Productivity growth rate of return to education, on the basis of in market and nonmarket services was less Mincer equation estimates, around the world. dynamic (with annual average growth rates For a detailed explanation on how the data of 1.6 percent and 1.1 percent, respectively). are compiled and used to construct the index, In spite of its faster growth, labor pro- see “Human Capital in PWT 9.0” (https:// ductivity is lower in agriculture than in the www​. rug.nl/ggdc/docs/human_capital_in​ _pwt_90.pdf). region’s nonagricultural activities—namely,   4. Throughout the “Development Accounting” manufacturing, nonmanufacturing, and section, the sample of Sub-Saharan African market and nonmarket services. By 2016, countries varies by the type of productivity, the ratio of value added per worker relative ratio, or index being measured because of to that of agriculture was 2.9 in market ser- the countries’ varying data availability. For vices, 5.7 in manufacturing, and 10.4 in non- example, a TFP calculation requires complete manufacturing activities (figure 2.9). information on output, inputs, and shares of The region’s sectoral productivity gaps labor in output, which several countries did relative to agriculture have remained slightly not have, resulting in a relatively low total invariant or have declined at a sluggish pace sample (37 countries) for that measurement. over the past quarter century. However,  5. Relative to the aspirational development benchmark (represented by the EAP5 coun- among the non-resource-rich countries, these tries), the output per worker in 35 (out of 45) gaps have been declining steadily. In contrast, countries in Sub-Saharan Africa is less than among resource-rich countries, they have half the EAP5 average—and less than one-fifth declined at a slower pace in all sectors but the EAP5 average among 24 of those countries. manufacturing.   6. The HCI has three components: probability Overall, sectoral labor productivity of survival, expected learning-adjusted years growth in Sub-Saharan Africa is consistent of school, and health. It reflects the human with the process of structural change and capital of the next generation given the risks aggregate performance. However, there is of inadequate education and health in the substantial heterogeneity across countries country where they live (World Bank 2019). and over time. A standard structural trans-   7. Outside Sub-Saharan Africa, only 11 coun- tries had such low relative TFP—although US formation model shows that low growth TFP is, on average, no more than eight times in agricultural productivity translates into that of this group. weak structural change—although faster  8. Appendix B of this report, “Country Pro- productivity growth since 1995 has almost ductivity Analysis of Sub-Saharan Africa,” doubled the pace of reallocation out of agri- presents a visual analysis of the develop- culture. The presence of medium-term cycles ment accounting exercises for all countries in trended productivity across countries and in the region, with data available on output, 38   Boosting Productivity in Sub-Saharan Africa employment, physical capital, human capital, Africa: A TFP Boost Is Needed.” Unpublished and the labor share of output. manuscript, World Bank, Washington, DC.   9. Here we depict the development accounting Barrot, Luis-Diego, César Calderón, and Luis exercise for Sub-Saharan African countries Servén. 2018b. “Sectoral Productivity Shifts in 1980–89 rather than 1960–69 because the in Sub-Saharan Africa.” Unpublished manu- 1980–89 period (a) increased the regional script, World Bank, Washington, DC. coverage from 21 countries to 37 countries, Duarte, Margarida, and Diego Restuccia. 2010. and (b) includes some of the largest coun- “The Role of the Structural Transformation in tries in the region (for example, Angola and Aggregate Productivity.” Quarterly Journal of Sudan). Economics 125 (1): 129–73. 10. This section is based largely on Barrot, Duarte, Margarida, and Diego Restuccia. 2018. Calderón, and Servén (2018a). “Structural Transformation and Productivity 11. “Fragile” refers to fragile and conflict- in Sub-Saharan Africa.” Unpublished manu- a f fec ted st ates (FC S), def i ned on t he script, University of Toronto. basis of financial and security status by Feenstra, Robert C., Robert Inklaar, and Marcel the World Bank’s Fragile, Conflict and P. Timmer. 2015. “The Next Generation of V iolence group. For more information, see ­ the Penn World Table.”  American Economic the Bank’s online topical overview: https:// Review 105 (10): 3150–82. w w w.worldba n k.org /en /topic /f rag i l it y​ Herrendorf, B erthold, R ichard Rogerson, conflictviolence/overview. and Ákos Valentinyi. 2014. “Growth and 12. The calibration of the elasticity of output to Structural Transformation.” In Handbook of public and private capital as well as the meth- Economic Growth Vol. 2, edited by Philippe odology to compute TFP growth are discussed Ag hion and Steven Du rlauf, 855 –941. in appendix A, “Output per Worker, Factor Amsterdam: Elsevier. Accumulation, and Total Productivity.” IMF (International Monetary Fund). 2015. 13. Note that the relative contribution of public “Making Public Investment More Efficient.” capital accumulation and TFP to growth per Staff report, IMF, Washington, DC. worker is similar among industrial countries Keefer, Philip, and Stephen K nack. 2007. and the EAP5 countries (about 25 percent and “Boondoggles, Rent-Seeking, and Political 24 percent, respectively). Checks and Balances: Public Investment under Unaccountable Governments.” Review of Economics and Statistics 89 (3): 566–72. References World Bank. 2019. World Development Report Barrot, Luis-Diego, César Calderón, and Luis 2019: The C hanging Nature of Work . Servén. 2018a. “Growth in Sub-Saharan Washington, DC: World Bank. Resource Misallocation in Sub-Saharan Africa: Firm-Level 3 Evidence Introduction and productivity. This argument lies at the heart of the literature on resource misallo- Why are African countries considerably less cation and takes center stage in the report’s productive than high-income economies? analysis of low productivity in Sub-Saharan What accounts for these differences in (labor Africa. (Box 3.1 summarizes the theoretical and total factor) productivity? And why is underpinnings of the relationship between firm performance in Sub-Saharan Africa the number of operating firms and their lower and more volatile than that of their distribution, misallocation, and aggregate counterparts in either the “aspirational devel- productivity.) opment” or the “global efficiency” bench- Empirical research on resource misalloca- mark countries?1 This report argues that tion as a source of low aggregate productivity distortions in decision-making processes at in Sub-Saharan Africa is growing, but it is the firm level have implications for a coun- still incipient. The lack of available firm-level try’s aggregate output and productivity and census data (across countries and over time) might also help explain aggregate productiv- in the region is still a binding constraint. The ity differences across countries. It suggests existing evidence for the region focuses on that low-income countries are not as effective both direct and indirect approaches to quan- in allocating inputs of production to their tifying the extent of resource misallocation most efficient use. across Sub-Saharan African countries and its The effectiveness of resource alloca- influence on aggregate total factor produc- tion is indicated in relation to an efficiency tivity (TFP). While some papers measure the benchmark in a model economy where firms total net effects of distortions on aggregate maximize final output. It is characterized productivity (the indirect approach), others by (a) decisions about setting the number of assess specific sources of distortions. establishments to be operating in the indus- This report finds evidence of severe try, and (b) the allocation of capital and misallocation in agriculture and manufac- labor across those operating establishments. turing across Sub-Saharan African countries. Distortions in each of these two stages of Low agricultural productivity is primarily decision making will reduce aggregate output explained by inefficiencies in the allocation 39 40  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca BOX 3.1  Resource Misallocation: Theoretical Underpinnings Resource misallocation refers to distortions in the production units, will generate lower aggregate allocation of inputs (capital, land, and labor, among output. Given the constant aggregate amount of others) across production units of different sizes. This inputs (say, capital, land, and labor), the output loss misallocation typically occurs when the different associated with an inefficient allocation is also an production establishments are taxed at different aggregate TFP loss. In this context, misallocation rates. This strand of the literature assumes that refers to situations were resources are not allocated aggregate output is produced by several producers efficiently across production units, and the cost (N ) with different levels of productivity (A i). Firm i ’s of misallocation is typically measured in terms of technology is summarized by a production function aggregate output or TFP losses. ( f ) that is strictly increasing and strictly concave. Theoretically, inefficiencies in the allocation There is a fixed cost of operation (c) for any pro- of labor and capital across heterogeneous produc- ducer. Given an aggregate demand of labor (H ) and ers will affect aggregate output and productivity capital (K), there is a unique allocation of labor and through three different channels: capital across producers that maximizes total output • The technology channel—the higher the productiv- net of fixed operating costs. ity for all firms, the greater the output According to this framework, lower values of • The selection channel—based on the choice of A i reflect either slow adoption or inefficient use of operating producers technology. The efficient allocation in this economy • The misallocation channel—based on the allocation maximizes final output and is characterized by two of capital and labor among operating producers. decisions: (a) the number of operating establishments (that is, establishments that can pay the fixed cost, These three channels are not independent: any c); and (b) the allocation of capital (K ) and labor policy or institution that distorts the allocation of (H ) across the operating establishments. If either of resources across producers will potentially generate these decisions is distorted, the economy will have additional effects through both the selection and lower output and hence lower aggregate total factor technology channels. productivity (TFP)—because the aggregate factor If the misallocation of resources across these inputs (K and H ) in the industry are constant. different producers helps explain cross-country dif- The allocation of inputs that maximizes output ferences in aggregate productivity levels, it is then across production units (say, either firms or farms) crucial to investigate the sources of misallocation. takes place when, conditional upon their operation, Resource misallocation across different production the marginal (and average) products are equal across units might reflect the following (Restuccia and all production units. In this equilibrium, no out- Rogerson 2017): put gains would be acquired by reallocating inputs of production (say, capital, land, and labor) from • Statutory provisions, including some features of the production units with low marginal products to tax code and regulations—for instance, provisions those with high marginal products. In the efficient of the tax code that vary with firm characteristics allocation, the most productive operating establish- (say, age or size); tariffs targeting certain groups of ments will demand a greater amount of inputs. In goods; employment protection measures; and land other words, a production unit’s productivity and regulations, among others size are positively associated in the efficient allo- • Discretionary government (or bank) provisions cation. In addition, production units with similar that favor or penalize specific firms—for instance, levels of productivity command the same amount of subsidies, tax breaks, or low-interest loans granted inputs and are of identical size. to specific firms; preferential market access; and Deviations from the efficient allocation of unfair bidding practices for government contracts, resources across firms may have implications on among others aggregate output and productivity. Input choices • Market imperfections—for instance, monopoly that are different from the efficiency model, even if power; market frictions (such as in credit and land they allocate more factors to the more-productive markets); and enforcement of property rights. (Box continues next page) R e s o u r c e M i s a l l o c a t i o n i n S u b - S a h a r a n A f r i c a : F i rm - L e v e l E v i d e n c e    41 BOX 3.1  Resource Misallocation: Theoretical Underpinnings (continued) The resource misallocation literature examines approach faces an important challenge: efficient (a) the extent of factor misallocation in the econ- allocations may not require that marginal products omy, (b) its impact on TFP differences across coun- be equalized across production units at every point tries and over time, and (c) the underlying factors in time —especially if input choices precede the real- driving misallocation. Two main approaches have ization of the individual productivity shock or are in been followed to tackle these questions—the direct the presence of adjustment costs. approach and the indirect approach (Restuccia and In sum, resource misallocation is closely related Rogerson 2013, 2017).a to a specific model economy and to a benchmark The direct approach seeks to ascertain the under- allocation relative to that economy. In this model lying sources of misallocation and evaluates their economy, inputs are homogeneous, and the only consequences through a structural model. Assessing source of heterogeneity among productive units is the degree of misallocation requires computation of the productivity of their operating establishments. a counterfactual. However, the direct approach also The output-maximizing allocation of factors in the requires quantitative measures of the source of mis- model economy is the commonly used benchmark. allocation. If the main drivers of misallocation come It is not optimal to allocate the entire endowment from discretionary rather than statutory provisions, of inputs to any individual production unit—even if there will be severe measurement problems. Further- it is the most productive one—because the increase more, the complexity of measuring drivers such as in output for a given increase in inputs declines as regulation (especially its differences across indus- the size of the establishment increases. Resource tries) may make it complicated to build, calibrate, misallocation can arise both across production units and simulate a structural model. with different levels of productivity and across units The indirect approach quantifies the extent of with similar productivity. An important interpreta- resource misallocation and does not dig deeply tion of misallocation is that production units effec- into the underlying factors that generate the distor- tively face different prices or wedges to their inputs tions driving the inefficient allocation of resources. or output. That is, production units face idiosyn- It measures the total net effect of these distortions cratic distortions (Restuccia and Rogerson 2008). without fully identifying their main sources. The These wedges or distortions support the observed efficient allocation of resources equalizes their mar- allocation, which differs from the efficient alloca- ginal products across all operating production units. tion, as an equilibrium outcome. A direct examination of the dispersion of marginal products provides a measure of the degree of mis- Sources: Restuccia 2011; Restuccia and Rogerson 2008, 2013, 2017. allocation. This approach does not require a full a. The direct and indirect approaches have been commonly applied to census data on manufacturing firms. See Restuccia and Rogerson (2008), Hsieh and structural model; however, it needs the specification Klenow (2009), and the empirical literature that arose from those seminal of the technology of production. Still, the indirect papers. of resources across farmers rather than of severe misallocation across manufac- agronomic endowments. The most produc- turing firms in select Sub-Saharan African tive farms cannot command more factors countries. 2 This dispersion is greater than of production, and their growth is impeded. in other low- to middle-income countries Resource misallocation also has dynamic (China and India) as well as that of the global implications for agricultural productivity by efficiency benchmark (the United  States). disincentivizing the adoption of new tech- Furthermore, the positive correlation nologies and reducing the farmers’ ability to between TFPR and quantity (or physical) learn new techniques. productivity (TFPQ)3 across Sub-Saharan In manufacturing, the large dispersion of African manufacturing firms indicates that revenue productivity (TFPR) is an indication the region’s more-productive firms tend to 42  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca face higher distortions—especially output farmers may produce crops that may not be distortions. These higher distortions decel- suitable to the geographical features of the erate the growth of firms over their life land they operate (Adamopoulos and Restuc- cycles and discourage the adoption of new cia 2014; Gollin and Rogerson 2014). technologies. However, low agricultural productivity in low-income countries—and, notably, in Sub-Saharan Africa—is primarily attribut- Resource Misallocation in able to inefficiencies in the use of resources Agriculture rather than poor agronomic conditions This report has so far corroborated some of (such as low-quality land and unfavorable the stylized facts found in the literature: weather). Worldwide evidence shows that approximately 80 percent of agricultural • There are large and persistent differences productivity differences between poor and in real output per worker across countries rich countries can be attributed to produc- (Hsieh and Klenow 2010; Jones 2016; tion inefficiencies. In other words, agricul- Restuccia 2011). tural productivity in low-income countries • Poorer countries tend to allocate most is not low because they have lower potential of their labor to agriculture (Duarte and yields. It is low because the actual yields lie Restuccia 2010, 2018; Herrendorf, Roger- far from their potential ones (Adamopoulos son, and Valentinyi 2014). and Restuccia 2018). • The productivity of agriculture (relative to nonagriculture sectors) in poorer countries Counterfactual Exercise, with Crop tends to be lower than in richer countries Selection Constant (Adamopoulos and Restuccia 2014; Gollin, What would be the gains in agricultural Lagakos, and Waugh 2014; Restuccia, Yang, output in Sub-Saharan African countries and Zhu 2008). if actual yields were raised to their poten- These three stylized facts stress the import- tial ones? A spatial productivity growth ant role played by agriculture in understand- accounting in agriculture was conducted ing the large disparities in real output per for five large countries in the region: the worker across countries.4 Democratic Republic of Congo, Ethiopia, Kenya, Nigeria, and Tanzania. 6 The bene- fits of closing the actual-potential yield gap Differences in Agricultural Productivity: is conducted under three scenarios of input About Efficiency, not Geography use and water supply but holding constant Are the differences in agricultural pro- the farmers’ crop choices (table 3.1). The dif- ductivity between Sub-Saharan Africa ferent scenarios considered are (a) low input and the aspirational and global efficiency use under rainfed cultivation, (b) high input benchmarks explained by land quality and use under rainfed cultivation, and (c) high geography? Or are these differences in pro- input use under irrigated cultivation (Sinha ductivity (say, differences in yields) attribut- and Xi 2018). able to inefficient use of agricultural inputs? Under the least productive scenario (low Agricultural output and productivity input use under rainfed cultivation), actual can depend on the region’s geographical yields are higher than potential ones for the features—exogenous factors such as rainfall, Democratic Republic of Congo and Nige- temperature, and soil quality. 5 In many ria. On aggregate, this implies that both Sub-Saharan African countries, rural farm- countries have moved beyond the least ers who operate at subsistence levels and productive scenario. In contrast, Ethiopia, lack the appropriate infrastructure make Kenya, and Tanzania still can reap produc- up a larger share of the population than in tivity gains from closing the actual-poten- other world regions. Under these conditions, tial gap, even using the least sophisticated R e s o u r c e M i s all o ca t i o n i n S u b - Sa h a r a n A f r i ca : F i r m - L e v e l E v i d e n c e    43 TABLE 3.1  Gap between Actual and Potential Agricultural Yields, Selected Sub-Saharan African Countries, 2000 Change in yield (%) Low input use a High input usea High input usea Country Rainfed Rainfed Irrigated Congo, Dem. Rep. −36 88 102 Ethiopia 32 367 450 Kenya 40 314 380 Nigeria −16 174 230 Tanzania 47 347 442 Source: Sinha and Xi 2018. Note: The Global Agronomic Ecological Zones (GAEZ) database provided data on crop-specific yields, crop choices, and land endowments as well as poten- tial crop yields (at the grid level) under different scenarios of water supply and input use (at the farm level). The scenarios hold the farmers’ crop choices constant. The GAEZ data are complemented with the Food and Agriculture Organization’s (FAO) harmonized crop calendars and country-level crop prices. a. Input use is classified into (a) low (labor-intensive and subsistence agricultural practices); (b) intermediate (market participation, use of better seed vari- eties, hand tools, livestock, and preliminary methods of mechanization); and (c) high (modern practices, production for market purposes only, completely mechanized, no shortfalls in use of fertilizers and chemicals). method of cultivation. For instance, agri- productivity improvements in Sub-Saharan cultural yields in Tanzania can increase by African countries (Sinha and Xi 2018). about half under this scenario. Aggregate yield gains become larger as Counterfactual Exercise, with Crop agricultural production scenarios become Selection Optimized more sophisticated. However, there is a great What role does crop selection play in deal of heterogeneity in productivity gains explaining the changes in aggregate agricul- across countries. Under the intermediate tural yields? Table 3.2, column [1], reports scenario (high input use and rainfed cultiva- the yield gains when farmers’ optimal crop tion), agricultural yields nearly double for the choice is cultivated under actual levels of Democratic Republic of Congo—yet these input use and water supply (in short, under gains are much smaller than in the other coun- actual yields).7 Optimal crop selection raises tries, especially Ethiopia, Kenya, and Tanza- farmers’ productivity, although these gains nia. The contribution of irrigation to farmers’ vary widely across countries. The smallest productivity, on the other hand, is limited productivity gains are attained in the Demo- once they use inputs at their highest level. cratic Republic of Congo and Tanzania, while If farmers were to raise their input use from yields nearly double among Nigerian farmers. low to high (holding constant the nature of The largest gains are achieved by farmers in the water supply), their potential productivity Ethiopia and Kenya (about 5.3 and 6.5 times gains would increase between 7 and 11 times their actual output, respectively). for Ethiopia, Kenya, and Tanzania. If, on Moving from the actual-yield benchmark, the other hand, the cultivation method shifts the role of crop selection is evaluated under from rainfed to irrigation (while maintaining three scenarios: (a) low input use under high input use), the potential gains are sig- rainfed cultivation (table 3.2, column [2]); nificantly lower. For instance, the marginal (b) high input use under rainfed cultivation yield gains of using irrigation fluctuate from (column [3]); and (c) high input use under irri- a paltry 14 percentage points (the Democratic gation (column [4]). Under the least produc- Republic of Congo) to 95 percentage points tive scenario, optimal crop choice expands (Tanzania). Overall, changes in input use agricultural output by 25 percent in the appear to play a greater role than the nature Democratic Republic of Congo and 44 per- of water supply when explaining agricultural cent in Nigeria. Agricultural output grows 44  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca TABLE 3.2  Optimal Crop Choice and Aggregate Yield Gains, Selected Sub-Saharan African Countries, 2000 Percent change in yield Actual input use Low input usea High input usea High input usea Actual water supply Rainfed Rainfed Irrigated Country [1] [2] [3] [4] Congo, Dem. Rep. 116 25 217 238 Ethiopia 532 125 628 948 Kenya 645 275 943 1,110 Nigeria 197 44 421 450 Tanzania 108 217 838 1,330 Source: Sinha and Xi 2018. Note: “Optimal” crop choice refers to the selection of the possible crops cultivated on a farm that maximizes output. The Global Agronomic Ecological Zones (GAEZ) database provided data on crop-specific yields, crop choices, and land endowments as well as on potential crop yields (at grid level) under different scenarios of water supply and use of intermediate inputs (at the farm level). The GAEZ data are complemented with the Food and Agriculture Organization’s (FAO) harmonized crop calendars and country-level crop prices. a. Input use is classified into (a) low (labor-intensive and subsistence agricultural practices); (b) intermediate (market participation, use of better seed vari- eties, hand tools, livestock, and preliminary methods of mechanization); and (c) high (modern practices, production for market purposes only, completely mechanized, no shortfalls in use of fertilizers and chemicals). 1.25 times in Ethiopia and more than dou- Aggregate Consequences of Inefficient bles in Kenya and Tanzania. Resource Allocation across Farms Yet the productivity gains grow exponen- Measuring the misallocation of resources tially under high input use—even if we keep across farms requires the definition of a con- the nature of the water supply invariant. The ceptual efficiency benchmark. Two-sector meager gains registered in the Democratic general equilibrium models with heteroge- Republic of Congo and Nigeria under the neous production units argue that the effi- least productive scenario expand by eight to cient allocation of factors is achieved when nine times if input use is enhanced (column the marginal product of land and labor are [3]), while they are considerably higher for equal across farmers (Adamopoulos and the other countries—especially Kenya, which Restuccia 2014; Aragón and Rud 2018). The potentially would achieve a nearly tenfold optimal decision rules of farmers suggest increase in actual output. the following testable implications on factor Under the most productive scenario allocative efficiency: First, the more produc- (column [4]), the marginal returns from tive farmers should be able to demand more irrigation (while keeping input use con- intermediate inputs (say, labor, capital, and stant) remain modest compared with the land). Second, agricultural yields should be returns from increasing input use (while uncorrelated with farmers’ productivity.8 keeping constant the nature of the water Farm-level evidence for Ethiopia, Malawi, supply). The marginal gains from irrigation and Uganda shows the following (Aragón (measured as a percentage change in yields) and Rud 2018; Chen, Restuccia, and San- are particularly small for the Democratic taeulàlia-Llopis 2017; Restuccia and Santae- Republic of Congo and Nigeria (20 and ulàlia-Llopis 2017): 30 percentage points, respectively). In Tan- zania, however, the marginal returns from • The more productive farmers tend to have the use of irrigation techniques are larger: greater use of intermediate inputs; however, yields are more than 12 times as large as the relationship between input use and pro- those obtained with actual production ductivity is flatter than the one suggested by (table 3.2). the allocative efficiency criteria. R e s o u r c e M i s all o ca t i o n i n S u b - Sa h a r a n A f r i ca : F i r m - L e v e l E v i d e n c e    45 • The more productive farmers tend to misallocation across farmers. Figure 3.1 illus- exhibit greater agricultural yields. In other trates the misallocation of resources across words, yields and farmers’ productivity fail Ugandan farmers (Aragón and Rud 2018). to be uncorrelated. The lack of correlation between the actual allocation of land across farmers and the In sum, the relationships between input corresponding level of farmers’ productivity use and farmers’ productivity and between (figure 3.1, panel a) is consistent with land yields and farmers’ productivity do not con- allocation mechanisms that are governed form with the predicted implications of the by inheritance norms and redistribution, efficient allocation of resources. This implies whereas market mechanisms (or rent and that there is substantial evidence of factor sale) are severely more restricted. Farmers’ FIGURE 3.1  Farmers’ Productivity, by Input Use and Yields, in Uganda a. Land b. Labor 20 20 10 10 ln(labor) ln(labor) 0 0 −10 −10 −20 −20 −10 −5 0 5 10 −10 −5 0 5 10 Farmer productivity (s_i) Farmer productivity (s_i) c. Yields 20 15 ln(output per acre) 10 0 –5 −10 −10 −5 0 5 10 Farmer productivity (s_i) Data Estimated E cient Source: Aragón and Rud 2018. Note: The data on farmers’ productivity, yields, and input use are taken from the Uganda Panel National Survey (UNPS), a household-level panel dataset collected as part of the World Bank’s Living Standards Measurement Studies–Integrated Surveys on Agriculture (LSMS-ISA) project. This survey, with representative information at the urban/rural and regional levels, has four available rounds (2009–10, 2010–11, 2011–12, and 2013–14). It collects agricultural information for cropping seasons taking place in either the first or second semester each year. Given that the period of analysis is the cropping season, the time dimension of the panel consists of eight periods at best. ln = natural logarithm. s_i = time-invariant total factor productivity of farmer i.   46  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca insecure property rights and land market (2010) shows that there is pervasive misallo- restrictions limit their ability to raise capital cation of resources across Sub-Saharan Afri- for agricultural production (Besley and Gha- can manufacturing firms (Cirera, Fattal-Jaef, tak 2010; de Soto 2000). Eliminating resource and Maemir 2018). misallocation in agriculture may yield sig- A look at the distribution of quantity nificant aggregate output and productivity and revenue productivity (TFPQ and TFPR, gains in Sub-Saharan Africa—for example, respectively) shows that there is also substan- an increase of about 200 percent in Ethiopia tial variation in firm-level productivity in all (Chen, Restuccia, and Santaeulàlia-Llopis four of these Sub-Saharan African countries. 2017) and 260 percent in Malawi (Restuccia The productivity dispersion across manufac- and Santaeulàlia-Llopis 2017). turing firms in Sub-Saharan Africa is larger than in more productive benchmarks—say, China, India, and the United States (Hsieh Resource Misallocation in and Klenow 2009). Manufacturing The magnitude of this productivity disper- The inefficient allocation of inputs across sion is particularly striking in Kenya, where manufacturing firms plays an important role less-productive firms coexist with a few very in understanding underdevelopment: resource productive ones. Kenyan firms in the top misallocation can explain up to 60 percent decile of TFPQ are 290 percent more produc- of aggregate TFP differences between poor tive than firms in the bottom decile. The gap and rich countries (Bartelsman, Haltiwan- between the most and the least productive ger, and Scarpetta 2013; Hsieh and Klenow firms is about 87 percent in Ghana, 2009; Restuccia and Rogerson 2008). Firm- 39 percent in Ethiopia, and 26 percent in level evidence from select Sub-Saharan Côte d’Ivoire (table 3.3) (Cirera, Fattal-Jaef, African countries shows substantial misal- and Maemir 2018). location of capital—as reflected in a greater The dispersion of TFPR across manu- dispersion in marginal products of capital facturing firms in the selected Sub-Saharan (as well as domestic interest rates).9 In this African countries is significantly higher than context, smaller firms tend to display the that of manufacturing firms in China, India, largest degree of misallocation, which might and the United States. For instance, the gap be tied to their higher cost of capital relative between the most and the least productive to medium and large firms.10 More broadly, firms (as measured by the ratio of top to bot- there is severe misallocation of resources tom TFPR deciles) is equal to 51 in Kenya, across manufacturing firms as resources are 17 in Ghana, 13 in Ethiopia, and 7 in Côte shifted from the more-productive firms to the d’Ivoire. These gaps are substantially larger less-productive ones. This implies the coex- than those in China (4.9), India (5.0), and istence of few productive firms with many the United States (3.3). A plausible expla- low-productivity ones. nation for the excessive dispersion of TFPR The efficient allocation of resources (say, across manufacturing firms is that policies capital and labor) is achieved when the mar- and institutions in Sub-Saharan Africa may ginal products of the factors of production prevent the most-productive firms from are equal across manufacturing firms. In expanding and replacing the least-productive the presence of multiple intermediate inputs, ones. The potential productivity gains from efficiency is attained when TFPR is equal better allocation of resources across manu- across firms. Hence, dispersion in TFPR sig- facturing establishments would be substan- nals resource misallocation, which in turn tial. An equalization of TFPR across firms can be attributed to distortions in output and in each industry would raise manufacturing capital. Evidence from firm-level manufactur- productivity by 31.4 percent in Côte d’Ivo- ing census data of Côte d’Ivoire (2003–12), ire, 66.6 percent in Ethiopia, 75.5 percent in Ethiopia (2011), Ghana (2003), and Kenya Ghana, and 162.6 percent in Kenya.11 R e s o u r c e M i s all o ca t i o n i n S u b - Sa h a r a n A f r i ca : F i r m - L e v e l E v i d e n c e    47 TABLE 3.3  Dispersion of Revenue and Quantity Productivity across Manufacturing Firms, Selected Sub-Saharan African Countries Cote d’Ivoirea Kenyab Ghanac Ethiopiad TFPR TFPQ TFPR TFPQ TFPR TFPQ TFPR TFPQ Metric 2003–12 2003–12 2010 2010 2003 2003 2011 2011 Standard deviation 0.65 1.24 1.52 2.41 0.95 1.75 0.78 1.30 Ratio of percentiles 75-25 0.88 1.74 1.99 3.34 1.43 2.61 1.26 1.94 90-10 1.99 3.25 3.94 5.67 2.89 4.47 2.56 3.67 Cov(TFPQ, TFPR) 0.70 0.85 0.69 0.74 Reg. Coeff. 0.42 0.52 0.44 0.53 NObs 4,146 4,146 757 757 1151 1151 4,012 4,012 Source: Cirera, Fattal-Jaef, and Maemir 2018. Note: Output and input data were obtained from firm-level manufacturing censuses of Côte d’Ivoire (2003–12), Ethiopia (2011), Ghana (2003), and Kenya (2010). The censuses (specified in notes a.–d. below) are nationally representative, adequately including both small and large firms in the formal sector. Revenue and quantity productivity (TFPR and TFPQ, respectively) are expressed in logs and are demeaned by industry-specific averages. Industries are weighted by their value-added shares. NObs = number of observations; Cov = covariance; Reg. Coeff. = Regression coefficient; TFPQ = Quantity total factor productivity; TFPR = Revenue total factor productivity. a. The Côte d’Ivoire data are from the Registrar of Companies for the Modern Enterprise sector, collected by the National Statistics Institute. The Côte d’Ivoire statistics are calculated by taking the average for the years 2003–12. b. The Kenyan data come from the 2010 Census of Industrial Production (CIP), conducted by the Kenyan National Bureau of Statistics (KNBS). c. The Ghanaian data come from the 2003 National Industrial Census (NIC) dataset, conducted by the Ghana Statistical Service (GSS). It is similar in structure to the Ethiopian survey, covers the universe of establishments employing more than 10 workers, and takes a representative sample of firms employing fewer than 10 workers. d. The datasets used for Ethiopia are the Large and Medium Scale Manufacturing Industries Survey (LMSMI) and the Small-Scale Manufacturing Industries Survey (SSMI), both of which are conducted by the Ethiopian Central Statistical Agency (CSA). Within-industry dispersion of TFPR is 0.42 for Côte d’Ivoire, 0.53 for Ethiopia, across manufacturing firms in Sub-Saharan 0.44 for Ghana, and 0.52 for Kenya (Cirera, Africa is also quite substantial (table 3.3). Fattal-Jaef, and Maemir 2018). How do these The distortions associated with the observed elasticities compare with the global efficiency dispersion in TFPR would be costlier if they benchmark? The computed elasticity of were positively associated with the firms’ TFPR with respect to TFPQ for the US man- TFPQ—as noted by Restuccia and Rogerson ufacturing sector is 0.09 (Hsieh and Klenow (2008) and Bartelsman, Haltiwanger, and 2014). Hence, TFPR rises more steeply Scarpetta (2013). In other words, distortions among Sub-Saharan African manufacturing would have a more deleterious impact on firms than among their counterparts in the aggregate productivity if they were to “tax” United States. the most-productive firms relative to the The larger elasticity of the Sub-­ S aharan least-productive ones. African manufacturing sector suggests Firm-level evidence shows that there is a that the more-productive firms cannot strong positive relationship between TFPQ use more resources and use them more and TFPR for select countries in the region efficiently— thus worsening aggregate (figure 3.2). This finding confirms that the ­ p roductivity (Restuccia and Rogerson region’s most productive manufacturing firms 2008). That the region’s more-productive face the largest distortions to resource alloca- firms face higher distortions may also tion. The presence of these “correlated distor- decelerate the growth of firms over their tions” is consistent with evidence found for life cycles by discouraging them from manufacturing firms in other low- and mid- investment in productivit y- enhancing dle-income countries. For instance, the esti- technologies (Bento and Restuccia 2017; mated elasticity of log(TFPR) on log(TFPQ) Hsieh and Klenow 2014). 48   Boosting Productivity in Sub-Saharan Africa FIGURE 3.2  Quantity versus Revenue Productivity across Selected Sub-Saharan African Countries a. Côte d’Ivoire b. Ethiopia 4 4 2 2 Log TFPR Log TFPR 0 0 −2 −2 −4 −4 −8 −6 −4 −2 0 2 −6 −4 −2 0 2 4 Log TFPQ Log TFPQ c. Ghana d. Kenya 4 4 2 2 Log TFPR Log TFPR 0 0 −2 −2 −4 −4 −8 −6 −4 −2 0 2 −10 −5 0 5 Log TFPQ Log TFPQ Source: Cirera, Fattal-Jaef, and Maemir 2018. Note: Revenue total factor productivity (TFPR) and quantity total factor productivity (TFPQ) denote the revenue and physical productivity measures. They are computed for each establishment in industry using firm-level census data from Côte d’Ivoire, Ethiopia, Ghana, and Kenya. TFPR and TFPQ are expressed in logs and scaled by the industry-specific average. Notes agricultural labor productivity in the richest countries was approximately 78 times that   1. As further discussed in chapter 1, the “aspi- of the poorest ones. Additionally, 86 per- rational development” benchmark is repre- cent of workers in the poorest nations were sented by the five “East Asian dragons” (or employed in agriculture—as opposed to 4 “EAP5,” comprising Indonesia, the Republic percent in the richest nations (Restuccia, of Korea, Malaysia, Singapore, and Thai- Yang, and Zhu 2008). land). The “global efficiency” benchmark is   5. Certain regions might be more suitable for proxied by the United States. cultivation of particular crops based on geog-   2. Total factor productivity revenue (TFPR) is raphy but may yield dismal output if used to typically defined as the ratio of firms’ sales cultivate other crops that require significantly (or revenues) to input costs (appropriately different geographical attributes. weighted by their production elasticities).   6. These five countries jointly account for just   3. Total factor productivity quantity (TFPQ), under half of the region’s population, and also called physical productivity, is defined as agriculture is an important activity in terms the ratio of a firm’s physical output to physi- of both employment and value added. cal inputs, appropriately weighted according   7. This simulation restricts the optimal crop to their production elasticities. selection to those choices that are actually   4. The productivity gap between the world’s observed at the farm level. Consequently, the richest and poorest nations is even larger narrow set of crop choices is smaller than the in the agriculture sector. For instance, entire set of crop choices including changing R e s o u r c e M i s a l l o c a t i o n i n S u b - S a h a r a n A f r i c a : F i rm - L e v e l E v i d e n c e    49 yields constructed using the GAEZ crop Bento, Pedro, and Diego Restuccia. 2017. c ­ alendar information (Sinha and Xi 2018). “Misallocation, Establishment Size, and  8. An additional implication suggests that Productivity.” American Economic Journal: adverse local productivity shocks should Macroeconomics 9 (3): 267–303. decrease the use of intermediate inputs B esley, Timothy, and Maitreesh Ghatak. (Aragón and Rud 2018). 2010. “Proper t y R ights and E conomic   9. In the presence of well-functioning domestic Development.” In Handbook of Development capital markets, efficient allocation is char- Economics, Vol. 5, edited by Dani Rodrik acterized by each firm’s marginal product and Mark Rosenzweig, 4525–95. Oxford: of capital (MPK) being equal to the mar- North-Holland. ket interest rate. If firms instead borrow at Bigsten, Arne, Paul Collier, Stefan Dercon, different interest rates, capital is likely to be Marcel Fafchamps, Bernard Gauthier, Jan misallocated and the MPK will differ across Willem Gunning, Abena Oduro, et al. 2004. firms. Differential access to informal finance “Do African Manufacturing Firms Learn from or political connections are among the factors Exporting?” Journal of Development Studies that may explain the variance in interest rates 40 (3): 115–41. for firms (Kalemli-Ozcan and Sørensen 2016). Chen, Chaoran, Diego Restuccia, and Raül 10. Smaller firms tend to have more binding con- Santaeulàlia-Llopis. 2017. “The Effects straints than larger firms. Smaller firms are of Land Markets on Resource Allocation less likely to access credit at more favorable and Agricultural Productivity.” Working contract terms than larger firms can, given Paper 24034, National Bureau of Economic their profits and collateral (Bigsten et al. Research, Cambridge, MA. 2004; Paganini 2016). Cirera, Xavier, Roberto Fattal-Jaef, and Hibret 11. These productivity gains from reversing misallo- Maemir. 2018. “Taxing the Good? Distor- cation are still small relative to the development tions, Misallocation, and Productivity in gaps in the region; however, they are reason- Sub-Saharan Africa.” World Bank Economic able lower bounds to the overall cost associated Review 34 (1): 75–100. with the extent of misallocation in a country. de Soto, Hernando. 2000. The Mystery of These calculations do not consider propagation Capital: Why Capitalism Triumphs in the through intersectoral linkages, and they only West and Fails Everywhere Else. New York: account for static gains from reallocation. Basic Books. Duarte, Margarida, and Diego Restuccia. 2010. “The Role of the Structural Transformation in References Aggregate Productivity.” Quarterly Journal of Adamopoulos, Tasso, and Diego Restuccia. 2014. Economics 125 (1): 129–73. “The Size Distribution of Farms and Interna- Duarte, Margarida, and Diego Restuccia. 2018. tional Productivity Differences.” American “Structural Transformation and Productivity Economic Review 104 (6): 1667–97. in Sub-Saharan Africa.” Background paper, Adamopoulos, Tasso, and Diego Restuccia. 2018. World Bank, Washington, DC. “Geography and Agricultural Productivity: Gollin, Douglas, David Lagakos, and Michael Cross-Country Evidence from Micro Plot-Level E. Waugh. 2014. “The Agricultural Produc- Data.” Working Paper 24532, National Bureau tivity Gap.” Quarterly Journal of Economics of Economic Research, Cambridge, MA. 129 (2): 939–93. Aragón, Fernando M., and Juan Pablo Rud. Gollin, Douglas, and Richard Rogerson. 2014. 2018. “Weather, Productivity, and ­ Factor “Productivity, Transport Costs and Subsis- ­ M isallocation: Evidence from Ugandan tence Agriculture.” Journal of Development Farmers.” B ­ ­ ackground paper, World Bank, Economics 107: 38–48. Washington, DC. Herrendorf, Berthold, Richard Rogerson, and Bartelsman, Eric, John Haltiwanger, and ­ Stefano Ákos Valentinyi. 2014. “Growth and Struc- Scarpetta. 2013. “Cross-Country ­ Differences tural Transformation.” In Handbook of Eco- in Productivity: The Role of Allocation and nomic Growth, Vol. 2 , edited by Philippe Selection.” American Economic Review Aghion and Steven Durlauf, 855–941. Amster- 103 (1): 305–34. dam: Elsevier. 50  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Hsieh, Chang-Tai, and Peter J. Klenow. 2009. Restuccia, Diego. 2011. “Recent Developments “Misallocation and Manufacturing TFP in Economic Growth.” Federal Reserve Bank in China and India.” Quarterly Journal of of Richmond Economic Quarterly 97 (3): Economics 124 (4): 1403–48. 329–57. Hsieh, Chang-Tai, and Peter J. Klenow. 2010. Restuccia, Diego, and Richard Rogerson. 2008. “Development Accou nting.” A m e ric an “Policy Distortions and Aggregate Productivity Economic Journal: Macroeconomics 2 (1): with Heterogeneous Establishments.” Review 207–23. of Economic Dynamics 11 (4): 707–20. Hsieh, Chang-Tai, and Peter J. Klenow. 2014. Restuccia, Diego, and Richard Rogerson. 2013. “The Life Cycle of Plants in India and “Misallocation and Productivity.” Review of Mexico.” Quarterly Journal of Economics Economic Dynamics 16 (1): 1–10. 129 (3): 1035–84. Restuccia, Diego, and Richard Rogerson. 2017. Jones, Charles I. 2016. “The Facts of Economic “The Causes and Costs of Misallocation.” Growth.” In Handbook of Macroeconomics, Journal of Economic Perspectives 31 (3): Vol. 2A, edited by John B. Taylor, and Harald 151–74. Uhlig, 3–69. Amsterdam: Elsevier. Restuccia, Diego, and Raül Santaeulàlia-Llopis. Kalemli-Ozcan, Sebnem, and Bent E. Sørensen. 2017. “Land Misallocation and Productivity.” 2016. “Misallocation, Property Rights, and Working Paper 23128, National Bureau of Access to Finance: Evidence from within and Economic Research, Cambridge, MA. across Africa.” In African Successes, Volume Restuccia, Diego, Dennis Tao Yang, and Xia- III: Modernization and Development, edited odong Zhu. 2008. “Agriculture and Aggregate by Sebastian Edwards, Simon Johnson, and Productivity: A Quantitative Cross-Country David N. Weil, 183–211. Chicago: University Analysis.” Journal of Monetary Economics of Chicago Press for the National Bureau of 55 (2): 234–50. Economic Research. Sinha, Rishabh, and Xican Xi. 2018. “Agronomic Paganini, Monica. 2016. “An Efficiency Analysis Endowment, Crop Choice and Agricultural of Firms: Evidence from Sub-Saharan Africa.” Productivity.” Background paper, World Bank, Doctoral dissertation, University of Kent, Washington, DC. Canterbury, UK. Policies and Institutions that Distort Resource Allocation in 4 ­Sub-Saharan Africa Introduction This chapter (a) examines various poli- cies and institutions that affect the produc- Institutions and economic policies may tivity of farms and firms; (b) evaluates their making introduce distortions in the decision-­ (static) impact on resource misallocation; processes of production units (farms and and (c) assesses, to the extent possible, their firms) in Sub-Saharan Africa. In turn, these dynamic effects through distorted occupa- distortions in resource allocation across the tional choices or inefficient technological different production units may affect not decisions. Specifically, this chapter discusses a only the quantities they produce but also the comprehensive, but by no means exhaustive, economy’s aggregate level of output and pro- set of potentially distortionary policies and ductivity. The aggregate productivity losses institutions (summarized in table 4.1) that are associated with these distortionary policies classified by three potential sources of misallo- and institutions, therefore, are transmitted cation (Restuccia and Rogerson 2017): through three distinct and interdependent channels (Restuccia and Rogerson 2017): • Market imperfections. The analysis dis- cusses (a) credit market imperfections • The technology channel, which affects the (that is, restricted access to finance due productivity of various production units to the lack of collateral); (b) lack of land • The selection channel, which affects the titling, affecting the allocation of land; and number of operating production units1 (c) information frictions, affecting produc- • The misallocation channel, which drives ers that are not connected to markets or the allocation of capital and labor among farmers who have inadequate information operating production units away from an on weather forecasts. efficiency benchmark. • Statutory provisions. Also discussed are These three channels are not independent: size-dependent policies—more specifically, any policy or institution that misallocates tax provisions and regulations that depend resources can potentially generate additional on features of the different production units effects through both the selection and tech- (say, size and age) as well as trade policies nology channels. that protect specific categories of goods. 51 52  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca TABLE 4.1  Policy-Related Sources of Potential Resource Misallocation Affecting Farm and Firm Productivity Source Policies Market imperfections • Financial frictions (collateral) • Nonmarket land allocation (land titling, rentals) • Information asymmetries (price dispersion, EWS) Statutory provisions • Size-dependent policies (tax provisions) • Age-and size-dependent regulations • Targeted trade policies Discretionary provisions • Input subsidy programs for farmers • Preferential lending to specific firms • Preferential market access to certain producers Source: Restuccia and Rogerson 2017. Note: EWS = early warning systems. • Discretionary provisions. In addition, the other sectors (including manufacturing) but report captures government provisions also employs a larger share of the region’s that favor or penalize certain types of pro- population. This chapter explores the poten- duction units—for example, subsidies to tial institutional and policy-related sources farmers, low-interest loans to specific firms, of misallocation in agriculture and manu- and preferential market access for specific facturing, including land market imperfec- groups of producers, among others. tions, agricultural subsidies, size-dependent taxation and informality, preferential trade In focusing on the potential sources of policies, differential access to infrastruc- misallocation across agriculture and manu- ture, and financial market imperfections. facturing production units in Sub-Saharan These sources have likely led to allocative Africa, the chapter uses farm- and firm-level inefficiencies, primarily through suboptimal information to quantitatively assess the impact selection of operating production units, dis- of policies and institutions on aggregate pro- torted occupational choices, and disincentives ductivity. Most of the empirical evidence pre- to investment in technological upgrading. sented in this chapter uses the direct approach to resource misallocation (as further discussed in chapter 3, box 3.1). First, it directly mea- sures the specific policies, institutional fac- Land Market Imperfections tors, and market imperfections that are likely The underdevelopment of land market insti- sources of misallocation. Second, it calibrates tutions is one of the potential sources of and simulates a model of heterogeneous pro- resource misallocation in agriculture across duction units to evaluate the extent to which Sub-Saharan African countries. The analysis these factors can generate effects on aggre- of available household data, integrated with gate total factor productivity (TFP) through farm-level agricultural production data across misallocation.2 the region, has yielded strong evidence of Chapter 3 presented farm- and firm-level capital and land misallocation in the agricul- evidence of pervasive resource misallocation ture sectors (see chapter 3 and the references in agriculture and manufacturing among therein). In turn, it has been argued that insti- Sub-Saharan African countries—and these tutions governing land allocation mechanisms allocative inefficiencies are even greater than are connected to the severe misallocation of in other low- and middle-income countries. agricultural resources across farms in Sub-Sa- Agriculture still plays an important role in haran Africa and impede farm-size growth the region’s economic performance. This pri- among the most-productive farmers (Restuc- mary activity is not only less productive than cia and Santaeulàlia-Llopis 2017). P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   53 Distortions in farm size can hamper agri- lead not only to resource misallocation but cultural productivity and discourage the also to (a) distorted incentives for technolog- uptake of modern technologies. Farm size ical adoption (Aragón and Rud 2018; Chen, in low-income countries can be distorted Restuccia, and Santaeulàlia-Llopis 2017); by a wide variety of institutions and farm- and (b) distorted occupational choices by level policies. For instance, many countries individuals between farming and nonfarm- (including Bangladesh, Ethiopia, and the activities—because individuals opting to ing ­ Philippines) have imposed ceilings on land work in nonagriculture sectors may have to holdings, partitioning any farms that exceed forfeit their untitled land (Chen 2017). those ceilings. Others (Indonesia and Zim- babwe) have established both maximum and Aggregate Consequences of Inefficient minimum size constraints. Several coun- Resource Allocation across Farms tries have also levied progressive land taxes (Namibia and Zimbabwe) or steep progres- There is a strong relationship between the sive income taxes (Ethiopia) on farmers depth of land markets and the degree of out- (Restuccia 2016). put loss attributed to misallocation. In other In addition, institutional mechanisms for words, resource misallocation across coun- allocating land are tightly linked to inheri- tries in the region is greater among farmers tance norms and redistribution. They tend to without rental markets than those with devel- restrict access to land in underdeveloped rental oped rental markets—as shown in Ethiopia, and sale markets. Insecure property rights or Malawi, and Uganda (box 4.1). This dispar- inefficient land allocation mechanisms may ity is captured by the greater dispersion in BOX 4.1  Land Institutions in Selected Sub-Saharan African Countries Malawi: Dominance of Customary Land Tenure Ethiopia: State Ownership of Land Most of the land tenure in Malawi is customary, From 1974 until the early 1990s, the Ethiopian gov- with user rights assigned locally by village chiefs. a ernment expropriated and uniformly redistributed The country’s Customary Land Act (No. 19 of the country’s rural land and legally prohibited land 2016) grants the village head (or superior chiefs transactions. Although land ownership still resides administering several villages) the power to allow or with the state and many of the restrictions to land ban land transactions and to resolve disputes across transactions remain in place, some reforms were villages associated with land limits (Kishindo 2011; implemented in the 2000s to grant land certificates Morris 2016). Although the Malawi Land Bill (also to farmers and to allow rentals of the use rights (up passed in 2016) looks to reduce these powers, it has to a certain limit). Because land sales are prohibited not yet been enacted. in Ethiopia, land rentals are the only channel for Most household farms in Malawi (83.4 percent) reallocating farms’ operational scale, and hence they do not operate any marketed land (rented-in or pur- constitute a measure of the depth of land markets. chased). Of the 16.6 percent of household farms However, the extent of land rentals began to dif- that do operate part of their land from the market, fer substantially across subregions as these reforms 3 percent rent-in land informally (borrowed for were decentralized to local governments (Deininger, free or moved into without permission); 9.5 percent Ayalew Ali, and Alemu 2008). For instance, the per- rent-in land formally (through leaseholds, short-term centage of rented land varies from 0 percent to more rentals, or farming as a tenant); 1.8 percent pur- than 73 percent among the 69 zones (with avail- chase land without a title; and 1.3 percent purchase able data) across the country. Among 234 woredas land with a title (Restuccia and ­Santaeulàlia-Llopis districts), the percentage of rented land varies from (­ 2017). 0 percent to 91 percent. These large differences in (Box continues next page) 54   Boosting Productivity in Sub-Saharan Africa BOX 4.1  Land Institutions in Selected Sub-Saharan African Countries (continued) land rentals across zones and districts reflect sub- Buganda (Central region) in 1900. Although Mailo stantial heterogeneity among local land market landowners hold their land in perpetuity and have institutions. similar rights to freeholders, tenants have secu- Despite the comprehensive land certification rity of occupancy as in common-law arrangements reform intended to provide tenure security to farm- (sometimes backed by a certificate) and can only be ers, land markets remain highly underdeveloped. removed if the land is unattended for at least three Among Ethiopian household farms, 67.6 percent years (Coldham 2000). neither rent-in nor rent-out any land; 24.3 percent In regions where noncustomary tenure systems are formally or informally rent-in some land for pro- more prevalent, 47 percent of land holdings have been duction; 10.6 percent rent-out land; and 2.5 percent marketed (purchased or rented). In regions where either rent-in or rent-out some land. For a more customary land tenure is more common, 27 percent extensive institutional background on the allocation of land holdings have been marketed. These tenure and use of land in Ethiopia, see Chen, Restuccia, systems are spatially concentrated as follows: (a) more and Santaeulàlia-Llopis (2017). than 90 percent of land holdings are under customary land tenure in the Northern and Eastern regions, and Uganda: Multiple Land Systems (b) noncustomary systems are mostly found in the Uganda provides four types of land tenure: freehold, Western and Central regions. leasehold, Mailo (a form of freehold), and custom- Finally, differences in land tenure appear to mat- ary land. The first three systems offer some degree ter for economic activity in Uganda: customary land of formal and secure property rights, while custom- is associated with lower agricultural investment ary systems are less secure and lack formal land reg- (Place and Otsuka 2002). istries (Coldham 2000; Place and Otsuka 2002). Sources: Aragón and Rud 2018; Chen, Restuccia, and Santaeulàlia-Llopis 2017; The Mailo territory (8,000 square miles) was Restuccia and Santaeulàlia-Llopis 2017. allocated to chiefs and notables after an agreement a. “Customary” land tenure (as opposed to “statutory” tenure) refers to owner- ship by indigenous communities, administered in accordance with their cus- between the British government and the Kingdom of toms. Common ownership is one form of customary land ownership. total factor productivity revenue (TFPR) and among farms operating on nonmarketed in the marginal product of land for farm- lands. Evidence at the household farm level ers who cannot rent land (table 4.2). 3 For in Ethiopia shows that the efficiency gains instance, evidence from Ethiopian farmers from reallocation for farmers who do not confirms the following (Chen, Restuccia, and participate in rentals are larger than those of Santaeulàlia-Llopis 2017): farmers who rent land (table 4.2). The same finding is obtained for farms in Malawi: the • The dispersion of TFPR is greater for farm- output level of farms without marketed land ers without rental markets (1.10) than for (about 84 percent of the sample) would be farmers with rental markets (0.96). 4.2-fold, compared with the 3.6-fold out- • The standard deviation of the marginal put gains for the entire sample of farmers in product of land is also significantly lower Malawi (Restuccia and Santaeulàlia-Llopis for farmers who rent-in or rent-out land 2017).4 (0.86) than for those who cannot do so The degree of association between farm (1.05). size and farm TFP is higher among farms In sum, firms with any portion of mar- operating with marketed land than among ket land tend to display less resource those operating without marketed land: misallocation. these correlations amount to 0.30 and The output gains from eliminating dis- 0.14, respectively, among Malawi farmers tortions in land allocation would be larger (Restuccia and Santaeulàlia-Llopis 2017). P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   55 TABLE 4.2  Impact of Land Rental on Resource Misallocation among Farmers in Ethiopia, 2013/14 Full Metric sample No rentals Rentals Efficiency gain (nationwide) 3.07 3.18 2.61 Std Dev (log TFPR) 1.06 1.10 0.96 Std Dev (log MP land) 0.99 1.05 0.86 Observations (no.) 2,887 1,951 936 Sample (%) 100.0 67.6 32.4 Source: Chen, Restuccia, and Santaeulàlia-Llopis 2017, using data from the Ethiopia Integrated Survey of Agriculture (ISA) 2013/14. Note: A baseline nationwide reallocation is conducted to compute efficiency gains separately for each group of farmers: those with no rental land and those with any percentage of rented-in or rented-out land. MP land = the marginal product of land; Std Dev = standard deviation; TFPR = revenue productivity. However, the weakness of these correlations levels of aggregate productivity (Restuccia suggests that land markets are still limited, and Santaeulàlia-Llopis 2017). even for farmers with access to marketed The actual distribution of income (and land; that is, these farmers are still far from productivity) is widely dispersed despite operating at their efficient scale. the relative equalization of inputs (say, cap- Despite the alleviating role played by land ital and land) across farmers in Malawi. rentals, the degree of misallocation is still For example, the ratio of the top to bottom severe, even among farms that rent land. Land quintile of agricultural output (a proxy of rentals help reassign land from less-productive farm income) is a factor of 34-fold (4.78 for to more-productive farms. However, these the top quintile and 0.14 for the bottom quin- farms are still operating far from the sec- tile) although the corresponding ratios of torally efficient allocation. This finding sug- top to bottom quintiles for capital and land gests that land markets in Sub-Saharan Africa use are within a factor of 1-fold to 2-fold. remain subject to various frictions. Weak legal In other words, equal access to land across institutions may also hinder rental activity. households does not necessarily translate into income equalization because these farmers differ substantially in their productivity. Distributional Implications of Assessing the distributional income effects Resource Reallocation of the reallocation of inputs across farm- ers requires the computation of a counter- Resource reallocation among farmers to factual income level that (a) considers the achieve efficient operational scales may actual distribution of factors as endowments, have distributional implications. Empirical and (b) allows the efficient allocation to be ­ evidence shows the actual versus the efficient achieved through perfectly competitive rental distribution of factors, output, and income markets. Such a simulation yields the follow- across farmers by productivity quintile in ing findings for household farms in Malawi: Malawi (table 4.3). The actual land distribution across farm • First, the least productive farmers reap TFP is flat: most farms are operating on the largest benefits from the higher factor less than 2 acres of land. The estimated effi- returns, and overall inequality declines. The cient land distribution, on the other hand, income ratio between farmers in the top suggests that the most productive farm (top and bottom quintiles would decline from quintile) should operate on almost 6 acres a factor of 34-fold (actual allocation) to on average (representing 97 percent of total about 3.4-fold (efficient allocation)—that land). These findings point to a substan- is, a decline of income inequality among tial redistribution of land to achieve higher these farmers by a factor of 10-fold. 56  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca TABLE 4.3  Effects of Actual and Efficient Distribution of Land, Capital, MPL, and MPK among Farms in Malawi Productivity partition Variable 1st 2nd 3rd 4th 5th var(ln x) Farm productivity (s) 0.75 2.10 3.72 6.39 21.50 1.435 Land (l): Actual 1.19 0.87 1.01 1.03 1.99 0.749 Efficient 0.00 0.01 0.04 0.14 5.91 6.782 Capital (k) Actual 55.93 25.35 21.81 24.70 26.71 1.820 Efficient 0.04 0.32 1.10 3.60 149.52 6.782 MP Land (yield) Actual 4.21 11.00 17.82 29.10 82.04 1.485 Efficient 76.30 76.30 76.30 76.30 76.30 0.000 MP Capital Actual 0.73 2.19 3.94 7.25 24.54 2.154 Efficient 6.03 6.03 6.03 6.03 6.03 0.000 Output (y) - level Actual 0.14 0.39 0.69 1.20 4.78 1.824 Efficient 0.00 0.05 0.18 0.60 25.06 6.782 % of total Actual 2.01 5.46 9.57 16.67 66.26 .. Efficient 0.02 0.20 0.71 2.33 96.71 .. Agricultural income - level: Actual 0.14 0.39 0.69 1.20 4.78 1.824 Efficient 4.28 2.22 2.17 2.56 14.65 1.228 Income gain 23.70 3.88 2.27 1.58 1.99 .. % of total: Actual 2.01 5.46 9.57 16.67 66.26 .. Efficient 16.55 8.58 8.41 9.88 56.56 .. Source: Restuccia and Santaeulàlia-Llopis 2017. Note: Household quintiles are shown in order of farm productivity (1st quintile the lowest, 5th quintile the highest). Land, capital, and output are in per hours terms. “MP land” (MPL) is the marginal product of land and “MP capital” (MPK) the marginal product of capital computed for each bin. “Actual” income is equal to actual agricultural output, whereas “efficient” income is computed assuming that (a) actual allocations are the endowments, and (b) the efficient allocation is achieved via perfectly competitive rental markets. “Income gain” is the ratio of efficient to actual income. .. = not calculated. • Second, the ratio of efficient to actual This counterfactual suggests that having income increases for all household farms. well-functioning rental markets for capital and However, this increase is largest for the land to achieve the efficient allocation of oper- poorest households (a 24-fold increase for ational scales can lead to substantial increases the bottom quintile of farmers as opposed to in agricultural productivity as well as sharp only a 2-fold increase for the top quintile). reductions in inequality levels and poverty. P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   57 Better Technology, Better Practices 18.7 percent to 35.1 percent). The agricul- tural productivity gains can be decomposed In an efficient allocation, access to marketed into (a) 41.9 percentage points from static land implies that farmers can command more gains associated with removing static misal- inputs and produce more output. Farmers location among farmers, (b) 103.6 percentage operating on marketed land are also more points from eliminating distortions in occu- likely to have greater access to other markets pational choices, and (c) 2 percentage points (say, credit markets) and are likely to be sub- associated with the learning component stantially more educated than farmers with- of the static productivity gain (Chen and out marketed land. In addition, the women of Restuccia 2018). these farms are more empowered in terms of From a dynamic perspective, removing labor force participation and market wages, distortions in land allocation will increase and a larger proportion of these farmers the percentage of farmers who learn and invest in intermediate inputs and technology hence will shift the technology frontier faster. adoptions (Restuccia 2016; Restuccia and Labor productivity growth in agriculture Santaeulàlia-Llopis 2017). For instance, a then will depend not only on the general farm with average productivity in Ethiopia is equilibrium effects associated with selection 18.3 percent more likely to use fertilizer and and structural transformation but also on the 20.6 percent more likely to use livestock if it growth of the technological frontier. operates on marketed land (Chen, Restuc- Faster productivity growth in agriculture cia, and Santaeulàlia-Llopis 2017). Although will also accelerate the structural transfor- fertilizers can boost agricultural productiv- mation of the Ethiopian economy—by reduc- ity almost independently of the size of the ing the country’s agricultural employment cultivated plot, this is not the case for large share by 0.16 percentage points per year animals, tractors, and other sizable capital, (Chen and Restuccia 2018). On a regional which, unless rented on a daily or hourly level, Sub-Saharan Africa’s rapid population basis, are more likely to pay off only on large growth (3 percent per year) tends to slow operational scales (Chen 2020). down the structural transformation process: Removing distortions in land allocation larger population implies a greater demand generates not only static productivity gains for agricultural goods, and as a result, more but also dynamic ones. Policies that deepen people work in agriculture, and the sector’s rental markets in Sub-Saharan Africa will TFP decreases because of standard selection positively influence farmers’ decisions on effects (Lagakos and Waugh 2013). technology upgrading to boost productiv- Population growth also affects both ity. Spillovers and learning-by-doing effects agricultural productivity and learning in operate among farmers: the more farmers this way: if distortions to resource allo- who learn in the village and the more their cation were eliminated and population improved techniques spill over to neighbors growth were reduced to an annual rate of who may not have learned, the more farm 1 percent, agricultural productivity would productivity will improve. Conversely, misal- grow faster (by 2.75 percent per year), location can reduce the returns from learning which is more than 1.5 percentage points among the more-talented farmers (Chen and under the high-­ population-growth scenario Restuccia 2018). (Chen and Restuccia 2018). The agricul- Simulations for the Ethiopian agriculture tural employment share would also decline sector shows that removing land distortions faster (by 0.35 percentage points per year)— would raise agricultural productivity (by and this decline is twice as fast as the sce- 264 percent), reduce the share of agricultural nario with high population growth. Faster employment (from 60 percent to 21 percent), agricultural productivity growth, resulting and increase the percentage of farmers from slower population growth, generates who are learning new techniques (from agricultural productivity that is 40 percent 58   Boosting Productivity in Sub-Saharan Africa higher than that of an economy with rapid How Allocative Inefficiencies Exacerbate population growth. Slower population the Impact of Climatic Shocks growth could also offset the gains in agri- Climate shocks have detrimental effects cultural productivity by lowering the price on agricultural productivity—impacts of agricultural goods and reducing the per- that are exacerbated in an environment centage of farmers who learn. with allocative distortions, thus raising the Resource misallocation in agriculture costs associated with climate adaptation. 6 has led to an excessive amount of inputs Extreme temperature induces a negative used to produce a certain minimum level of shock to productivity: an increase of 1 degree value added per capita in agriculture across Celsius in the average temperature above the Sub-Saharan African countries. This chap- optimal threshold reduces agricultural pro- ter so far suggests that improving the quality ductivity by 9 percentage points among farm- of institutions supporting the functioning of ers in Uganda. These harmful effects take land markets can help reduce misallocation place in all regions of the country regardless (Aragón and Rud 2018; Chen, Restuccia, and of the system of land rights. Santaeulàlia 2017; Restuccia 2016). Resource misallocation is particularly Finally, insecure property rights or inef- worrisome in light of the current climate ficient mechanisms to allocate land may change predictions (Aragón, Oteiza, and lead not only to resource misallocation but Rud 2018; Carleton and Hsiang 2016; Chen, also to (a) distorted incentives of technol- Chen, and Xu 2016; Zhang, Zhang, and ogy adoption (Chen, Restuccia, and San- Chen 2017). In response to high-temperature taeulàlia-Llopis 2017); and (b) distorted events, farmers may tend to increase (instead occupational choices by individuals between of decreasing) their land use without reduc- farming and nonagricultural activities ing labor use, especially in the regions with because individuals opting to work in the less-developed land markets (say, the Eastern nonagriculture sectors may have to forfeit and Northern regions of Uganda) (table 4.4). their untitled land (Chen 2017).5 TABLE 4.4  Impact of Weather Shocks on Input Use and Output on Farmers in Uganda, 2009–14 Land Labor Land Labor Output ln(T) In(L) ln(T) ln(L) ln(Y) Dependent variable (1) (2) (3) (4) (5) HDD 0.038** −0.001 0.047*** 0.017 −0.075“ (0.015) (0.014) (0.016) (0.015) (0.033) HDD x ... ... −0.070* −0.139*** −0.077 Western/Central ... ... (0.039) (0.035) (0.065) N 13.113 13.113 13.113 13.113 13.113 R2 0.021 0.025 0.022 0.027 0.050 Source: Aragón and Rud 2018. Note: ... = variable was not included in the regression specification. Standard errors are clustered at household level (in parentheses). All regressions include household fixed effects; fixed effects by growing season (year) and by cropping season (first and second semester); degree days (DD); harmful degree days (HDD); and natural logarithm ln(precipitation). DD and HDD are two measures of cumulative exposure to heat during the growing seasons. DD measures cumulative exposure to temperatures between a lower bound (usually 8 degrees Celsius) and an upper threshold, while HDD captures exposure to extreme temperature (above the threshold). Columns (3) and (4) include interactions of HDD with an indicator of being in the Western/Central region. N = number of observations. R2 = R-squared, the proportion of variance for a dependent variable explained by an independent variable or variables. Significance level: * = 1 percent, ** = 5 percent, *** = 10 percent. P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   59 The greater use of inputs, especially land, History and Effects of ISPs in response to negative supply shocks is Sub-Saharan Africa phased out most ISPs interpreted as a risk management strategy of throughout the 1990s except in Malawi and farmers in the context of imperfect markets. Zambia, where modest ISPs have been imple- That is, absent insurance or access to credit, mented sporadically during the 2010s. Fertil- subsistence farmers may need to increase izer subsidy programs were largely ineffective input use to offset the loss of agricultural in contributing to agricultural productivity output and avoid an undesirable reduction in growth, food security, or poverty reduction consumption (Aragón, Oteiza, and Rud 2018; in the 1980s and 1990s. Instead, they placed Aragón and Rud 2018). In other words, neg- a major fiscal burden on African govern- ative climate shocks can exacerbate allocative ments (Kherallah et al. 2002; Morris et al. inefficiency in environments with imperfect 2007; World Bank 2008). input markets. Reallocating resources from Fertilizer subsidy programs in the region agriculture to other productive uses may also led to corruption and state paternal- attenuate the negative productivity effects of ism, often hindering the development of climate change, but imperfect markets might commercial input distribution systems hinder this reallocation. and contributing to local supply gluts that Climate mitigation and adaptation pol- put political pressure on governments to icies may help reduce the frequency of implement costly grain purchases and extreme temperature events and hence their price-support policies for farmers. For these potential impact on agricultural productiv- reasons, international lenders and bilateral ity. The introduction of digital technologies donors tended to discourage African govern- to implement early warning systems (EWS) ments from relying on ISPs during this period and provide timely information on flood of aid conditionality. alerts, drought warnings, wildfires, and pest The landscape, however, changed quickly outbreaks can also help farmers manage cli- and profoundly since 2005. After African mate shocks.7 Additionally, property rights governments committed to increase their agri- appear to matter for adaptive behavior by culture expenditures under the 2003 Maputo farmers exposed to weather shocks. Policies declaration,9 at least 10 countries introduced that foster property rights and increase the or reintroduced fertilizer subsidy programs, competitiveness in the allocation of land at a collective cost of roughly US$1 billion markets may allow farmers to better cope annually (figure 4.1).10 Large-scale input sub- with climatic shocks. sidy programs often became the centerpiece of governments’ agricultural development Agricultural Subsidies programs. Skepticism based on the past per- Targeted input subsidy programs (ISPs) are formance of these programs was swept aside one of the main tools for many African gov- by arguments that a new vintage of “smart” ernments to boost fertilizer use. ISPs have subsidy programs (further discussed in the yielded short-term benefits for national pro- next section) could take account of past les- duction and food security. However, their sons to maximize the benefits and minimize impacts have been weakened by poor crop the problems of prior programs. response to fertilizer implementation features W hat has b e en t he ex p er ience of that weaken the programs’ contribution to Sub-Saharan African countries with ISPs? broader fertilizer use. 8 Low crop response Large-scale ISPs have tended to raise benefi- to fertilizer has also impeded the growth of ciary households’ crop yields and production commercial demand for fertilizer in Africa, levels, at least in the year that they receive and the ISPs have further crowded out the the subsidy. However, the production effects development of commercial distribution of ISPs are smaller than expected because of channels (Goyal and Nash 2017). low crop-yield responses to fertilizer by most 60  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 4.1  Government Spending on Agricultural Input Subsidies, by Type, in Sub-Saharan African Countries with the 10 Largest ISPs, 2014 200 180 160 140 US dollars, millions 120 100 180 183 80 167 166 60 92 89 40 20 44 43 32 16 0 i i l a o a a a ia ria ga aw al ny an bi ni as op M ge za ne m aF al Ke Gh hi Ni Za M n Se in Et Ta rk Bu Universal subsidy Targeted subsidy Other subsidy Source: Goyal and Nash 2017. Note: A “universal” input subsidy program (ISP) is universal in the sense that (in theory) any farmer can access it. The quantity available to a given farmer is determined roughly based on farm size. “Targeted” programs provide subsidies to selected households based on some observable criteria. The “other subsidy” category pertains to Ethiopia, whose government officially states that it does not have an ISP, yet fertilizer is typically made available to farmers at prices roughly 20–25 percent lower than the commercial distributors’ price to other countries of the region. smallholder-managed fields and the tendency and income appear to decay the year after of ISPs to partially crowd out commercial fer- the farmers receive the subsidies (Carter, tilizer demand. Subsidies have also had rela- Laajaj, and Yang 2014). In Malawi, their tively small, transitory effects on the incomes impact on fertilizer use or crop production of recipient households. was limited even one year after farmers grad- The lack of persistent yield response and uated from the subsidy program following the crowding-out effect are directly linked three years of participation (Ricker-Gilbert to the natural effects of ISPs on incomes and and Jayne 2012). The lack of effectiveness poverty (Goyal 2018). Furthermore, fertilizer might be partly attributed to the influence of subsidy programs have only a modest, if not political and election-related motives on the negligible, impact on food prices (for exam- geographic distribution of subsidies. ple, maize in Malawi) (Ricker-Gilbert et al. 2013). In other cases, the production effects Potential Benefits from Reform of subsidy programs are not large enough to have a significant impact on local food mar- A more systematic strategy for raising small- kets or rural wage rates. holder crop productivity—focusing on Moreover, fertilizer subsidy programs fail sustainably raising the efficiency and quan- to kick-start dynamic growth processes. In tity of fertilizer used—will more effectively Mozambique, their impact on production achieve the region’s agricultural, food P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   61 security, and poverty reduction goals. Such by the efficient allocation. This “taxing the a comprehensive strategy may include ISPs good” misallocation is starker in Sub-Saharan if they can be implemented according to African countries relative to countries in other “smart subsidy” criteria. Other important regions. Firm-level data suggest estimated dis- elements of such a strategy include (a) greater tortion elasticities of productivity of 0.53 for public investment in coordinated systems Ethiopia, 0.44 for Ghana, and 0.52 for Kenya of agricultural research and development (Cirera, Fattal-Jaef, and Maemir 2018). (R&D); (b) water management and exten- These “correlated distortions” alter the sion that emphasize bidirectional learning allocation of inputs and reduce the incentives between farmers of varying resource con- to invest in innovation (Gabler and Poschke straints; and (c) input from agroecologists, 2013; Ranasinghe 2014). This dynamic researchers, and agrodealers (Goyal 2018). inefficiency is likely to both further depress Overall, reforming the design and imple- aggregate productivity and widen the effi- mentation of ISPs while rebalancing govern- ciency gap between high-income and low- to ment spending in favor of high-return core middle-income economies. public goods and policies could deliver high Most distortions caused by tax systems returns for Sub-Saharan African agricul- in low- and middle-income countries arise ture systems. Effective science and extension from size-dependent policies and informality. programs are also necessary to interactively These distortions do not emerge from built-in work with farmers to identify best practices differentiated effective taxation across assets for maintaining and increasing crop pro- or sources of financing but rather reflect ductivity amid changes in economic and either de jure statutory provisions (related to biophysical environments. the scale of operations) or de facto differen- tiated treatment (resulting from incomplete enforcement). In the context of low-­ income Taxation and Informality countries, size-dependent policies and Taxation constitutes an important source of informality issues are therefore the most misallocation, because it interferes with the salient mechanisms through which taxation equalization of marginal products across firms. causes misallocation and hinders productiv- Tax systems may induce sizable productivity ity. These distortions hamper the expansion losses. The productivity cost of tax-induced of efficient firms and contribute to the sur- distortions is particularly high in emerging vival of inefficient ones, resulting in lower and low- to middle-income economies, where aggregate productivity. resource misallocation is more pervasive and firm-level productivity more sensitive to distor- Size-Dependent Tax Policies tions (Bento and Restuccia 2017). This implies that tax systems are an even more sizable source Size-dependent policies interfere with factor of productivity losses in low-income countries demand by (implicitly or explicitly) subsi- than in high-income countries. Tax dispari- dizing or taxing firms based on their scale ties between productive capital and real estate of production, thus distorting the size dis- alone account for as much as a 5–7 percent loss tribution of firms. These policies feature in industrial TFP in emerging and low-income pervasively in tax codes of high-income and economies (IMF 2017). low- to middle-income economies alike. Production inefficiencies induced by taxes Size-dependent provisions also gener- are amplified if the tax wedges in marginal ate implicit marginal taxes that vary with products across firms are positively associ- the scale of operations. For instance, if ated with their productivity. These production size-dependent regulations are being phased inefficiencies take place if, as a result of tax- in as firms expand while licensing, they ation, the size of the most (least) productive act effectively as quotas. They create dis- firms is smaller (larger) than the one indicated incentives for firms to grow and take full 62  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca advantage of scale economies, inducing Mozambique were offered a simplified tax small-business traps. on gross turnover, called the Simplified Tax The resulting negative impact on pro- for Small Taxpayers (ISPC). Since 2009, this duction is sizable: estimates from calibrated tax replaced the corporate income tax, the models suggest that interventions reduc- personal income tax, and the value added ing the average size of production units by tax. A flat tax rate of 3 percent was imposed 20 percent lead to an output contraction of on taxpayers with annual business volumes 8 percent (Guner, Ventura, and Xu 2008). below Mt 2.5 million, while those with busi- Furthermore, although market failures some- ness volumes below 36 times the minimum times warrant size-dependent interventions wage were exempt from the tax. A significant in the short or medium run, the dynamic bunching of small firms emerged below the inefficiency resulting from inertia amplifies eligibility threshold in Mozambique after the the long-run reduction in aggregate produc- ISPC was introduced, as shown in figure 4.2 tion because of the negative impact of such (Swistak, Liu, and Varsano 2017). well-intended policies on firm growth (Buera, Removing size-dependent tax enforce- Moll, and Shin 2013). ment in low- and middle-income countries Preferential tax regimes for small taxpayers would increase TFP, with productivity gains can create disincentives to firm growth. Evi- amounting to 0.8 percent (Bachas, Fat- dence shows that, to minimize their tax liabil- tal-Jaef, and Jensen 2018). Making tax ities, firms tend to bunch below the regulatory systems size-neutral would significantly thresholds created by these regimes (Asatryan attenuate inefficiencies in the allocation of and Peichl 2017; Brockmeyer and Hernan- resources. Size-dependent policies that gen- dez 2016). For example, small taxpayers in erate small-business traps must be elimi- nated because they implicitly subsidize the least-productive firms and tax the most-pro- FIGURE 4.2  Distribution of ISPC Taxpayers in Mozambique: 2010 ductive ones. If governments are to provide versus 2015 tax incentives to spur growth, they should target productivity-enhancing investments to 120 minimize the adverse allocative impact of dis- torting marginal products. If aimed at reliev- ing the proportionally high fixed costs faced 90 by start-ups, policies should aim at facilitat- Number of ISPC taxpayers ing entry rather than subsidizing small firms. In Ghana, switching from a size-dependent 60 tax to a uniform rate would substantially raise per capita income. Gollin (2006) cali- brates the span-of-control model of Lucas 30 (1978) with self-employment technology and dynamics that match the manufacturing sector’s firm-size distribution in Ghana. Gha- na’s tax policy environment, which includes 0 10 15 20 25 30 35 40 a three-tiered tax scheme, is compared with a revenue-neutral counterfactual in which Turnover (Mt 100,000) the rate of taxation is uniform across firm 2010 2015 size. The simulation generates substantial labor reallocation from own-account to wage Source: IMF 2017, using data from Swistak, Liu, and Varsano 2017. Note: The horizontal axis designates the turnover bins (as multiples of Mt 100,000) by which the employment and estimated efficiency gains business taxpayers are classified. The vertical rule designates the business eligibility threshold for that amount to an increase of 6.5 percent the Simplified Tax for Small Taxpayers (ISPC), to the right of which taxpayers are ineligible. A small number of ISPC taxpayers appear above the threshold, possibly because the registration require- in per capita income. However, it does not ment is applied to turnover in the previous year instead. greatly reduce the prevalence of small firms. P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   63 Tax-induced distortions can have sig- I n for ma l, noncompl ia nt f i r ms a re nificant dynamic implications on resource significantly less productive than formal, misallocation through the impact that tax-compliant ones. At the country level, the corresponding allocative inefficiencies there is a significant negative correlation have on the firms’ incentives to invest in between country TFP and the extent of ­ productivity-enhancing technologies. As a informality. Firm-level evidence suggests result, these firms will exhibit slower life-­ that noncompliant manufacturing firms in cycle productivity growth and hence slower low- and middle-income economies have employment growth. Empirical evidence lower productivity than their compliant shows that productivity growth over a firm’s counterparts; for example, businesses that life cycle is flatter among Sub-Saharan Afri- only report 30 percent of their sales have, on can countries relative to more-efficient bench- average, a 4 percent lower TFP than those marks. For example, revenue productivity reporting a greater proportion of their sales (TFPR) increases steadily with the age of the (IMF 2017). establishments in Ethiopia but at a slow pace. Informality is a source of misalloca- This suggests that older firms face bigger dis- tion and hinders productivity. The relative tortions. In contrast to the Ethiopian exam- cost advantage enjoyed by noncompliant ple, older firms in Ghana and Kenya tend to firms affects business dynamism by distort- exhibit smaller TFPR as firms age. In sum, ing creative destruction and growth. In this tax-induced distortions tend to decelerate context, informality enables the survival of firm growth over the cycle and discourage the unprofitable businesses—thus increasing adoption of productivity-enhancing technol- their participation in aggregate output at the ogies (Cirera, Fattal-Jaef, and Maemir 2018). expense of more profitable and compliant Overall, leveling the playing field firms. Differences in the size distribution of regardless of firm size would also gener- all firms relative to formal firms signal the ate large productivity gains. Independently misallocation that arises from such inefficient of government revenue considerations, growth dynamics. The pervasiveness of this better-functioning tax administration would ­ resource misallocation is illustrated in the generate productivity gains by putting an end size distribution of manufacturing firms in to the distortive implicit subsidies enjoyed by Cameroon, Rwanda, and Zambia (figure 4.3) informal, typically low-productivity firms. (Cirera, Fattal-Jaef, and Maemir 2018). Moreover, reducing compliance costs would The size and scope of the informal sector play a part in spurring growth by reallocat- plays an important role in explaining produc- ing resources toward productive activities. tivity differences within sectors and across countries. General equilibrium model simu- lations suggest that countries with high entry Informality-Related Issues and operation costs in the formal sector as Informality is widespread in low- and well as weak debt enforcement tend to have middle-income economies, especially in greater allocative inefficiencies and a larger Sub-Saharan Africa. Informal firms account share of output produced by low-productivity for up to half of aggregate output in low-­ informal firms. These frictions tend to gener- income countries, and they typically circum- ate large informal sectors and exacerbate the vent taxation. Tax avoidance or evasion and misallocation of capital, thus explaining a other nonremitted contributions constitute decline in TFP of up to 25 percent (D’Erasmo the main benefit from informality (Fajnzyl- and Moscoso Boedo 2012). In other words, ber 2007). Tax compliance increases with the model yields a strong negative association development, as the gradual construction of between income per capita and the size of the functioning legal and regulatory frameworks informal sector. makes it more attractive for firms to operate Finally, incomplete tax enforcement can in the formal economy. reduce the capital intensity of informal 64  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca FIGURE 4.3  Size Distribution of Formal Firms versus All Firms and US Benchmark, Selected Sub-Saharan African Countries a. All Firms, Cameroon, 2009 c. All Firms, Rwanda, 2014 e. All Firms, Zambia, 2012 100 100 100 80 80 80 Percentage of rms Percentage of rms Percentage of rms 60 60 60 40 40 40 20 20 20 0 0 0 4 10 9 20 19 50 49 0– 9 0 9 0– 9 1, 99 0+ 4 10 9 20 19 50 9 0 9 0– 9 0– 9 1, 99 0+ 4 9 20 19 50 9 0 9 0 9 0– 9 1, 99 0+ 1– 5– 10 –9 25 24 50 –49 1– 5– –4 10 –9 25 –24 50 49 1– 5– –4 10 –9 25 –24 50 –49 – – 9 – 9 – 9 00 00 00 10 Number of workers Number of workers Number of workers b. Formal Firms, Cameroon, 2009 d. Formal Firms, Rwanda, 2014 f. Formal Firms, Zambia, 2012 40 40 40 30 30 30 Percentage of rms Percentage of rms Percentage of rms 20 20 20 10 10 10 0 0 0 4 10 9 20 9 50 49 0 9 0 9 0– 9 1, 99 0+ 4 9 20 9 50 49 0 9 0 9 0– 9 1, 99 0+ 4 10 –9 20 19 50 9 0 9 0 9 0– 9 1, 99 0+ 1– 5– –1 10 –9 25 –24 50 –49 1– 5– –1 10 –9 25 –24 50 –49 1– –4 10 –9 25 –24 50 –49 5 – 9 – 9 – 9 00 00 00 10 Number of workers Number of workers Number of workers Source: Cirera et al. 2018. Note: The figure shows the distribution of manufacturing firm size (as measured by the number of workers) in selected Sub-Saharan countries. Outlined bars represent the size dis- tribution in United States, which proxies a global efficiency benchmark. firms, induce excess entry of less-productive An individual’s choice between formal entre- businesses, and lead to the reallocation of preneurship, informal entrepreneurship, and inputs toward less-productive firms (Leal nonentrepreneurial work can be influenced Ordóñez 2014). by both personal features (skill level and ini- The coexistence of a sizable informal tial wealth) and institutional factors (such as sector with a formal one poses serious chal- entry costs, taxation enforcement, or finan- lenges to the design of policies to foster entre- cial frictions). The institutional environment preneurship. Policies that impose barriers on of African economies is characterized by high the formal sector tend to lower aggregate pro- registration costs, imperfect credit markets, ductivity through distortions in occupational and low-enforcement tax collection. Entre- choice. Different types of frictions (financial, preneurs pay a registration fee to become institutional, and others) and their interplay formal. Afterward, these entrepreneurs pay may lead to suboptimal occupational choices. taxes and have better access to credit. In P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   65 contrast, informal entrepreneurs evade tax an efficient skill allocation and significant payments and are more likely to face borrow- income gains can be obtained by reducing ing constraints. registration costs and selecting the optimal Taxation and registration costs reduce the tax rate while fostering entrepreneurial skills entry of low-productivity entrepreneurs into and enterprise creation through business the formal sector and induce a larger number training and improved access to credit. of unproductive firms to enter the informal sector. In Cameroon, barriers to entry drive the choice of whether entrepreneurs join the Trade Policy formal sector (Nguimkeu 2015). A counter- Trade policies can affect firm performance factual exercise shows that cutting registra- through (a) mechanisms that induce changes tion costs by half will double the share of within firms and hence affect firm-level com- formal enterprises through the formalization ponents of profitability, and (b) mechanisms of informal firms and new entrants to the that induce the reallocation of economic industry (figure 4.4, panel a). It also increases activity across firms in an industry. The first aggregate income by 15 percent, and the gov- channel’s impact is summarized in box 4.2. ernment’s total net tax revenues more than In the case of the second channel, trade may double (figure 4.4, panel b). In sum, the coun- not affect firm-level profitability, but the terfactual exercises for Cameroon show that trade-induced factor reallocation from the FIGURE 4.4  Simulated Impact of Business Registration Reform on Occupational Choice and Income in Cameroon a. Change in occupational choicea b. Change in income and tax revenuesb 0.09 0.08 0.08 0.07 Proportion of entrepreneurs 0.06 0.06 Percent change 0.05 0.04 0.04 0.03 0.02 0.02 0.01 0 0.02 0.04 0.06 0.08 1.00 0 0.02 0.04 0.06 0.08 1.00 b/c0 b/c0 Formal enterprises Tax revenue gains Informal enterprises Net tax revenue gains Enterprise creation Source: Nguimkeu 2015. Note: The simulated reform is to cut business registration costs. On the x-axes, b represents the reduction of the entry cost implied by the reform, and c0 is the fixed entry cost for the entrepreneur to join the formal sector. 0.05 = 5%. a. Panel a denotes the fraction of formal enterprises, informal enterprises, and new enterprise creation simulated by the corresponding change in entry cost (b/c0). b. Panel b indicates the simulated variation in aggregate income gains (computed as the total income gain from all sectors) and the tax revenue gains (computed as the total tax revenues net from forgone registration fees due to reform). 66  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca BOX 4.2  Trade Liberalization and Within-Firm Changes Firms participate in world markets as producers or competitions and constant elasticity of substitution sellers of goods and as buyers of intermediate inputs (CES) preferences render constant markups. Under used in the production of these goods. Trade poli- alternative demand systems, prices and markups cies, therefore, can potentially affect all phases of tend to respond to trade liberalization (Arkolakis the firm’s production and expenditure decisions: et al. 2019; Feenstra and Weinstein 2017; Mayer, Melitz, and Ottaviano 2014; Melitz and Ottaviano • Transformation of physical inputs to output 2008). • Upgrade (or downgrade) in the quality of the pro- Multiproduct firms can improve revenue produc- ducers’ outputs and inputs tivity (TFPR) by reallocating within-firm resources • Remuneration of workers of different skills from the production of the least to the most • A firm’s locational choices. profitable products. This mechanism improves firm- The channels through which trade reforms affect level performance, and it is analogous to the role of firms will depend on the specific nature of the reallocation in raising aggregate industry perfor- trade policy changes and, particularly, on whether mance. This mechanism only increases TFPR, and these policy changes affect output relative to input this increase is attributed mostly to the reshuffling markets. of resources across products of varying profitability In response to trade shocks, firms are expected (Bernard, Redding, and Schott 2010). to raise their productivity as they undertake actions Effects of Tariff Cuts to become more efficient—say, by adopting better management practices or appointing better manag- Empirical evidence shows that an industry’s prof- ers (Bloom et al. 2013; Schmidt 1997). Productiv- itability increases with its exposure to foreign ity improvements are usually linked to investment competition. Trade liberalization studies focus on in new technologies, research and development episodes of output and input tariff reductions (Amiti (R&D), and entry in export markets. Productivity-­ and Konings 2007; Pavcnik 2002). The effects of enhancing actions are associated with inputs; that is, input tariff cuts are larger than those of output tariff investment will affect not only productivity but also reductions in low- and middle-income countries. the capital stock (De Loecker 2013). They typically operate through two channels— within-firm performance and factor reallocation— Input Market Costs and the relative importance of each channel depends Exposure to international trade can affect the firm’s on the industry’s setting (Melitz and Redding 2014; performance through changes in the trade cost of Melitz and Trefler 2012). inputs. Lower trade costs lead to the import of new The effects of trade liberalization on performance intermediate inputs and an increase in production is heterogeneous across firms. Firms with different beyond what the increase in expenditures would characteristics—such as initial profit level, R&D predict. This increase will be more pronounced if expenditure, and capital intensity, among others— the new inputs are of higher quality than those pre- tend to cope differently with trade shocks (Aw, viously used. If the production technology exhibits a Roberts, and Xu 2011; Bustos 2011; Lileeva and taste for variety, a larger number of imported inputs Trefler 2010). “Learning by exporting” (the mech- will translate into higher output (Halpern, Koren, anism by which firms’ productivity improves after and Szeidl 2015). This mechanism is likely to under- entering export markets) appears to play an import- lie the large within-firm productivity gains found ant role when controlling for the fact that entering in studies that examine the effects of input tariff export markets comes along with higher investment liberalization in India and Indonesia (Amiti and (De Loecker 2013). Konings 2007; Topalova and Khandelwal 2011). In High-Productivity Export Firms fact, input tariff liberalization led to large increases in the number of imported inputs in India (Goldberg High-productivity firms are more likely to enter et al. 2009, 2010). international markets and continue raising their Firms’ prices and markups will adjust in response productivity—as in the case of export firms in to trade shocks. Trade models with monopolistic nine A frican countries: Burundi, Cameroon, (Box continues next page) P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   67 BOX 4.2  Trade Liberalization and Within-Firm Changes (continued) Côte d’Ivoire, Ethiopia, Ghana, Kenya, Tanzania, productive firms are more likely to enter export Zambia, and Zimbabwe (Van Biesebroeck 2005). markets, and (b) learning by exporting is not fully Postexpor t productivit y grow th (learning by supported by the data. exporting) for these firms is attributed to the Finally, the destination of exports also matters to reduction of credit and contract enforcement con- productivity among firms engaged in international straints. However, evidence of postexport growth trade. Firms exporting to other African coun- is not robust. The finding of a positive correlation tries tend to exhibit lower productivity than firms between exports and firm productivity in Ghana, exporting to the rest of the world (Bresnahan et al. Kenya, and Tanzania suggests that (a) highly 2016; Mengistae and Teal 1998). least to the most profitable firms can still Preferential incentives. Trade policies that improve industry-level performance. This support prioritized subsectors and regions section reviews the evidence of trade policy’s may also distort the allocation of factors impact on misallocation in select African and hence lower aggregate productivity. countries. In Ethiopia, import substitution policies Trade reforms can contribute to the real- provided a series of incentives to firms in pri- location of factors of production from the ority sectors from 1996 to 2002—for exam- least-productive to the most-productive firms, ple, subsidies such as tax exemptions and thus boosting the industry’s performance and loss carry-forwards as well as easier access that of the overall economy. From an individ- to credit. These policies led to the entry or ual producer perspective, these reforms are survival of inefficient firms, and the later exogenous and can affect the market shares removal of these policies facilitated the exit of a particular industry. Changes in market of the less-productive firms and incentivized shares toward more-productive firms have firms to grow at a faster pace. the potential to raise aggregate productiv- Such policies were followed by export ity (Collard-Wexler and De Loecker 2015). promotion policies from 2003 to 2012, com- The reallocation process plays an important plemented by the 2002 investment procla- role in improving performance, and its mation that removed the classification of impact depends on the initial dispersion of subsectors as “pioneer” and “promoted.”11 productivity—that is, the dispersion before The export promotion policies granted the reform (Pavcnik 2002). incentives based on the export capability of Tariff policies. Tariff policies have led agroindustry and manufacturing firms (Gebre- ­ to distortions in the allocation of resources silasse 2016). The export-based eligibility cri- across manufacturing firms in Sub-Saharan teria of the 2003–12 policies could reduce the African countries. Changes in output and misallocation of resources by compelling firms input tariffs create wedges that lead to a dis- to be efficient—because exposure to foreign persion in marginal revenue products across competition requires a high level of efficiency. firms. Output tariffs distort competition, Firms in sectors that had been targeted while input tariffs create distortions in cap- during the import substitution period ital and other intermediate inputs markets. (1996–2002) tended to have lower marginal Alleviating or eliminating these distortions products and lower TFPR than nontargeted through trade reforms is conducive to a more firms. Ethiopian firms eligible for these efficient allocation of factors across firms pioneer and promoted sector benefits also (De Loecker and Goldberg 2014). exhibited lower physical productivity (TFPQ). 68  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca In contrast, the export promotion policies development to growth and productivity.13 had no significant effects on the firms’ TFPR Although there are several methods to assess and TFPQ.12 Still, they had less-distortionary the infrastructure-growth nexus empirically, effects on firm productivity as a result of a consensus has emerged that, under the right changes in the eligibility criteria as well as the conditions, infrastructure development can overall export promotion goal (Gebresilasse play a major role in promoting growth and 2016; Gebrewolde and Rockey 2015). equity—and, through both channels, help Trade reforms. The impact of trade reduce poverty. reforms on firm-level productivity is mixed. Sub-Saharan Africa ranks at the bottom On the one hand, trade reforms have tended of all low- and middle-income regions in to raise firm-level productivity in Ethiopia, virtually all dimensions of infrastructure per- but the impact is heterogeneous. Average formance. It also has inherent characteristics firm productivity increased by 2 percent that may enhance the potential importance of after tariffs were cut by 10 percentage points. infrastructure to its economic development— However, productivity grew faster among notably, the large number of landlocked coun- exporting firms than nonexporting firms tries (home to a large proportion of the region’s after liberalization. Productivity gains were total population) and the remoteness of most also higher if input tariffs were cut rather of the region’s economies from global market than output tariffs. Resource reallocation centers. Sub-Saharan Africa’s geographic dis- as a result of lowering tariffs accounted for advantages result in high transportation costs 73 percent of the improvement in Ethiopian that hinder intraregional and interregional manufacturing productivity (Zenebe 2018). trade (Behar and Manners 2008; Elbadawi, On the other hand, the labor productiv- Mengistae, and Zeufack 2006; Limao and ity of manufacturing plants in Swaziland Venables 2001). Other things being equal, (renamed Eswatini in 2018) declined, on the landlocked countries’ limited openness average, by 3 percent during that country’s to trade appears to be the main drag on their trade liberalization period (1994–2003). The growth. The region’s poor infrastructure only productivity effects, however, were hetero- adds to its geographic disadvantages.14 How- geneous across sectors: labor productivity ever, adequate transportation and communi- increased in the apparel sectors but decreased cation facilities can help to overcome them. in pulp and paper and basic metals (Mhlanga Few academic or policy experts would dis- and Rankin 2015). The lower productivity pute the view that infrastructure development of manufacturing firms in Swaziland was fosters growth, but there is no consensus on attributed to the fact that the positive impact the magnitude of the effect or the factors that of reallocating resources to higher-activity shape it. Empirical research initially focused producers was offset by the lack of comple- on the long-term effects of infrastructure on mentary investments to enhance production aggregate output and productivity.15 There is efficiency through innovation and technol- ample evidence, for instance, of a long-term ogy adoption. relationship between infrastructure and out- put in Nigeria and South Africa (Ayogu 1999; Kuralatne 2006; Perkins, Fedderke, and Infrastructure Luiz 2005). This might partly reflect more In the academic literature and in policy data availability compared with other coun- circles, an adequate supply of infrastructure tries in the region. Panel data evidence also services has long been viewed as a key ingre- reveals a significant contribution of trans- dient for economic development (Aschauer portation infrastructure to output (Boopen 1989; IMF 2014; World Bank 1994). Over the 2006; Kamara 2006). Finally, roads, power, past 30 years, researchers have devoted con- and telecommunications infrastructure have siderable effort to theoretical and empirical a significant impact on Africa’s long-run analyses of the contribution of infrastructure growth (Estache 2005). P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   69 At the firm level, there are two distinct Transportation improvements, therefore, approaches to assess the growth and pro- strengthen agglomeration economies if ductivity effects of infrastructure. The first they increase connectivity within the spa- approach evaluates the impact of infrastruc- tial economy (Eberts and McMillen 1999; ture on the firm’s revenues and productivity Graham 2007). relative to its counterparts within the indus- The likely economic benefits of invest- try and in other countries—and this impact is ments in the transportation sector have transmitted through differences in the ability justified funding for new and improved trans- to adopt more efficient technologies or to portation infrastructure. Inadequate trans- operate technologies efficiently across indus- portation infrastructure adds 30–40 percent tries and countries. The second approach to the cost of goods traded among African examines the effects of infrastructure connec- countries (Sinate et al. 2018). Since Africa is tivity (or lack thereof) on resource misalloca- home to 16 landlocked countries, poor and tion and hence on cross-country differences underdeveloped transportation infrastructure in TFP. Using both approaches, this section limits accessibility to consumers, hampers looks at the impacts on firm output and intraregional trade, and drives up import productivity of three specific infrastructure and export costs. For instance, the expense sectors: transportation, energy, and the dig- of moving Africa’s imports to customers ital economy. inland is, on average, 50 percent higher than shipping costs in other low-income regions (Sinate et al. 2018). Transportation Infrastructure A poor road network, in terms of its con- Different strands of economic theory have nectivity and quality, can have deleterious extensively investigated the impact of effects on economic activity. Low-quality transportation infrastructure on economic road networks, along with inefficient trans- activity. Economic geography suggests that portation and trade services, raise logisti- transportation costs play a role in deter- cal and transaction costs—thus restricting mining the location of economic activities producers’ access to markets. Africa’s road (Weber  1928), especially in a context of infrastructure gap, along with high logis- imperfect competition and varying degrees tics costs, have a detrimental impact on the of labor mobility across regions (Fujita and region’s productivity and overall competitive- Thisse 2002). Endogenous growth theory ness (Escribano, Guasch, and Pena 2010). provides a framework that posits public infra- Transportation policies have so far been structure (including transportation infra- insufficient to introduce more competition structure) as an engine of growth through and attract foreign investors, while the region its contribution to TFP (Garcia-Milà and continues to trail the world in both con- McGuire 1992; Hulten and Schwab 1991; nectivity and quality of road infrastructure Munnell 1992). (Calderón, Cantú, and Chuhan-Pole 2018). Transportation improvements, along At the production unit level, there is evi- with lower transportation costs, can poten- dence of the impact of infrastructure on tially reduce firms’ input costs and hence agriculture and manufacturing. The follow- increase productivity. The lower produc- ing discusses the impact of transportation tion and distribution costs induced by the infrastructure on manufacturing activity; the improvements in the transportation sector effects on agriculture are presented in box 4.3. can lead to scale effects and enhance com- Improved transportation networks can petition (Baldwin and Okubo 2006; Melitz spur factor mobility and boost productivity. and Ottaviano 2008). Transportation infra- In Ethiopia, an improved road network can structure can also contribute to productivity influence the entry decisions and entry sizes through agglomeration effects: firms and of manufacturing firms. The effects of infra- workers benefit from being close to others. structure on firm decisions were analyzed 70  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca BOX 4.3  The Role of Transportation Infrastructure in Agriculture The size of subsistence agriculture can be character- infrastructure spending. In Uganda, high transpor- ized as the outcome of the interplay between sectoral tation costs are reflected in substantial price dis- productivities (in agriculture and manufacturing) persion: these high costs incentivize individuals to and transportation productivity. Agriculture takes choose locations that minimize transportation costs place in near or remote rural areas, while manufac- for their agricultural goods. This explains the larger turing goods are produced in urban areas. Economic share of subsistence agriculture because people live in models assume that people devoted to subsistence remote areas to be close to their food source (Gollin agriculture live in remote areas and that labor is and Rogerson 2016). This finding is consistent with mobile across regions. evidence that poor transportation facilities constrain The model calibration for Sub-Saharan Africa agricultural growth (Diao and Yanoma 2003) and finds that agricultural productivity improvements that higher transportation costs alter the incentives and lower costs of intermediate inputs free up labor for agricultural investment (Renkow, Hallstrom, and from the agriculture sector (Gollin and Rogerson Karanja 2004; Stifel and Minten 2008). 2014). Improved transportation productivity helps Improvement in rural road infrastructure can individuals move from subsistence agriculture into reduce crop prices in rural markets, and these price manufacturing, leaving the share of workers living effects are stronger in markets farther from major in the near rural areas unchanged. If productivity urban centers and in low-productivity areas. After improves only in manufacturing, the share of pop- the European Union’s feeder rehabilitation program ulation in subsistence agriculture still declines but in Sierra Leone improved the quality of small rural more slowly than if the boost were in agricultural roads, transportation costs declined for traders pur- or transportation productivity. These findings imply chasing agricultural produce from rural markets that structural transformation at low levels of devel- as well as for farmers bringing their crops to these opment is primarily driven by productivity surges markets (Casaburi, Glennerster, and Suri 2013). a in agriculture and transportation. Economically The better quality of rural roads helped reduce the speaking, a 10 percent increase in agricultural TFP price of the main staples cultivated domestically in combined with a 10 percent reduction in transpor- rural markets along the rehabilitated roads—that is, tation costs leads to a 14 percentage point reduc- rice and cassava. The price reductions for cassava tion in the labor share in subsistence agriculture were larger owing to idiosyncratic factors associ- (Gollin and Rogerson 2014). The welfare effects are ated with the crop: (a) cassava sales are less affected significant—comparable to raising consumption per by seasonal factors, and (b) cassava is bulkier than capita in the economy by 62 percent. other crops to transport. The pattern of labor allocation observed in a. The 2009–11 rehabilitation program targeted four districts in three different low-income countries (that is, a large share of provinces: Kambia and Port Loko (Northern Province), Kenema (Eastern Prov- labor in low-productivity agricultural employment) ince), and Pujehun (Southern Province). These four districts cover 27 percent of the country’s area and 30 percent of its population (Casaburi, Glennerster, and is influenced by high transportation costs and low Suri 2013). using geographic information system (GIS)- are captured in (a) and (b), while (c) measures based panel data on road accessibility of the connectivity of firms with local or distant Ethiopian towns and census-based panel data markets. for manufacturing firms from 1996 to 2008 The quality of local road infrastructure (Shiferaw et al. 2015). Three measures of was positively associated with the number of road infrastructure were considered: (a) total firms in the locality. For instance, a 1 percent distance traveled during a 60-minute drive, improvement in road infrastructure was asso- (b) total area accessible during the 60-minute ciated with a 1.1–1.2 percent increase in the drive, and (c) total travel time from a partic- number of firms. The number of firms had no ular locality to major economic destinations. significant relationship with the connectivity Local improvements in road infrastructure of the road infrastructure. However, the size P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   71 of new entrants was more strongly associated (Alby, Dethier, and Straub 2013). An unre- with connectivity than with the quality of the liable power supply has direct and indirect local road infrastructure. impacts on productivity. The lack of access In sum, better road networks can influence to electricity and the unreliability of its pro- the entry decisions and entry sizes of manu- vision directly hamper the manufacturing facturing firms. Evidence for Ethiopia shows process (Allcott, Collard-Wexler, and that higher-quality local road infrastructure O’Connell 2016; Reinikka and Svens- enables the entry of firms and that more-ex- son 2002). Productivity drops even when tensive market connectivity is important to firms can smooth electricity supply through determine the entry of larger firms. In other self-generation. For instance, scheduled words, improved road infrastructure affects blackouts reduce firms’ productivity by aggregate productivity through its impact on forcing them to shift resources away from the number and size of operating firms—the productivity-enhancing activities (Poczter selection channel (Shiferaw et al. 2015). 2017; Reinikka and Svensson 2002). The unreliable provision of electricity also leads to allocative inefficiencies. Perva- Energy Sector sive inefficiency in the allocation of factors is The availability and reliability of electricity found in Ghanaian manufacturing, as cap- services is key for the economic development tured by the wide dispersion in revenue and of the African continent. No country in the quantity productivity across firms. If these world has ever developed without having inefficiencies were eliminated, the potential access to energy. Energy is needed to oper- TFP gains of Ghanaian manufacturing firms ate industrial machinery and contributes would be in the range of 35–65 percent. These to human capital productivity by providing distortions are partly explained by electricity power to essential facilities such as schools shortages and insufficient power-generating and hospitals as well as for information and capacity (Ackah, Asuming, and Abudu 2018; communication technologies (ICTs). Estache and Vagliasindi 2017). Additional Insufficient and unreliable availability of evidence shows that unreliable electricity electric power is one of the biggest challenges supply affects firm performance through the facing African firms. According to World reduction of firm-level investments (Lumbila Bank Enterprise Surveys, (a) 78 percent of 2005; Reinikka and Svensson 1999). African firms experienced power outages in Power outages have a significant negative 2018; (b) 41 percent of African firms (com- impact on productivity. Eliminating power pared with 30 percent worldwide) identified outages could potentially increase the electricity as a major obstacle to business productivity of Ghanaian manufactur- operation; and (c) the average power interrup- ing establishments by 10 percent.16 Firms’ tion faced by African firms exceeds 50 hours various strategies to cope with power per month—amounting to a 25-day loss of outages (such as using generators, switch- economic activity per year. The economic ing to less-electricity-intensive production cost of the poor energy infrastructure net- processes, changing production times, and work is sizable. For instance, African firms temporarily suspending production) cannot lose nearly 5 percent of their total annual shield them from the negative productivity sales because of power outages (Oseni 2019). effects of these outages. If firms are willing to Poor access to a reliable electric power pay a premium for uninterrupted electricity, infrastructure hinders manufacturing governments can invest in the electric power production in low- and middle-income sector even if it raises electricity prices. How- countries. For instance, manufacturers may ever, that premium is bounded by the addi- have problems connecting to the power tional cost of generating their own electricity grid or, when they do, experience shortages relative to purchasing it from the public grid or fluctuations in voltage and frequency (Abeberese, Ackah, and Asuming 2019). 72  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Access to electric power also increases The  application of digital technologies in the ability of poorer households to allocate finance across African countries will be their labor resources for market production. presented when discussing financial markets. In postapartheid South Africa, the rural elec- trification rollout contributed significantly to Agricultural Applications employment growth in rural communities. By lowering transaction costs (both the The rollout, which expanded rapidly into pecuniary and time costs of access to rural areas and low-capacity household use, and exchange of information), ICTs are had a positive and causal impact on com- facilitating the diffusion of information and munity employment in the rural province knowledge in agriculture. Specifically, ICTs of KwaZulu-Natal.17 The evidence shows a can help improve agriculture in low- and substantial increase of 9.0–9.5 percentage middle-income countries—and, notably, in points in female employment—translating Africa—through three different mechanisms into 15,000 more women participating in the (Deichmann, Goyal, and Mishra 2016): labor force. The increase in female employ- • Promotion of market transparency by ment takes place at the intensive margin: reducing informational frictions and women work about 8.9 more hours per increasing the capacity to assess market week in districts with an average increase information. For instance, access to mobile in electrification between 1995 and 2001 phones reduces information asymmetries (Dinkelman 2011). resulting from intermediaries with mar- Rural electrification has a substantial ket power. Inexpensive mobile technology impact on home production activities as enables rural and often marginalized farm- well: there is evidence of a significant shift ers to join regional and national markets. away from burning wood at home to using • Stimulation of increased demand for timely, electric cooking and lighting in communi- high-quality information on inputs. Agri- ties that were recently electrified. Household culture education and extension services electrification becomes a labor-saving tech- can potentially facilitate the technology nological shock to home production in rural transfer process by assisting farmers in areas, helping women reallocate their time problem solving and by becoming more from home to market production activities. inserted within the agricultural knowledge Electricity may have also cut the production and information systems (Asenso-Okyere costs of new, home-based services for the and Mekonnen 2012). Information deliv- market and provided individuals with other ery about better agricultural practices, new ways to use their labor in self-employment seeds, or new tools is helping to raise the and microenterprises (Dinkelman 2011). productivity of other factors of produc- tion and thus boosting the efficiency of the Digital Infrastructure production process. • Reduction of logistics costs in the different Information and communication technologies stages of the agriculture supply chain. The (ICTs) are key ingredients of a country’s mechanism includes platforms that connect development strategy because of their buyers and sellers along the production (a) inclusiveness (expanding market access to chain, coordinate product delivery, and individuals and firms); (b) efficiency (boost- facilitate secure payments, among others. ing the productivity of different inputs); and (c) innovation capacity (through the creation The rapid expansion of digital tech- of new business models) (Deichmann, Goyal, nologies, as captured by the surge of and Mishra 2016). This subsection focuses mobile-phone and internet penetration, has on the insertion of digital technologies into helped reduce farmers’ and traders’ search agricultural practices and their impact on costs—even in environments with poor wire- development through different channels. line infrastructure or road quality. This lower P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   73 cost to access knowledge and information impact on (producer) price dispersion of can potentially raise rural incomes. Mobile- products that are typically stored by farmers, phone coverage is strongly associated with such as millet and sorghum (Aker and Faf- greater market efficiency because it lowers champs 2015). the price dispersion of agricultural goods. For Mobile-phone coverage can also help instance, mobile phones helped reduce grain increase the market participation of farm- price dispersion across markets in Niger by ers who are in remote areas and producing at least 6.5 percent and intra-annual price perishable goods. After the expansion of variation by 10 percent (Aker 2008). Specif- mobile-phone coverage, the share of Ugandan ically, the largest decline in price dispersion farmers selling bananas increased in the com- across Niger’s grain markets occurred in munities located more than 20 miles from the first four months after getting mobile- district centers (Muto and Yamano 2009). phone coverage, and the marginal impact Lower price dispersion was also observed has decreased over time (figure 4.5). The among sardine fishers and wholesalers in lower price dispersion is partly attributed the Indian state of Kerala because of greater to reduced search costs because grain trad- mobile-phone coverage (Jensen 2007), as ers with mobile-phone coverage had infor- well as among smallholder Ghanaian farm- mation about and access to more markets ers who used multiple data sources including (Aker 2008, 2010). open government data provided by Esoko18 Mobile-phone coverage is more likely to (Schalkwyk, Young, and Verhulst 2017). lower the spatial price dispersion of agricul- In addition, access to digital technolo- tural products that are more perishable (for gies enables farmers to connect with agents example, cowpeas), and this reduction is the and traders to estimate market demand and largest for remote markets in certain periods the selling price of their products, but the of the year. In contrast, it has no significant impact on farm gate prices is not conclusive. FIGURE 4.5  Changes in Price Dispersion before and after Mobile-Phone Coverage in Niger’s Grain Markets 4 3 Monthly difference, CFAF per kilogram 2 1 0 –1 –2 –3 –4 –5 –6 –7 –8 –9 –5 –4 –3 –2 –1 0 1 2 3 4 5 6 Months pre- and post-mobile-phone coverage With year trend Lower con dence interval Upper confidence interval Source: Aker 2008. Note: Price dispersion is regressed on a series of dummy variables pre- and post-mobile-phone coverage. Upper and lower confidence intervals are shown. CFAF = CFA franc. 74  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca On  t he  one ha nd , access to ma rket Digital technologies can also connect information (through radio or mobile phone) farmers with capital goods—especially for was related to higher farm prices in Uganda, those smallholder farmers in remote rural especially for more-perishable goods (Muto areas who can use machinery to improve pro- and Yamano 2009; Svensson and Yanag- ductivity but cannot afford to purchase it. For izawa 2009). On the other hand, there is instance, Hello Tractor in Nigeria is an Uber- evidence that access to information did not like service that enables farmers to request, significantly change average produce prices schedule, and prepay for tractor services from (Fafchamps and Minten 2012), although the nearby owners through short message service effect could be restricted to specific prod- (SMS) texts and using mobile money. The ucts (Aker and Fafchamps 2015; Tadesse smart, two-wheeled tractors are equipped and Bahiigwa 2015). The lack of robust- with Global Positioning System (GPS) anten- ness in the relationship between farm gate nae that collect and transfer necessary data. prices and access to market information The prepayment is released to the owner once could be attributed to differences in the the service is completed (IFC 2018). degree of information asymmetries, type of Digital technologies are also used to information or platform used for delivery, implement EWS, particularly climate mod- and the presence of other market failures els that provide public information on flood (Deichmann, Goyal, and Mishra 2016). alerts, drought warnings, wildfires, and pest Access to information may encourage outbreaks. Timely provision of this infor- farmers, including poor smallholders, to mation can help farmers manage these cli- invest in new technologies. Information on mate shocks. EWS use data from a wide technology transfers and advisory services array of sources, including satellite images is communicated by specialists through and surveys. Satellite images provide cli- agricultural extension services. Digital tech- matic parameters in almost real time (for nology has reenergized such advisory services. example, rainfall, temperature, evaporation, For example, Digital Green, the Grameen vegetation, and land cover) that can reach Foundation, and TechnoServe deliver timely, remote areas without measurement stations actionable information and advice to farmers and allow farmers to manage crop growth. in South Asia and Sub-Saharan Africa (Naka- Automated systems provide early warning of sone, Torero, and Minten 2014). Transaction deviations from normal growth or other fac- costs associated with traditional agricultural tors. Examples of EWS include the following extension services are reduced through a (Ekekwe 2017): mix of voice, text, videos, and the internet. • Zenvus, a Nigerian precision farming Governments are also partnering with mobile start-up, provides soil data (on temperature, operators to coordinate the distribution of nutrients, and vegetative health) to farm- better seeds and subsidized fertilizers in ers so they can optimally apply fertilizers remote areas through e-vouchers (for example, and irrigate their farms. These data-driven Nigeria’s large-scale e-wallet initiative). farming practices are improving farm pro- Electronic extension systems differ in ductivity and reducing waste. their complexity, range of tools, platforms, • UjuziKilimo, a Kenyan start-up, uses big and devices used to transmit information. data and analytics to transform farmers into Digital Green, a global nongovernmental knowledge-based communities and boost organization (NGO), used a participatory productivity by identifying the needs of indi- process to allow farmers’ access to agricul- vidual crops. tural advice by linking them with experts • SunCulture, founded in Kenya and with through local social networks in Ethiopia operations across the region, sells afford- and India (Gandhi et al. 2009). Such ICT able, high-efficiency drip irrigation kits approaches have led to greater adoption of that use solar energy to pump water from agricultural practices by reducing the dis- any source. tance between instructors and farmers. P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   75 Timely delivery of digital extension services tracking and safety in China. Decentralized and EWS information on a large scale requires ledger technology will be able to trace back content development and maintenance. The the origin of food products in shorter intervals cost is high in low-productivity agricultural of time, making it easier to prevent food areas populated by smallholder farmers with scandals and build trust among domestic food poor infrastructure and low skills. Hence, it producers and distributors (Aitken 2017). is essential to develop low-cost tools to make the delivery of agricultural production advice Aggregate Employment Impacts more efficient. ICT developments that reduced Finally, the adoption and use of digital tech- costs for farmers in rural areas lacking appro- nologies may improve employment rates, shift priate infrastructure improved the efficiency occupational employment shares, and reduce of extension service delivery. However, the job inequality across African countries. returns from ICT for farmers in poorer coun- Recent research compares the economic tries were nearly half of those in richer coun- performance of individuals and firms in tries (Lio and Liu 2006). African locations that were on the terrestrial Digital technologies can potentially network of internet cables with those that improve agriculture supply chain manage- were not during the gradual coastal arrival ment as well. They enhance the coordination of submarine cables from Europe (Hjort and of product transportation and delivery, secure Poulsen 2019).19 Employment increased when food safety in global agricultural produc- fast internet arrived, but the higher employ- tion chains, and facilitate secure payments ment in connected areas did not occur at the D eichmann, Goyal, and Mishra 2016). (­ expense of jobs in unconnected areas. Access The nature of food production, along with to fast internet also increased the likelihood greater awareness of foodborne diseases, of individuals being employed in skilled has emphasized the need to guarantee food jobs and had no significant impact on the safety in global food supply chains. Techno- likelihood of individuals being employed in logical products that can trace products from unskilled jobs across African countries. farm to market effectively are being put in However, fast internet did shift employment place, especially among farmers in low- and shares to higher-productivity occupations. middle-income countries trying to reach or As  a result, job inequality declined: the expand to new export markets (Karippach- ­ p ercentage-point increase in the probabil- eril, Rios, and Srivastava 2011). For exam- ity of having a job was comparable between ple, radio frequency identification (RFID) those who only completed primary school- chips are being placed on crates of produce ing and those with secondary or tertiary or in the ears of livestock to collect data on schooling (Hjort and Poulsen 2019). The motion, temperature, spoilage, density, and increase in skilled employment was the larg- light, among other data. The Namibian est for those with tertiary education, while Livestock Identification and Traceability Sys- those with primary schooling joined the tem implements a system that facilitates the unskilled labor force. control, risk management, and eradication of After the arrival of submarine internet bovine disease. The use of RFID rather than cables in Africa, net firm entry increased in paper-based recording has increased data sectors that use ICT extensively (for exam- accuracy and its speed of dissemination, thus ple, finance in South Africa), while produc- contributing to a more dynamic livestock tivity grew among operating manufacturing market (World Bank 2012). firms (Ethiopia). After getting access to fast Technology may also improve food safety internet, firms in Ghana, Kenya, Nigeria, in value chains. For example, IBM, Walmart, Senegal, and Tanzania tend to export more, and the Chinese retailer JD.com together with increase online communication with clients, Tsinghua University have announced a block- and boost employee training (Hjort and chain food safety alliance to improve food Poulsen 2019). 76  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Financial Market Imperfections Financial Frictions and Aggregate Productivity: The Channels of Financial systems across the world have Transmission become deeper, more efficient, and more sta- ble over recent decades. However, the devel- Financial frictions play an important role in opment of domestic financial systems has influencing entrepreneurs’ ability to enter the been uneven across the world’s countries and industry or expand their scale of operations. regions as well as among users within a coun- Firms are heterogeneous in their level of pro- try. Countries with sound macroeconomic ductivity (and hence their optimal scale of policy frameworks and robust growth have operation) or in the sector in which they oper- deepened their financial systems considerably ate. Financial frictions can affect firm-level (Beck, Demirgüç-Kunt, and Levine 2009). and aggregate growth dynamics, and their However, low- and lower-middle-income impact on aggregate productivity will depend countries still exhibit low levels of financial on the number and type of entrepreneurs as depth, lower access to formal financial ser- well as their distribution across the different vices (such as savings accounts and bank types (Buera, Kaboski, and Shin 2015). loans), and restricted access to external Financial frictions can affect TFP through finance (Banerjee and Duflo 2005; Demirgüç- several different channels. At the intensive Kunt, Feyen, and Levine 2013; Lane and margin, financial frictions introduce dis- Milesi-Ferretti 2007, 2017). tortions in the allocation of capital (capital Financial development, in turn, enhances misallocation) among heterogeneous oper- growth at both the country and firm lev- ating production units. The inefficient allo- els. It fosters economic growth through cation of capital among active entrepreneurs improvements in the allocation of resources (as captured by the dispersion in their mar- and higher TFP growth (Beck, Levine, and ginal product of capital) would lower TFP. Loayza 2000). Financial deepening stimu- At the extensive margin, financial frictions lates the growth of those industries that are affect aggregate TFP through two different more dependent on external finance (Rajan channels: the number and the composition and Zingales 1998) and helps reduce the of entrepreneurs. In other words, financial financing constraints on firms—in particular, frictions introduce distortions in (a) the selec- small enterprises (Beck, Demirgüç-Kunt, and tion into entrepreneurship, as productive but Maksimovic 2005). It has a transformative poor individuals delay their entry while rich impact on economic activity: it shapes the and low-productivity entrepreneurs remain structure of industries, the size distribu- in business (misallocation of talent); and tion of firms, and organizational structures (b) the number of production units for a given (Demirgüç-Kunt, Love, and Maksimovic distribution of entrepreneurial talent in an 2006). Countries with deeper financial sys- economy (Buera, Kaboski, and Shin 2015). tems also tend to experience faster poverty reduction as the income shares of their poor- The Intensive Margin est quintiles tend to grow at the fastest pace Financial frictions can influence produc- (Beck, Demirgüç-Kunt, and Levine 2007). tivity through their impact on the level and Allocative efficiency and productivity dispersion of the marginal product of capital growth are enhanced if the financial sys- (MPK). Firms’ MPK tends to be higher than tem (a) enhances the quality of information it would be otherwise amid binding credit about firms; (b) exerts sound corporate gov- constraints (and holding constant wages and ernance over the resource-borrowing firms; rental rates of capital). There is evidence of (c) provides effective mechanisms to manage, large returns to capital among small-scale pool, and diversify risks; (d) mobilizes savings retailers in Mexico—about 20–30 percent from surplus units toward the most promis- per month or three to five times the market ing projects in the economy; and (e) facilitates interest rates (McKenzie and Woodruff trade (Levine 2005). 2008). Similarly, randomized grants to P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   77 microentrepreneurs in Ghana render very number; hence the impact of tighter credit large returns to capital: 7–10 percent and constraints on firm size is ambiguous. approximately 25 percent per month for cash T h e s h i f t i n t h e c omp o sit ion of and in-kind grants, respectively (Fafchamps entrepreneurs leads to higher average et al. 2011). productivity and, therefore, greater labor In equilibrium, the MPK is higher if the demand. However, financial frictions reduce entrepreneur’s productivity is greater (given the amount of capital used by entrepre- any amount of capital). In this context, the neurs with binding credit constraints. Poor, MPK of constrained entrepreneurs is higher marginal-ability entrepreneurs will switch if their levels of productivity increase (for their occupations from entrepreneur to worker any given amount of wealth), even in cases in the presence of financial constraints. The of pure collateral constraints. The empirical demand for labor declines while the supply evidence shows a great degree of concentra- increases, and lower wages will clear the tion in the returns of Mexican entrepreneurs market. The constrained demand for capital who report themselves as financially con- declines, thus lowering the interest rate and the strained (McKenzie and Woodruff 2008). In cost of capital. Lower labor and capital costs addition, the larger returns of women entre- lead to an increase in the firm’s profitability as preneurs in Ghana correspond to those firms well as in the threshold for entry or survival. that were already more profitable (Fafchamps Relative to a frictionless environment, some et al. 2011). These findings suggest that pro- high-wealth, low-­ productivity individuals will ductivity leads to higher returns to capital. enter and replace poor, marginal-productivity The MPK across firms also varies greatly. entrepreneurs. Empirical evidence suggests that manufac- Although financial frictions have an turing firms in Sub-Saharan Africa with ambiguous net impact on entrepreneurship less access to finance have higher MPK in partial equilibrium, they unequivocally and that, conditional on access to finance, lead to higher entrepreneurship rates in gen- small firms have lower MPK. These find- eral equilibrium (Moll 2014). Amid financial ings imply that higher efficiency could be constraints, wealthier individuals are more attained by allocating more capital to larger likely to become entrepreneurs while lower firms (Kalemli-Ozcan and Sørensen 2016). wages and capital rental rates translate into Restricted access to finance has important higher firm profits. The lower input prices, in real effects: for instance, moving firms from turn, lead to a larger unconstrained scale of environments with easy access to finance to production for all entrepreneurs. The region those with poor access will increase their of high-productivity, low-wealth entrepre- MPK by 45 percent. Therefore, financial con- neurs who are capital-constrained expands straints may help explain resource misalloca- with lower input prices in general equilib- tion within countries, and this misallocation rium. Those high-productivity, low-wealth is significantly associated with the strength individuals who remain as entrepreneurs of property rights. In other words, firms may are more constrained than they would be in be reluctant to reinvest their profits in the partial equilibrium, whereas the high-wealth, absence of secure property rights. low-productivity entrepreneurs who enter because of the lower input prices tend to be The Extensive Margin unconstrained. Financial frictions can also affect productivity through their impact on individuals’ occupa- Quantitative Analysis of the Channels of tional choices. They tend to operate through Transmission the occupational choices of low-wealth, marginal-ability entrepreneurs. In partial Financial frictions tend to have sizable equilibrium, financial frictions increase the effects on labor productivity, aggregate and activity of entrepreneurs but lower their sector-level TFP, and capital-output ratios. 78  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Financial frictions may lead to an estimated models enable financial intermediaries to decline of 20–30 percent in aggregate TFP select the amount of labor devoted to mon- at equilibrium in one-sector, closed economy itoring loan activity. In these models, the models (Buera, Kaboski, and Shin 2011; likelihood of detecting malfeasance depends Buera and Shin 2013). The intensive margin on this decision and the technology used in (capital misallocation) explains nearly the financial sector (Greenwood, Sanchez, 40 percent of the TFP reduction, and finan- and Wang 2010). cial frictions tend to reduce entrepreneurship Model simulations suggest that techno- rates (Midrigan and Xu 2014). logical improvements in financial interme- Financial frictions have more deleterious diation account for about 29 percent of US effects on manufacturing activities than on growth (Greenwood, Sanchez, and Wang services. They tend to reduce the TFP of manu­ 2010). Further analysis shows that 45 percent facturing sectors by more than 50 percent, of per capita growth in Taiwan, China, while the TFP of services sectors fall by less from 1974 to 2004 (6.3 percent per year) than 30 percent. The differential impact across is attributed to financial development (and sectors might reflect the higher relative price of about 16 percent to TFP growth). Finally, manufacturing goods to services in financially the evidence suggests that the output per cap- underdeveloped economies. At the same time, ita of Uganda could more than double if the the capital-output ratio declines by 15 percent country were to adopt global best practices in the presence of financial frictions—an effect in financial intermediation. However, this driven primarily by the higher relative price of impact amounts to only 29 percent of the gap manufactured investment goods in financially between Uganda’s potential and actual output underdeveloped economies (Buera, Kaboski, (Greenwood, Sanchez, and Wang 2013). and Shin 2011). Alleviation of financial constraints has Wealth, Self-Financing, and dynamic effects on entrepreneurship rates Financial Frictions and aggregate TFP. For instance, quadrupling access to financial services will increase entre- Wealth is an important driver of occupational preneurship rates in Thailand by 4 percentage choice in the presence of financial frictions. points (Giné and Townsend 2004). Financial Savings are incentivized by the higher rates of deepening explains 70 percent of the overall entrepreneurship stemming from financially TFP growth in Thailand from 1976 to 1996 constrained environments and the role played (Jeong and Townsend 2007). Financial fric- by wealth in relaxing these constraints. In tions also have an impact on the transition turn, self-financing motives will depend on dynamics after growth-enhancing reforms: the persistence of the firm’s productivity. output growth converges slowly to a new There is ample evidence of highly persistent equilibrium after reforms that trigger an capital (and other asset) returns in Thailand efficient reallocation of resources (at half the (Pawasutipaisit and Townsend 2011); among speed of a neoclassical Solow model). Addi- manufacturing firms in the Republic of tionally, investment rates and TFP tend to be Korea (Midrigan and Xu 2014); and among initially low and increase over time, consis- industrial plants in Chile and Colombia tent with the experience of miracle economies (Moll 2014). (Buera and Shin 2013). Financial frictions also play an important Finally, financial market development role in business entry decisions in low- and could ensure that capital is channeled to those middle-income countries. Wealth and access firms that are more productive and whose to finance greatly influenced the business survival is highly dependent on the availabil- entry decisions of individuals in rural and ity of finance. Technological improvements in semiurban regions of Thailand before the financial intermediation may increase under- 1997 Asian financial crisis (Paulson and standing of this channel. Costly verification Townsend 2005). However, wealth did not P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   79 play a significant role in the occupational on services sector TFP (figure 4.6, panel a), choices of Thai individuals in the aftermath while talent misallocation accounts for more of the crisis (Nyshadam 2014). This finding than 50 percent of the effect on manufactur- points to a substantial relaxation of financial ing sector TFP (figure 4.6, panel b) (Buera, constraints. Kaboski, and Shin 2011). Financial frictions tend to hinder factor Financial frictions can act like an adjust- reallocation. Self-financing with internal ment cost that prevents credit-constrained funds or through forward-looking behavior firms from fully adjusting their capital in can potentially alleviate these constraints response to productivity shocks, thus lower- (Buera, Kaboski, and Shin 2011). It can ing aggregate TFP. The evidence shows that potentially reduce capital misallocation financial frictions can reduce TFP levels by resulting from financial frictions. But its up to 40 percent, and these losses are asso- impact may not be as large in sectors with ciated primarily with distortions on entry larger-scale, substantial financing needs, and technology adoption decisions in the such as manufacturing (Buera, Kaboski, modern sector of the economy (that is, man- and Shin 2015). On average, manufactur- ufacturing). TFP losses attributed to capital ing firms are more vulnerable to financial misallocation across manufacturing firms frictions than services firms because the are smaller and account for a fraction of the former have larger scale and financing needs, overall efficiency losses related to the tighten- greater misallocation of capital and entre- ing of borrowing constraints (Midrigan and preneurial talent, and larger distortions on Xu 2014). entry and exit decisions. The evidence sug- The inability of financial frictions to gests that capital misallocation accounts for generate large losses from misallocation 90 percent of the effect of financial frictions might be attributed to the fact that relatively FIGURE 4.6  Modeling the Impact of Financial Frictions on Sector-Level TFP a. Services sector TFP b. Manufacturing sector TFP 1.1 1.1 1.0 1.0 E ect of financial frictions (%) E ect of financial frictions (%) 0.9 0.9 0.8 0.8 0.7 0.7 0.6 0.6 0.5 0.5 0.4 0.4 0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5 External finance to GDP External finance to GDP Total effect Capital reallocation Capital and talent reallocation Source: Buera, Kaboski, and Shin 2011. Note: The solid lines trace the total effect of financial frictions on the measured total factor productivity (TFP) of the service sector (panel a) and the manu- facturing sector (panel b). Sector-level TFPs are normalized by their respective levels in the perfect-credit benchmark. 80  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca more-productive firms accumulate inter- intending to become entrepreneurs would nal funds over time, enabling them to relax prefer to dissave (Banerjee and Moll 2010; or overcome borrowing constraints. Entry Buera 2008). and technology adoption decisions, on the Productivity losses and poverty traps other hand, entail large, long-lived invest- resulting from financial frictions have led ments with gradual payoffs that are diffi- to several antipoverty policy interventions. cult to finance using internal funds. Here, Asset grant programs have become com- well-developed financial markets play a monplace in the policy agenda of low-income critical role in generating efficient entry and countries. Grant programs can help identify technology adoption—and thus increasing potential high-growth entrepreneurs and aggregate productivity. facilitate their growth. Self-financing can help eliminate the In Nigeria, for example, grants with a capital misallocation arising from financial competition component (for example, the frictions if idiosyncratic productivity shocks You Win! program in 2011) increased entre- are persistent. However, these efficiency gains preneurial activity, including entry, survival will depend on the entrepreneurs’ productiv- employment, and profits (McKenzie 2017). ity levels and asset variation. Entrepreneurs Nearly 6,000 (out of 24,000 applicants) who generate wealth out of previous busi- were selected for a four-day business-plan ness success can accumulate enough internal training course, and each winner received an funds to self-finance their investment pro- average award of US$50,000. Grants were grams only if their high-productivity episodes provided to both new and existing firms. are protracted. In the presence of persistent After three years, new-firm winners were idiosyncratic productivity shocks, capital 37 percentage points more likely than the misallocation and TFP losses from financial control group to be operating a business and frictions are small, but their speed of transi- 23 percentage points more likely to employ tion is slow. In fact, it takes a prolonged time 10 or more workers. Existing-firm win- to achieve allocative efficiency as the initial ners were 20 percentage points more likely capital misallocation slowly unwinds over to have survived and 21 percentage points time (Moll 2014). more likely to employ 10 or more workers. Firms that received these grants tended to innovate more than the control group and Financial Frictions and Poverty Traps earned higher sales and profits. They also Financial frictions may lead to (individual acquired more inputs (capital and labor) and aggregate) poverty traps by distorting without changes in business networks, men- the entry decisions of entrepreneurs. Pov- tors, self-efficacy, or uses of other sources of erty traps can be driven by either lower finance (McKenzie 2017). wages (as individuals join the labor supply On the other hand, the impact of micro- because they cannot afford the fixed costs to finance programs in either urban or rural become entrepreneurs) or lower interest rates environments has been widely examined. (as excess capital supply lowers interest rates Evidence comes from countries ­ i ncluding and limits individuals’ ability to save their Bangladesh (Pitt and Khandker 1998); way out of poverty over time) (Aghion and Ethiopia (Tarozzi, Desai, and Johnson Bolton 1997; Banerjee and Newman 1993). 2015); India (Banerjee et al. 2014; Field Poverty traps are driven not only by lower et al. 2013); Kenya, Tanzania, and Uganda input prices but also by the self-financing (Greaney, Kaboski, and Van Leemput 2013); motive. Initial wealth levels influence how the Philippines (Karlan and Zinman 2010); rapidly self-financing materializes, while and Thailand (Kaboski and Townsend 2011, some individuals might not find it opti- 2012). These programs have low take-up mal to save for prolonged periods. In a rates, and they fail to report large, dramatic, low-interest-rate environment, those not or sustained increases in entrepreneurship, P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   81 income, investment, or consumption. There users remained invariant (Jack and Suri 2014). is a large variation on the impact across The consumption-smoothing abilities of programs, partly attributed to differences M-Pesa users were attributed to their greater in the program design (Attanasio et al. likelihood of receiving remittances in response 2015; Field and Pande 2008; Field et al. to shocks (not only in greater amounts but 2013; Greaney, Kaboski, and Van Leemput also from more different types of people). This 2013). Finally, microloans have a substan- greater risk-sharing ability translated into tial long-run impact on earnings for estab- increased saving, higher consumption, and lished entrepreneurs, while entrants tend to occupational changes for user households. be marginal borrowers (Banerjee et al. 2014; Mobile money services have also led Field et al. 2013; Greaney, Kaboski, and Van to changes in the composition of house- ­Leemput 2013). hold assets in Kenya. Financial savings (that is, self-reported cash plus balances in bank accounts, savings clubs, and mobile Introduction of Digital Technologies money accounts) have increased in areas in Banking and Finance with a growing network of agents, espe- Mobile money services are an innovation that cially among households headed by women. is boosting financial inclusion in Sub-Saharan M-Pesa account holders tend to be less prone Africa as they bring unbanked people into to using informal saving mechanisms (such the formal financial system. Kenya’s M-Pesa as rotating savings and credit associations application, now the country’s dominant [ROSCAs]) and are more likely to access for- retail payment platform, has been one of mal banking services (Mbiti and Weil 2016). the most successful deployments of mobile M-Pesa registered users are also more likely money in the world. 20 Nearly 70 percent of to save than those who are not registered Kenya’s adult population adopted M-Pesa’s (Demombynes and Thegeya 2012). mobile money services within four years after In Burkina Faso, mobile money users are its launch in 2007. more likely to save for health emergencies— especially among the rural population, Africa’s Mobile Money Frontier women, and less-educated individuals. This The rapid adoption of M-Pesa in Kenya greater incidence of saving is attributed to has been attributed to the fast expansion of the possibility of transferring money within mobile-phone networks and the swift deploy- subregions of the country using a secure ment and growth of a dense network of platform (Ky, Rugemintwari, and Sauviat agents (end distributors of the service), which 2021). Greater access to mobile money ser- are small business outlets that transform vices has also raised the likelihood of using cash into e-money and vice versa for custom- a bank account rather than other financial ers (Jack and Suri 2014). The rapid uptake products, possibly because banking insti- of digital finance in Kenya has also been tutions started collaborating or competing attributed to the dominant position of the with M-Pesa. (For example, in 2012, M-Pesa mobile network operator Safaricom, a pro- launched the M-Shwari account, a micro- gressive financial regulator (the Central Bank credit and microsavings product, further of Kenya), and multiple densely populated discussed below.) areas (Babcock 2015). Saving can help microentrepreneurs Kenyan households have been able to increase their ability to respond to unexpected strengthen their informal risk-sharing net- shocks and finance lumpy investments. In works and respond better to shocks by using Tanzania, policy interventions that promote mobile money services. For example, in access to new mobile accounts increased sav- response to an adverse income shock, con- ings and access to finance among women sumption declined by 7 percent among non­ microentrepreneurs (Gautam et al. 2018). users of M-Pesa while the consumption of The Business Women Connect program, in 82  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca partnership with an international nonprofit, also declined in areas with increased access TechnoServe, was evaluated through two to agents. The diffusion of mobile money interventions: The first was a registration services in Kenya helped lift about 194,000 and training session on M-Pawa—a mobile households out of extreme poverty and finance product that enables customers to induced 185,000 women to change their (a) save money in an interest-bearing mobile main occupation to business or retail services savings account, and (b) access microloans (Suri and Jack 2016). based on good savings performance. The Digital credit is also emerging as an option second intervention provided microentrepre- to short-term banking for microfinance neurs with intensive business skills training. loans. Mobile operators are partnering with One year after the interventions, women financial institutions to provide small, short- saved substantially more through the mobile term loans directly to customers through accounts, had greater access to microloans their existing mobile money ecosystem. through the accounts, expanded their busi- M-Pesa partnered with Commercial Bank of ness portfolios through the creation of new Africa Ltd. (CBA) to launch M-Shwari prod- businesses, and reported higher levels of ucts in November 2012. M-Shwari users can empowerment and well-being (Gautam et al. earn interest on savings products and qualify 2018). In turn, the women’s business and for CBA-backed loans. Digital loans have not financial literacy has further bolstered the only led to an expansion of credit for eligible use of mobile savings accounts—thus incen- households but also have strengthened house- tivizing greater capital investment, labor hold resilience. In response to a negative effort, new products, and better business shock, households eligible for M-Shwari are practices. These short-term impacts have yet less likely to forgo expenditures (Bharadwaj, to translate into greater profits, but the evi- Jack, and Suri 2019). The successful update dence suggests that access to mobile savings of M-Shwari in Kenya has led to the devel- accounts has a greater impact on perfor- opment of similar products in other African mance if it comes along with measures that countries—say, M-Pawa in Tanzania, which alleviate the complementary human capital serviced 4.9 million borrowers in the first constraints faced by women. two years (Aglionby 2015), and MoKash in Uganda, which registered 1 million users in Ripple Effects of Digital Banking the first three months of its launch in 2016. and Finance Digital credit has some advantages rela- Mobile money services may have facilitated tive to traditional loans. It is approved more occupational choice in Kenya. Individuals liv- quickly and is readily available to customers ing in areas with a larger increase in mobile without requiring an in-person vetting by money agents are more likely to work in the banking institution. Telecommunications business or sales and less likely to work in data are used to develop alternative credit farming or have a secondary occupation. The scores, thus facilitating the extension of loans expansion of M-Pesa has also enabled women without collateral or the traditional credit to graduate from subsistence agriculture, cut scores computed by credit bureaus. Digi- down their reliance on multiple part-time tal-based credit scores may grant financial jobs, and reduce the average household size inclusion to individuals without credit scores (Suri and Jack 2016). in environments that lack verifiable credit Access to mobile money in Kenya has also history data or have nonexistent or ineffec- had a long-term impact on household wel- tive credit bureaus. fare. Consumption per capita grew substan- On the other hand, digital credit also tially among households living in areas with poses some challenges. The size of these loans increased access to mobile money agents, is not as large, and they often have relatively and this effect was twice as large for house- high interest rates, multiple fees, and short holds headed by women. Extreme poverty repayment periods. For example, the average P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A l l o c a t i o n i n ­S u b - S a h a r a n A f r i c a    83 M-Shwari loan is about US$12 with maturity increased use of digital financial services. In of no more than 30 days (Cook and McKay this context, efforts to foster digital literacy 2015). Users are charged a fixed facilitation would help potential users to understand fee (instead of an interest rate). These fees are the interface with digital financial systems. typically high—for instance, a monthly fee Training sessions to understand the benefits of 7.5 percent for M-Shwari (or 138 percent of digital financial products and, more annually), or 10 percent per week for some importantly, how to use them will increase Malawian digital loans (an annualized rate the uptake of digital accounts and deposits of 1,000 percent). Repayment of digital (Holloway, Niazi, and Rouse 2017). loans on time raises the probability of the user being granted larger loans with lower fees and longer maturity. It remains an open Notes question whether the uptake of digital loans   1. The selection channel also involves distor- would decline if borrowers had more infor- tions that can affect individuals’ occupational mation on these products or were already choices, such as (a) joining the formal sector fully informed about their costs (Francis, as an entrepreneur (instead of as an informal Blumenstock, and Robinson 2017). entrepreneur or worker); and (b) agriculture Access to credit among women entrepre- versus nonagriculture jobs. neurs is more restricted than among men   2. So far, the academic literature has identified some factors that can account for large effects because of inequality in the ownership of of misallocation in agriculture; however, that fixed assets (say, land or property) to serve is not the case for the extent of misallocation as collateral to secure loans. However, found in manufacturing (Restuccia and Rog- developments in the financial technology erson 2017). industry can be harnessed to unlock the col-   3. Revenue total factor productivity (TFPR) is lateral challenge facing women entrepreneurs. typically defined as the ratio of firms’ sales ­ Psychometric loan appraisal technologies— (or revenues) to input costs (appropriately which predict the likelihood of loan repay- weighted by their production elasticities). The ment by entrepreneurs—have been used as an marginal product of land is the additional alternative to traditional collateral in Ethio- output gained from adding another unit of pia (Alibhai et al. 2018). Specifically, they test land. This might apply to a farmer who pur- chases a field adjacent to the existing prop- the ability (business skills and intelligence) erty or to a factory owner who increases the and willingness (ethics, honesty, attitudes, square footage of a facility. and beliefs) to repay a loan. Borrowers take   4. Restuccia and Santaeulàlia-Llopis (2017) an interactive, tablet-based test consisting of use the 2010–11 Malawi Integrated Survey games, puzzles, and questions. If they score of Agriculture (ISA). This survey has ample above a certain cutoff, they can obtain an information on agricultural production (phys- uncollateralized loan of up to US$7,500. ical amounts by drop and plot) and the inputs Customers scoring at a high threshold on used in all agricultural activities at the plot the psychometric test were seven times more level. The data are representative at the likely than lower-performing customers to national level, with a sampling frame based repay their loans. This pilot is being cur- on the census and an original sample that includes 12,271 households (56,397 individ- rently scaled up in Madagascar and Zimba- uals), of whom 81 percent live in rural areas. bwe and will be implemented next in Côte Household land is measured as the sum of the d’Ivoire, Nigeria, and Zambia. In the absence size of each cultivated household plot, includ- of collateral, and with limited information ing rented-in land (about 12.5 percent of all available on the creditworthiness of women cultivated land). Household farms, on aver- borrowers, psychometric testing is a promis- age, cultivate 1.8 plots. Plot size is recorded ing solution. in acres using the Global Positioning Sys- Finally, cash still dominates the transac- tem (GPS) for 98 percent of plots. For each tions of many of the world’s poor despite the household, the amount of the land used for 84   Boosting Productivity in Sub-Saharan Africa agricultural production is measured regard- 10. The 10 countries in the region with the largest less of the tenure status. The operational scale ISPs (Burkina Faso, Ethiopia, Ghana, Kenya, of farms is small: 78.3 percent of households Malawi, Mali, Nigeria, Senegal, Tanzania, operate less than 1 hectare, 96.1 percent and Zambia) spent US$1.02 billion on fer- of households operate less than 2 hectares, tilizer subsidy programs in 2014 (Goyal and and only 0.3 percent of households operate Nash 2017). more than 4 hectares. The average farm size 11. The main criteria for the distinction between is 0.83 hectares. The data contain detailed “promoted” and “pioneer” are a sector’s information on the quality of land for each labor intensiveness and linkage to the agri- plot used in every household. It distinguishes culture sector. “Pioneer” activities are the top up to 11 dimensions of land, thus enabling tier of activities that are agriculture-based and the control for land quality when measuring require a large outlay or have strong linkage household-farm productivity. effects. “Promoted” activities are of second-   5. Chen (2017) builds a two-sector general equi- ary priority and include rainfed agriculture, librium model where untitled land cannot be livestock development, nonbasic industries, rented or traded across farmers, and it can and contracting. only be used by those who were originally 12. Bigsten, Gebreeyesus, and Söderbom (2016) assigned to the plot. Simulations of the model also found that output tariff reductions had show that titling all of the land raises agri- no impact on firms’ productivity in Ethiopia. cultural productivity by 51.8 percent. About 13. Increasing attention has been paid recently to 42.5 percent of this productivity gain arises the impact of infrastructure on poverty and from land reallocation, while 57.5 percent inequality (Calderón and Servén 2004, 2010; reflects lower distortions in occupational De Ferranti et al. 2004; Estache 2005; World choice. Bank 2005).  6. An increase in average temperature of 14. Infrastructure gaps in the Africa region are 2 degrees Celsius is expected to reduce agri- driven by a host of issues beyond the financ- cultural production by almost 25 percent ing gap—for instance, the lack of commit- (IPCC 2015). Attenuating the collateral ment to sustainable tariffs in infrastructure effects of climatic shocks (such as migration, services such as electric power, transporta- occupational change, and land changes) will tion, and water. Yet there is heavy reliance require reallocation of land and labor. on public subsidies. The gap is also attributed   7. Data from satellite images can provide a to the poor performance of public utilities, range of climatic parameters in almost real characterized by weak management and time (such as rainfall, temperature, and so on) political interference. There is also weak that can reach farmers in remote areas with- political support for sector reforms that can out measurement stations and enable them to crowd-in private infrastructure investment, better manage crop growth. such as the opposition of state-owned enter-   8. Fertilizer implementation features that have prises to p ­ ublic-private partnerships. An weakened the impacts of targeted ISPs include in-depth discussion of the issues mentioned frequent late delivery of vouchers, politicized above—although highly relevant to under- voucher allocation, and illegal collusion stand infrastructure gaps in the region—goes between leaders and agrodealers, among beyond the scope of this report. others. 15. See Ndulu (2006) and Ayogu (2007) for   9. At the Second Ordinary Assembly of the diagnostic views on infrastructure and its African Union in July 2003 in Maputo, long-term impact in Africa. Mozambique, African heads of state and 16. The estimation of a causal relationship government endorsed the “Maputo Dec- exploits exogenous variation in outages, laration on Agriculture and Food Security induced by an electricity rationing program, in Africa.” The Declaration’s important across small- and medium-size Ghanaian provisions included a “commitment to the manufacturing firms. allocation of at least 10 percent of national 17. Accounting for the likely endogeneity in budgetary resources to agriculture and rural the relationship between electrification and development policy implementation within employment growth can be affected by five years.” endogenous criteria to place infrastructure P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A l l o c a t i o n i n ­S u b - S a h a r a n A f r i c a    85 projects in certain regions. For instance, per- under Electricity Constraints in Developing formance criteria can introduce biases in the Countries.” World Bank Economic Review comparison between electrified and nonelec- 27 (1): 109–32. trified areas (Dinkelman 2011). Alibhai, Salman, Niklas Buehren, Rachel Coleman, 18. Esoko is an agricultural profiling and mes- Markus Goldstein, and Francesco Strobbe. 2018. saging service, managed on the internet, that “Disruptive Finance: Using Psychometrics to delivers market data via mobile phone. It is Overcome Collateral Constraints in Ethiopia.” headquartered in Accra, Ghana. Working paper, World Bank, Washington, DC. 19. Submarine internet cables at landing points Allcott, Hunt, Allan Collard-Wexler, and Stephen on the African coast arrived gradually in the D. O’Connell. 2016. “How Do Electricity late 2000s and early 2010s. Shortages Affect Industry? Evidence from 20. “M-Pesa” is derived from M for mobile and India.” American Economic Review 106 (3): “pesa” (Swahili for cash). 587–624. Amiti, Mary, and Jozef Konings. 2007. “Trade Liberalization, Intermediate Inputs, and References Productivity: Evidence from Indonesia.” Abeberese, A ma Baafra, Charles Godfred American Economic Review 97 (5): 1611–38. Ackah, and Patrick Opoku Asuming. 2019. Aragón, Fernando M., Francisco Oteiza, and “Productivity Losses and Firm Responses to Juan Pablo Rud. 2018. “Climate Change and Electricity Shortages: Evidence from Ghana.” Agriculture: Farmer Adaptation to Extreme World Bank Economic Review 35 (1): 1–18. Heat.” Working Paper W18/06, Institute for Ackah, Charles G odfred, Pat rick Opoku Fiscal Studies, London. A su m i ng , a nd D er r ick A budu . 2 018 . Aragón, Fernando M., and Juan Pablo Rud. “Misallocation of Resources and Productivity: 2018. “Weather, Productivity, and Factor The Case of Ghana.” Unpublished manuscript, Misallocation: Evidence from Ugandan Institute of Statistical, Social and Economic Farmers.” Background paper, World Bank, Research, University of Ghana. Washington, DC. Aghion, Philippe, and Patrick Bolton. 1997. Arkolakis, Costas, Arnaud Costinot, Dave “A Theory of Trickle-Down Growth and Donaldson, and Andrés Rodríguez-Clare. 2019. Development.” Review of Economic Studies “The Elusive Pro-Competitive Effects of Trade.” 64 (2): 151–72. Review of Economic Studies 86 (1): 46–80. Aglionby, John. 2015. “New Payments Service to Asatryan, Zareh, and Andreas Peichl. 2017. Launch in Kenya.” Financial Times, October “Responses of Firms to Tax, Administrative 27, 2015. and Accou nting Ru les: Evidence from Aitken, Roger. 2017. “IBM & Walmart Launching Armenia.” Working Paper 6754, Center for Blockchain Food Safety Alliance in China with Economic Studies and ifo Institute (CESifo), Fortune 500’s JD.com.” Forbes, December 14, Munich. 2017. Aschauer, David A. 1989. “Is Public Expenditure Aker, Jenny C. 2008. “Does Digital Divide or Productive?” Journal of Monetary Economics Provide? The Impact of Cell Phones on Grain 23 (2): 177–200. Markets in Niger.” Working Paper 177, Asenso-Okyere, Kwadwo, and Daniel Ayalew Bureau for Research and Economic Analysis of Mekonnen. 2012. “The Importance of ICTs Development (BREAD), London, UK. in the Provision of Information for Improving Aker, Jenny C. 2010. “Information from Markets Agricultural Productivity and Rural Incomes Near and Far: Mobile Phones and Agricultural in Africa.” Working Paper 2012-015, United Markets in Niger.” American Economic Nations Development Programme, Regional Journal: Applied Economics 2 (3): 46–59. Bureau for Africa, New York. Aker, Jenny C., and Marcel Fafchamps. 2015. Attanasio, Orazio, Britta Augsburg, Ralph De “Mobile Phone Coverage and Producer Haas, Emla Fitzsimons, and Heike Harmgart. Markets: Evidence from West Africa.” World 2015. “The Impacts of Microfinance: Evidence Bank Economic Review 29 (2): 262–92. from Joint-Liability Lending in Mongolia.” Alby, Philippe, Jean-Jacques Dethier, and American Economic Journal: Applied Eco­ Stéphane Straub. 2013. “Firms Operating nomics 7 (1): 90–122. 86  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Aw, Bee Yan, Mark J. Roberts, and Daniel Yi Beck, Thorsten, Asli Demirgüç-Kunt, and Ross Xu. 2011. “R&D Investment, Exporting, and Levine. 2009. “Financial Institutions and Productivity Dynamics.” American Economic Markets across Countries and over Time: Data Review 101 (4): 1312–44. and Analysis.” Policy Research Working Paper Ayogu, Melvin D. 1999. “Before Prebendalism: 4943, World Bank, Washington, DC. A Positive Analysis of Core Infrastructure Beck, Thorsten, Asli Demirgüç-Kunt, and Investment in a Developing Fiscal Federalism.” Vladimir Maksimovic. 2005. “Financial and African Development Review 11 (2): 169–98. Legal Constraints to Firm Growth: Does Ayogu, Melvin D. 2007. “Infrastructure and Firm Size Matter?” Journal of Finance 60 (1): Economic Development in Africa: A Review.” 137–77. Journal of African Economies 16 (Suppl. 1): Beck, Thorsten, Ross Levine, and Norman 75–126. Loayza. 2000. “Finance and the Sources of Babcock, Bruce A. 2015. “Using Cumulative Growth.” Journal of Financial Economics Prospect Theory to Explain Anomalous 58 (1–2): 261–300. Crop Insurance Coverage Choice.” American Behar, A lberto, and Phil Manners. 2008. Journal of Agricultural Economics 97 (5): “Logistics and Exports.” Working Paper 2008- 1371–84. 13, Centre for the Study of African Economies Bachas, Pierre, Roberto N. Fattal-Jaef, and (CSAE), Oxford University. A nders Jensen. 2018. “Size-Dependent Bento, Pedro, and Diego Restuccia. 2017. Tax Enforcement and Compliance: Global “Misallocation, Establishment Size, and Evidence and Aggregate Implications.” Policy Productivity.” American Economic Journal: Research Working Paper 8363, World Bank, Macroeconomics 9 (3): 267–303. Washington, DC. Bernard, Andrew B., Stephen J. Redding, and Baldwin, Richard E., and Toshihiro Okubo. Peter K. Schott. 2010. “Multiple-Product 2006. “Heterogeneous Firms, Agglomeration Firms and Product Switching.” American and Economic Geography: Spatial Selection Economic Review 100 (1): 70–97. and Sorting.” Journal of Economic Geography Bharadwaj, Prashant, William Jack, and Tavneet 6 (3): 323–46. Suri. 2019. “Fintech and Household Resilience Banerjee, Abhijit V., and Esther Duflo. 2005. to Shocks: Evidence from Digital Loans in “Grow th T heor y th roug h the L ens of Kenya.” Working Paper 25604, National Development Economics.” In Handbook of Bureau of Economic Research, Cambridge, Economic Growth, Vol. 1A, edited by Philippe MA. Aghion and Steven N. Durlauf, 473–552. Bigsten, Arne, Mulu Gebreeyesus, and Måns Amsterdam: Elsevier. S ö d e rb o m . 2 016 . “ Ta r i f f s a n d F i r m Banerjee, Abhijit V., Emily Breza, Esther Duflo, Performance in Ethiopia.” Journal of Develop­ Rachel Glennerster, and Cynthia G. Kinnan. ment Studies 52 (7): 986–1001. 2014. “Does Microfinance Foster Business Bloom, Nicholas, Benn Eifert, Aprajit Mahajan, Growth? The Importance of Entrepreneurial David McKenzie, and John Roberts. 2013. Heterogeneity.” Unpublished manuscript, “Does Management Matter? Evidence from Massachusetts Institute of Technology, India.” Quarterly Journal of Economics Cambridge, MA. 128 (1): 1–51. Banerjee, Abhijit V., and Andrew F. Newman. Boopen, Seetanah. 2006. “Transport Infra­ 1993. “Occupational Choice and the Process of structure and Economic Growth: Evidence Development.” Journal of Political Economy from Africa Using Dynamic Panel Estimates.” 101 (2): 274–98. Empirical Economics Letters 5 (1): 37–52. Banerjee, Abhijit V., and Benjamin Moll. 2010. Bresnahan, Lauren, Ian Coxhead, Jeremy Foltz, “Why Does Misallocation Persist?” American and Tewodaj Mogues. 2016. “Does Freer Economic Journal: Macroeconomics 2 (1): Trade Really Lead to Productivity Growth? 189–206. Evidence from Africa.” World Development Beck, Thorsten, Asli Demirgüç-Kunt, and Ross 86: 18–29. Levine. 2007. “Finance, Inequality and Brockmeyer, Anne, and Marco Hernandez. 2016. the Poor.” Journal of Economic Growth “Taxation, Information, and Withholding: 12: 27–49. Evidence from Costa Rica.” Policy Research P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   87 Working Paper 7600, World Bank, Washington, Intermediated Trade: Regression Discontinuity DC. Evidence from Sierra Leone.” Unpublished Buera, Francisco J. 2008. “Persistency of Poverty, manuscript, Harvard University, Cambridge, Financial Frictions, and Entrepreneurship.” MA. Unpublished manuscript, Northwestern C he n , C h aor a n . 2 017. “ U nt it le d L a nd , University, Evanston, IL. Occupational Choice, and Agricultural Buera, Francisco J., Joseph P. Kaboski, and Productivity.” American Economic Journal: Yo n g s e o k S h i n . 2 011. “ F i n a n c e a n d Macroeconomics 9 (4): 91–121. Development: A Tale of Two Sectors.” Chen, Chaoran. 2020. “Technology Adoption, A m e r i c a n E c o n o m i c R e v i e w 101 (5): C apit a l D e ep en i ng , a nd I nter nat iona l 1964–2002. P roduc t iv it y Di f ferenc e s.” Jo u r n al of Buera, Francisco J., Joseph P. Kaboski, and Development Economics 143: 10238. Yongseok Shin. 2015. “Entrepreneurship and Chen, Chaoran, and Diego Restuccia. 2018. Financial Frictions: A Macrodevelopment “Agricultural Productivity Growth in Africa.” Perspective.” Annual Review of Economics Background paper, World Bank, Washington, 7 (1): 409–36. DC. Buera, Francisco J., Benjamin Moll, and Yongseok Chen, Chaoran, Diego Restuccia, and Raül Shin. 2013. “Well-Intended Policies.” Review Santaeulàlia-Llopis. 2017. “The Effects of Economic Dynamics 16 (1): 216–30. of Land Markets on Resource Allocation Buera, Francisco J., and Yongseok Shin. 2013. and Agricultural Productivity.” Working “Financial Frictions and the Persistence of Paper 24034, National Bureau of Economic History: A Quantitative Exploration.” Journal Research, Cambridge, MA. of Political Economy 121 (2): 221–72. Chen, Shuai, Xiaoguang Chen, and Jintao Bustos, Paula. 2011. “Trade Liberalization, Xu. 2016. “Impacts of Climate Change on Exports, and Technology Upgrading: Evidence Agriculture: Evidence from China.” Journal of on the Impact of MERCOSUR on Argentinian Environmental Economics and Management Firms.” American Economic Review 101 (1): 76: 105–24. 304–40. Cirera, Xavier, Roberto Fattal-Jaef, Nicolas P. P. Calderón, César, Catalina Cantú, and Punam Gonne, and Hibret Maemir. 2018. “Elusive C hu h a n - Po l e . 2 018 . “ I n f r a s t r u c t u r e Missing Middle: The Size Distribution of Development in Sub-Saharan Africa: A Firms in Sub-Saharan Africa.” Unpublished Scorecard.” Policy Research Working Paper manuscript, World Bank, Washington, DC. 8425, World Bank, Washington, DC. Ci rera, X avier, Rober to Fat tal-Jaef, and Calderón, César, and Luis Servén. 2004. “The Hibret Maemir. 2018. “Taxing the Good? Effects of Infrastructure Development on Distortions, Misallocation, and Productivity in Growth and Income Distribution.” Policy Sub-Saharan Africa.” World Bank Economic Research Working Paper 270, World Bank, Review 34 (1): 75–100. Washington, DC. Coldham, Simon. 2000. “Land Reform and Calderón, César, and Luis Ser vén. 2010. Customary Rights: The Case of Uganda.” “Infrastructure and Economic Development Journal of African Law 44 (1): 65–77. in Sub-Saharan Africa.” Journal of African Collard-Wexler, Allan, and Jan De Loecker. 2015. Economies 19 (Suppl. 1): i13–i87. “Reallocation and Technology: Evidence from Carleton, Tamma A., and Solomon M. Hsiang. the US Steel Industry.” American Economic 2016. “Social and Economic Impacts of Review 105(1): 131–71. Climate.” Science 353 (6304): aad9837. Cook, Tamara, and Claudia McKay. 2015. Carter, Michael R., Rachid Laajaj, and Dean “How M-Shwari Works: The Story So Far.” Yang. 2014. “Subsidies and the Persistence Access to Finance Forum Report No. 10, of Technology Adoption: Field Experimental Consultative Group to Assist the Poor (CGAP), Evidence from Mozambique.” Working Washington, DC. Paper 20465, National Bureau of Economic De Ferranti, David, Guillermo E. Perry, Francisco Research, Cambridge, MA. H. G. Ferreira, and Michael Walton. 2004. Casaburi, Lorenzo, Rachel Glennerster, and Inequality in Latin America: Breaking with Tavneet Suri. 2013. “Rural Roads and History? Washington, DC: World Bank. 88  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Deichmann, Uwe, Aparajita Goyal, and Deepak Elbadawi, Ibrahim, Taye Mengistae, and Albert Mishra. 2016. “Will Digital Technologies G. Zeufack. 2006. “Market Access, Supplier Tra n sfor m A g r ic u lt u re i n D evelopi ng Access, and Africa’s Manufactured Exports: A Countries?” Agricultural Economics 47 (S1): Firm Level Analysis.” Journal of International 67–100. Trade & Economic Development 15  (4): Deininger, Klaus, Daniel Ayalew Ali, and Tekie 493–523. Alemu. 2008. “Assessing the Functioning of Escribano, Alvaro, J. Luis Guasch, and Jorge Pena. Land Rental Markets in Ethiopia. Economic 2010. “Assessing the Impact of Infrastructure Development and Cultural Change 57 (1): Quality on Firm Productivity in Africa Cross- 21–33. Country Comparisons Based on Investment: De Loecker, Jan. 2013. “Detecting Learning by Climate Surveys from 1999 to 2005.” Policy Exporting.” American Economic Journal: Research Working Paper 5191, World Bank, Microeconomics 5 (3): 1–21. Washington, DC. De Loecker, Jan, and Pinelopi Koujianou Estache, Antonio. 2005. “What Do We Know Goldberg. 2014. “Firm Performance in a about Sub-Saharan Africa’s Infrastructure and Global Market.” Annual Review of Economics the Impact of Its 1990s Reforms?” Working 6 (1): 201–27. paper, World Bank, Washington, DC. Demirgüç-Kunt, Asli, Erik Feyen, and Ross Estache, Antonio, and Maria Vagliasindi. 2007. Levine. 2013. “The Evolving Importance of “Infrastructure for Accelerated Growth for Banks and Securities Markets.” World Bank Ghana: Needs and Challenges.” Background Economic Review 27 (3): 476–90. paper for “Ghana: Meeting the Challenge Demirgüç-Kunt, Asli, Inessa Love, and Vladimir of Accelerated and Shared Growth.” Ghana Maksimovic. 2006. “Business Environment Country Economic Memorandum, Report No. and the Incorporation Decision.” Journal of 40934-GH, World Bank, Washington, DC. Banking and Finance 30 (11): 2967–93. Fafchamps, Marcel, David McKenzie, Simon R. Demombynes, Gabriel, and Aaron Thegeya. Quinn, and Christopher Woodruff. 2011. 2012 . “Kenya’s Mobile Revolution and “When Is Capital Enough to Get Female the Promise of Mobile Savings.” Policy Microenterprises Growing? Evidence from a Research Working Paper 5988, World Bank, Randomized Experiment in Ghana.” Working Washington, DC. Paper 17207, National Bureau of Economic D’Erasmo, Pablo, and Hernan Moscoso Boedo. Research, Cambridge, MA. 2012. “Financial Structure, Informality Fafchamps, Marcel, and Bart Minten. 2012. and Development.” Journal of Monetary “Impact of SMS-Based Agricultural Infor­ Economics 59 (3): 286–302. mation on Indian Farmers.” World Bank Diao, X i nshen , a nd Yu k it sug u Ya nom a. Economic Review 26 (3): 383–414. 2003. “Exploring Regional Dynamics in Fajnzylber, Pablo. 2007. “Informality, Productivity, Sub-Saharan African Agriculture.” Discussion and the Firm.” In Informality: Exit and Paper No. 2 , Development Strategy and Exclusion , edited by Guillermo E. Perry, Governance Division, International Food William F. Maloney, Omar S. Arias, Pablo Policy Research Institute, Washington, DC. Fajnzylber, Andrew D. Mason, and Jaime Dinkelman, Taryn. 2011. “The Effects of Rural Saavedra-Chanduvi, 157–78. Washington, Electrification on Employment: New Evidence DC: World Bank. from South Africa.” American Economic Feenstra, Robert C., and David E. Weinstein. Review 101 (7): 3078–3108. 2017. “Globalization, Markups, and the U.S. Eberts, Randall W., and Daniel P. McMillen. Price Level.” Journal of Political Economy 1999. “Agglomeration Economies and Urban 125 (4): 1040–74. Public Infrastructure.” In Handbook of Field, Erica, and Rohini Pande. 2008. “Repayment Regional and Urban Economics, Volume 3: Frequency and Default in Microfinance: Applied Urban Economics , edited by Paul Evidence from India.” Journal of the European C. Cheshire and Edwin S. Mills, 1455–95. Economic Association 6 (2–3): 501–09. Amsterdam: Elsevier Science B.V. Field, Erica, Rohini Pande, John Papp, and Natalia E kek we , Ndubu i si. 2017. “How Dig it a l Rigol. 2013. “Does the Classic Microfinance Technology Is Changing Farming in Africa.” Model Discourage Entrepreneurship among Harvard Business Review, May 18, 2017. the Poor? Experimental Evidence from P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   89 India.” American Economic Review 103 (6): Growth: Evidence from India.” Quarterly 2196–2226. Journal of Economics 125 (4): 1727–67. Francis, Eilin, Joshua Blumenstock, and Jonathan Gollin, Douglas. 2006. “Do Taxes on Large Firms Robinson. 2017. “Digital Credit: A Snapshot Impede Growth? Evidence from Ghana.” of the Current Landscape and Open Research Bulletins No. 7488, Center for Economic Questions.” Working Paper No. 516, BREAD, Development, University of Minnesota, Duluth. London. Gollin, Douglas, and Richard Rogerson. 2014. Fujita, Masahisa, and Jacques-François Thisse. “Productivity, Transport Costs and Subsistence 2002. Economics of Agglomeration: Cities, Ag riculture.” Jour n al of De velopme nt Industrial Location, and Regional Growth. Economics 107: 38–48. Cambridge: Cambridge University Press. G ollin, Douglas, and R ichard Rogerson. Gabler, Alain, and Markus Poschke. 2013. 2016. “Agriculture, Roads, and Economic “Experimentation by Firms, Distortions, and D evelopment i n Uga nd a.” I n A f r i c a n Aggregate Productivity.” Review of Economic Successes, Volume IV: Sustainable Growth, Dynamics 16 (1): 26–38. edited by Sebastian Edwards, Simon Johnson, Gandhi, Rikin, Rajesh Veeraraghavan, Kentaro and David N. Weil, 69 –110. Chicago: Toyama, and Vanaja Ramprasad. 2009. University of Chicago Press for the National “Digital Green: Participatory Video and Bureau of Economic Research. Med iated I nst r uc tion for Ag ricu lt u ral G oya l , Apa rajit a. 2018. “Subsid ie s a nd Extension.” Information Technologies & Productivity.” Background note for the International Development 5 (1): 1–15. Boosting Productivity in Sub-Saharan Africa Garcia-Milà, Teresa, and Therese J. McGuire. project, World Bank, Washington, DC. 1992. “The Contribution of Publicly Provided Goyal, Aparajita, and John Nash. 2017. Reaping Inputs to States’ Economies.” Regional Science Richer Returns: Public Spending Priorities and Urban Economics 22 (2): 229–41. fo r A f r i c a n A g r i c u lt u re Pro d u c t i v i t y Gautam, Bastian, Iacopo Bianchi, Markus Growth. Africa Development Forum Series. Goldstein, and Joao Montalvao. 2018. “Short- Washington, DC: World Bank. Term Impacts of Improved Access to Mobile Graham, Daniel J. 2007. “Agglomeration, Savings, with and without Business Training: Productivity and Transport Investment.” Experimental Evidence from Tanzania.” Journal of Transport Economics and Policy Working Paper 478, Center for Global 41 (3): 317–43. Development, Washington, DC. Greaney, Brian, Joseph P. Kaboski, and Eva Gebresilasse, Mesay M. 2016. “Industrial Van Leemput. 2013. “Can Self-Help Groups Policy and Misallocation in the Ethiopian Really Be ‘Self-Help’?” Working Paper 18970, Manufacturing Sector.” Unpublished manu- National Bureau of Economic Research, script, Boston University. Cambridge, MA. Gebrewolde, Tewodros Makonnen, and James Greenwood, Jeremy, Juan M. Sanchez, and Rockey. 2015. “The Global Gender Gap Cheng Wang. 2010. “Financing Development: in Labor Income.” Discussion Papers in The Role of Information Costs.” American Economics 17/14, Division of Economics, Economic Review 100 (4): 1875–91. School of Business, University of Leicester. Greenwood, Jeremy, Juan M. Sanchez, and Giné, Xavier, and Robert Townsend. 2004. Cheng Wang. 2013. “Quantifying the Impact “Evaluation of Financial Liberalization: A of Financial Development on Economic General Equilibrium Model with Constrained D e ve lop m e nt .” R e v i e w o f E c o n o m i c Occupation Choice.” Journal of Development Dynamics 16 (1): 194–215. Economics 74 (2): 269–307. Guner, Nezih, Gustavo Ventura, and Yi Xu. Goldberg, Pinelopi K., Amit Khandelwal, Nina 2008. “Macroeconomic Implications of Pavcnik, and Petia Topalova. 2009. “Trade Size-Dependent Policies.” Review of Economic Liberalization and New Imported Inputs.” Dynamics 11 (4): 721–44. American Economic Review 99: 494–500. Halpern, László, Miklós Koren, and Adam Szeidl. Goldberg, Pinelopi K., Amit Khandelwal, Nina 2015. “Imported Inputs and Productivity.” Pavcnik, and Petia Topalova. 2010. “Imported A me ric an Economic Re vie w 105 (12): Intermediate Inputs and Domestic Product 3660–3703. 90   Boosting Productivity in Sub-Saharan Africa Hjort, Jonas, and Jonas Poulsen. 2019. “The in Sub-Saharan Africa.” Unpublished manu- Arrival of Fast Internet and Employment in script, University of Cape Town, South Africa. Africa.” American Economic Review 109 (3): Kalemli-Ozcan, Sebnem, and Bent E. Sørensen. 1032–79. 2016. “Misallocation, Property Rights, and Holloway, Kyle, Zarah Niazi, and Rebecca Rouse. Access to Finance: Evidence from within and 2017. “Women’s Economic Empowerment across Africa.” In African Successes, Volume Through Financial Inclusion: A Review of III: Modernization and Development, edited Existing Evidence and Remaining Knowledge by Sebastian Edwards, Simon Johnson, and Gaps.” Paper, Financial Inclusion Program, David N. Weil, 183–211. Chicago: University Innovations for Poverty Action, New Haven, of Chicago Press for the National Bureau for CT. Economic Research. Hulten, Charles R., and Robert M. Schwab. 1991. Karippacheril, Tina George, Luz Diaz Rios, “Public Capital Formation and the Growth of and Lara Srivastava. 2011. “Module 12: Regional Manufacturing Industries.” National Improving Food Safety and Traceability.” Tax Journal 44 (4): 121–34. I n e-sourcebook, “IC T in Agriculture: IFC (International Finance Corporation). 2018. Connecting Smallholders to Knowledge, Working with Smallholders: A Handbook for Networks, and Institutions.” Report No. Firms Building Sustainable Supply Chains. 64605, World Bank, Washington, DC. Washington, DC: World Bank. Karlan, Dean, and Jonathan Zinman. 2010. IMF (International Monetary Fund). 2014. World “Expanding Credit Access: Using Randomized Economic Outlook, October. Washington, Supply Decisions to Estimate the Impacts.” DC: IMF. Review of Financial Studies 23 (1): 433–64. IMF (International Monetary Fund). 2017. Fiscal Kherallah, Mylene, Christopher L. Delgado, Monitor, April 2017: Achieving More with Eleni Zaude Gabre-Madhin, Nicholas Minot, Less. Washington, DC: IMF. and Michael Johnson. 2002. Reforming Agri­ IPCC (Intergovernmental Panel on Climate cultural Markets in Africa: Achievements and Change). 2015. Clim ate C hange 2014: Challenges. Washington, DC: International Synthesis Report. Contribution of Working Food Policy Research Institute. Groups I, II and III to the Fifth Assessment Kishindo, Paul. 2011. “The Village Head and Report of the Intergovernmental Panel on the Problem of Role Relevance in the Context Climate Change. Geneva: IPCC. of Declining Rural Land Availability in Jack, William, and Tavneet Suri. 2014. “Risk Malawi.” Working Paper No. 2, Centre for Sharing and Transactions Costs: Evidence from Social Research, Chancellor College, Zomba, Kenya’s Mobile Money Revolution.” American Malawi. Economic Review 104 (1): 183–223. Kuralatne, Chandi. 2006. “Social and Economic Jensen, Robert. 2007. “The Digital Provide: Infrastructure Impacts on Economic Growth I n f o r m a t i o n ( Te c h n o l o g y ) , M a r k e t in South Africa.” Paper presented at the Performance, and Welfare in the South Development Policy Research Unit (DPRU) Indian Fisheries Sector.” Quarterly Journal of Conference, Johannesburg, South Africa, Economics 122 (3): 879–924. October 18–20. Jeong, Hyeok, and Robert M. Townsend. 2007. Ky, Serge, Clovis Rugemintwari, and Alain “Sources of TFP Growth: Occupational Sauviat. 2021. “Friends or Foes? Mobile Money Choice and Financial Deepening.” Economic Interaction with Formal and Informal Finance.” Theory 32 (1): 179–221. Telecommunications Policy 45 (1): 102057. Kaboski, Joseph P., and Robert M. Townsend. Lane, Philip R., and Gian-Maria Milesi-Ferretti. 2011. “A Structural Evaluation of a Large- 2007. “The External Wealth of Nations Mark Scale Quasi-Experimental Microfinance II: Revised and Extended Estimates of Foreign Initiative.” Econometrica 79 (5): 1357–1406. Assets and Liabilities, 1970–2004.” Journal of Kaboski, Joseph P., and Robert M. Townsend. International Economics 73 (2): 223–50. 2012. “The Impact of Credit on Village Lane, Philip R., and Gian-Maria Milesi-Ferretti. Economies.” American Economic Journal: 2017. “International Financial Integration Applied Economics 4 (2): 98–133. in the Aftermath of the Global Financial Kamara, Ibrahim B. 2006. “Economic Growth Crisis.” Working Paper 17/115, International and Government Infrastructure Expenditure Monetary Fund, Washington, DC. P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A l l o c a t i o n i n ­S u b - S a h a r a n A f r i c a    91 Lagakos, David, and M ichael E . Waugh. Melitz, Marc J., and Stephen J. Redding. 2013. “Selection, Agriculture, and Cross- 2014. “Heterogenous Firms and Trade.” Country Productivity Differences.” American In Handbook of International Economics, Economic Review 103 (2): 948–80. Vol. 4, edited by Gita Gopinath, Elhanan Leal Ordóñez, Julio C. 2014. “Tax Collection, the Helpman, and Ken net h Rogof f, 1–54. Informal Sector, and Productivity.” Review of Amsterdam: Elsevier. Economic Dynamics 17 (2): 262–86. Melitz, Marc J., and Daniel Trefler. 2012. “Gains Levine, Ross. 2005. “Finance and Growth: from Trade When Firms Matter.” Journal of Theory and Evidence.” In Handbook of Economic Perspectives 26 (2): 91–118. Economic Growth, Vol. 1A , edited by Mengistae, Taye A., and Francis Teal. 1998. “Trade Philippe Aghion and Steven Durlauf, 865–934. Liberalization, Regional Integration and Amsterdam: Elsevier. Firm Performance in Africa’s Manufacturing Lileeva, Alla, and Daniel Trefler. 2010. “Improved Sector: A Report to European Commission.” Access to Foreign Markets Raises Plant-Level REP98-1, Centre for the Study of African Productivity . . . for Some Plants.” Quarterly Economies, University of Oxford. Journal of Economics 125 (3): 1051–99. Mhlanga, Samuel V., and Neil A. Rankin. Limao, Nuno, and Anthony J. Venables. 2001. 2015. “Does Technical Efficiency Dominate “Infrastructure, Geographical Disadvantage, R e s o u r c e R e a l l o c at i o n i n A g g r e g at e Transport Costs, and Trade.” World Bank Productivity Growth? Firm-Level Evidence Economic Review 15 (3): 451–79. from the Manufacturing Sector in Swaziland.” Lio, Monchi, and Meng-Chun Liu. 2006. “ICT Unpublished manuscript, Department of and Agricultural Productivity: Evidence from E conomics, Universit y of Stellenbosch, Cross-Country Data.” Agricultural Economics South Africa. 34 (3): 221–28. Midrigan, Virgiliu, and Daniel Y. Xu. 2014. Lucas, Robert E ., Jr. 1978. “On the Size “Finance and Misallocation: Evidence from Distribution of Business Firms.” Bell Journal Plant-Level Data.” American Economic of Economics 9 (2): 508–23. Review 104 (2): 422–58. Lumbila, K. N. 2005. “What Makes DFI Work? Moll, Benjamin. 2014. “Productivity Losses from A Panel Analysis of the Growth Effects of FDI Financial Frictions: Can Self-Financing Undo in Africa.” African Region Working Paper Capital Misallocation?” American Economic Series, No. 80, World Bank, Washington, DC. Review 104 (10): 3186–3221. Mayer, Thierry, Marc J. Melitz, and Gianmarco I. Morris, Brian. 2016. An Environmental History P. Ottaviano. 2014. “Market Size, Competition, of Southern Malawi: Land and People of the and the Product Mix of Exporters.” American Shire Highlands. Cham, Switzerland: Palgrave Economic Review 104 (2): 495–536. Macmillan. Mbit i , I saac , a nd Dav id N . Wei l. 2016. Morris, Michael, Valerie A. Kelly, Ron J. Kopicki, “Mobile Banking: The Impact of M-Pesa in and Derek Byerlee. 2007. Fertilizer Use in Kenya.” In African Successes, Volume III: African Agriculture: Lessons Learned and Good Modernization and Development , edited Practice Guidelines. Directions in Development by Sebastian Edwards, Simon Johnson, and Series. Washington, DC: World Bank. David N. Weil, 247–93. Chicago: University Munnell, Alicia H. 1992. “Policy Watch: of Chicago Press. Infrastructure Investment and Economic McKenzie, David J. 2017. “Identifying and Growth.” Journal of Economic Perspectives Spurring High-Growth Entrepreneurship: 6 (4): 189–98. Experimental Evidence from a Business Plan Muto, Megumi, and Takashi Yamano. 2009. “The Competition.” American Economic Review Impact of Mobile Phone Coverage Expansion 107 (8): 2278–2307. on Market Participation: Panel Data Evidence McKenzie, David J., and Christopher Woodruff. from Uganda.” World Development 37 (12): 2008. “Experimental Evidence on Returns to 1887–96. Capital and Access to Finance in Mexico.” Nakasone, Eduardo, Maximo Torero, and Bart World Bank Economic Review 22 (3): 457–82. Minten. 2014. “The Power of Information: The Melitz, Marc J., and Gianmarco I. P. Ottaviano. ICT Revolution in Agricultural Development.” 2008. “Market Size, Trade, and Productivity.” Annual Review of Resource Economics 6 (1): Review of Economic Studies 75 (1): 295–316. 533–50. 92  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Ndulu, Benno J. 2006. “Infrastructure, Regional Ranasinghe, Ashantha. 2014. “Impact of Policy Integration and Growth in Sub-Saharan Distor tions on Firm-L evel I nnovation, Africa: Dealing with the Disadvantages of Productivity Dynamics and TFP.” Journal Geography and Sovereign Fragmentation.” of Economic Dynamics & Control 46 (3): Journal of African Economies 15 (suppl 2): 114–29. 212–44. Reinikka, Ritva, and Jakkob Svensson. 1999. Nguimkeu, Pierre E . 2015. “An Estimated “How I nadequate Prov ision of P ubl ic Model of Informality with Constrained Infrastructure and Services Affects Private Entrepreneurship.” Unpublished manuscript, I nvestment.” Policy Research Working Department of Economics, Georgia State Paper 2262, World Bank, Washington, DC, University, Atlanta. December. Nyshadam, Anant. 2014. “Learning about Reinikka, Ritva, and Jakob Svensson. 2002. Comparative Advantage in Entrepreneurship: “Coping with Poor Public Capital.” Journal of Evidence from Thailand.” Unpublished man- Development Economics 69 (1): 51–69. uscript, Department of Economics, University Renkow, Mitch, Daniel G. Hallstrom, and Daniel of Southern California, Los Angeles. D. Karanja. 2004. “Rural Infrastructure, Oseni, Musiliu O. 2019. “Costs of Unreliable Transactions Costs and Market Participation Electricity to African Firms.” Brief, Energy in Kenya.” Journal of Development Economics Hub for Grow th, Nigeria n Elec t ricit y 73 (1): 349–67. Regulatory Commission. Restuccia, Diego. 2016. “Resource Allocation Paulson, Anna L., and Robert M. Townsend. and Productivity in Agriculture.” Unpublished 20 05. “Fi nancial Const rai nts and manuscript, University of Toronto. Entrepreneurship: Evidence from the Thai Restuccia, Diego, and Richard Rogerson. 2017. Financial Crisis.” Federal Reserve Bank of “The Causes and Costs of Misallocation.” Chicago, Economic Perspectives 3Q/2005: Journal of Economic Perspectives 31 (3): 34–48. 151–74. Pavcnik, Nina. 2002. “Trade Liberalization, Exit, Restuccia, Diego, and Raül Santaeulàlia-Llopis. and Productivity Improvements: Evidence from 2017. “Land Misallocation and Productivity.” Chilean Plants.” Review of Economic Studies Working Paper 23128, National Bureau of 69 (1): 245–76. Economic Research, Cambridge, MA. Pawasutipaisit, Anan, and Robert M. Townsend. Ricker-Gilbert, Jacob, Nicole M. Mason, Francis 2011. “Wealth Accumulation and ­ Factors A. Darko, and Solomon T. Tembo. 2013. A c c ou nt i ng for S u c c e s s .” Jo u r n a l of “What Are the Effects of Input Subsidy Econometrics 161 (1): 56-81. Programs on Maize Prices? Evidence from Perkins, Peter, Johann Fedderke, and John Luiz. Malawi and Zambia.” Agricultural Economics 2005. “An Analysis of Economic Infrastructure 44 (6): 671–86. Investment in South Africa.” South African Ricker-Gilbert, Jacob, and Thomas S. Jayne. Journal of Economics 73 (2): 211–28. 2012. “Do Fertilizer Subsidies Boost Staple Pitt, Mark M., and Shahidur R. Khandker. 1998. Crop Production and Reduce Poverty across “The Impact of Group-Based Credit Programs the Distribution of Smallholders in Africa? on Poor Households in Bangladesh: Does the Quantile Regression Results from Malawi.” Gender of Participants Matter?” Journal of Selected Paper for the Triennial Meeting of Political Economy 106 (5): 958–96. the International Association of Agricultural Place, Frank, and Keijiro Otsuka. 2002. “Land Economists, Foz Do Iguaçu, Brazil, August Tenu re Systems and T heir I mpacts on 18–24. Agricultural Investments and Productivity in Schalkwyk, Francois van, Andrew Young, and Uganda.” Journal of Development Studies 38 Stefaan Verhulst. 2017. “Ghana. Esoko: (6): 105–28. Leveling the Information Playing Field for Poczter, Sharon. 2017. “You Can’t Count on Smallholder Farmers in Ghana.” Open Data’s Me: The Impact of Electricity Unreliability Impact case study, The Governance Lab, on Productivity.” Agricultural and Resource Tandon School of Engineering, New York Economics Review 46 (3): 579–602. University. Rajan, Raghuram G., and Luigi Zingales. 1998. Schmidt, Klaus M. 1997. “Managerial Incentives “Power in a Theory of the Firm.” Quarterly and Product Market Competition.” Review of Journal of Economics 113 (2): 387–432. Economic Studies 64 (2): 191–213. P o l i c i e s a n d I n s t i t u t i o n s t h a t D i s t o r t   R e s o u r c e A ll o ca t i o n i n ­S u b - Sa h a r a n A f r i ca   93 Shiferaw, Admasu, Måns Söderbom, Eyerusalem Topalova, Petia and Amit Khandelwal. 2011. Siba, and G etnet A lemu. 2015. “Road “Trade Liberalization and Firm Productivity: Infrastructure and Enterprise Dynamics in The Case of India.” Review of Economics and Ethiopia.” Journal of Development Studies Statistics 93 (3): 995–1009. 51 (11): 1541–58. Van Biesebroeck, Johannes. 2005. “Exporting Sinate, David, Vanlalruata Fanai, Snehal Bangera, R aises Produc tivit y i n Sub - Saha ra n and Sara Joy. 2018. “Connecting Africa: African Manufacturing Firms.” Journal of Role of Transport Infrastructure.” Working International Economics 67 (2): 373–91. Paper No. 72, Export-Import Bank of India, Weber, M. 1928. Theory of the Location of Mumbai. Industries. Chicago: University of Chicago. Stifel, David, and Bart Minten. 2008. “Isolation World Bank. 1994. World Development Report and Agricultural Productivity.” Agricultural 1994: Infrastructure for Development . Economics 39 (1): 1–15. New  York: Oxford University Press for the Suri, Tavneet, and William Jack. 2016. “The World Bank. Long-Run Poverty and Gender Impacts World Bank. 2005. World Development Report of Mobile Money.” Science 354 (6317): 2006: Equity and Development. Washington, 1288–92. DC: World Bank. Svensson, Jakob, and David Yanagizawa. 2009. World B a n k. 20 08. World De velop m e nt “Getting Prices Right: The Impact of the Report 2008: Agriculture for Development. Market Information Service in Uganda.” Washington, DC: World Bank. Journal of the European Economic Association World B a n k. 2012 . “A g r ic u lt u re S e c tor 7 (2–3): 435–45. Report.” In “Transformation-Ready: The Swistak, Arthur, L. Liu, and Ricardo Varsano. Strategic Application of Information and 2017. “Towa rd s More E f f icient Non- Communication Technologies in Africa.” Resource Taxation: A Strategy for Reform.” World Bank, Washington, DC. Technical Assistance Report, Fiscal Affairs Zenebe, Dawit. 2018. “Trade Liberalization, Department, International Monetary Fund, Productivit y, and Resou rce A llocation Washington, DC. in M anufac t u ring Fi rms in Ethiopia.” Tadesse, Getaw, and Godfrey Bahiigwa. 2015. Unpublished manuscript, Universit y of “Mobile Phones and Farmers’ Marketing Wisconsin–Madison. Decisions in Ethiopia.” World Development Zhang, Peng, Junjie Zhang, and Minpeng Chen. 68: 296–307. 2017. “Economic Impacts of Climate Change Tarozzi, Alessandro, Jaikishan Desai, and Kristin on Agriculture: The Importance of Additional Johnson. 2015. “The Impacts of Microcredit: Climatic Variables Other than Temperature Evidence from Ethiopia.” American Economic and Precipitation.” Journal of Environmental Journal: Applied Economics 7 (1): 54–89. Economics and Management 83: 8–31. Agenda for Future Research 5 The report has documented the low aggregate distinguish productivity shocks (or techni- labor productivity of Sub-Saharan Africa cal efficiency) from demand shocks in the relative to other world regions. The region’s measures of revenue productivity (TFPR) scarce resources, compounded by inefficien- among Sub-Saharan African production cies in their allocation, have exacerbated the establishments. This research requires the problem, as reflected by timely availability and recurrent produc- tion of high-quality data on output and • Cross-country differences in total fac- input prices at the establishment level. This tor productivity (TFP), which over- does not preclude improving the country whelmingly explain the cross-country coverage as well as the methodology and differences in income per worker at the periodicity of firm-level censuses. However, aggregate level; the quest for new and more data faces other • Marked delays in structural transfor- challenges: mation, as captured by the agriculture sector’s high employment share and • Output price data is more widely avail- low productivity; and able than input price data at the estab- • Pervasive misallocation of resources lishment level. across farms and firms, with deleteri- • The reported output prices are, in most ous consequences for aggregate output cases, unit values. and productivity. • Surveys should be undertaken at the product level if most of a specific sec- Although this report has focused on the tor’s manufacturing establishments are role played by misallocation in explaining multiproduct. the low productivity of Sub-Saharan Africa, several avenues of research could provide fur- Having greater data availability on output ther insights on the dynamics of productivity and input prices does not prevent the need to in the region as well as different channels of impose more structure in identifying the role policy transmission to boost productivity. played by demand shocks in the measured Impacts of productivity shocks ver- TFPR. Recent research using firm-level cen- sus demand shocks. Future work needs to suses with price data shows that there is still 95 96   Boosting Productivity in Sub-Saharan Africa a large dispersion of TFPR across manufac- Other internal drivers to explore include turing firms in Ethiopia, and this is mirrored greater input quality, product innovation by large differences in physical productivity and research and development (R&D) invest- (TFPQ). Prices tend to vary significantly less ments, and firm structure decisions. than productivity levels and do not constitute a major driving factor of TFPR differences (Söderbom 2018). Notes Policy impact at the firm level. Fur-   1. Cusolito and Maloney (2018) provide evi- ther analysis should be undertaken on the dence on the policy impacts on the “within” impact of policies on the within rather component of aggregate productivity growth than the between component of aggregate in emerging markets outside Africa—namely, Chile, Colombia, and Malaysia. (As chapter productivity growth using longitudinal 1 discusses further, the “within” component data.1 Trade liberalization, for instance, accounts for the productivity growth within may affect both components of aggregate firms. The “between” component reflects productivity. However, the elements that the role of factor reallocation across firms in may boost productivity at the firm level aggregate productivity growth.) (rather than at the industry level) have not   2. X-inefficiency is the divergence of a firm’s been adequately discussed—for example, observed behavior in practice (influenced by the reduction of X-inefficiencies, 2 invest- a lack of competitive pressure) from efficient ment in new technologies, quality upgrade behavior assumed or implied by economic of products and inputs, and locational theory. The concept, introduced by Leiben- decisions, among others. In Sub-Saharan stein (1966),  refers to the result of inputs not producing their maximum output as a conse- Africa, recent work has estimated these dif- quence of an “X” factor. This translates into ferent components or sources of productiv- failures of both cost minimization and pro- ity growth in manufacturing (Dennis et al. duction maximization and, hence, implies a 2016; Jones et al. 2019). loss of efficiency and refers to all nonalloca- Drivers of productivity improvements tive inefficiencies. from managerial practices. Finally, there is greater need to deepen research in the region on the internal drivers of productivity at the References establishment level in Sub-Saharan Africa. Bender, Stefan, Nicholas Bloom, David Card, A growing field of research focuses on the John Van Reenen, and Stefanie Wolter. 2018. productivity improvements of adopting bet- “Management Practices, Workforce Selection, ter managerial practices. For instance, firms and Productivity.” Journal of Labor Econom- with better management practices tend to ics 36 (S1): S371–S409. perform better along several dimensions: Bloom, Nicholas, Raffaella Sadun, and John Van Reenen. 2016. “Management as a Technol- they are larger and grow faster, they are more ogy?” Working Paper 22327, National Bureau productive, and they have higher survival of Economic Research, Cambridge, MA. rates (Bloom and Van Reenen 2007). And Bloom, Nicholas, and John Van Reenen. 2007. better-managed firms also recruit and retain “Measuring and Explaining Management workers with higher average human capital Practices across Firms and Countries.” (Bender et al. 2018). Quarterly Journal of Economics 122 (4): To those ends, more-flexible labor market 1351–1408. regulations are associated with better use of Bloom, Nicholas, and John Van Reenen. 2010. incentives by management (Bloom and Van “Why Do Management Practices Differ across Reenen 2010). Increased product compe- Firms and Countries?” Journal of Economic tition also tends to improve firm manage- Perspectives 24 (1): 203–24. Cusolito, Ana Paula, and William F. Maloney. ment, including through the reallocation of 2018. Productivity Revisited: Shifting Para- economic activity toward better-managed digms in Analysis and Policy. Washington, firms (Bloom, Sadun, and Van Reenen 2016). DC: World Bank. A g e n da f o r F u t u r e R e s e a r c h    97 Dennis, Allen, Taye Mengistae, Yutaka Yoshino, Growth in Africa.” Policy Research Working and Albert Zeufack. 2016. “Sources of Pro- Paper 8980, World Bank, Washington, DC. ductivity Growth in Uganda: The Role of Leibenstein, Harvey. 1966. “Allocative Efficiency Interindustry and Intraindustry Misallocation vs.  ‘X-Efficiency.’” American Economic in the 2000s.” Policy Research Working Paper Re­view 56 (3): 392–415. 7909, World Bank, Washington, DC. Söderbom, Måns. 2018. “Productivity Dispersion Jones, Patricia, Emmanuel K. K. Lartey, Taye and Firm Dynamics in Ethiopia’s Manufactur- Mengistae, and A lber t Z eufack. 2019. ing Sector.” Unpublished manuscript, Univer- “Sources of Manufacturing Productivity sity of Gothenburg, Sweden. Output per Worker, Factor Accumulation, and Total A Productivity This appendix describes the concepts, data, Trends. Second, the study tracks the evo- and methodologies used to generate the lution over time of output per worker, the ­ statistics reported in appendix B (“Country capital-output ratio, and the PWT index of Productivity Analysis”), which presents fac- human capital from 1960 to 2017 (or from tor accumulation, output, and productivity a later starting year according to data avail- for a wide array of Sub-Saharan countries ability across countries). All these series are from 1960 to 2017.1 Appendix B reports the expressed relative to the benchmark country evolution of output per worker, factor accu- that approximates the world technological mulation, and total factor productivity (TFP) frontier—that is, the United States. The time from three different dimensions: (a) latest series are used to conduct a basic develop- data on output, population, and sectoral ment accounting exercise: it computes the shares of employment and output; (b) trends share of factor accumulation and the share in labor productivity, capital-output ratios, of TFP that explain the output differences and human capital; and (c) growth decompo- between any Sub-Saharan African country sitions under different assumptions. and the global efficiency benchmark (the Latest data. First, each country reports the United States). latest figures on (a) output and population, Decompositions. Third, the results are and (b) sectoral shares in value added and computed for three different growth account employment. The data on output and popu- exercises—a traditional Solow decomposi- lation are collected from Penn World Table tion, a Solow decomposition that incorpo- (PWT) 9.0 (which contains annual informa- rates the accumulation of public and private tion from 1950 to 2014, at best), and these physical capital, and a Solow decomposition series were updated using PWT 9.1 data from that includes natural capital—for each coun- 2015 to 2017. Additionally, each country try in the Sub-Saharan Africa sample. reports the output and employment shares This report gathered data on output per across five sectors of economic activity in worker for a sample of 45 Sub-Saharan 2016: agriculture, manufacturing, nonmanu- African countries; however, the data avail- facturing industry, market services, and non- ability of the different inputs of produc- market services. tion (employment, physical capital, human 99 100  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca capital, and natural capital) was more lim- used. Finally, it is assumed that there are ited. Thirty-seven of the countries had data no adjustment costs in the accumulation of on gross domestic product (GDP), physical, capital and that there is perfect competition and human capital, while only 24 of the in the markets of production factors so that countries had data on natural capital (the their remuneration is equal to their social stock of all extractable resources such as marginal products. geology, soils, air, water, and living organ- In the spirit of Klenow and Rodríguez- isms). Finally, the report gathered data on Clare (1997), the production function in sectoral shares of output and employment for equation (A.1) is expressed in its intensive only 29 countries in the region. form: α Development Accounting  K  1−α yt = Zt  t  ht , (A.2) Development accounting exercises have been Y  t undertaken as early as the late 1960s, albeit for a limited number of countries (Denison where y is the real output per worker 1967; Walters 1968). Subsequent efforts K (y = Y/L), t is the capital-output ratio, and integrated Jorgenson’s growth accounting Yt framework with the work of structure pro- 1 posed by Griliches and Christensen 2 to com- Zt ≡ ( At )1−α is TFP measured in labor aug- pare the levels of output per worker between menting units. the United States and other high-income Equation (A.2) is at the core of the develop- countries (Christensen, Cummings, and ment accounting framework. It is compatible Jorgenson 1981; Jorgenson and Nishimizu with the steady state of a neoclassical growth 1978). K model where (a) the capital output ratio, t , More-recent applications of this frame- Yt work have calculated the sources of the large is proportional to the investment rate; and and persistent income differences observed (b) the level of human capital or T FP between the world’s richest and poorest has no direct effects on the steady-state countries (Hall and Jones 1999; Hsieh and capital-output ratio (Mankiw, Romer, and Klenow 2010; Jones 2016; Klenow and Weil 1992). When expressing the produc- Rodríguez-Clare 1997). tion function in per worker terms, changes in effective labor per worker or residual TFP are accompanied by changes in cap- Exposition of the Framework ital per worker (Hsieh and Klenow 2010; The development accounting framework Klenow and Rodríguez-Clare 1997). assumes that the relationship between output The development accounting framework and the factors of production is captured by uses equation (A.2) to decompose the dis- the following production function (Caselli tance of the different Sub-Saharan African 2005; Hall and Jones 1999): countries to the United States (the benchmark typically used in the literature to proxy the frontier of production possibilities) into two Yt = At Ktα ( hL )t 1−α , (A.1) distinct components—(a) the distance to the where Y is the country’s GDP in the period frontier in terms of physical and human cap- t, K is the aggregate capital stock, and hL ital (that is, factor accumulation); and (b) the is the “quality adjusted” labor force—that distance in terms of TFP—as described in is, the number of workers L multiplied by equation (A.3): their average human capital h. Furthermore, α a is the sensitivity of output with respect ytj  Ztj   κ tj  1−α  htj  to capital, and A represents the efficiency =  US ,(A.3) with which the factors of production are US yt  Zt κt  US    ht  US  O u t p u t p e r W o r k e r , Fac t o r A cc u m u la t i o n , a n d T o t al P r o d u c t i v i t y    101 ytj Sub-Saharan Africa and the United States. where is the output per worker of Sub- ytUS Given the differences in years of schooling Saharan African country j relative to that of for adults over 15 years old and differ- the United States in period t. This measure of ences in the returns to education, the rela- distance to the frontier can be decomposed tive human capital index for Sub-­ Saharan into (a) a composite factor that accounts for Africa is 0.42 in 1960–69 (column [3]). the differences in the stock α of physical and That is, human capital in the region is  κ tj  1−α  htj  about 42 percent of that in the United human capital,  US   US  ; and (b) States.  κ t   ht  Sub-Saharan Africa’s TFP relative to the •  a portion of the differences in output per United States is about 0.56 (column [4] worker that are attributed to the relative dis- of table A.1). This implies that productive tance in TFP,  Zt  . The comparison of j processes in the region are slightly more  Zt   US  than half as productive as those in the large and persistent cross-country differences United States. in productivity per worker may require a In other words, the table shows that real •  steady-state approximation. The larger expo- output per worker in the United States was nent on human capital and TFP (relative to about 18 times higher than that of Sub-Sa- the per worker expression of the production haran Africa (17.7) in 1960–69.3 A factor function) reflects the impact of these vari- of 8.8 of this difference is due to inputs, ables on output both directly and indirectly and a factor of 2 is due to TFP. This implies through capital per worker. that the distance to the frontier in terms of output per worker is 8.8 parts due to Development Accounting for inputs and 2 parts due to TFP. Hence, the Sub-Saharan Africa share due to TFP is 25 percent (column [5] of table A.1). Sub-Saharan A frica and the group of In 2010–17, real output per worker in •  low- and middle-income countries in non-­ Sub-Saharan Africa was about 8 per- African regions show diverging paths in cent that of the United States (0.083). real output per worker over time in spite of The relative gap in terms of the region’s quite similar initial conditions in the 1960s capital-output ratio has narrowed sig- (table A.1): nificantly; in fact, its relative capital-out- • T  he gap in output per worker in Sub- put ratio increased from 0.4 in 1960–69 Saharan Africa relative to the United States to 0.85 in 2010–17. In terms of human is 0.12 from 1960 to 1969; that is, the capital, the relative h index increased region’s labor productivity is about 12 per- only from 0.42 in 1960–69 to 0.47 in cent that of the United States (column [1] 2010–17. This is a small improvement in of table A.1). This gap is the product of the reducing the gap in human capital. The contribution of the gap in capital-output (implied) TFP differences between Sub- ratios, human capital, and TFP (columns Saharan Africa and the United States in [2], [3], and [4], respectively). 2010–17 are even larger than those in •  The capital-output ratio in Sub-Saharan 1960–69; that is, production processes Africa relative to that of the United States are not even one-tenth as efficient as is approximately 0.4, and the differ- those in the United States (0.06). In other ence in capital-output ratio that matters words, real output per worker in the for output per worker is about 0.51 in United States was about 23 times more 1960–69 (column [2] of table A.1). This productive than in Sub-Saharan Africa, implies that differences in physical capi- with more than 75 percent of the distance tal help explain about 44 percent of the to the frontier attributed to differences in differences in output per worker between TFP levels (0.781). 102  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca TABLE A.1  Development Accounting in Sub-Saharan Africa and in Non-African Developing Countries, Relative to the United States, 1960–2017 Contribution Output per Capital- Human Share due to worker [1] GDP [2] Capital [3] TFP [4] TFP [5] Sub-Saharan Africa 1960–69 0.118 0.511 0.416 0.554 0.251 (SSA) 1970–79 0.136 0.742 0.381 0.481 0.285 1980–89 0.103 1.070 0.382 0.253 0.483 1990–99 0.062 1.344 0.409 0.113 0.663 2000–09 0.067 1.407 0.442 0.108 0.758 2010–17 0.083 2.842 0.474 0.061 0.781 Developing Countries 1960–69 0.088 0.773 0.455 0.250 0.553 excluding SSA 1970–79 0.111 0.743 0.469 0.317 0.503 1980–89 0.110 0.899 0.503 0.243 0.577 1990–99 0.130 1.776 0.569 0.128 0.663 2000–09 0.149 1.244 0.616 0.194 0.703 2010–17 0.215 1.647 0.653 0.200 0.674 Source: Penn World Table (PWT) 9.0 and PWT9.1 updates (Feenstra, Inklaar, and Timmer 2015). Note: Output per worker, contribution of capital-GDP ratios, human capital, and total factor productivity (TFP) are expressed in terms relative to US efficiency benchmarks, following equation (A.3) in the text. Regional or group figures are employment-weighted averages. “Developing” countries are low- and middle-income countries according to World Bank country income classifications. The Diverging Path of Non-African In these non-African countries, relative Low- and Middle-Income Countries human capital grew much faster than in Sub-Saharan Africa: the relative h index of In spite of having similar starting labor non-African low- and middle-income coun- productivity levels in the 1960s, low- and tries increased from 0.45 in 1960 – 69 to middle-income countries in non-African 0.65 in 2010–17. The implied TFP differ- regions evolved differently from those in Sub- ence between such countries and the United Saharan Africa (table A.1). From 1960 to States is 0.2 in 2010–17 (down from 0.25 in 1969, the relative output per worker in Sub- 1960–69); that is, production processes are Saharan Africa was 12 percent that of the about one-fifth as efficient across these low- United States, while that of the other regions’ and middle-income countries as in the United low- and middle-income countries was about States. About half of output per worker dif- 9 percent (table A.1). The latter group had a ferences between these countries and the lower gap in capital-output and human cap- United States were attributed to TFP differ- ital relative to the United States during this ences in 1960–69. The efficiency narrative period. became more marked in 2010–17: differences Although relative labor productivity in in TFP levels now explain two-thirds of the Sub-Saharan Africa declined over the past output per worker gap (0.674). 50 years, it increased among the low- and Appendix B in this report shows the middle-income countries in other regions— trends of the different factors that make especially over the past two decades. Relative up the development accounting analysis output per worker of such countries increased for each Sub-Saharan African country: the from 0.09 in 1960–69 to 0.22 in 2010–17. In time series of the real output per worker, addition, their capital-output ratios caught the capital-output ratio, the human cap- up with those of the United States (increasing ital index, and TFP. All these series are from 0.61 in 1960–69 to 1.12 in 2010–17). expressed as a ratio of the corresponding O u t p u t p e r W o r k e r , Fac t o r A cc u m u la t i o n , a n d T o t al P r o d u c t i v i t y    103 series for the efficiency benchmark. (In prac- accumulation of physical capital and 10–30 tice, the United States benchmark is equal percent by human capital. Hence, differ- to 1.) Appendix B also presents the evolu- ences in TFP may account for 50–70 percent tion of the share of labor productivity differ- of country income differences (Hsieh and ences explained by factor accumulation and Klenow 2010). TFP as well as the TFP gaps for each Sub- This appendix conducts the growth Saharan African country by decade. accounting analysis under different techno- logical specifications. Hence, it computes the TFP growth using the traditional Solow Growth Accounting decomposition, a decomposition accounting Assessing the sources of economic growth for public and private capital accumulation, dates back to the late 1950s. Growth in real and a decomposition including natural cap- output was decomposed as the weighted ital. The appendix shows the estimation of average of the growth rate of labor and cap- TFP growth for Sub-Saharan African coun- ital as well as a residual labeled total factor tries using these three different specifications productivity (TFP) growth (Abramovitz provided that there is data availability. 1956; Solow 1957; Tinbergen 1942). The The technology of production function so-called Solow residual was nothing but is represented by a Cobb-Douglas produc- the unexplained part of economic growth tion function with constant returns to scale that was interpreted as a measure of tech- (Caselli 2005; Hall and Jones 1999): nological change. Subsequent contributions −α ) (A.4) Yt = At Ktα (hL)(1 , in the 1960s and 1970s led to the applica- t tion of more general production functions where Y is the country’s GDP, K is the aggre- and more accurate measurement of inputs gate capital stock, and hL is the “quality and outputs (Denison 1962; Denison, adjusted” labor force—that is, the number of Griliches, and Jorgenson 1972; Jorgenson workers L multiplied by their average human a nd G r i l i c h e s 19 6 7 ) — fo r i n s t a n c e , capital h. Furthermore, a is the (constant) accounting for changes in both the quan- sensitivity of output with respect to capital, tity and quality of labor and capital inputs and A represents the level of TFP or the effi- (Denison 1962).4 In spite of these adjust- ciency with which factors of production are ments, from 1947 to 1973, the estimated used or combined. In addition, it is assumed contribution of TFP to economic growth that there are no adjustment costs in capital was still about one-third of GDP growth accumulation and that there is perfect com- in the United States, 42 percent in Japan, petition in the markets of production factors, and more than half in several European so that they are paid their social marginal economies (Christensen, Cummings, and products. Jorgenson 1981). The differences in output per worker across the world’s countries—especially the Traditional Solow Decomposition large and protracted differences documented in the literature between high-income coun- The technology described in equation (A.4) tries and low- to middle-income countries— can be expressed in per worker terms: are overwhelmingly attributed to differences in TFP rather than to differences in the yt = At ktα ht(1−α ) , (A.5) levels of physical or human capital (see, among others, Caselli 2005; Hall and Jones where k is the capital labor ratio (k = K/L). 1999; Hsieh and Klenow 2010; Klenow and dx If we define xˆ t = t , then TFP growth is Rodríguez-Clare 1997). 5 The consensus in xt the literature points to 20 percent of country ˆ =y A ˆ (A.6) ˆ − (1 − α ) h ˆt − αk income differences being explained by the t t t 104   Boosting Productivity in Sub-Saharan Africa The definition and the construction of countries (Gupta et al. 2011). The estimated human capital are explained in the next output elasticities of private capital and pub- section. lic capital are summarized in columns [1] and [2] of table A.2. These estimated elastic- ities are used to compute the relative income Solow Decomposition Accounting for shares of private and public capital for Private and Public Capital Stock high-income countries, middle-income coun- The stock of capital of the economy is decom- tries, and low-income countries. The relative posed into the stocks of private and public income share of private capital is computed capital (denoted by the subindexes p and g, as ap /(ag + ap), as shown in the last column of respectively). The production function in table A.2. The income shares for private and equation (A.1) now becomes public capital (a p and a g, respectively) are then computed using the estimated relative Yt = At Kptp K gtg ( ht Lt ) income shares (which varies across groups) α α 1−α p −α g , (A.7) and the PWT 9.0 labor share (which may where Kp and Kg represent the private and vary across countries and over time). public capital stock, respectively. TFP growth can be expressed as Solow Decomposition Accounting for Natural Capital ˆ =y At ˆ −L ˆt − αp K pt ˆ −α K t g ( ˆ −L gt ˆ t ) ( ) The technology described in equation (A.1) ( ˆ − 1− αp − αg h t ) (A.8) now incorporates the use of natural resources (Monge-Naranjo, Sánchez, and Santaeulàlia- Llopis 2019): The values of a p and a g are calibrated ( ) (h L ) α 1−α following Lowe, Papageorgiou, and Pérez- Yt = At Ktγ Tt1−γ t t ,(A.9) Sebastián (2012): ap and ag cannot be directly derived from national income and product where K is the aggregate stock of capital, T accounts data. However, the share of repro- represents the service flows of the natural ducible capital, ag + ap, can be calculated from capital, and a (1– g ) represents the natural the labor share of income of the economy resource share in GDP. (labsh in the PWT) as (1-labsh).6 The income share of natural resources The estimation of the composition of cap- is computed using data on the rents from ital is not trivial: this report uses estimates of natural resources. These data are collected the production function augmented by public from the World Bank’s World Development capital for high-income economies (Kamps Indicators database. In this context, TFP 2004) as well as for low- and middle-income growth is TABLE A.2  Estimated Output Elasticities to Private and Public Capital, by Country Income Group Elasticities Relative share of Private capital Public capital private capital Sample [1] [2] [1]/([1]+[2]) Low-Income Countries 0.23 0.25 0.48 Middle-Income Countries 0.29 0.17 0.63 High-Income Countries 0.26 0.22 0.54 Sources: Gupta et al. 2011 (low- and middle-income countries); Kamps 2004 (high-income countries). O u t p u t p e r W o r k e r , Fac t o r A cc u m u la t i o n , a n d T o t al P r o d u c t i v i t y    105 ˆ =y At ( ˆ −L ˆ t − αγ K t ˆ − {α (1 − γ )} T t ) ( ˆ −L t t ) ˆ. ˆ − (1 − α ) h t heavily on the dataset’s companion paper by Feenstra, Inklaar, and Timmer (2015). (A.10) One of the goals of this appendix is to compare differences in aggregate labor pro- With information on a (from the share of ductivity and TFP across countries and over labor force) and a (1- g ) (as computed by the time, with emphasis on the productivity ratio of natural resource rents to GDP), we and growth performance of Sub-Saharan can implicitly compute g . For the purposes of African countries. Labor productivity is our calculation, assume a natural-resource-­ an indicator that is related to economic augmented production function similar to growth, competitiveness, and living stan- equation (A.7): Yt = At Ktα Ttγ ( ht Lt ) dards in an economy. It is typically defined 1−α −γ , where a and g represent the share of reproducible as the total volume of output (as mea- and natural capital in GDP. Equation (A.6) sured by GDP) produced per unit of labor defines Ât as TFP growth without accounting (or number of people employed) during a for natural resources. Then, TFP growth in a period of time. GDP captures the monetary technology that accounts for natural capital, value of goods and services produced in a Aˆ NR , is equal to determined country during a period of time. t Employment consists of all working-age ( ) people who either have paid employment or ˆ + γ  αk ˆ NR = A A ˆ + (1 − α ) h ˆ  , (A.11) ˆ −T t t  t t t  are self-employed.8 Labor productivity only partially reflects where the difference between the traditional the productivity of labor in terms of work- TFP growth (equation [A.3]) and the mea- ers’ personal capacit y or intensit y of sure of TFP growth including natural capital effort. TFP, on the other hand, is the ratio (equation [A.11]) depends on the growth rate of aggregate output to aggregate inputs of the composite input index from the clas- (Sickles and Zelenyuk 2019). It measures sical model, the growth in the use of natu- the impact of technological change and ral capital T, and the share of natural capital changes in worker knowledge on the long- rents in production (see Brandt, Schreyer, and term output of an economy. It is derived Zipperer 2017). from increases in the levels of efficiency and technology. It is considered to be that portion of the growth of an economy that Definitions and Data is not explained by the amount of inputs Description used (say, labor and capital). The comparison of productivity levels and Output. The levels of output are prox- sources of productivity growth in Sub- ied using the data on real GDP estimated Saharan Africa relative to other world regions from the output side (GDP O) from the requires a dataset with ample coverage across PWT. This indicator is a better approxima- countries and over time. To conduct this tion of the total production of the economy. analysis at the aggregate level, this report uses Additionally, the PWT uses the expenditure PWT 9.0 data with annual information from approach to measure the level of economic 1960 to 2014 for a wide array of countries activity (GDPE). In contrast, GDPE captures in the world. This information is updated the standards of living of the different coun- using the PWT 9.1 with information for the tries in the world. According to the PWT years 2015 to 2017.7 This discussion of the methodology, countries with strong terms data, presented as a preamble to appendix of trade will have a higher real GDPE than B (the country-specific data analyses), relies GDPO. 106  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca Real output. The PWT distinguishes real a piece-wise linear function (as in Caselli output measures that are constant across 2005): countries but depend on the current year (CGDP O) from those that are constant  0.134 ⋅ s    if   s ≤ 4 across countries and also constant over   time (RGDP O). The former indicator, also φ ( s) =  0.134 ⋅ 4 + 0.101 ⋅ ( s − 4 )  if  4 < s ≤ 8 known as current price real GDP (CGDPO),   0.134 ⋅ 4 + 0.101 ⋅ 4 + 0.068 ⋅ ( s − 8)  if   s > 8. is used to conduct country comparisons  in a particular year, whereas the latter (A.12) one, known as constant price real GDP (RGDP O), is used for comparisons across If we assume that the wage-schooling rela- countries and over time. CGDP O —the tionship is log-linear (as in the empirical lit- output-side real GDP at current purchas- erature), then the relationship between h and ing power parities (PPPs) (in US$, mil- s should also be log-linear. The PWT con- lions, at 2011 prices)—is used to conduct structs the h index for 150 countries using development accounting exercises. Finally, data on schooling years from Barro and Lee the PWT uses the real GDP at constant (2013) for 95 countries, and from Cohen and national prices (RGDP NA) (also expressed Soto (2007) and Cohen and Leker (2014) for in US$, millions, at 2011 prices) to conduct an additional 55 countries. growth accounting exercises. International data on education-wage Physical capital. The stock of physical profiles suggest that the return to an addi- capital is estimated based on the accumu- tional year of education in Sub-Saharan lation and depreciation of past investments Africa, the region with the lowest level of using the perpetual inventory method education, is about 13.4 percent. In con- (PIM). One of the novel aspects of the esti- trast, the return to an additional year of mation of the aggregate capital stock in education is 10.1 percent for the world PWT 9.0 is the use of investments disaggre- and 6.8 percent for the Organisation for gated by type of asset. The data on invest- Economic Co-operation and Development ments by asset type is obtained from either (OECD) countries (Psacharopoulos 1994). the national accounts or partly estimated This measure of human capital tries to rec- using the commodity-flow method in the oncile the properties of a log-linear relation- spirit of Caselli and Wilson (2004). The ship between education and income at the average depreciation rate in PWT 9.0 var- country level with the concavity of that rela- ies across countries and over time because tionship across countries. The h index from the asset composition differs across coun- the PWT assumes homogeneous returns tries and the depreciation rate is not simi- across countries. lar across assets. In addition, PWT 9.0 uses The relationship between h and s has information on the asset composition of the also been characterized in the empirical lit- capital stock to compute the relative price erature as f (sit) = f i sit for each country i in of investment. period t. This specification assumes that Human capital per worker. The index of the returns to education are heterogeneous human capital per worker, h, is constructed across countries. This report has used two using the average years of schooling in different sets of country estimates for the the population over 25 years old (Barro returns to education to construct the human and Lee 2013). Following Hall and Jones capital index: (a) the estimated Mincerian (1999), the years of schooling are con- returns from Caselli, Ponticelli, and Rossi verted into a measure of h through the for- (2017) and Caselli (2017);9 and (b) the esti- mula h = exp{f (s)}, with s representing the mated Mincerian returns from Montenegro average years of schooling and f (s) being and Patrinos (2014).10 These two sets of O u t p u t p e r W o r k e r , Fac t o r A cc u m u la t i o n , a n d T o t al P r o d u c t i v i t y    107 country estimates of Mincerian returns lead The main sources of sectoral data are to two different human capital indexes with the United Nations National Accounts heterogeneous returns to education across (UN-NAC) database, the World Bank’s countries. World Development Indicators (WDI) data- The degree of association between the base, and International Labour Organization indexes with homogeneous and hetero- (ILO) statistics. The value-added data at geneous returns to education—expressed in the sector level are first obtained from five-year growth rates, f (sit)– f (sit-5)—is quite UN-NAC from 1990 to 2016. It is expressed high, and it fluctuates between 0.75 and in US dollars at current prices and at 2010 0.80. Note that the correlation among the prices. Economic activity is disaggregated three indexes of human capital expressed in into seven large sectors: agriculture, mining 10-year growth rates, f (sit)– f (sit-10), fluctu- and utilities, construction, manufacturing, ates between 0.73 and 0.77. trade and hospitality, transport and com- Labor share of income. Finally, the PWT munication, and other activities. For the has estimated the labor share (or the share of purposes of the analysis conducted in this labor income in economic activity) for a wide report, the International Standard Industrial array of countries and years. There is broad Classification (ISIC) Revision 3.0 data are availability of information on labor compen- sectors (table A.3): reclassified into five larger ­ sation of employees; however, a separate esti- agriculture, manufacturing, nonmanufac- mation is needed for the labor compensation turing, market services, and nonmarket of self-employed workers. The cross-country services. estimates of the labor share yield some styl- S e c ond , we obt a i n d i s a g g re g at e d ized facts (Feenstra, Inklaar, and Timmer s ector-level data on total workers from ­ 2015): the  ILO and grouped the data using the same five-­ s ector classification outlined in  he global average of the labor share in • T table A.3. The sector-level data on employ- income is about 0.52 (significantly lower ment from the ILO range from 1990 to 2016. than the two-thirds typically assumed in Hence, employment data constitute the most the macroeconomic literature). binding constraint in terms of data availabil- There is no systematic relationship between •  ity over time. labor shares and income per capita levels. The list of 28 Sub-Saharan African Labor shares have declined over time in •  countries with sectoral data is presented most of the countries covered.11 in table A.4. The analysis of sectoral productivity in the main text of the report Sectoral Productivity: Sources is undertaken not only for Sub-Saharan of Data Africa but also for country groups within the region, classified as follows: This report has assembled a large database of value-added, employment, and productiv- • I ncome level: Low- and lower-middle- ity indicators at the sectoral level for a wide income countries (LLMCs) and upper- array of countries from 1990 to 2016. This middle-income countries (UMICs) dataset allows for the examination of trends • Degree of natural resource abundance: in labor productivity and employment shares 0 for non-resource abundance, 1 for non-oil across sectors in Sub-Saharan Africa relative resource abundance, and 2 for oil abundance to other regions. In other words, it helps to • Geographical subregion: West, East, Cen- document the patterns of structural trans- tral, and Southern African countries. formation among countries in the region. 108   Boosting Productivity in Sub-Saharan Africa TABLE A.3  Classification of Sectors of Economic Activity ISIC sector of economic activity Sector group for analysis A. Agriculture, hunting, and forestry Agriculture B. Fishing Agriculture C. Mining and quarrying Nonmanufacturing D. Manufacturing Manufacturing E. Electricity, gas, and water supply Nonmanufacturing F. Construction Nonmanufacturing G. Wholesale and retail trade; repair of motor vehicles, motorcycles Market services H. Hotels and restaurants Market services I. Transport, storage, and communications Market services J. Financial intermediation Market services K. Real estate, renting, and business activities Market services L. Public administration and defense; compulsory social security Nonmarket services M. Education Nonmarket services N. Health and social work Nonmarket services O. Other community, social, and personal service activities Nonmarket services P. Private households with employed persons Nonmarket services Source: Original table for this publication. Note: ISIC = International Standard Industrial Classification provided by UNIDO. TABLE A.4  Classification of Sub-Saharan African Countries Code Name Income Resources Subregion AGO Angola LLMC 2 Southern Africa BEN Benin LLMC 0 West Africa BWA Botswana UMIC 1 Southern Africa BFA Burkina Faso LLMC 0 West Africa BDI Burundi LLMC 0 East Africa CMR Cameroon LLMC 2 Central Africa CAF Central African Republic LLMC 0 Central Africa COG Congo, Rep. LLMC 2 Central Africa GAB Gabon UMIC 2 Central Africa GMB Gambia, The LLMC 0 West Africa KEN Kenya LLMC 0 East Africa LSO Lesotho LLMC 0 Southern Africa MDG Madagascar LLMC 0 Southern Africa MWI Malawi LLMC 0 Southern Africa MLI Mali LLMC 1 West Africa MRT Mauritania LLMC 1 West Africa MUS Mauritius UMIC 0 East Africa MOZ Mozambique LLMC 1 Southern Africa NAM Namibia UMIC 1 Southern Africa NER Niger LLMC 0 West Africa Table continued next page O u t p u t p e r W o r k e r , F a c t o r A c c u m u l a t i o n , a n d To t a l P r o d u c t i v i t y    109 TABLE A.4  Classification of Sub-Saharan African Countries (Continued) Code Name Income Resources subregion NGA Nigeria LLMC 2 West Africa RWA Rwanda LLMC 0 East Africa SLE Sierra Leone LLMC 1 West Africa ZAF South Africa UMIC 0 Southern Africa SWZ Eswatini LLMC 0 Southern Africa TGO Togo LLMC 0 West Africa UGA Uganda LLMC 0 East Africa ZMB Zambia LLMC 1 Southern Africa Source: Barrot, Calderón, and Servén 2018. Note: The indicator variable under “Resources” classifies countries into oil-rich countries (2); non-oil-rich countries (1); and resource-poor countries (0). LLMC = low- or lower-middle-income country; UMIC = upper-middle-income country. Notes  9. The data on Mincerian returns collected by Caselli, Ponticelli, and Rossi (2017) are   1. The start and end of the time series for each ­ available for download from http://personal​ country will depend on data availability. .lse​.ac.uk/casellif/papers/references_table.pdf.   2. Discussion of this framework refers to the 10. In both cases, whenever there are data on years work of Christensen and Jorgenson (1970); of schooling and no data on returns for a spe- Christensen, Jorgenson, and Lau (1973); cific country, we input the average returns to Griliches and Jorgenson (1966); and Jorgen- education of its corresponding region. son and Griliches (1967). 11. The labor share in income for many coun-   3. This corresponds to the ratio of output per tries has been declining over the past two worker in the United States to that of each decades (IMF 2017). In industrial countries, Sub-Saharan African country and aggregated technological progress accounts for about using labor-force weighted averages. half of the overall decline in the labor share.   4. Denison (1962) adjusted the measurement This progress is manifested by sharp reduc- of the labor input for changes in the size of tions in the relative price of investment goods the labor force and shifts related to age, and varying exposure to routine-based occu- gender, hours worked, and unemployment. pations. For instance, 47 percent of US work- These improvements, as well as others in the ers are at risk of automation over the next basic growth-accounting methodology, led two decades (Frey and Osborne 2017), while to estimates of the contribution of TFP to US that percentage increases to 57 percent of growth that were much lower than Solow’s. jobs in the OECD (World Bank 2016). This   5. In a comprehensive survey, Caselli (2005) reduces the earnings of middle-skilled work- finds that factor accumulation cannot explain ers. In low- and middle-income countries, more than half of the differences in income the declining labor share is mainly driven by per capita across countries. global integration forces—in particular, the   6. Note that labsh is heterogeneous across coun- expansion of global value chains that have tries and displays some time variation within contributed to increasing the overall capital each country. intensity in production. Trade and finan-  7. The data can be downloaded from the cial integration grew sharply over the past PWT 9.1 website: https://www.rug.nl/ggdc​ quarter century, thanks to the removal of /productivity/pwt/. restrictions on international trade and capi-   8. Labor productivity can also be measured as tal mobility and the decline in transportation GDP per hour worked. The labor input here and communication costs—the latter being is defined as total hours worked of all peo- facilitated by technological progress. In the ple engaged in employment. Empirically, short term, policy makers should implement the availability of hours worked of people policies to provide workers access to growth employed in the economy is more limited for opportunities and design mechanisms to low- and lower-middle-income countries. share growth benefits more broadly. 110  B o o s t i n g P r o d u c t i v i t y i n S u b - Sa h a r a n A f r i ca References Denison, Edward F. 1962. “Sources of Growth in the United States and the Alternatives before Abramovitz, Moses. 1956. “Resource and Output Us.” Supplement Paper No. 13, Committee for Trends in the United States Since 1870.” Amer- Economic Development, New York. ican Economic Review 46 (2): 5–23. Denison, Edward F. 1967. Why Growth Rates Barro, Robert J., and Jong-Wha Lee. 2013. “A Differ: Postwar Experience in Nine Western New Data Set of Educational Attainment in Countries. Washington, DC: Brookings Insti- the World, 1950–2010.” Journal of Develop- tution Press. ment Economics 104: 184–98. Denison, Edward F., Zvi Griliches, and Dale W. Barrot, Luis-Diego, César Calderón, and Luis Jorgenson. 1972. “The Measurement of Pro- Servén. 2018. “Sectoral Productivity Shifts ductivity.” Survey of Current Business 52 (5) in Sub-Saharan Africa.” Background paper, Part II: 3–111. World Bank, Washington, DC. Feenstra, Robert C., Robert Inklaar, and Marcel Brandt, Nicola, Paul Schreyer, and Vera Zipperer. P. Timmer. 2015. “The Next Generation of 2017. “Productivity Measurement with Natural the Penn World Table.”  American Economic Capital.” Review of Income and Wealth Review 105 (10): 3150–82. 63 (S1): S7–S21. Frey, Carl Benedikt, and Michael A. Osborne. Caselli, Francesco. 2005. “Accounting for 2017. “The Future of Employment: How Cross-Country Income Differences.” In Hand- Susceptible Are Jobs to Computerisation?” book of Economic Growth, Vol. 1, Part A, Technological Forecasting and Social Change edited by Philippe Aghion and Steven Durlauf, 114: 254–80. 679–741. Amsterdam: Elsevier. Griliches, Zvi, and Dale W. Jorgenson. 1966. Caselli, Francesco. 2017. Technology Differences “Sources of Measured Productivity Change: over Space and Time. Princeton, NJ: Univer- Capital Input.” American Economic Review sity Press. 56 (1/2): 50–61. Caselli, Francesco, Jacopo Ponticelli, and Fed- Gupta, Sanjeev, Alvar Kangur, Chris Papageor- erico Rossi. 2017. “A New Dataset on Mince- giou, and Abdoul Wane. 2011. “Efficiency-Ad- rian Returns.” In Technology Differences over justed Public Capital and Growth.” Working Space and Time, by Francesco Caselli, 108–18. Paper No. 11/217, International Monetary Princeton, NJ: Princeton University Press. Fund, Washington, DC. Caselli, Francesco, and Daniel J. Wilson. 2004. Hall, Robert E., and Charles I. Jones. 1999. “Importing Technology.” Journal of Monetary “Why Do Some Countries Produce So Much Economics 51 (1): 1–32. More Output per Worker than Others?” Christensen, Laurits R., Dianne Cummings, and Quarterly Journal of Economics 114 (1): Dale W. Jorgenson. 1981. “Relative Productiv- 83–116. ity Levels, 1947–1973: An International Com- Hsieh, Chang-Tai, and Peter J. Klenow. 2010. parison.” European Economic Review 16 (1): “Development Accounting.” American Eco- 61–94. nomic Journal: Macroeconomics 2 (1): 207–23. Christensen, Laurits R., and Dale W. Jorgenson. IMF (International Monetary Fund). 2017. 1970. “U.S. Real Product and Real Factor “Understanding the Downward Trend in Input, 1929–1967.” Review of Income and Labor Income Shares.” In World Economic Wealth 16 (1): 19–50. Outlook, April 2017: Gaining Momentum?, Christensen, Laurits R., Dale W. Jorgenson, and 121–72. Washington, DC: IMF. Lawrence J. Lau. 1973. “Transcendental Loga- Jones, Charles I. 2016. “The Facts of Economic rithmic Production Frontiers.” Review of Eco- Growth.” In Handbook of Macroeconomics, nomics and Statistics 55 (1): 28–45. Vol. 2A, edited by John B. Taylor and Harald Cohen, Daniel, and Laura Leker. 2014. “Health Uhlig, 3–69. Amsterdam: Elsevier. and Education: Another Look with the Proper Jorgenson, Dale W., and Zvi Griliches. 1967. Data.” Discussion Paper 9940, Centre for Eco- “The Explanation of Productivity Change.” nomic Policy Research, London. Review of Economic Studies 34 (3): 249–83. Cohen, Daniel, and Marcelo Soto. 2007. “Growth Jorgenson, Dale W., and Mieko Nishimizu. 1978. and Human Capital: Good Data, Good “U.S. and Japanese Economic Growth, 1952– Results.” Journal of Economic Growth 12 (1): 1974: An International Comparison.” Eco- 51–76. nomic Journal 88 (352): 707–26. O u t p u t p e r W o r k e r , Fac t o r A cc u m u la t i o n , a n d T o t al P r o d u c t i v i t y    111 Kamps, Christophe. 2004. “New Estimates of Gov- Montenegro, Claudio E., and Harry Anthony ernment Net Capital Stocks for 22 OECD Coun- Patrinos. 2014. “Comparable Estimates of tries, 1960–2001.” Working Paper No. 04/67, Returns to Schooling Around the World.” International Monetary Fund, Washington, DC. Policy Research Working Paper 7020, World Klenow, Peter J., and Andrés Rodríguez-Clare. Bank, Washington, DC. 1997. “The Neoclassical Revival in Growth Psacharopoulos, George. 1994. “Returns to Economics: Has It Gone Too Far?” In NBER Investment in Education: A Global Update.” Macroeconomics Annual 1997, edited by Ben World Development 22 (9): 1325–43. S. Bernanke and Julio J. Rotemberg, 73–103. Sickles, Robin C., and Valentin Zelenyuk. 2019. Cambridge, MA: MIT Press. Measurement of Productivity and Efficiency: Lowe, Matt, Chris Papageorgiou, and Fidel Theory and Practice. Cambridge: Cambridge Pérez-Sebastián. 2012. “The Public and Pri- University Press. vate MPK.” Paper No. c017_021, International Solow, Robert M. 1957. “Technical Change and Conference on Dynamics, Economic Growth, the Aggregate Production Function.” Review and International Trade (DEGIT), Milan, of Economics and Statistics 39 (3): 312–20. September 13–14. Ti nbergen, Jan. 1942 . “Zu r T heorie der Mankiw, N. Gregory, David Romer, and David ­L angfristigen Wirtschaftsent wicklung.” N. Weil. 1992. “A Contribution to the Empir- ­ Weltwirtschaftliches Archiv 55 (1): 511–49. ics of Economic Growth.” Quarterly Journal Walters, Dorothy. 1968. “Canadian Income Lev- of Economics 107 (2): 407–37. els and Growth: An International Perspective.” Monge-Naranjo, Alexander, Juan M. Sánchez, Staff Study No. 23, Economic Council of Natural and Raül Santaeulàlia-Llopis. 2019. “­ ­Canada, Ottawa. Resou rces and Global M isallocation.” World Bank. 2016. World Development 2016: ­ American Economic Journal: Macroeconomics Digital Dividends. Washington, DC: World 11 (2): 79–126. Bank. Country Productivity Analysis in Sub-Saharan Africa B 113 114   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA ANGOLA Country Profile (2017) 12 GDP growth 10 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 166,402 US$, millions 8 6 GDP per capita (2011 PPP): 6,056 US$ 4 2 Population: 26.8 million 0 -2 Employment: 9.1 million -4 -6 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 8.3 5.8 53.1 14.7 18.2 Employment 51.1 1.5 7.4 24.1 15.9 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 2.0 0.6 0.20 1.5 0.15 0.4 1.0 0.10 0.2 0.5 0.05 0 0 0 19 0 65 70 19 5 80 19 5 90 95 20 0 05 10 15 15 60 65 70 75 80 85 90 95 00 05 10 15 60 80 85 90 95 00 05 10 70 75 65 6 7 8 0 19 19 19 19 19 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 20 20 20 19 19 19 AGO SSA East Asian Dragons AGO SSA East Asian Dragons AGO SSA East Asian Dragons Real output per worker (US = 1.0) Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) 0.20 100 Ratio relative to the United States 0.80 Ratio relative to the United States 80 0.15 0.60 60 Percent 0.10 0.40 40 0.05 0.20 20 0 0 0 69 79 89 99 09 7 9 9 9 9 9 7 9 9 9 9 9 7 –1 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 10 60 70 80 90 00 10 60 70 80 90 00 10 6 7 8 9 0 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 Factor accumulation TFP 20 Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.39 −1.30 −1.95 2.84 0.39 −1.30 −1.95 2.84 Physical capital 0.26 1.24 −1.13 1.10 0.19 0.88 −0.81 0.78 Human capital 0.79 0.25 0.90 0.86 0.23 0.07 0.26 0.25 TFP .. .. .. .. −0.02 −2.26 −1.40 1.81 II. Growth accounting: private and public capital accumulation Output 0.39 −1.30 −1.95 2.84 0.39 −1.30 −1.95 2.84 Physical capital .. .. .. .. .. .. .. .. - Public 1.61 1.47 −0.64 3.48 0.42 0.39 −0.17 0.92 - Private −1.52 1.09 −1.48 −2.38 −0.68 0.49 −0.67 −1.07 Human capital 0.79 0.25 0.90 0.86 0.23 0.07 0.26 0.25 TFP .. .. .. .. 0.43 −2.25 −1.38 2.75 III. Growth accounting including the natural capital Output .. .. .. 2.84 .. .. .. 2.84 Physical capital .. .. .. 1.10 .. .. .. 0.22 Natural capital .. .. .. 5.14 .. .. .. 1.60 Human capital .. .. .. 0.86 .. .. .. 0.18 TFP .. .. .. .. .. .. .. 0.84 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    115 BENIN Country Profile (2017) 8 GDP growth 7 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 23,544 US$, millions 6 GDP per capita (2011 PPP): 1,882 US$ 5 4 Population: 11.5 million 3 2 Employment: 4.8 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 24.5 14.5 8.8 24.3 27.9 Employment 40.9 15.7 3.5 28.1 11.8 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.15 0.9 0.4 0.10 0.6 0.2 0.05 0 0.3 0 15 60 80 85 90 95 00 05 10 60 65 70 75 80 85 90 95 00 05 10 15 70 75 65 60 65 70 75 80 85 90 95 00 05 10 15 20 19 19 19 19 19 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 BEN SSA East Asian Dragons BEN SSA East Asian Dragons BEN SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.08 100 Ratio relative to the United States 0.35 Ratio relative to the United States 80 0.30 0.06 0.25 60 Percent 0.20 0.04 40 0.15 0.02 0.10 20 0.05 0 0 0 69 79 89 99 09 17 9 9 9 9 9 7 9 9 9 9 9 7 0– 0– 0– 0– 0– 0– –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 6 7 8 9 0 1 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.21 .. 1.03 1.33 1.21 .. 1.03 1.33 Physical capital −0.75 .. −1.96 0.07 −0.27 .. −0.71 0.03 Human capital 0.97 .. 1.13 1.48 0.91 .. 0.86 0.94 TFP .. .. .. .. 0.57 .. 0.88 0.36 II. Growth accounting: private and public capital accumulation Output 1.21 .. 1.03 1.33 1.21 .. 1.03 1.33 Physical capital .. .. .. .. .. .. .. .. - Public −2.33 .. −2.74 −2.05 −0.58 .. −0.60 −0.39 - Private 1.49 .. −0.07 2.56 0.34 .. −0.01 0.44 Human capital 0.97 .. 1.13 1.48 0.81 .. 0.83 0.94 TFP .. .. .. .. 0.64 .. 0.81 0.34 III. Growth accounting including the natural capital Output .. .. .. 1.33 .. .. .. 1.33 Physical capital .. .. .. 0.07 .. .. .. 0.03 Natural capital .. .. .. −3.66 .. .. .. 0.00 Human capital .. .. .. 1.48 .. .. .. 0.95 TFP .. .. .. .. .. .. .. 0.36 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 116   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA BOTSWANA Country Profile (2017) 10 GDP growth 8 population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 35,951 US$, millions 6 4 GDP per capita (2011 PPP): 15,896 US$ 2 0 Population: 2.3 million -2 -4 Employment: 1.0 million -6 -8 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 1.9 7.0 22.2 28.4 40.4 Employment 27.2 1.3 11.5 24.6 35.4 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.35 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.7 0.30 1.5 0.6 0.25 0.5 0.20 1.0 0.4 0.3 0.15 0.5 0.2 0.10 0.1 0.05 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 BWA SSA East Asian Dragons BWA SSA East Asian Dragons BWA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.35 100 Ratio relative to the United States 0.5 Ratio relative to the United States 0.30 80 0.4 0.25 60 Percent 0.20 0.3 0.15 40 0.2 0.10 20 0.1 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 2.74 .. 3.79 2.02 2.74 .. 3.79 2.02 Physical capital 4.25 .. 3.03 5.09 2.88 .. 2.05 3.44 Human capital 1.57 .. 2.88 1.14 0.67 .. 1.10 0.37 TFP .. .. .. .. −0.81 .. 0.64 −1.79 II. Growth accounting: private and public capital accumulation Output 2.74 .. 3.79 2.02 2.74 .. 3.79 2.02 Physical capital .. .. .. .. .. .. .. .. - Public 4.36 .. 3.65 4.84 1.15 .. 0.95 1.21 - Private 4.15 .. 2.62 5.20 1.88 .. 1.17 2.22 Human capital 1.57 .. 2.88 1.14 0.54 .. 0.98 0.37 TFP .. .. .. .. −0.84 .. 0.69 −1.78 III. Growth accounting including the natural capital Output .. .. .. 2.02 .. .. .. 2.02 Physical capital .. .. .. 5.09 .. .. .. 2.68 Natural capital .. .. .. 17.81 .. .. .. 0.82 Human capital .. .. .. 1.14 .. .. .. 0.31 TFP .. .. .. .. .. .. .. −1.80 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    117 BURKINA FASO Country Profile (2017) 10 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 32,506 US$, millions 8 GDP per capita (2011 PPP): 1,679 US$ 6 Population: 19.2 million 4 2 Employment: 7.0 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 28.6 8.3 14.3 18.8 30.0 Employment 26.3 15.8 16.1 29.9 11.9 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.9 0.15 0.4 0.6 0.10 0.2 0.3 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 15 60 80 85 90 95 00 05 10 70 75 65 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 20 20 20 19 19 19 BFA SSA East Asian Dragons BFA SSA East Asian Dragons BFA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.25 0.040 Ratio relative to the United States Ratio relative to the United States 0.035 80 0.20 0.030 0.025 60 Percent 0.15 0.020 40 0.10 0.015 0.010 20 0.05 0.005 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 2.00 1.57 0.87 3.26 2.00 1.57 0.87 3.26 Physical capital 3.67 2.55 4.45 3.91 1.35 0.94 1.63 1.44 Human capital 0.39 0.03 0.20 0.83 0.25 0.02 0.13 0.52 TFP .. .. .. .. 0.40 0.61 −0.89 1.29 II. Growth accounting: private and public capital accumulation Output 2.00 1.57 0.87 3.26 2.00 1.57 0.87 3.26 Physical capital .. .. .. .. .. .. .. .. - Public 3.65 2.59 3.63 4.49 0.70 0.50 0.70 0.86 - Private 3.81 2.51 5.13 3.73 0.67 0.44 0.90 0.65 Human capital 0.39 0.03 0.20 0.83 0.25 0.02 0.13 0.52 TFP .. .. .. .. 0.38 0.61 −0.85 1.21 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 118   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA BURUNDI Country Profile (2017) 30 GDP growth population growth (%) 25 Population growth GDP growth vis-à-vis GDP (2011 PPP): 9,437 US$, millions 20 GDP per capita (2011 PPP): 753 US$ 15 10 Population: 11.9 million 5 Employment: 5.4 million 0 -5 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 32.4 11.6 3.3 13.7 39.0 Employment 91.3 1.8 0.7 2.8 3.5 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.5 0.6 0.20 0.15 1.0 0.4 0.10 0.5 0.2 0.05 0 0 0 15 60 80 85 90 95 00 05 10 60 65 70 75 80 85 90 95 00 05 10 15 70 75 65 60 65 70 75 80 85 90 95 00 05 10 15 20 19 19 19 19 19 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 BDI SSA East Asian Dragons BDI SSA East Asian Dragons BDI SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) TFP gap (US = 1.0) 100 Ratio relative to the United States 0.12 0.25 Ratio relative to the United States 80 0.10 0.20 60 0.08 Percent 0.15 0.06 40 0.10 0.04 20 0.05 0.02 0 0 0 69 79 89 99 9 17 –0 9 9 9 9 9 7 0– 0– 0– 0– 0– 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 00 6 7 8 9 1 – – – – – – 60 70 80 90 00 10 19 19 19 19 20 20 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.00 .. −0.16 1.80 1.00 .. −0.16 1.80 Physical capital 0.01 .. 0.19 −0.11 0.00 .. 0.05 −0.03 Human capital 0.42 .. 0.36 0.78 0.47 .. 0.31 0.57 TFP .. .. .. .. 0.54 .. −0.53 1.26 II. Growth accounting: private and public capital accumulation Output 1.00 .. −0.16 1.80 1.00 .. −0.16 1.80 Physical capital .. .. .. .. .. .. .. .. - Public 2.55 .. 0.75 3.77 0.43 .. 0.08 0.54 - Private −0.97 .. 0.04 −1.66 −0.15 .. 0.00 −0.22 Human capital 0.42 .. 0.36 0.78 0.36 .. 0.20 0.57 TFP .. .. .. .. 0.36 .. −0.45 0.91 III. Growth accounting including the natural capital Output .. .. .. 1.80 .. .. .. 1.80 Physical capital .. .. .. −0.11 .. .. .. −0.02 Natural capital .. .. .. 126.99 .. .. .. 0.34 Human capital .. .. .. 0.78 .. .. .. 0.30 TFP .. .. .. .. .. .. .. 1.17 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    119 CABO VERDE Country Profile (2017) GDP growth Population growth 5 population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 3,906 US$, millions 4 3 GDP per capita (2011 PPP): 6,783 US$ 2 Population: 0.5 million 1 0 Employment: 0.2 million -1 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.3 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 1.5 0.6 0.2 1.0 0.4 0.1 0.5 0.2 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 CPV SSA East Asian Dragons CPV SSA East Asian Dragons CPV SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.14 100 Ratio relative to the United States 1.0 Ratio relative to the United States 0.12 80 0.8 0.10 60 Info. not available Percent 0.08 0.6 Info. not available 0.06 40 0.4 0.04 20 0.2 0.02 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 120   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA CAMEROON Country Profile (2017) GDP growth Population growth 6 population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 69,451 US$, millions 5 GDP per capita (2011 PPP): 2,785 US$ 4 3 Population: 24.6 million 2 Employment: 10.9 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 15.4 15.3 15.0 29.1 25.2 Employment 62.4 5.8 3.4 22.9 5.5 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.7 0.25 1.2 0.6 0.20 0.5 0.9 0.15 0.4 0.6 0.3 0.10 0.2 0.05 0.3 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 CMR SSA East Asian Dragons CMR SSA East Asian Dragons CMR SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.10 100 Ratio relative to the United States 0.5 Ratio relative to the United States 0.08 80 0.4 60 Percent 0.06 0.3 0.04 40 0.2 0.02 20 0.1 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 9 9 7 69 79 89 99 09 17 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 0– 0– 0– 90 00 10 19 19 19 19 20 20 6 7 8 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.77 2.37 −0.42 0.50 0.77 2.37 −0.42 0.50 Physical capital 0.48 0.96 0.91 −0.24 0.22 0.45 0.42 −0.11 Human capital 0.90 0.67 1.41 0.67 0.48 0.36 0.75 0.36 TFP .. .. .. .. 0.06 1.57 −1.60 0.26 II. Growth accounting: private and public capital accumulation Output 0.77 2.37 −0.42 0.50 0.77 2.37 −0.42 0.50 Physical capital .. .. .. .. .. .. .. .. - Public 0.90 0.90 0.85 0.94 0.22 0.22 0.21 0.23 - Private 0.33 0.97 0.93 −0.66 0.07 0.22 0.21 −0.15 Human capital 0.90 0.67 1.41 0.67 0.48 0.36 0.75 0.36 TFP .. .. .. .. −0.01 1.58 −1.59 0.06 III. Growth accounting including the natural capital Output .. .. .. 0.50 .. .. .. 0.50 Physical capital .. .. .. −0.24 .. .. .. −0.11 Natural capital .. .. .. 0.39 .. .. .. 0.03 Human capital .. .. .. 0.67 .. .. .. 0.42 TFP .. .. .. .. .. .. .. 0.17 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    121 CENTRAL AFRICAN REPUBLIC Country Profile (2017) GDP growth Population growth 5 0 population growth (%) GDP (2011 PPP): 3,544 US$, millions GDP growth vis-à-vis -5 GDP per capita (2011 PPP): 707 US$ -10 -15 Population: 5.0 million -20 -25 Employment: 2.2 million -30 -35 -40 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 31.2 23.0 4.3 20.9 20.6 Employment 85.6 7.5 0.5 3.4 3.0 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.5 0.6 0.20 0.15 1.0 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 CAF SSA East Asian Dragons CAF SSA East Asian Dragons CAF SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.40 100 Ratio relative to the United States 0.025 Ratio relative to the United States 0.35 80 0.020 0.30 0.25 60 Percent 0.015 0.20 40 0.010 0.15 0.10 20 0.005 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output −1.44 .. −1.23 −1.58 −1.44 .. −1.23 −1.58 Physical capital −1.80 .. −1.78 −1.82 −1.41 .. −1.39 −1.42 Human capital 0.64 .. 1.03 0.69 0.19 .. 0.25 0.15 TFP .. .. .. .. −0.22 .. −0.09 −0.31 II. Growth accounting: private and public capital accumulation Output −1.44 .. −1.23 −1.58 −1.44 .. −1.23 −1.58 Physical capital .. .. .. .. .. .. .. .. - Public −1.96 .. −1.61 −2.20 −0.77 .. −0.64 −0.90 - Private −1.64 .. −2.11 −1.32 −0.59 .. −0.77 −0.49 Human capital 0.64 .. 1.03 0.69 0.13 .. 0.22 0.15 TFP .. .. .. .. −0.21 .. −0.04 −0.34 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 122   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA CHAD Country Profile (2017) 15 GDP growth Population growth population growth (%) 12 GDP growth vis-à-vis GDP (2011 PPP): 24,358 US$, millions 9 GDP per capita (2011 PPP): 1,738 US$ 6 3 Population: 14.9 million 0 Employment: 5.3 million -3 -6 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.8 Ratio relative to the United States 0.30 1.5 Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 TCD SSA East Asian Dragons TCD SSA East Asian Dragons TCD SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.05 100 Ratio relative to the United States 1.0 Ratio relative to the United States 0.04 80 0.8 60 Info. not available 0.6 Percent 0.03 Info. not available 40 0.4 0.02 20 0.2 0.01 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    123 COMOROS Country Profile (2017) 5 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 1,264 US$, millions 4 GDP per capita (2011 PPP): 1,473 US$ 3 Population: 0.8 million 2 Employment: 0.2 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.5 0.6 0.20 0.15 1.0 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 COM SSA East Asian Dragons COM SSA East Asian Dragons COM SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 1.0 0.20 Ratio relative to the United States Ratio relative to the United States 80 0.8 0.15 60 Info. not available Percent 0.6 Info. not available 0.10 40 0.4 0.05 20 0.2 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 124   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA DEMOCRATIC REPUBLIC OF CONGO Country Profile (2017) GDP growth Population growth 10 population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 92,195 US$, millions 8 GDP per capita (2011 PPP): 1,157 US$ 6 Population: 82.6 million 4 Employment: 27.5 million 2 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 2.0 0.6 0.20 1.5 0.15 0.4 1.0 0.10 0.2 0.05 0.5 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 COD SSA East Asian Dragons COD SSA East Asian Dragons COD SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.8 0.12 Ratio relative to the United States Ratio relative to the United States 80 0.7 0.10 0.6 0.08 60 0.5 Percent 0.06 0.4 40 0.3 0.04 20 0.2 0.02 0.1 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output −1.53 −0.57 −4.20 −0.08 −1.53 −0.57 −4.20 −0.08 Physical capital 0.20 1.13 1.31 −1.43 0.09 0.49 0.57 −0.62 Human capital 0.78 0.48 1.35 0.54 0.44 0.28 0.77 0.31 TFP .. .. .. .. −2.06 −1.34 −5.53 0.23 II. Growth accounting: private and public capital accumulation Output −1.53 −0.57 −4.20 −0.08 −1.53 −0.57 −4.20 −0.08 Physical capital .. .. .. .. .. .. .. .. - Public 4.28 1.16 1.36 9.09 0.97 0.26 0.31 2.05 - Private −0.17 1.13 1.31 −2.40 −0.04 0.23 0.27 −0.49 Human capital 0.78 0.48 1.35 0.54 0.44 0.28 0.77 0.31 TFP .. .. .. .. −2.90 −1.34 −5.54 −1.94 III. Growth accounting including the natural capital Output .. .. .. −0.08 .. .. .. −0.08 Physical capital .. .. .. −1.43 .. .. .. −0.99 Natural capital .. .. .. 12.52 .. .. .. 3.39 Human capital .. .. .. 0.54 .. .. .. 0.66 TFP .. .. .. .. .. .. .. −3.13 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    125 REPUBLIC OF CONGO Country Profile (2017) GDP growth 9 Population growth GDP (2011 PPP): 16,987 US$, millions 6 population growth (%) GDP growth vis-à-vis 3 GDP per capita (2011 PPP): 3,904 US$ 0 –3 Population: 4.9 million –6 –9 Employment: 1.8 million –12 –15 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 5.3 4.6 65.3 13.5 11.4 Employment 37.3 19.9 5.3 28.1 9.5 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 2.0 0.6 0.20 1.5 0.15 0.4 1.0 0.10 0.2 0.5 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 19 0 19 0 19 0 19 0 20 0 20 0 65 75 85 95 05 15 19 0 19 5 19 0 19 5 19 0 85 19 0 20 5 20 0 05 20 0 15 6 7 8 9 0 1 6 6 7 7 8 9 9 0 1 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 20 20 19 19 20 COG SSA East Asian Dragons COG SSA East Asian Dragons COG SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.6 0.12 Ratio relative to the United States Ratio relative to the United States 80 0.4 0.08 60 Percent 0.3 40 0.2 0.04 20 0.1 0 0 0 69 79 89 9 09 7 –9 –1 9 9 9 9 9 7 9 9 9 9 9 7 0– 0– 0– 0– –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 90 10 6 7 8 0 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.10 .. 0.64 −0.27 0.10 .. 0.64 −0.27 Physical capital 2.72 .. 3.18 2.40 1.79 .. 2.10 1.58 Human capital 1.13 .. 2.19 0.32 0.36 .. 0.72 0.11 TFP .. .. .. .. −2.05 .. −2.18 −1.96 II. Growth accounting: private and public capital accumulation Output 0.10 .. 0.64 −0.27 0.10 .. 0.64 −0.27 Physical capital .. .. .. .. .. .. .. .. - Public 4.47 .. 1.03 6.81 1.22 .. 0.34 2.35 - Private 1.75 .. 4.10 0.15 0.44 .. 1.24 0.05 Human capital 1.13 .. 2.19 0.32 0.31 .. 0.72 0.11 TFP .. .. .. .. −1.87 .. −1.66 −2.77 III. Growth accounting including the natural capital Output .. .. .. −0.27 .. .. .. −0.27 Physical capital .. .. .. 2.40 .. .. .. 0.21 Natural capital .. .. .. 4.77 .. .. .. 1.57 Human capital .. .. .. 0.32 .. .. .. 0.12 TFP .. .. .. .. .. .. .. −2.17 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 126   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA CÔTE D’IVOIRE Country Profile (2017) 12 GDP growth 10 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 93,958 US$, millions 8 6 GDP per capita (2011 PPP): 3,937 US$ 4 2 Population: 23.9 million 0 -2 Employment: 9.0 million -4 -6 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.9 0.4 0.15 0.6 0.2 0.10 0.3 0.05 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 CIV SSA East Asian Dragons CIV SSA East Asian Dragons CIV SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.12 1.0 Ratio relative to the United States Ratio relative to the United States 0.10 80 0.8 0.08 60 Percent 0.6 0.06 40 0.4 0.04 20 0.02 0.2 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.93 4.17 −2.10 0.89 0.93 4.17 −2.10 0.89 Physical capital −0.32 1.60 −1.39 −0.93 −0.16 0.80 −0.69 −0.46 Human capital 0.82 0.26 1.22 0.94 0.42 0.13 0.61 0.47 TFP .. .. .. .. 0.67 3.24 −2.02 0.88 II. Growth accounting: private and public capital accumulation Output 0.93 4.17 −2.10 0.89 0.93 4.17 −2.10 0.89 Physical capital .. .. .. .. .. .. .. .. - Public 0.54 2.15 1.73 −1.67 0.14 0.56 0.45 −0.43 - Private −0.93 1.37 −3.63 −0.50 −0.22 0.32 −0.86 −0.12 Human capital 0.82 0.26 1.22 0.94 0.42 0.13 0.61 0.47 TFP .. .. .. .. 0.59 3.16 −2.30 0.97 III. Growth accounting including the natural capital Output .. .. .. 0.89 .. .. .. 0.89 Physical capital .. .. .. −0.93 .. .. .. −0.59 Natural capital .. .. .. 15.47 .. .. .. 0.95 Human capital .. .. .. 0.94 .. .. .. 0.66 TFP .. .. .. .. .. .. .. −0.13 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    127 EQUATORIAL GUINEA Country Profile (2017) 20 GDP growth population growth (%) 15 Population growth GDP growth vis-à-vis GDP (2011 PPP): 17,750 US$, millions 10 GDP per capita (2011 PPP): 22,582 US$ 5 0 Population: 0.9 million -5 Employment: 0.4 million -10 -15 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 1.0 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.8 1.5 0.6 0.6 1.0 0.4 0.4 0.5 0.2 0.2 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 GNQ SSA East Asian Dragons GNQ SSA East Asian Dragons GNQ SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 100 1.0 0.8 Ratio relative to the United States Ratio relative to the United States 80 0.8 0.6 60 Info. not available Percent 0.6 Info. not available 0.4 40 0.4 0.2 20 0.2 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 19 19 19 19 20 20 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 128   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA ESWATINI Country Profile (2017) 5 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 10,384 US$, millions 4 GDP per capita (2011 PPP): 7,943 US$ 3 Population: 1.3 million 2 1 Employment: 0.3 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 7.3 37.0 3.7 21.6 30.3 Employment 68.2 11.0 1.8 11.5 7.4 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.5 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.4 1.5 0.6 0.3 1.0 0.4 0.2 0.5 0.2 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 SWZ SSA East Asian Dragons SWZ SSA East Asian Dragons SWZ SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 1.2 0.40 Ratio relative to the United States Ratio relative to the United States 0.35 80 1.0 0.30 60 0.8 0.25 Percent 0.20 0.6 40 0.15 0.4 0.10 20 0.2 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 9 9 9 9 9 7 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.96 .. 0.76 1.10 0.96 .. 0.76 1.10 Physical capital 1.13 .. 3.52 −0.50 0.40 .. 1.24 −0.18 Human capital 0.90 .. 0.79 0.69 0.41 .. 0.37 0.45 TFP .. .. .. .. 0.15 .. −0.85 0.83 II. Growth accounting: private and public capital accumulation Output 0.96 .. 0.76 1.10 0.96 .. 0.76 1.10 Physical capital .. .. .. .. .. .. .. .. - Public 3.38 .. 4.76 2.44 0.38 .. 0.52 0.32 - Private 0.23 .. 3.26 −1.84 0.04 .. 0.61 −0.41 Human capital 0.90 .. 0.79 0.69 0.50 .. 0.43 0.45 TFP .. .. .. .. 0.05 .. −0.81 0.75 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natual capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    129 ETHIOPIA Country Profile (2017) 15 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 160,054 US$, millions 12 GDP per capita (2011 PPP): 1,444 US$ 9 Population: 104.5 million 6 3 Employment: 49.2 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 20 20 20 20 19 ETH SSA East Asian Dragons ETH SSA East Asian Dragons ETH SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.03 100 Ratio relative to the United States 0.25 Ratio relative to the United States 80 0.20 0.02 60 Percent 0.15 40 0.10 0.01 20 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.54 −0.05 −0.90 4.77 1.54 −0.05 −0.90 4.77 Physical capital 1.06 0.40 −2.07 4.14 0.46 0.17 −0.89 1.79 Human capital 0.58 0.06 0.54 1.03 0.33 0.03 0.31 0.58 TFP .. .. .. .. 0.75 −0.25 −0.32 2.40 II. Growth accounting: private and public capital accumulation Output 1.54 −0.05 −0.90 4.77 1.54 −0.05 −0.90 4.77 Physical capital .. .. .. .. .. .. .. .. - Public 1.11 0.40 −1.91 4.13 0.25 0.09 −0.43 0.93 - Private 1.12 0.40 −2.26 4.45 0.23 0.08 −0.47 0.92 Human capital 0.58 0.06 0.54 1.03 0.33 0.03 0.31 0.58 TFP .. .. .. .. 0.73 −0.25 −0.31 2.34 III. Growth accounting including the natural capital Output .. .. .. 4.77 .. .. .. 4.77 Physical capital .. .. .. 4.14 .. .. .. 1.61 Natural capital .. .. .. 125.11 .. .. .. 0.64 Human capital .. .. .. 1.03 .. .. .. 0.53 TFP .. .. .. .. .. .. .. 1.99 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 130   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA GABON Country Profile (2017) 8 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 23,004 US$, millions 6 GDP per capita (2011 PPP): 12,108 US$ 4 Population: 1.8 million 2 Employment: 0.5 million 0 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 4.7 4.9 43.3 12.5 34.6 Employment 41.3 3.7 9.2 31.7 14.0 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.8 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.6 0.6 1.0 0.4 0.4 0.5 0.2 0.2 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 GAB SSA East Asian Dragons GAB SSA East Asian Dragons GAB SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 2.0 0.6 Ratio relative to the United States Ratio relative to the United States 0.5 80 1.5 0.4 60 Percent 0.3 1.0 40 0.2 20 0.5 0.1 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.40 .. 0.43 0.37 0.40 .. 0.43 0.37 Physical capital −0.23 .. −1.83 0.86 −0.14 .. −1.14 0.53 Human capital 1.69 .. 2.37 1.67 0.73 .. 0.87 0.63 TFP .. .. .. .. −0.19 .. 0.70 −0.79 II. Growth accounting: private and public capital accumulation Output 0.40 .. 0.43 0.37 0.40 .. 0.43 0.37 Physical capital .. .. .. .. .. .. .. .. - Public 0.71 .. −0.76 1.72 0.21 .. −0.17 0.40 - Private −0.67 .. −2.19 0.37 −0.34 .. −0.82 0.14 Human capital 1.69 .. 2.37 1.67 0.83 .. 0.85 0.63 TFP .. .. .. .. −0.30 .. 0.57 −0.80 III. Growth accounting including the natural capital Output .. .. .. 0.37 .. .. .. 0.37 Physical capital .. .. .. 0.86 .. .. .. 0.14 Natural capital .. .. .. 6.11 .. .. .. 1.30 Human capital .. .. .. 1.67 .. .. .. 0.39 TFP .. .. .. .. .. .. .. −1.45 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    131 THE GAMBIA Country Profile (2017) GDP growth Population growth 7 population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 3,983 US$, millions 5 GDP per capita (2011 PPP): 1,563 US$ 3 1 Population: 2.1 million 0 -1 Employment: 0.6 million -3 -5 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 20.0 4.9 9.2 44.3 21.6 Employment 27.9 9.2 6.2 44.6 12.1 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 0.8 Ratio relative to the United States 1.5 Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 GMB SSA East Asian Dragons GMB SSA East Asian Dragons GMB SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.09 100 Ratio relative to the United States 0.6 Ratio relative to the United States 80 0.5 0.06 0.4 60 Percent 0.3 40 0.03 0.2 20 0.1 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.33 .. 0.15 0.46 0.33 .. 0.15 0.46 Physical capital 1.64 .. 1.30 1.88 0.71 .. 0.56 0.81 Human capital 0.76 .. 0.93 1.10 0.61 .. 0.57 0.63 TFP .. .. .. .. −0.98 .. −0.99 −0.98 II. Growth accounting: private and public capital accumulation Output 0.33 .. 0.15 0.46 0.33 .. 0.15 0.46 Physical capital .. .. .. .. .. .. .. .. - Public 0.58 .. 0.49 0.64 0.27 .. 0.16 0.14 - Private 3.87 .. 3.85 3.88 1.65 .. 1.11 0.80 Human capital 0.76 .. 0.93 1.10 0.90 .. 0.74 0.63 TFP .. .. .. .. −2.49 .. −1.85 −1.11 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 132   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA GHANA Country Profile (2017) 15 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 100,580 US$, millions 12 GDP per capita (2011 PPP): 3,543 US$ 9 Population: 28.6 million 6 3 Employment: 13.8 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 0.8 Ratio relative to the United States 2.0 Ratio relative to the United States Ratio relative to the United States 1.5 0.6 0.20 1.0 0.4 0.10 0.5 0.2 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 GHA SSA East Asian Dragons GHA SSA East Asian Dragons GHA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.16 100 Ratio relative to the United States 0.30 Ratio relative to the United States 0.14 80 0.25 0.12 0.20 0.10 60 Percent 0.08 0.15 40 0.06 0.10 0.04 20 0.05 0.02 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.64 −1.24 0.40 2.29 0.64 −1.24 0.40 2.29 Physical capital −0.69 −2.14 −1.89 1.41 −0.30 −0.92 −0.82 0.61 Human capital 1.43 1.56 1.94 0.91 0.81 0.89 1.10 0.52 TFP .. .. .. .. 0.13 −1.20 0.11 1.16 II. Growth accounting: private and public capital accumulation Output 0.64 −1.24 0.40 2.29 0.64 −1.24 0.40 2.29 Physical capital .. .. .. .. .. .. .. .. - Public −0.35 −1.96 −1.26 1.63 −0.08 −0.44 −0.28 0.37 - Private −0.81 −2.22 −2.20 1.40 −0.17 −0.46 −0.45 0.29 Human capital 1.43 1.56 1.94 0.91 0.81 0.89 1.10 0.52 TFP .. .. .. .. 0.07 −1.23 0.03 1.12 III. Growth accounting including the natural capital Output .. .. .. 2.29 .. .. .. 2.29 Physical capital .. .. .. 1.41 .. .. .. 0.51 Natural capital .. .. .. 22.32 .. .. .. 1.68 Human capital .. .. .. 0.91 .. .. .. 0.52 TFP .. .. .. .. .. .. .. −0.42 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    133 GUINEA Country Profile (2017) 12 GDP growth Population growth population growth (%) 10 GDP growth vis-à-vis GDP (2011 PPP): 23,868 US$, millions 8 GDP per capita (2011 PPP): 1,625 US$ 6 4 Population: 13.2 million 2 Employment: 6.2 million 0 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 GIN SSA East Asian Dragons GIN SSA East Asian Dragons GIN SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 0.09 100 Ratio relative to the United States 1.0 Ratio relative to the United States 0.08 80 0.8 0.07 0.06 60 0.6 Percent Info. not available Info. not available 0.05 0.04 40 0.4 0.03 0.02 20 0.2 0.01 0 0 0 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 0– 0– 0– 0– 0– 0– –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 6 7 8 9 0 1 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 134   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA GUINEA-BISSAU Country Profile (2017) 10 GDP growth population growth (%) 8 Population growth GDP growth vis-à-vis GDP (2011 PPP): 3,016 US$, millions 6 GDP per capita (2011 PPP): 1,623 US$ 4 2 Population: 1.9 million 0 Employment: 0.8 million -2 -4 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 GNB SSA East Asian Dragons GNB SSA East Asian Dragons GNB SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.35 Ratio relative to the United States 0.05 Ratio relative to the United States 80 0.30 0.04 0.25 60 Info. not available Percent 0.03 0.20 Info. not available 40 0.15 0.02 0.10 20 0.01 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 69 79 9 9 9 7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 0– 0– 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 6 7 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    135 KENYA Country Profile (2017) 10 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 155,045 US$, billions 8 GDP per capita (2011 PPP): 3,058 US$ 6 Population: 48.4 million 4 2 Employment: 19.2 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 25.0 11.0 9.9 21.0 33.0 Employment 37.3 11.3 3.4 32.5 15.5 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 1.5 0.8 Ratio relative to the United States 0.30 Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.15 0.9 0.4 0.10 0.6 0.2 0.05 0 0.3 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 15 60 80 85 90 95 00 05 10 70 75 65 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 20 20 20 19 19 19 KEN SSA East Asian Dragons KEN SSA East Asian Dragons KEN SSA East Asian Dragons Real output per worker (US = 1.0) Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) 0.10 0.35 Ratio relative to the United States Ratio relative to the United States 100 0.30 0.08 80 0.25 0.06 60 0.20 Percent 0.04 0.15 40 0.10 0.02 20 0.05 0 0 0 9 9 9 9 9 7 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.79 1.26 0.13 0.96 0.79 1.26 0.13 0.96 Physical capital −0.08 −0.60 −1.19 1.22 −0.03 −0.21 −0.42 0.43 Human capital 1.16 0.98 1.41 1.09 0.75 0.64 0.92 0.71 TFP .. .. .. .. 0.06 0.83 −0.37 −0.17 II. Growth accounting: private and public capital accumulation Output 0.79 1.26 0.13 0.96 0.79 1.26 0.13 0.96 Physical capital .. .. .. .. .. .. .. .. - Public 0.34 −0.47 −0.43 1.60 0.06 −0.09 −0.08 0.29 - Private −0.19 −0.65 −1.55 1.27 −0.03 −0.11 −0.26 0.21 Human capital 1.16 0.98 1.41 1.09 0.75 0.64 0.92 0.71 TFP .. .. .. .. 0.01 0.81 −0.44 −0.25 III. Growth accounting including the natural capital Output .. .. .. 0.96 .. .. .. 0.96 Physical capital .. .. .. 1.22 .. .. .. 0.46 Natural capital .. .. .. 135.42 .. .. .. 0.15 Human capital .. .. .. 1.09 .. .. .. 0.77 TFP .. .. .. .. .. .. .. −0.42 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 136   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA LESOTHO Country Profile (2017) 8 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 6,439 US$, millions 6 GDP per capita (2011 PPP): 2,533 US$ 4 Population: 2.2 million 2 0 Employment: 0.7 million -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 5.3 10.8 16.9 24.1 42.8 Employment 7.8 12.8 29.6 21.4 28.4 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 LSO SSA East Asian Dragons LSO SSA East Asian Dragons LSO SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.09 100 Ratio relative to the United States 0.25 Ratio relative to the United States 80 0.20 0.06 60 0.15 Percent 40 0.10 0.03 20 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 9 9 9 9 9 7 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 3.37 .. 3.95 2.98 3.37 .. 3.95 2.98 Physical capital 4.71 .. 7.08 3.09 1.47 .. 2.22 0.97 Human capital 0.38 .. 0.99 −0.37 0.14 .. 0.71 −0.25 TFP .. .. .. .. 1.76 .. 1.02 2.27 II. Growth accounting: private and public capital accumulation Output 3.37 .. 3.95 2.98 3.37 .. 3.95 2.98 Physical capital .. .. .. .. .. .. .. .. - Public 8.88 .. 14.65 4.95 1.40 .. 2.42 0.81 - Private 2.83 .. 4.78 1.50 0.41 .. 0.72 0.22 Human capital 0.38 .. 0.99 −0.37 0.25 .. 0.69 −0.25 TFP .. .. .. .. 1.31 .. 0.13 2.20 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    137 LIBERIA Country Profile (2017) 15 GDP growth population growth (%) 12 Population growth GDP growth vis-à-vis GDP (2011 PPP): 3,931 US$, millions 9 GDP per capita (2011 PPP): 841 US$ 6 Population: 4.7 million 3 Employment: 1.6 million 0 -3 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 0.8 Ratio relative to the United States 4.0 Ratio relative to the United States Ratio relative to the United States 0.25 3.0 0.6 0.20 0.15 2.0 0.4 0.10 1.0 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 LBR SSA East Asian Dragons LBR SSA East Asian Dragons LBR SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.05 100 Ratio relative to the United States 0.15 Ratio relative to the United States 0.04 80 0.12 60 Percent 0.03 0.09 40 0.06 0.02 20 0.03 0.01 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 19 19 19 19 20 20 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.34 .. −11.46 10.06 1.34 .. −11.46 10.06 Physical capital −1.55 .. −2.70 −0.77 −0.67 .. −1.16 −0.33 Human capital 0.94 .. 1.33 0.82 0.60 .. 0.80 0.46 TFP .. .. .. .. 1.41 .. −11.10 9.93 II. Growth accounting: private and public capital accumulation Output 1.34 .. −11.46 10.06 1.34 .. −11.46 10.06 Physical capital .. .. .. .. .. .. .. .. - Public −1.04 .. −2.24 −0.23 −0.25 .. −0.50 −0.05 - Private −1.57 .. −3.19 −0.47 −0.34 .. −0.65 −0.10 Human capital 0.94 .. 1.33 0.82 0.56 .. 0.75 0.46 TFP .. .. .. .. 1.36 .. −11.06 9.75 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 138   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA MADAGASCAR Country Profile (2017) GDP growth Population growth 4 population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 32,785 US$, millions 3 2 GDP per capita (2011 PPP): 1,297 US$ 1 0 Population: 25.6 million -1 -2 -3 Employment: 11.2 million -4 -5 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 24.2 11.9 11.9 26.8 25.1 Employment 72.3 6.4 2.8 10.4 8.1 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.5 0.6 0.20 0.15 1.0 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 MDG SSA East Asian Dragons MDG SSA East Asian Dragons MDG SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.05 100 Ratio relative to the United States 0.25 Ratio relative to the United States 0.04 80 0.20 60 Percent 0.03 0.15 0.02 40 0.10 0.01 20 0.05 0 0 0 69 79 89 99 09 17 9 9 9 9 9 7 9 9 9 9 9 7 0– 0– 0– 0– 0– 0– –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 6 7 8 9 0 1 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output −0.68 0.08 −1.86 −0.31 −0.68 0.08 −1.86 −0.31 Physical capital −0.90 −1.06 −1.94 0.08 −0.39 −0.46 −0.84 0.03 Human capital 0.59 0.08 0.87 0.75 0.33 0.04 0.50 0.43 TFP .. .. .. .. −0.63 0.49 −1.52 −0.77 II. Growth accounting: private and public capital accumulation Output −0.68 0.08 −1.86 −0.31 −0.68 0.08 −1.86 −0.31 Physical capital .. .. .. .. .. .. .. .. - Public −1.95 −1.04 −2.20 −2.45 −0.44 −0.23 −0.50 −0.55 - Private 0.95 −1.12 −1.03 4.16 0.20 −0.23 −0.21 0.86 Human capital 0.59 0.08 0.87 0.75 0.33 0.04 0.50 0.43 TFP .. .. .. .. −0.77 0.50 −1.65 −1.04 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    139 MALAWI Country Profile (2017) 8 GDP growth 7 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 17,923 US$, millions 6 5 GDP per capita (2011 PPP): 939 US$ 4 3 Population: 18.2 million 2 1 Employment: 7.1 million 0 -1 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 29.9 10.2 5.3 26.9 27.7 Employment 84.9 7.4 0.7 3.5 3.5 Development Accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.9 0.15 0.4 0.6 0.10 0.2 0.05 0.3 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 MWI SSA East Asian Dragons MWI SSA East Asian Dragons MWI SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.05 Ratio relative to the United States 0.35 Ratio relative to the United States 80 0.30 0.04 0.25 60 Percent 0.03 0.20 40 0.15 0.02 0.10 20 0.01 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 19 19 19 19 20 20 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.99 5.65 −0.86 1.50 1.99 5.65 −0.86 1.50 Physical capital 1.37 6.78 0.04 −1.72 0.59 2.93 0.02 −0.74 Human capital 0.69 0.05 0.51 1.33 0.39 0.03 0.29 0.76 TFP .. .. .. .. 1.01 2.70 −1.16 1.49 II. Growth accounting: private and public capital accumulation Output 1.99 5.65 −0.86 1.50 1.99 5.65 −0.86 1.50 Physical capital .. .. .. .. .. .. .. .. - Public 2.09 6.84 0.97 −0.66 0.47 1.54 0.22 −0.15 - Private 1.08 6.76 −0.45 −2.07 0.22 1.39 −0.09 −0.43 Human capital 0.69 0.05 0.51 1.33 0.39 0.03 0.29 0.76 TFP .. .. .. .. 0.91 2.69 −1.28 1.32 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 140   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA MALI Country Profile (2017) 10 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 31,887 US$, millions 8 GDP per capita (2011 PPP): 1,742 US$ 6 Population: 18.7 million 4 2 Employment: 6.1 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 45.2 16.2 6.4 20.0 12.2 Employment 66.1 4.2 3.1 18.4 8.2 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.5 0.6 0.20 0.15 1.0 0.4 0.10 0.5 0.2 0.05 0 0 0 65 70 75 80 85 90 95 00 05 10 15 60 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 MLI SSA East Asian Dragons MLI SSA East Asian Dragons MLI SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.05 100 Ratio relative to the United States 0.30 Ratio relative to the United States 0.04 80 0.25 0.20 0.03 60 Percent 0.15 0.02 40 0.10 0.01 20 0.05 0 0 0 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.45 1.19 1.68 1.48 1.45 1.19 1.68 1.48 Physical capital 1.30 0.98 2.45 0.61 0.56 0.42 1.06 0.26 Human capital 0.49 0.13 0.44 0.81 0.28 0.07 0.25 0.46 TFP .. .. .. .. 0.62 0.69 0.37 0.75 II. Growth accounting: private and public capital accumulation Output 1.45 1.19 1.68 1.48 1.45 1.19 1.68 1.48 Physical capital .. .. .. .. .. .. .. .. - Public 0.45 0.97 1.59 −0.89 0.10 0.22 0.36 −0.20 - Private 2.16 0.98 3.70 1.80 0.44 0.20 0.76 0.37 Human capital 0.49 0.13 0.44 0.81 0.28 0.07 0.25 0.46 TFP .. .. .. .. 0.63 0.69 0.31 0.85 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    141 MAURITANIA Country Profile (2017) 8 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 14,474 US$, millions 6 GDP per capita (2011 PPP): 3,397 US$ 4 Population: 4.3 million 2 0 Employment: 0.9 million -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 14.1 5.6 50.6 11.9 17.9 Employment 75.8 6.0 1.7 11.1 5.4 Development Accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.3 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.6 0.2 1.0 0.4 0.1 0.5 0.2 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 MRT SSA East Asian Dragons MRT SSA East Asian Dragons MRT SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.25 100 Ratio relative to the United States 0.6 Ratio relative to the United States 0.20 80 0.5 0.4 60 Percent 0.15 0.3 0.10 40 0.2 0.05 20 0.1 0 0 0 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.13 .. −0.86 0.94 0.13 .. −0.86 0.94 Physical capital 1.59 .. −0.79 3.54 0.68 .. −0.34 1.52 Human capital 0.70 .. 0.65 1.06 0.50 .. 0.37 0.60 TFP .. .. .. .. −1.05 .. −0.89 −1.18 II. Growth accounting: private and public capital accumulation Output 0.13 .. −0.86 0.94 0.13 .. −0.86 0.94 Physical capital .. .. .. .. .. .. .. .. - Public 2.25 .. 0.00 4.10 2.08 .. 0.00 0.92 - Private 1.54 .. −1.09 3.69 1.30 .. −0.22 0.75 Human capital 0.70 .. 0.65 1.06 1.65 .. 0.37 0.60 TFP .. .. .. .. −4.89 .. −1.01 −1.34 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 142   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA MAURITIUS Country Profile (2017) 5 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 27,962 US$, millions 4 GDP per capita (2011 PPP): 21,357 US$ 3 Population: 1.3 million 2 1 Employment: 0.6 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 3.7 14.1 6.5 30.3 45.4 Employment 7.1 13.1 11.4 43.4 24.8 Development accounting Capital output ratio (US = 1.0) Relative labor productivity (US = 1.0) Human Capital Index (US = 1.0) 0.5 1.50 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.7 0.4 0.6 1.00 0.3 0.5 0.4 0.2 0.3 0.50 0.2 0.1 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 MUS SSA East Asian Dragons MUS SSA East Asian Dragons MUS SSA East Asian Dragons Real output per worker (US = 1.0) Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) 100 1.0 0.40 Ratio relative to the United States Ratio relative to the United States 80 0.8 0.30 60 Percent 0.6 0.20 40 0.4 0.10 20 0.2 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 3.20 .. 3.57 2.95 3.20 .. 3.57 2.95 Physical capital 2.65 .. 1.91 3.15 1.35 .. 0.97 1.60 Human capital 1.12 .. 1.38 0.85 0.53 .. 0.68 0.42 TFP .. .. .. .. 1.33 .. 1.92 0.92 II. Growth accounting: private and public capital accumulation Output 3.20 .. 3.57 2.95 3.20 .. 3.57 2.95 Physical capital .. .. .. .. .. .. .. .. - Public 2.64 .. 2.78 2.55 0.49 .. 0.52 0.48 - Private 2.65 .. 1.32 3.56 0.84 .. 0.42 1.14 Human capital 1.12 .. 1.38 0.85 0.55 .. 0.68 0.42 TFP .. .. .. .. 1.32 .. 1.94 0.91 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    143 MOZAMBIQUE Country Profile (2017) 8 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 39,075 US$, millions 6 GDP per capita (2011 PPP): 1,123 US$ 4 Population: 29.7 million 2 Employment: 12.8 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 23.6 9.4 10.1 28.3 28.7 Employment 73.2 0.5 3.8 18.4 4.1 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.3 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.6 0.2 1.0 0.4 0.1 0.5 0.2 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 MOZ SSA East Asian Dragons MOZ SSA East Asian Dragons MOZ SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.03 100 0.20 Ratio relative to the United States Ratio relative to the United States 80 0.15 0.02 60 Percent 0.10 40 0.01 20 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 2.34 1.97 −0.89 5.27 2.34 1.97 −0.89 5.27 Physical capital 2.87 −0.03 0.97 6.67 1.66 −0.02 0.56 3.86 Human capital 0.21 0.17 0.01 0.41 0.09 0.07 0.00 0.17 TFP .. .. .. .. 0.59 1.92 −1.46 1.25 II. Growth accounting: private and public capital accumulation Output 2.34 1.97 −0.89 5.27 2.34 1.97 −0.89 5.27 Physical capital .. .. .. .. .. .. .. .. - Public 3.65 −0.04 2.32 7.59 1.10 −0.01 0.70 2.29 - Private 3.05 −0.02 −0.05 7.98 0.84 −0.01 −0.01 2.20 Human capital 0.21 0.17 0.01 0.41 0.09 0.07 0.00 0.17 TFP .. .. .. .. 0.31 1.92 −1.59 0.61 III. Growth accounting including the natural capital Output .. .. .. 5.27 .. .. .. 5.27 Physical capital .. .. .. 6.67 .. .. .. 3.34 Natural capital .. .. .. 70.49 .. .. .. 1.22 Human capital .. .. .. 0.41 .. .. .. 0.15 TFP .. .. .. .. .. .. .. 0.56 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 144   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA NAMIBIA Country Profile (2017) 8 GDP growth 7 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 26,751 US$, millions 6 5 GDP per capita (2011 PPP): 10,905 US$ 4 3 Population: 2.6 million 2 1 Employment: 0.7 million 0 -1 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 6.4 10.3 16.1 22.9 44.3 Employment 25.2 5.1 11.9 31.0 26.7 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.4 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.7 0.3 0.6 1.0 0.5 0.2 0.4 0.3 0.5 0.1 0.2 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 NAM SSA East Asian Dragons NAM SSA East Asian Dragons NAM SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.35 Ratio relative to the United States 0.8 Ratio relative to the United States 0.30 80 0.6 0.25 60 Percent 0.20 0.4 0.15 40 0.10 20 0.2 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.76 .. −1.71 2.43 0.76 .. −1.71 2.43 Physical capital 1.92 .. −2.50 4.93 0.81 .. −1.06 2.09 Human capital 0.69 .. 0.95 0.44 0.37 .. 0.55 0.25 TFP .. .. .. .. −0.43 .. −1.20 0.09 II. Growth accounting: private and public capital accumulation Output 0.76 .. −1.71 2.43 0.76 .. −1.71 2.43 Physical capital .. .. .. .. .. .. .. .. - Public 1.47 .. −2.01 3.84 0.22 .. −0.31 0.60 - Private 2.58 .. −2.98 6.38 0.67 .. −0.80 1.71 Human capital 0.69 .. 0.95 0.44 0.38 .. 0.55 0.25 TFP .. .. .. .. −0.52 .. −1.14 −0.12 III. Growth accounting including the natural capital Output .. .. .. 2.43 .. .. .. 2.43 Physical capital .. .. .. 4.93 .. .. .. 1.75 Natural capital .. .. .. 17.68 .. .. .. 0.35 Human capital .. .. .. 0.44 .. .. .. 0.22 TFP .. .. .. .. .. .. .. 0.11 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    145 NIGER Country Profile (2017) 12 GDP growth population growth (%) 10 Population growth GDP growth vis-à-vis GDP (2011 PPP): 18,851 US$, millions 8 GDP per capita (2011 PPP): 853 US$ 6 4 Population: 21.4 million 2 Employment: 7.7 million 0 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 40.5 6.6 15.1 19.0 18.8 Employment 76.0 6.9 0.7 11.3 5.1 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.5 0.6 0.20 0.15 1.0 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 19 0 19 5 19 0 19 5 80 19 5 19 0 20 5 00 20 5 20 0 15 60 65 70 75 80 85 90 95 00 05 10 15 6 6 7 7 8 9 9 0 1 19 19 19 19 19 19 19 19 20 20 20 20 19 19 20 19 19 19 19 19 19 19 19 20 20 20 20 NER SSA East Asian Dragons NER SSA East Asian Dragons NER SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.10 100 Ratio relative to the United States 0.20 Ratio relative to the United States 0.08 80 0.15 60 Percent 0.06 0.10 40 0.04 20 0.05 0.02 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 69 79 89 99 09 17 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 6 7 8 9 0 1 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output −0.58 −0.56 −2.15 0.68 −0.58 −0.56 −2.15 0.68 Physical capital −1.90 −2.19 −2.45 −1.22 −0.82 −0.95 −1.07 −0.53 Human capital 0.32 0.04 0.37 0.50 0.18 0.03 0.21 0.28 TFP .. .. .. .. 0.06 0.37 −1.30 0.93 II. Growth accounting: private and public capital accumulation Output −0.58 −0.56 −2.15 0.68 −0.58 −0.56 −2.15 0.68 Physical capital .. .. .. .. .. .. .. .. - Public 0.32 −1.98 1.73 0.95 0.07 −0.45 0.39 0.21 - Private −2.40 −2.21 −3.31 −1.81 −0.50 −0.46 −0.69 −0.38 Human capital 0.32 0.04 0.37 0.50 0.18 0.03 0.21 0.28 TFP .. .. .. .. −0.34 0.32 −2.07 0.56 III. Growth accounting including the natural capital Output .. .. .. 0.68 .. .. .. 0.68 Physical capital .. .. .. −1.22 .. .. .. −0.39 Natural capital .. .. .. 22.42 .. .. .. 0.51 Human capital .. .. .. 0.50 .. .. .. 0.22 TFP .. .. .. .. .. .. .. 0.34 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 146   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA NIGERIA Country Profile (2017) 8 GDP growth 7 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 832,983 US$, millions 6 5 GDP per capita (2011 PPP): 4,391 US$ 4 3 Population: 192.0 million 2 1 0 Employment: 65.4 million -1 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 24.5 9.3 12.6 30.9 22.8 Employment 33.6 8.6 4.0 35.0 18.8 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.9 0.15 0.4 0.6 0.10 0.2 0.3 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 NGA SSA East Asian Dragons NGA SSA East Asian Dragons NGA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.25 100 Ratio relative to the United States 4.0 Ratio relative to the United States 3.5 0.20 80 3.0 60 2.5 Percent 0.15 2.0 0.10 40 1.5 20 1.0 0.05 0.5 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 69 79 89 99 09 17 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 6 7 8 9 0 1 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.65 1.87 0.19 2.68 1.65 1.87 0.19 2.68 Physical capital 3.70 6.09 3.38 2.12 2.57 4.23 2.35 1.47 Human capital 0.91 0.15 0.69 1.67 0.28 0.04 0.21 0.51 TFP .. .. .. .. −1.19 −2.41 −2.37 0.70 II. Growth accounting: private and public capital accumulation Output 1.65 1.87 0.19 2.68 1.65 1.87 0.19 2.68 Physical capital .. .. .. .. .. .. .. .. - Public 3.39 6.13 4.04 0.74 1.23 2.23 1.47 0.27 - Private 3.93 6.05 2.84 3.19 1.31 2.01 0.94 1.06 Human capital 0.91 0.15 0.69 1.67 0.28 0.04 0.21 0.51 TFP .. .. .. .. −1.16 −2.41 −2.43 0.85 III. Growth accounting including the natural capital Output .. .. .. 2.68 .. .. .. 2.68 Physical capital .. .. .. 2.12 .. .. .. 0.80 Natural capital .. .. .. −0.61 .. .. .. −0.09 Human capital .. .. .. 1.67 .. .. .. 0.39 TFP .. .. .. .. .. .. .. 1.58 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    147 RWANDA Country Profile (2017) 10 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 23,143 US$, millions 8 GDP per capita (2011 PPP): 1,830 US$ 6 Population: 12.2 million 4 2 Employment: 6.0 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 26.4 6.5 12.4 18.1 36.6 Employment 66.9 2.3 6.2 16.0 8.6 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 90 00 10 15 60 65 70 75 80 85 95 05 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 RWA SSA East Asian Dragons RWA SSA East Asian Dragons RWA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.04 0.30 Ratio relative to the United States Ratio relative to the United States 80 0.25 0.03 60 0.20 Percent 0.02 0.15 40 0.10 0.01 20 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 69 79 89 99 09 17 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– – – – – – – 19 19 19 19 20 20 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 2.21 .. −1.69 4.87 2.21 .. −1.69 4.87 Physical capital 3.92 .. 3.43 4.25 0.94 .. 0.82 1.02 Human capital 0.99 .. 0.68 1.58 0.92 .. 0.52 1.20 TFP .. .. .. .. 0.35 .. −3.03 2.65 II. Growth accounting: private and public capital accumulation Output 2.21 .. −1.69 4.87 2.21 .. −1.69 4.87 Physical capital .. .. .. .. .. .. .. .. - Public 3.66 .. 3.89 3.51 0.50 .. 0.49 0.44 - Private 4.02 .. 2.97 4.74 0.50 .. 0.34 0.54 Human capital 0.99 .. 0.68 1.58 0.82 .. 0.51 1.20 TFP .. .. .. .. 0.40 .. −3.03 2.69 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 148   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA SÃO TOMÉ AND PRÍNCIPE Country Profile (2017) GDP growth 5 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 709 US$, millions 4 GDP per capita (2011 PPP): 3,456 US$ 3 Population: 0.2 million 2 Employment: 0.1 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development Accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 2.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 2.0 0.6 0.20 1.5 0.15 0.4 1.0 0.10 0.2 0.05 0.5 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 STP SSA East Asian Dragons STP SSA East Asian Dragons STP SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.12 100 Ratio relative to the United States 1.0 Ratio relative to the United States 0.10 80 0.8 0.08 60 Info. not available Percent 0.6 Info. not available 0.06 40 0.4 0.04 20 0.2 0.02 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    149 SENEGAL Country Profile (2017) 8 GDP growth 7 Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 40,828 US$, millions 6 5 GDP per capita (2011 PPP): 2,423 US$ 4 Population: 16.0 million 3 2 Employment: 5.4 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 SEN SSA East Asian Dragons SEN SSA East Asian Dragons SEN SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.15 100 Ratio relative to the United States 0.35 Ratio relative to the United States 80 0.30 0.12 0.25 60 Percent 0.09 0.20 40 0.15 0.06 0.10 20 0.03 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 19 19 19 19 20 20 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output −0.45 −1.38 −1.68 1.28 −0.45 −1.38 −1.68 1.28 Physical capital −1.62 −4.12 −2.65 1.16 −1.03 −2.59 −1.63 0.68 Human capital 0.72 0.15 0.81 1.08 0.29 0.06 0.31 0.45 TFP .. .. .. .. 0.29 1.15 −0.36 0.15 II. Growth accounting: private and public capital accumulation Output −0.45 −1.38 −1.68 1.28 −0.45 −1.38 −1.68 1.28 Physical capital .. .. .. .. .. .. .. .. - Public −1.59 −4.09 −2.82 1.36 −0.53 −1.35 −0.91 0.41 - Private −1.59 −4.13 −2.58 1.18 −0.48 −1.24 −0.76 0.33 Human capital 0.72 0.15 0.81 1.08 0.29 0.06 0.31 0.45 TFP .. .. .. .. 0.27 1.15 −0.33 0.09 III. Growth accounting including the natural capital Output .. .. .. 1.28 .. .. .. 1.28 Physical capital .. .. .. 1.16 .. .. .. 0.71 Natural capital .. .. .. 42.71 .. .. .. 0.65 Human capital .. .. .. 1.08 .. .. .. 0.48 TFP .. .. .. .. .. .. .. −0.55 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 150   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA THE SEYCHELLES Country Profile (2017) 10 GDP growth Population growth population growth (%) 8 GDP growth vis-à-vis GDP (2011 PPP): 3,200 US$, millions 6 GDP per capita (2011 PPP): 27,836 US$ 4 Population: 0.1 million 2 Employment: 0.0 million 0 -2 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development Accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.8 2.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.7 0.7 0.6 1.5 0.6 0.5 0.5 0.4 1.0 0.4 0.3 0.3 0.2 0.5 0.2 0.1 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 SYC SSA East Asian Dragons SYC SSA East Asian Dragons SYC SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.5 100 1.0 Ratio relative to the United States 0.4 80 0.8 60 Info. not available Percent 0.3 0.6 Percent Info. not available 0.2 40 0.4 0.1 20 0.2 0 0 0 9 9 9 9 9 7 69 79 9 99 09 7 –6 –7 –8 –9 –0 –1 –8 –1 69 79 89 99 09 17 0– 0– 0– 0– 60 70 80 90 00 10 80 10 0– 0– 0– 0– 0– 0– 6 7 9 0 19 19 19 19 20 20 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. II. Growth accounting: private and public capital accumulation Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. - Public .. .. .. .. .. .. .. .. - Private .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    151 SIERRA LEONE Country Profile (2017) 25 GDP growth Population growth population growth (%) GDP growth vis-à-vis GDP (2011 PPP): 9,567 US$, millions 15 GDP per capita (2011 PPP): 1,458 US$ 5 0 Population: 6.7 million -5 -15 Employment: 2.5 million -25 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 53.0 2.1 9.5 14.6 20.8 Employment 59.1 3.1 3.2 29.4 5.3 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.7 0.6 0.20 1.0 0.5 0.15 0.4 0.5 0.3 0.10 0.2 0.05 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 SLE SSA East Asian Dragons SLE SSA East Asian Dragons SLE SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 0.06 100 Ratio relative to the United States 1.0 Ratio relative to the United States 0.05 80 0.8 0.04 60 Percent 0.6 0.03 40 0.4 0.02 20 0.2 0.01 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 9 9 9 9 9 7 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.00 .. −0.99 0.67 0.00 .. −0.99 0.67 Physical capital −0.28 .. −0.17 −0.35 −0.13 .. −0.08 −0.16 Human capital 0.74 .. 0.85 0.99 0.51 .. 0.48 0.53 TFP .. .. .. .. −0.38 .. −1.39 0.31 II. Growth accounting: private and public capital accumulation Output 0.00 .. −0.99 0.67 0.00 .. −0.99 0.67 Physical capital .. .. .. .. .. .. .. .. - Public 0.23 .. 0.81 −0.16 0.00 .. 0.19 −0.04 - Private −0.16 .. −0.97 0.40 0.00 .. −0.21 0.09 Human capital 0.74 .. 0.85 0.99 0.02 .. 0.44 0.53 TFP .. .. .. .. −0.02 .. −1.42 0.10 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 152   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA SOUTH AFRICA Country Profile (2017) 3.5 GDP growth 3.0 population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 666,245 US$, millions 2.5 2.0 1.5 GDP per capita (2011 PPP): 12,105 US$ 1.0 0.5 Population: 56.1 million 0 -0.5 Employment: 19.6 million -1.0 -1.5 -2.0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 2.4 13.7 14.3 24.6 45.0 Employment 4.5 11.1 12.1 40.9 31.3 Development Accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 1.5 0.8 Ratio relative to the United States Ratio relative to the United States 0.6 Ratio relative to the United States 0.5 0.6 0.4 1.0 0.3 0.4 0.2 0.5 0.2 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 ZAF SSA East Asian Dragons ZAF SSA East Asian Dragons ZAF SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 100 0.5 1.2 Ratio relative to the United States Ratio relative to the United States 80 1.0 0.4 60 0.8 Percent 0.3 0.6 40 0.2 0.4 20 0.1 0.2 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 69 79 89 99 09 17 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 6 7 8 9 0 1 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.16 3.62 −0.48 0.60 1.16 3.62 −0.48 0.60 Physical capital 1.70 4.66 0.08 0.75 0.71 1.94 0.03 0.31 Human capital 0.85 0.35 0.44 1.57 0.49 0.20 0.26 0.91 TFP .. .. .. .. −0.05 1.47 −0.76 −0.63 II. Growth accounting: private and public capital accumulation Output 1.16 3.62 −0.48 0.60 1.16 3.62 −0.48 0.60 Physical capital .. .. .. .. .. .. .. .. - Public 2.07 5.34 1.87 −0.29 0.32 0.82 0.29 −0.04 - Private 1.65 4.46 −0.66 1.36 0.43 1.17 −0.17 0.36 Human capital 0.85 0.35 0.44 1.57 0.49 0.20 0.26 0.91 TFP .. .. .. .. −0.09 1.42 −0.85 −0.63 III. Growth accounting including the natural capital Output .. .. .. 0.60 .. .. .. 0.60 Physical capital .. .. .. 0.75 .. .. .. 0.18 Natural capital .. .. .. 5.28 .. .. .. 0.24 Human capital .. .. .. 1.57 .. .. .. 0.64 TFP .. .. .. .. .. .. .. −0.46 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    153 SUDAN Country Profile (2017) 10 GDP growth population growth (%) 8 Population growth GDP growth vis-à-vis GDP (2011 PPP): 213,717 US$, millions 6 GDP per capita (2011 PPP): 4,027 US$ 4 2 Population: 54.0 million 0 Employment: 13.7 million -2 -4 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development Accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.9 0.15 0.4 0.6 0.10 0.2 0.3 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 SDN SSA East Asian Dragons SDN SSA East Asian Dragons SDN SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 1.0 0.15 Ratio relative to the United States Ratio relative to the United States 80 0.8 0.10 60 Percent 0.6 40 0.4 0.05 20 0.2 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 69 79 89 99 09 17 69 79 89 99 09 17 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 6 7 8 9 0 1 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.16 3.59 −0.65 1.87 1.16 3.59 −0.65 1.87 Physical capital 3.28 −1.92 −1.55 8.89 0.69 −0.40 −0.32 1.86 Human capital 0.82 0.37 0.89 0.91 0.65 0.29 0.71 0.72 TFP .. .. .. .. −0.17 3.70 −1.03 −0.71 II. Growth accounting: private and public capital accumulation Output 1.16 3.59 −0.65 1.87 1.16 3.59 −0.65 1.87 Physical capital .. .. .. .. .. .. .. .. - Public 5.79 −1.91 −0.96 13.77 0.63 −0.21 −0.10 1.50 - Private 2.79 −1.92 −1.61 7.89 0.28 −0.19 −0.16 0.79 Human capital 0.82 0.37 0.89 0.91 0.65 0.29 0.71 0.72 TFP .. .. .. .. −0.40 3.70 −1.09 −1.14 III. Growth accounting including the natural capital Output .. .. .. 1.87 .. .. .. 1.87 Physical capital .. .. .. 8.89 .. .. .. 0.57 Natural capital .. .. .. 80.20 .. .. .. 9.82 Human capital .. .. .. 0.91 .. .. .. 0.65 TFP .. .. .. .. .. .. .. −9.16 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 154   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA TANZANIA Country Profile (2017) 8 GDP growth Population growth population growth (%) 7 GDP growth vis-à-vis GDP (2011 PPP): 134,398 US$, millions 6 GDP per capita (2011 PPP): 2,368 US$ 5 4 Population: 55.3 million 3 2 Employment: 24.5 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 3.0 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 2.5 0.6 0.20 2.0 0.15 1.5 0.4 0.10 1.0 0.2 0.05 0.5 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 TZA SSA East Asian Dragons TZA SSA East Asian Dragons TZA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 100 0.08 0.25 Ratio relative to the United States Ratio relative to the United 80 0.20 0.06 60 Percent 0.15 0.04 40 0.10 20 0.02 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 9 9 9 9 9 7 69 79 89 99 09 17 19 19 19 19 20 20 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.69 1.97 0.12 2.74 1.69 1.97 0.12 2.74 Physical capital 1.38 2.69 −1.11 2.41 0.67 1.30 −0.54 1.16 Human capital 0.46 0.01 0.39 0.85 0.24 0.01 0.20 0.44 TFP .. .. .. .. 0.79 0.67 0.46 1.14 II. Growth accounting: private and public capital accumulation Output 1.69 1.97 0.12 2.74 1.69 1.97 0.12 2.74 Physical capital .. .. .. .. .. .. .. .. - Public 0.58 2.70 −0.62 −0.08 0.15 0.68 −0.16 −0.02 - Private 1.61 2.69 −1.41 3.26 0.37 0.62 −0.32 0.75 Human capital 0.46 0.01 0.39 0.85 0.24 0.01 0.20 0.44 TFP .. .. .. .. 0.93 0.67 0.40 1.57 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    155 TOGO Country Profile (2017) 8 GDP growth 7 population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 12,101 US$, millions 6 5 GDP per capita (2011 PPP): 1,515 US$ 4 Population: 7.7 million 3 2 Employment: 3.3 million 1 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 27.7 10.1 6.9 22.6 32.7 Employment 39.1 13.3 4.2 31.6 11.8 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.30 1.5 0.8 Ratio relative to the United States Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 TGO SSA East Asian Dragons TGO SSA East Asian Dragons TGO SSA East Asian Dragons Share explained by factor accumulation and TFP (%) TFP gap (US = 1.0) Real output per worker (US = 1.0) 100 0.06 Ratio relative to the United States 0.25 Ratio relative to the United States 0.05 80 0.20 0.04 60 Percent 0.15 0.03 40 0.10 0.02 20 0.05 0.01 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 19 19 19 19 20 20 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output −0.92 .. −2.24 −0.01 −0.92 .. −2.24 −0.01 Physical capital −1.25 .. −2.90 −0.12 −0.24 .. −0.56 −0.02 Human capital 0.95 .. 1.67 0.50 0.77 .. 1.31 0.40 TFP .. .. .. .. −1.45 .. −2.99 −0.39 II. Growth accounting: private and public capital accumulation Output −0.92 .. −2.24 −0.01 −0.92 .. −2.24 −0.01 Physical capital .. .. .. .. .. .. .. .. - Public −2.31 .. −2.77 −1.99 −0.23 .. −0.29 −0.20 - Private 0.02 .. −3.07 2.14 0.00 .. −0.29 0.20 Human capital 0.95 .. 1.67 0.50 0.76 .. 1.37 0.40 TFP .. .. .. .. −1.44 .. −3.04 −0.41 III. Growth accounting including the natural capital Output .. .. .. −0.01 .. .. .. −0.01 Physical capital .. .. .. −0.12 .. .. .. 0.00 Natural capital .. .. .. 114.33 .. .. .. 0.38 Human capital .. .. .. 0.50 .. .. .. 0.01 TFP .. .. .. .. .. .. .. −0.41 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 156   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA UGANDA Country Profile (2017) 10 GDP growth population growth (%) Population growth GDP growth vis-à-vis GDP (2011 PPP): 76,579 US$, millions 8 GDP per capita (2011 PPP): 1,800 US$ 6 Population: 41.7 million 4 2 Employment: 16.1 million 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 23.0 9.6 12.2 27.4 27.8 Employment 70.3 4.2 2.8 15.1 7.6 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 1.5 0.8 Ratio relative to the United States 0.30 Ratio relative to the United States Ratio relative to the United States 0.25 1.2 0.6 0.20 0.9 0.15 0.4 0.6 0.10 0.2 0.3 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 UGA SSA East Asian Dragons UGA SSA East Asian Dragons UGA SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 100 Ratio relative to the United States 0.25 0.05 Ratio relative to the United States 80 0.20 0.04 60 Percent 0.03 0.15 40 0.10 0.02 20 0.05 0.01 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 69 79 89 99 09 17 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 0– 0– 0– 0– 0– 0– 60 70 80 90 00 10 19 19 19 19 20 20 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 1.32 0.69 0.03 2.86 1.32 0.69 0.03 2.86 Physical capital 3.34 3.63 0.70 5.27 1.44 1.57 0.30 2.27 Human capital 1.25 0.44 1.35 1.79 0.71 0.25 0.77 1.02 TFP .. .. .. .. −0.83 −1.13 −1.04 −0.43 II. Growth accounting: private and public capital accumulation Output 1.32 0.69 0.03 2.86 1.32 0.69 0.03 2.86 Physical capital .. .. .. .. .. .. .. .. - Public 4.96 3.68 6.07 5.03 1.12 0.83 1.37 1.13 - Private 2.98 3.62 −0.59 5.40 0.61 0.75 −0.12 1.11 Human capital 1.25 0.44 1.35 1.79 0.71 0.25 0.77 1.02 TFP .. .. .. .. −1.12 −1.14 −1.98 −0.40 III. Growth accounting including the natural capital Output .. .. .. .. .. .. .. .. Physical capital .. .. .. .. .. .. .. .. Natural capital .. .. .. .. .. .. .. .. Human capital .. .. .. .. .. .. .. .. TFP .. .. .. .. .. .. .. .. Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. C O U N T R Y P R O D U C T I V I T Y A N A L Y S I S I N S U B - S A H A R A N A F R I C A    157 ZAMBIA Country Profile (2017) 12 GDP growth population growth (%) Population growth GDP growth vis-à-vis 10 GDP (2011 PPP): 64,419 US$, millions 8 GDP per capita (2011 PPP): 3,913 US$ 6 Population: 17.2 million 4 Employment: 4.8 million 2 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares, 2016 (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added 7.8 8.5 24.1 33.0 26.6 Employment 51.7 4.6 7.2 20.5 16.0 Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 5 0.8 Ratio relative to the United States Ratio relative to the United States 0.30 Ratio relative to the United States 0.7 0.25 4 0.6 0.20 0.5 3 0.15 0.4 2 0.3 0.10 1 0.2 0.05 0.1 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 20 20 20 20 ZMB SSA East Asian Dragons ZMB SSA East Asian Dragons ZMB SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 100 0.40 Ratio relative to the United States Ratio relative to the United States 0.15 80 0.35 0.12 0.30 Percent 60 0.25 0.09 0.20 40 0.06 0.15 20 0.10 0.03 0.05 0 0 0 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 69 79 89 99 09 17 60 70 80 90 00 10 –6 –7 –8 –9 –0 –1 0– 0– 0– 0– 0– 0– 19 19 19 19 20 20 60 70 80 90 00 10 6 7 8 9 0 1 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.81 1.09 −1.92 2.82 0.81 1.09 −1.92 2.82 Physical capital −1.48 −0.27 −4.33 −0.08 −0.64 −0.12 −1.87 −0.03 Human capital 1.26 1.02 1.82 0.98 0.72 0.58 1.04 0.56 TFP .. .. .. .. 0.73 0.63 −1.09 2.29 II. Growth accounting: private and public capital accumulation Output 0.81 1.09 −1.92 2.82 0.81 1.09 −1.92 2.82 Physical capital .. .. .. .. .. .. .. .. - Public −1.59 −0.14 −3.25 −1.34 −0.36 −0.03 −0.73 −0.30 - Private −1.04 −0.42 −5.93 2.48 −0.21 −0.09 −1.22 0.51 Human capital 1.26 1.02 1.82 0.98 0.72 0.58 1.04 0.56 TFP .. .. .. .. 0.67 0.64 −1.00 2.05 III. Growth accounting including the natural capital Output .. .. .. 2.82 .. .. .. 2.82 Physical capital .. .. .. −0.08 .. .. .. −0.02 Natural capital .. .. .. 26.46 .. .. .. 3.30 Human capital .. .. .. 0.98 .. .. .. 0.44 TFP .. .. .. .. .. .. .. −0.91 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. 158   BOOSTING PRODUCTIVITY IN SUB-SAHARAN AFRICA ZIMBABWE Country Profile (2017) 60 GDP growth population growth (%) Population growth GDP growth vis-à-vis 50 GDP (2011 PPP): 31,184 US$, millions 40 GDP per capita (2011 PPP): 1,945 US$ 30 Population: 16.4 million 20 Employment: 7.3 million 10 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sectoral shares (%) Agriculture Manufacturing Nonmanufacturing activities Market services Nonmarket services Value added .. .. .. .. .. Employment .. .. .. .. .. Development accounting Relative labor productivity (US = 1.0) Capital output ratio (US = 1.0) Human Capital Index (US = 1.0) 0.8 Ratio relative to the United States 0.30 1.5 Ratio relative to the United States Ratio relative to the United States 0.25 0.6 0.20 1.0 0.15 0.4 0.10 0.5 0.2 0.05 0 0 0 60 65 70 75 80 85 90 95 00 05 10 15 15 60 65 70 75 80 85 90 95 00 05 10 60 65 70 75 80 85 90 95 00 05 10 15 19 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 20 20 20 20 19 19 19 19 19 19 19 19 19 20 20 20 20 ZWE SSA East Asian Dragons ZWE SSA East Asian Dragons ZWE SSA East Asian Dragons Share explained by factor accumulation and TFP (%) Real output per worker (US = 1.0) TFP gap (US = 1.0) 100 1.0 Ratio relative to the United States 0.15 Ratio relative to the United States 80 0.8 0.10 60 Percent 0.6 40 0.4 0.05 20 0.2 0 0 0 9 9 9 9 9 7 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 9 9 9 9 9 7 –6 –7 –8 –9 –0 –1 –6 –7 –8 –9 –0 –1 60 70 80 90 00 10 60 70 80 90 00 10 60 70 80 90 00 10 19 19 19 19 20 20 19 19 19 19 20 20 19 19 19 19 20 20 Factor accumulation TFP Growth accounting (% per year) Observed annual growth rates Contribution to output growth per worker 1961–2017 1961–1977 1978–1995 1996–2017 1961–2017 1961–1977 1978–1995 1996–2017 I. Traditional growth accounting Output 0.33 .. −0.82 1.11 0.33 .. −0.82 1.11 Physical capital −1.89 .. −3.17 −1.02 −0.72 .. −1.21 −0.39 Human capital 1.34 .. 1.60 1.29 0.91 .. 1.07 0.80 TFP .. .. .. .. 0.15 .. −0.67 0.71 II. Growth accounting: private and public capital accumulation Output 0.33 .. −0.82 1.11 0.33 .. −0.82 1.11 Physical capital .. .. .. .. .. .. .. .. - Public −1.53 .. −3.02 −0.51 −0.41 .. −0.55 −0.10 - Private −2.02 .. −3.21 −1.21 −0.49 .. −0.53 −0.22 Human capital 1.34 .. 1.60 1.29 1.10 .. 0.90 0.80 TFP .. .. .. .. 0.13 .. −0.63 0.64 III. Growth accounting including the natural capital Output .. .. .. 1.11 .. .. .. 1.11 Physical capital .. .. .. −1.02 .. .. .. −0.23 Natural capital .. .. .. −0.81 .. .. .. −0.04 Human capital .. .. .. 1.29 .. .. .. 0.57 TFP .. .. .. .. .. .. .. 0.82 Note: .. = insufficient or no data to perform the calculation. East Asian Dragons = five East Asian economic “dragons”: Indonesia, the Republic of Korea, Malaysia, Singapore, and Thailand; PPP = purchasing power parity; PPP = public-private partnerships; SSA = Sub-Saharan Africa; TFP = total factor productivity. ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, we leverage electronic publishing options and print-on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping dis- tances decreased, resulting in reduced paper consumption, chemical use, green- house gas emissions, and waste. We follow the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC) –certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine–free (PCF), or enhanced elemental c ­ hlorine–free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility. Economic growth in the Sub-Saharan Africa region has been plagued by a series of shocks—wars, political instability, natural disasters, epidemics, terms-of-trade deterioration, and sudden stops in capital inflows—that have had lingering effects on productivity and growth. Within the overall productivity gap of the region are substantial differences across the sectors of economic activity and production units. Boosting Productivity in Sub-Saharan Africa: Policies and Institutions to Promote Efficiency documents the productivity trends in Sub-Saharan Africa in three different dimensions, assessing productivity at the aggregate level, the sectoral level, and the establishment level. It characterizes the evolution of productivity in the region relative to other countries and regions, as well as country groups in Africa, classified by their degree of natural resource abundance and condition of fragility. The volume suggests that the persistence of the productivity gap in Africa vis-à-vis the technological frontier can be attributed to the slow accumulation of physical and human capital relative to the region’s growing population, as well as the poor allocation of these resources. These allocative inefficiencies are the outcome of policies and institutions that introduce distortions in the decision-making process of individuals. Hence, the volume assesses the implications of production decisions across agricultural farms and manufacturing firms. It presents evidence on aggregate productivity from the perspective of production units, using recent household surveys for farmers and firm-level surveys for select countries, as well as frontier estimation techniques. It documents the extent of severe resource misallocation across agricultural and manufacturing production units. These distortions decelerate the growth of the production units, disincentivize their adoption of productivity-enhancing technologies, and reduce the ability of their peers to learn new techniques. Boosting Productivity in Sub-Saharan Africa highlights the adoption of digital technologies to reduce some of these market frictions. Mobile money has increased financial inclusion in several countries, and digital financial technologies have given individuals access to savings instruments and loan products. Enhancing access to credit can help individuals invest in schooling and overcome the costs of formality. The volume discusses further avenues of research that may provide additional insights on the productivity dynamics across countries in the region, and it identifies the different channels of policy transmission to enhance productivity. The empirical work presented can help to guide the design of policy in the region. ISBN 978-1-4648-1550-8 SKU 211550