A World Bank Group Flagship Report WORLD DEVELOPMENT REPORT The Middle Income Trap The Middle Income Trap A World Bank Group Flagship Report WORLD DEVELOPMENT REPORT The Middle Income Trap © 2024 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 27 26 25 24 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, links/footnotes, and other information shown 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. The citation of works authored by others does not mean The World Bank endorses the views expressed by those authors or the content of their works. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions 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: World Bank. 2024. World Development Report 2024: The Middle-Income Trap. Washington, DC: World Bank. doi:10.1596/978-1-4648-2078-6. 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 attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. 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, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. ISSN, ISBN, e-ISBN, and DOI: Softcover Hardcover ISSN: 0163-5085 ISSN: 0163-5085 ISBN: 978-1-4648-2078-6 ISBN: 978-1-4648-2097-7 e-ISBN: 978-1-4648-2096-0 DOI: 10.1596/978-1-4648-2097-7 DOI: 10.1596/978-1-4648-2078-6 Cover and interior design: Creative Services, Global Corporate Solutions, World Bank. Library of Congress Control Number: 2024941787 Contents xv Foreword xvii Acknowledgments xxi Summary xxv Glossary xxix Abbreviations 1 Overview: Making a Miracle 1 In brief 2 ‘To get rich is glorious’ 4 One trap or two? 5 Investment, infusion, and innovation—additively and progressively 10 The economics of creative destruction 17 Striking the right balance 26 The road ahead 28 Notes 28 References 31 Part 1: Middle-Income Transitions 33 Chapter 1: Slowing Growth 33 Key messages 33 Introduction 34 Growth in middle-income countries 37 Measuring progress through the middle stages of development 40 Growth in middle-income countries is slower 45 Notes 45 References 47 Spotlight 53 Chapter 2: Structural Stasis 53 Key messages 53 Introduction 55 Economic development = structural change 56 Infuse first, then innovate 67 Notes 67 References v 69 Chapter 3: Shrinking Spaces 69 Key messages 69 Introduction 72 Fragmenting international trade 73 Elevated debt 75 Climate action 78 Notes 78 References 81 Part 2: Creative Destruction 85 Chapter 4: Creation 85 Key messages 85 Creation: The protagonist of economic growth, where incumbents create value alongside entrants 88 Creative destruction: Three decades of increasingly refined analysis 91 In middle-income countries, too few small entrants disrupt, and too few large incumbents innovate or infuse global technologies 96 How governments stifle firms’ incentives to grow, infuse global technologies, and innovate 104 Modernizing data and diagnostic tools to understand and regulate creative destruction—from X-rays to MRIs 106 Notes 106 References 109 Chapter 5: Preservation 109 Key messages 109 Preservation is an antagonist of creation because it is also an antagonist of destruction 111 Talent drives economic progress, but social immobility holds back the development of talent 115 Elite pacts perpetuate social immobility and preserve the status quo 123 Patriarchal gender norms hold back a large proportion of the population 127 The cost of social immobility and preservation: Holding back the energies that drive creation 129 Notes 130 References 135 Chapter 6: Destruction 135 Key messages 135 Destruction: To be expected, managed, and mitigated 137 The climate and energy crises could trigger restructuring and reallocation 151 Destruction without creation: The risks of becoming stranded nations 155 Notes 156 References vi | Contents 159 Part 3: Making Miracles 161 Chapter 7: Disciplining Incumbency 161 Key messages 161 Balancing incumbents’ innovation and abuse of dominance 163 Updating institutions to weaken the forces of preservation 172 Incentives for incumbents to strengthen creation 179 Interventions to correct errant behavior by incumbents 183 Notes 184 References 189 Chapter 8: Rewarding Merit 189 Key messages 189 Moving forward by promoting merit activities 190 The economic and social mobility of people 198 The value added by firms 207 Reducing an economy’s greenhouse gas emissions 213 Notes 214 References 221 Chapter 9: Capitalizing on Crises 221 Key messages 221 Using crises to destroy outdated arrangements 222 Globalizing decarbonization 228 Expanding low-carbon infrastructure 232 Decoupling economic growth and emissions 239 Notes 240 References Boxes 12 O.1 Who and what are 65 2.3 The magic of investment incumbents? Leading firms, accelerations technologies, nations, elites— 70 3.1 Graying growth and men 82 P2.1 Joseph Schumpeter and 37 1.1 Misunderstanding through creative destruction misclassification 93 4.1 Vibrant corporate R&D, 41 1.2 A growth superstar: How the connected places, mobile Republic of Korea leveraged people, and ­successful markets foreign ideas and innovation for patents: How the United 42 1.3 Identifying growth slowdowns States nurtured an ­innovation 60 2.1 The Meiji Restoration ecosystem reconnected Japan with global 97 4.2 Examples of size-dependent knowledge policies 61 2.2 Three ways to evade the 100 4.3 The productivity effects of middle-income trap: Swiftly credit misallocation and (Estonia), steadily (Poland), or capital market slowly (Bulgaria) underdevelopment Contents | vii 113 5.1 Firms with better-educated 179 7.5 Tackling anticompetitive managers adopt more practices increases technology incumbents’ innovation 120 5.2 Living in favelas makes it more incentives difficult to get a job 192 8.1 Developing foundational skills: 128 5.3 Global Gender Distortions Learning from Finland and Index: Measuring economic Chile growth lost to gendered 195 8.2 Promoting better student barriers choices with digital tools 138 6.1 The diffusion of low-carbon 196 8.3 Improving students’ test scores technologies as defined and by using online studying measured in this chapter assistance from the Khan 172 7.1 A digital tool helps female Academy entrepreneurs obtain capital 203 8.4 Catching up by opening up and training in rural Mexico and modernizing firms: The 174 7.2 Technology for market access Spanish growth miracle  175 7.3 Supplier development 207 8.5 Productivity growth can slow programs to connect small deforestation in Brazil firms with large firms 212 8.6 Correcting abuses of 177 7.4 Turning brain drain into dominance in electricity brain gain markets 237 9.1 Technologies that can act as “stabilizers” of energy supply Figures 3 O.1 Income per capita of middle- 9 O.6 From infusion to innovation income countries relative to in the Republic of Korea that of the United States has 11 O.7 Over the last four decades, as been stagnant for decades the Republic of Korea’s labor 5 O.2 If capital accumulation were productivity relative to that enough, work in middle- of the United States income countries would be continued to climb, Brazil’s nearly three-quarters as peaked—and then sagged rewarding as in the United 12 O.8 Three views of creative States, not just one-fifth destruction 6 O.3 Economies become more 14 O.9 Creation is a weak force in sophisticated as they middle-income countries, transition from middle- where it is characterized by a income to high-income rampant misallocation of status resources 7 O.4 Middle-income countries 17 O.10 Middle-income countries must engineer two successive have to strike a balance transitions to move to among creation, high-income status preservation, and 8 O.5 In the Republic of Korea, destruction Poland, and Chile, the rapid 21 O.11 In emerging market and growth from middle- to developing economies, few high-income status has been companies are funded by interspersed with economic venture capital or private crises equity viii | Contents 23 O.12 Countries with large, 60 2.5 Calibrating policies to a successful diasporas have the country’s stage of highest potential for development: From imitation knowledge transfers to innovation in the Republic 26 O.13 In low- and middle-income of Korea countries, the cost of capital 64 2.6 The innovation gap between for renewables is high high-income countries and 35 1.1 A handful of economies have others is substantial transitioned from middle- 64 2.7 Middle-income countries income to high-income significantly lag behind status over the last three high-income countries in decades research capacity 36 1.2 Income per capita of middle- 66 B2.3.1 Investment growth income countries relative to accelerations: Colombia, that of the United States has Republic of Korea, and been stagnant for decades Türkiye 40 1.3 Sustained growth periods 70 B3.1.1 Today’s middle-income are short-lived, even in countries are aging more rapidly growing economies rapidly than high-income 43 1.4 Growth slowdowns are most countries did in the past frequent when countries’ 72 3.1 Globally, harmful trade GDP per capita is less than policies outnumber helpful one-fourth of the United trade policies States’ 73 3.2 Harmful interventions in the 44 1.5 Growth is expected to slow global semiconductor trade down as countries approach have skyrocketed since 2019 the economic frontier 74 3.3 Most developing economies (United States) are more severely indebted 44 1.6 Weak institutions hasten than ever and worsen growth 75 3.4 Debt service payments in slowdowns emerging markets and 55 2.1 As economies develop, middle-income countries capital accumulation brings may skyrocket as the cost of diminishing returns borrowing soars 56 2.2 A middle-income country 76 3.5 In middle-income countries, will need to engineer two the energy intensity and successive transitions to carbon intensity of energy achieve high-income status: consumption are quite high Infusion, followed by 77 3.6 In middle-income countries, innovation the weighted average cost of 58 2.3 The demand for highly capital for utility-scale solar skilled workers increases in power projects is middle-income countries substantially higher than the 59 2.4 STEM graduates are cost in high-income countries increasingly concentrated in 78 3.7 Low- and middle-income middle-income countries, countries are exposed to thereby increasing similar levels of risk from opportunities for technology climate change, and they infusion have less adaptive capacity Contents | ix 83 P2.1 Rebalancing the forces of 97 4.10 Young firms—not small creation, preservation, and firms—create the most jobs destruction to advance (net) in the United States infusion and innovation 100 B4.3.1 Productivity-dependent 86 4.1 Both entrants and financial distortions, by GDP incumbents create value and per capita reinforce one another’s 101 4.11 Productivity-dependent growth through competition distortions are more severe in India’s computing services in low- and middle- industry income countries 87 4.2 The interactions between 102 4.12 Management practices are entrants and incumbents set worse in economies with the pace of creative more policy distortions destruction 112 5.1 The share of skilled workers 88 4.3 Entrants drive growth: in large firms increases with Insights from Aghion and GDP per capita Howitt’s seminal paper on 113 B5.1.1 Better-educated managers creative destruction are more likely to adopt 89 4.4 Entrants and incumbents technology in middle- drive growth through income countries turnover and upgrading: 114 5.2 Higher inequality is Insights from Akcigit and associated with higher Kerr’s refined approach to intergenerational immobility creative destruction 115 5.3 Intergenerational mobility of 90 4.5 Contrasting examples of skilled workers matters more innovation: Growth is driven for middle-income countries by entrants in the United than for low-income States and by incumbents in countries Germany 118 5.4 High inequality within cities 91 4.6 Entrants and incumbents is associated with low social can reinforce one another’s mobility from one generation growth: The case of the US to the next business services industry 120 B5.2.1 Slum residents in Rio de 92 4.7 A cartelized industry Janeiro identified their suppresses innovation and residence in a favela as the dynamism: Evidence from largest impediment to the Japanese auto parts getting a job sector 121 5.5 In many middle-income 93 B4.1.1 The number of patents filed countries, movement of by corporations with the US workers from one part of the Patent and Trademark country to another is more Office has skyrocketed limited than in high-income since 1880 countries such as France and 95 4.8 In middle-income countries, the United States the growth rate of firms 122 5.6 In many middle-income across their life cycles is countries, migration costs much lower than in the are higher for individuals United States without high levels of 96 4.9 Microenterprises dominate education firm size distributions in India, Mexico, and Peru x | Contents 124 5.7 There is a substantial gap 162 7.1 In Italy, market leaders between low- and high- increase their political income countries in female connections while educational attainment reducing innovation 124 5.8 Female labor force 163 7.2 Promoting contestability participation is low in the through institutions, Middle East and North incentives, and interventions Africa and in South Asia 164 7.3 In many middle-income 125 5.9 Female labor force countries, markets are participation has evolved dominated by a few differently across countries business groups, as a survey 125 5.10 The share of female suggests professionals has risen in 165 7.4 In middle-income countries, some countries but not restrictive product market others regulations are pervasive 126 5.11 Globally, women own a 166 7.5 In middle-income countries, smaller share of firms than both economywide and men sectoral input and product 127 5.12 Women lag behind men in market regulations are more having financial accounts restrictive than in high- 139 6.1 Learning by doing in the income countries manufacture of key low- 167 7.6 The BRICS and large carbon technologies has middle-income countries resulted in rapid cost have a significant presence of declines publicly owned enterprises 140 6.2 The diffusion of low-carbon and governance technologies is rapidly frameworks that stifle accelerating competition 142 6.3 Low-carbon innovation is 168 7.7 A state presence has driving the emergence of important effects on firm new spatial clusters, start- entry, market ups, and financing concentration, and 145 6.4 The rate of adoption of clean preferential treatment energy technologies is 169 7.8 State-owned enterprises growing more rapidly in dominate coal power middle-income countries generation, while the private than in high-income sector leads in modern countries, but the level of renewable energy adoption is lower 169 7.9 In low- and middle-income 147 6.5 Clean energy technology countries, state-owned value chains are still enterprises are the largest dominated by high-income investors in fossil fuel energy countries and China generation 148 6.6 Costa Rica and China are 173 7.10 Foreign technology licensing the global front-runners in is limited among middle- jobs related to low-carbon income country firms technologies 178 B7.4.1 Some countries are strongly 153 6.7 Most of the countries positioned to benefit from currently “locked in” to knowledge spillovers from declining brown industries their diaspora are middle-income countries Contents | xi 180 B7.5.1 In Colombia, after a cartel is 223 9.1 Use of globalized value sanctioned, market chains for solar panels outcomes improve results in faster learning and through the entry and lower global prices growth of previously lagging 224 9.2 Middle-income countries firms can support global 180 B7.5.2 In Colombia, after an abuse decarbonization by of dominance case, positive becoming global suppliers of market outcomes are driven “granular” (type 1 and type 2) by improvements in leading energy technologies firms 225 9.3 Extraction and processing of 181 7.11 Competition authorities in critical minerals for the clean middle-income countries energy transition remain need more capacity to deal highly concentrated in with sophisticated policy certain countries problems 226 9.4 Many middle-income 191 8.1 Middle-income countries countries have untapped that transitioned to high- potential to manufacture income status first focused green products on foundational skills 227 9.5 All industrial policy 200 8.2 Countries at lower levels of implementation and green development have more industrial policy opportunities for potentially implementation are correlated productivity-enhancing job with GDP per capita reallocation 229 9.6 Countries must clear hurdles 201 8.3 The number of countries for both efficient domestic creating special enforcement investment and foreign units for large taxpayers has investment in renewable increased energy 202 8.4 Improvements in allocative 230 9.7 In many middle-income efficiency in Chile, China, countries, it is economically and India have been efficient to expand renewable driven by reducing energy productivity-dependent 231 9.8 In low- and middle-income distortions countries, the cost of capital 206 8.5 In emerging market and for renewables is high developing economies, few 232 9.9 Today’s upper-middle- companies are funded income countries are more through venture capital or energy efficient than upper- private equity middle-income countries in 208 B8.5.1 Amazon deforestation falls the past when Brazilian productivity 233 9.10 Carbon emissions per unit of rises GDP have been declining 209 8.6 Indirect carbon pricing such worldwide as energy taxes is the 234 9.11 High-income countries have strongest price signal succeeded in reducing overall 211 8.7 In some middle-income emissions by curbing energy countries, the prices of intensity renewable energy through 235 9.12 The world is slowly competitive auctions have transitioning away from reached record lows fossil fuels xii | Contents Maps 141 6.1 In 2022, one-third of online 154 6.3 Low-carbon technology jobs job postings related to in China are growing in low-carbon technologies manufacturing hubs on the were in middle-income southeast coast, whereas countries fossil fuel jobs are close to 150 6.2 Limited or outdated coal mines electricity transmission networks serve as barriers to the entry of renewable sources Tables 4 O.1 World Bank country 49 S.1 Suggested indicators classifications and selected provide a clear picture of the global indicators, 2022 underlying structure of an 7 O.2 To achieve high-income economy status, countries will need to 54 2.1 Middle-income countries will recalibrate their mix of need to engineer two investment, infusion, and successive transitions to innovation develop economic structures 27 O.3 The 3i strategy: What that can sustain high-income countries should do at status different stages of 103 4.1 Examples of possible effects of development market power on development 31 P1.1 World Bank country outcomes classifications and selected global indicators, 2022 Contents | xiii Foreword In 2007 the World Bank published An East Asian Renaissance: Ideas for Economic Growth— the report that coined the phrase “middle-income trap.” This was during a decade of boom- ing growth and poverty reduction in developing countries. Yet it was clear by then that many economies—particularly in Latin America and the Middle East—had remained stuck for decades, despite their efforts to rise to high-income status. “Middle-income trap” is now a popular phrase: it results in tens of thousands of Google search references. And it is fre- quently on the tongues of academics and politicians from developing countries—in Latin America and South Asia and just about every place in between. A decade ago, in “The Middle-Income Trap Turns Ten,” Brookings Institution econo- mist Homi Kharas and I reviewed the burgeoning literature that An East Asian Renaissance had inspired. We found that economists had yet to provide a reliable theory of growth to help policy makers navigate the transition from middle- to high-income status. Some had attempted to develop models, but they were poor substitutes for a well-constructed growth framework on which policy makers could build effective development strategies. Meanwhile, the ranks of middle-income countries continued to grow. Five years later, in “Growth Strategies to Avoid the Middle-Income Trap,” we proposed that Schumpeterian growth models emphasizing creative destruction and institutional change had the poten- tial to provide the analytical foundations for a fuller understanding of middle-income economies. But to be useful they had to be made a lot more accessible to policy makers. This is what World Development Report 2024 sets out to do: provide a simple but reliable growth framework for avoiding or escaping the trap. It identifies lessons from more than 50 years of successes and failures among developing countries while they were climbing the income ladder. Based on these ideas and evidence, it proposes a sequenced, three-pronged approach for today’s 100-odd middle-income countries: first investment, then infusion of new technology from around the world, and then innovation. Each shift requires a new mix of policies that, if implemented reasonably well, result in increasingly dynamic enterprises, an increasingly productive workforce, and an increasingly energy-efficient economy. It is an approach that can benefit all countries—low-, middle-, and high-income—seeking high-quality growth. We are not naive enough to think this will be easy. Middle-income countries will have to work miracles—not only to lift themselves up to high-income status but also to shift away from carbon-intensive growth paths that will lead to environmental ruin. Income levels in Sub-Saharan Africa, where more than half the population lives in middle-income coun- tries, are the same as they were a decade ago. Economic growth rates in middle-income countries have been falling and are expected to average just 4 percent in the 2020s, down from 5 percent in the 2010s and more than 6 percent in the 2000s. xv This has implications for the whole world. Middle-income countries are home to three out of every four people—and nearly two-thirds of those who struggle in extreme poverty. They are responsible for 40 percent of the world’s total economic output—and nearly two- thirds of global carbon emissions. In short, the global effort to end extreme poverty and spread prosperity and livability will largely be won or lost in these countries. The road ahead has even stiffer challenges than those seen in the past: rapidly aging populations and burgeoning debt, fierce geopolitical and trade frictions, and the growing difficulty of speeding up economic progress without fouling the environment. Yet many middle-income countries still use a playbook from the last century, relying mainly on pol- icies designed to expand investment. That is like driving a car just in first gear and trying to make it go faster. If they stick with the old playbook, most developing countries will lose the race to create reasonably prosperous societies by the middle of this century. At current trends, it will take China more than 10 years just to reach one-quarter of US income per capita, Indonesia nearly 70 years, and India 75 years. The team that has written this report hopes to radically alter this arithmetic. Our hope is that World Development Report 2024 will, in short order, make the expression “middle- income trap” completely obsolete. Indermit Gill Chief Economist of the World Bank Group and Senior Vice President for Development Economics xvi | Foreword Acknowledgments World Development Report 2024 was prepared by a World Bank team led by Somik V. Lall. Ufuk Akcigit served as the Academic Lead, and Joyce Antone Ibrahim was the Report ­ Manager. Overall guidance was provided by Indermit Gill, Senior Vice President and Chief Economist. The Report was sponsored by the Development Economics Vice Presidency. The core team comprised Roberto Fattal Jaef, Maria Marta Ferreyra, Kenan Karakülah, Tatjana Kleineberg, Mathilde Lebrand, Martha Martinez Licetti, Dino Merotto, Forhad Shilpi, Katherine Stapleton, Maria Vagliasindi, and Ekaterina Vostroknutova. Victor Ajayi, Deniz Aycan, Narcisse Cha’ngom, Dong Phuong Dao, Matteo Gasparini, Juan Holguín Posada, Karry Jiao, Yonatan Litwin, Theodore Naff, Juan Porras Lopez, Mariana Santi, Zeki Berkay Saygin, Karthik Sridhar, Gabriel Suárez Obando, Adesola Sunmoni, Facundo Ulivarri, and Natalia Valdebenito Contreras were research analysts. Selome Missael Paulos provided the team with administrative support through May 2023, and Sandi Soe Lwin provided administrative support to the team beginning in May 2023. Pia Andres, Oya Pinar Ardic Alper, Sina Ates, Christopher Bataille, Tania Begazo Gomez, Bhavna Bhatia, Tanuj Bhojwani, Julia Bird, Craig Chikis, Xavier Cirera, Fernando Dancausa Diaz, Yuheng Ding, Maciej Drozd, Alice Evans, Nisan Gorgulu, Soulange Gramegna Mesa, Rogelio Granguillhome Ochoa, Michael Grubb, Federico Haslop, Sheirin Iravantchi, Gautam Jain, Aidara Janulaityte, Noah Kaufman, Joohyun Lee, Munseob Lee, Ming Lu, Antonio Martins Neto, Penelope Mealy, Nandan Nilekani, Stefanie Onder, Paul Phumpiu Chang, Denisse Pierola Castro, Brian Pinto, Laurent Porte, Gaël Raballand, Ana Belen Ruival, Sagatom Saha, Yongseok Shin, Hassan Soumya, Sowjanya, Yana R. Ukhaneva, Harald Walkate, Diane Zovighian, and María Pluvia Zúñiga Lara provided inputs to the report at various stages. Nicolas Moschovakis, Bruce Ross-Larson, and Timothy Taylor provided developmental guidance in drafting the Report. Anthony Venables was a member of the extended team. Anwar Aridi and Hoon Sahib Soh contributed to box 1.2 on the Republic of Korea. Karsten Staehr and Sebastian Stolorz authored box 2.2 on Bulgaria, Estonia, and Poland. Tristan Reed and Kersten Stamm authored box 2.3 on investment accelerations. Pinelopi Goldberg, Michael Peters, and Aishwarya Ratan authored box 5.3 on the gendered barriers to growth. Oscar Calvo-González authored box 8.4 on the Spanish growth miracle. Marek Hanusch authored box 8.5 on Brazil. Nurana Ahmadova, Narcisse Cha’ngom, Karry Jiao, Gabriel Suárez Obando, and Renato Schwambach Vieira assisted with the reviews of translations. xvii The communications and engagement strategy was led by a team comprising Chisako Fukuda, Kristen Milhollin, Karolina Ordon, Joseph Rebello, Mikael (Kelly) Reventar, Shane Romig, and Mariana Teixeira. Roula Yazigi provided web and online services. Special thanks are extended to Mark McClure, who coordinated and oversaw formal production of the Report, and to the World Bank’s Formal Publishing Program, including Cindy Fisher and Patricia Katayama. Mary C. Fisk facilitated the multiple translations of the overview and main messages by the Translations and Interpretation team, coordinated by Bouchra Belfqih. Yaneisy Martinez and Orlando Mota managed the printing and elec- tronic conversions of the Report and the many ancillary products. The Report was edited by Sabra Ledent and Nancy Morrison and proofread by Gwenda Larsen and Catherine Farley. Robert Zimmermann verified the Report’s extensive citations and assisted with the copyediting. The design team of the World Bank’s Global Corporate Solutions unit designed the cover and the interior layout. Bill Pragluski of Critical Stages designed some of the Report’s figures and infographics. Datapage supplied typesetting services. Dayana Leguizamon provided the team with resource management support. The team would also like to thank colleagues at various World Bank country offices who assisted with logistics and stakeholder engagements. Special thanks to Marcelo Buitron, Gabriela Calderon Motta, and Grace Soko for their help with coordination and high-level engage- ment strategies. The team is grateful for the guidance, comments, and inputs provided by other World Bank Group colleagues, particularly those with the Development Economics Vice Presidency; East Asia and Pacific Region; Eastern and Southern Africa Region; Economics and Private Sector Development Vice Presidency (International Finance Corporation); Education Global Practice; Environment, Natural Resources, and Blue Economy Global Practice; Equitable Growth, Finance, and Institutions Vice Presidency; Europe and Central Asia Region; External and Corporate Relations; Gender Group; Human Development Vice Presidency; Independent Evaluation Group; Infrastructure Vice Presidency; Latin America and the Caribbean Region; Middle East and North Africa Region; Poverty and Equity Global Practice; South Asia Region; Sustainable Development Vice Presidency; and Western and Central Africa Region. The team thanks the many World Bank colleagues who offered written comments during the formal Bankwide review process. These comments provided invaluable guidance at a crucial stage in the production of the Report. The team gratefully received suggestions and guidance from a High-Level Advisory Panel: Masood Ahmed, President, Center for Global Development; Ann Bernstein, Executive Director, Centre for Development and Enterprise (South Africa); Poonam Gupta, Director General, National Council of Applied Economic Research, and mem- ber of the Economic Advisory Council to the Prime Minister of India; Homi Kharas, Senior Fellow–Global Economy and Development, Center for Sustainable Development, Brookings Institution; Mario Marcel Cullell, Minister of Finance, Chile; Mustapha Kamel Nabli, former Minister of Economic Development and Minister of Planning and Regional Development and former Central Bank Governor, Tunisia; Njuguna Ndung’u, Minister of Finance, Kenya; José Antonio Ocampo, former Minister of Finance, Colombia; Normunds Popens, Acting Director General, Directorate-General for Regional and Urban Policy (DG REGIO), European Commission (through January 2024); and Omar Razzaz, former Minister of Education and former Prime Minister, Jordan. xviii | Acknowledgments The team received suggestions and inputs from an Academic Advisory Committee: Daron Acemoglu (Massachusetts Institute of Technology), Philippe Aghion (Collège de France, INSEAD, London School of Economics and Political Science), Gerardo Esquivel (El  Colegio de México), Ricardo Hausmann (Harvard Kennedy School), Robert Pindyck (Sloan School of Management, Massachusetts Institute of Technology), Danny Quah (Lee Kuan Yew School of Public Policy, National University of Singapore), Jahen F. Rezki (Universitas Indonesia), Qiyuan Xu (Institute of World Economy and Politics, Chinese Academy of Social Sciences), and Fabrizio Zilibotti (Yale University). The team consulted other academics, including Manuel Agosin (Universidad de Chile), Belinda Archibong (Barnard College, Columbia University), Iwan Azis (Cornell University and Universitas Indonesia), Nicolas Bottan (Cornell University), Andrea Bubula (Columbia University), John Carruthers (Cornell University), Julieta Caunedo (Cornell University), Nancy Chau (Cornell University), Abigayle Davidson (Aspen Institute), Jan Eeckhout (Pompeu Fabra University), Gary Fields (Cornell University), Caroline Flammer (Columbia University), Rodrigo Fuentes (Pontificia Universidad Católica de Chile), Alvaro Garcia-Marin (Universidad de los Andes), Ravi Kanbur (Cornell University), David Kohn (Central Bank of Chile and Pontificia Universidad Católica de Chile), Saurabh Lall (University of Glasgow), Patricia Mosser (Columbia University), Cristian Pop-Eleches (Columbia University), Jan Svejnar (Columbia University), and Eric Verhoogen (Columbia University). Thanks are due to the following academics who participated with presentations during a series of seminars hosted by the World Development Report 2024 team: Craig Chikis (University of Chicago), Sebastián Gallegos (Business School, Universidad Adolfo Ibáñez, Santiago, Chile), Pulak Ghosh (Indian Institute of Management, Bangalore), Munseob Lee (University of California, San Diego), Ming Li (Chinese University of Hong Kong, Shenzhen), Javier Miranda (Halle Institute for Economic Research, Halle, Germany), Ricardo Paredes (former President of Duoc and current Minister, Free Trade Special Court, Santiago, Chile), Marta Prato (Bocconi University, Milan), Younghun Shim (International Monetary Fund), and María Pluvia Zúñiga Lara (United Nations University–Maastricht Economic and Social Research Institute on Innovation and Technology, Maastricht, the Netherlands). The team thanks the Chinese Academy of Social Sciences (China), Columbia University (United States), Cornell University (United States), Fudan University (China), the National Council of Applied Economic Research (India), Políticas Públicas (Colombia), Shanghai Institute for International Studies (China), and Shenzhen Finance Institute (Chinese University of Hong Kong) (China) for organizing and hosting a roundtable discussion with academics. The team conducted a series of bilateral consultations and field visits with several gov- ernments, including Brazil, Chile, China, Colombia, the European Commission, India, Italy, Kenya, Mexico, Morocco, and the United Kingdom. The team benefited from the inputs of several think tanks, research institutes, academic institutions, civil society organizations, private sector organizations, and other organiza- tions, including Accenture, African Centre for Economic Transformation, African Economic Research Consortium, ANT Group, Banco de Crédito e Inversiones S.A., Betterfly, Boston Consulting Group, Celulosa Arauco y Constitución, Center for Advanced Economic Studies, Center for International Knowledge on Development, Centre for Social and Economic Progress, China Academy of Information and Communications Technology, China Center Acknowledgments | xix for International Economic Exchanges, China Center for Macroeconomic Research, China International Capital Corporation, China Pacific Insurance Company, Chinese Academy of Social Sciences, Chinese University of Hong Kong, Economic Commission for Latin America and the Caribbean, Einaudi Institute for Economics and Finance, Empresas Copec, Enel, Essence Securities Company, Fudan University, Fundación Chile, Haitong International Securities Group Limited, Harvard Growth Lab, Huatai Securities, Inria, Institute for World Economy Studies Shanghai, Institute of Comparative Politics and Public Policy of Shanghai University, Institute of Economy Pontificia Universidad Católica de Chile, Institute of Quantitative and Technical Economics, Institute of Statistics and Applied Economics, International Centre for China Development Studies, Inversiones SB, Lal Bahadur Shastri National Academy of Administration, National Council of Applied Economic Research, National Evaluation and Productivity Commission, NTT Data Chile, Pan American Association of Student Loan Institutions, Peking University, Renmin University of China, Shanghai Artificial Intelligence Laboratory, Shanghai Institutes for International Studies, Shanghai Jiao Tong University, Sky Airlines, Sonda S.A., Songhe Chuangzhi Venture Capital Investment Partnership Enterprise, Tencent Research Institute, Thomas B. Fordham Institute, Tsinghua University, UBS Wealth Management, United Nations (Morocco), Universidad Andrés Bello, Universidad de Chile School of Economic Sciences and Business, Universidad de los Andes, Universidad del Desarrollo, and VIDA Security. The team apologizes to any individuals or organizations inadvertently omitted from this list. It is grateful for the help received from all who contributed to this Report, includ- ing those whose names may not appear here. Team members would also like to thank their families for their support throughout the preparation of this Report. xx | Acknowledgments Summary Part 1: Middle-Income applicable to lower-middle-income Transitions countries. • The second transition is to switch to a Chapter 1: Slowing Growth 3i strategy, which entails paying more Is growth in middle-income countries attention to innovation—a process more slower than that in countries at other applicable to upper-middle-income income levels? countries. • Yes. Growth slowdowns occur more frequently in middle-income countries Chapter 3: Shrinking Spaces than in low- or high-income countries. • Development strategies that served Is growth in middle-income countries countries well in their low-income now harder to achieve? phase—capital investment, in • Yes. Foreign trade and investment are particular—yield diminishing returns. ­ in danger of becoming constricted by • Countries with weaker institutions— geopolitical tensions, and populism is and especially those with lower levels shrinking the room for governments to of economic and political freedom—are act. susceptible to slowdowns at even lower • Rising debt and adverse demographics levels of income. are crowding out private investors and reducing public investment. • Accelerating climate action will require Chapter 2: Structural Stasis large investments in infrastructure Is growth in middle-income countries and regulatory reforms that may stall different from that in countries at other productivity. income levels? • Yes. Successful middle-income coun- Part 2: Creative tries will have to engineer two succes- Destruction sive transitions to develop economic structures that can eventually sustain Chapter 4: Creation high-income levels. Who creates value? • The first transition is from a 1i strat- • Both incumbents and entrants can egy for accelerating investment to a 2i create value. Incumbents bring scale. strategy focusing on both investment They can compete with entrants and infusion in which a country brings in the market to jointly expand a technologies from abroad and diffuses country’s technological capabilities, them domestically—a process broadly thereby moving the country closer xxi to the global frontier. Entrants bring Discrimination can be pervasive, change—enterprises with new products affecting the businesses women own, or production processes, workers with the jobs they get, the pay they receive, new skills and ideas, or energy sources what their families spend on educating such as renewables that embody new them, and their ability to manage technologies. By doing so, they expand financial accounts. a country’s technology frontier. What is the implication of having both Chapter 6: Destruction incumbents and entrants as value Why is destruction important for creators? ­structural change? • Policy makers will have to stop relying • The destruction of outdated arrange- on superficial measures of structural ments—enterprises, jobs, technologies, efficiency such as firm size, income private contracts, policies, and public inequality, and energy sources. The institutions—is essential to creating imperative for today’s middle-income value through infusion and innovation. economies is “efficiency”—in the use of capital, labor, and energy. Policy mak- Who are the antagonists blocking creative ers will need to heed the value added destruction in response to today’s energy of firms, social mobility, and emissions crisis? intensity. They are more reliable and • Incumbents, usually state-owned more realistic metrics for policy mak- enterprises, have the strongest incen- ing, but they also require collecting tive to maintain the status quo and more information. limit competition from low-carbon energy providers. • Many G20 economies are introducing Chapter 5: Preservation incentives for producing and deploying How do incumbents preserve the status low-carbon technologies.  Some mea- quo? sures may unintentionally preserve • Incumbents’ dominance can buy enterprises in advanced economies economic, social, and political power. and destroy them in middle-income By capturing political and social countries. institutions, incumbents have an outsize say in who learns where and what, who gets a sought-after job and Part 3: Making Miracles what they are paid, and who gets to Chapter 7: Disciplining start a business. Incumbency How do discrimination and patriarchal How can middle-income countries gender norms hold back the potential of weaken the forces of preservation women? that protect incumbents from healthy • Patriarchal norms and systems of competition? belief that give men greater status • By promoting contestable markets, and authority and define strict gender middle-income countries can strike a roles and responsibilities hold back balance between supporting incum- women from benefiting from attractive bents and ensuring that they do not educational and job opportunities. abuse their market power. xxii | Summary • Institutional arrangements that pro- • To decouple carbon emissions from mote contestability include retract- a growing economy, middle-income ing protection of incumbents such as countries can effectively price carbon market leaders and state-owned enter- emissions and scale up deployment of prises and norms that work against low-carbon energy by respecting the women. merit order—the sequence followed • Openness to foreign trade, investment, by grid operators selling power to the and talent helps with technological market. upgrading. • Interventions that target errant incum- Chapter 9: Capitalizing on Crises bents to destroy harmful arrangements include adopting competition laws and How can middle-income countries ensuring the effectiveness of competi- capitalize on crises to destroy outdated ­ tion authorities, as well as using fiscal arrangements and make way for creation? policy to make elites contestable. • Because middle-income countries need to recalibrate their mix of investment, infusion, and innovation, crises can Chapter 8: Rewarding Merit become a necessary evil because they How can middle-income countries provide the momentum to weaken the strengthen the forces of creation by status quo. rewarding merit—that is, those forces • To capitalize on today’s climate and that aid in the efficient use of talent, capi- energy crises, middle-income countries tal, and energy? can support global decarbonization by • To reward merit, middle-income coun- infusing global technologies domesti- tries can upgrade their talent pools, cally to join low-carbon value chains for select efficient learners, and tap the pro- global markets. They can also invest in ductive power of women. deploying low-carbon energy if it reaps • To efficiently use capital, middle-­ economic returns. income countries can move away from • Middle-income countries face critical coddling small firms or vilifying large needs: growth, decarbonization, and firms, let go of unproductive firms, energy security. Solutions will require modernize the management of firms, decoupling emissions from a growing and connect entrepreneurs with men- economy while extending affordable, tors and markets. secure energy to all firms and families. Summary | xxiii Glossary The following general descriptions of the terms and phrases commonly used in this Report reflect their context in the Report. brain drain The movement of educated carbon intensity A measure of carbon or professional people from one place or dioxide and other greenhouse gases emit- profession to another to gain better pay or ted per unit of activity. living conditions. contestability  An environment in which brain gain  An increase in the number of incumbents feel pressure to compete and highly trained foreign-born professionals upgrade because their products and pro- entering a country to live and work and cesses could be displaced by technologi- benefit from the greater opportunities cally sophisticated producers in their own offered. country or from other countries. business of the state (BOS) An enter- creative destruction  A concept intro- prise with majority or minority state duced by economist Joseph Schumpeter shareholdings. that refers to the process of innovation capital accumulation  An increase in and technological change that leads to the assets from investments or profits. destruction of existing economic struc- tures such as industries, firms, and jobs. capitalizing on crises  The process of This destruction paves the way for new using a crisis as an opportunity to imple- structures to emerge, thereby creating ment major reforms that otherwise would long-term economic growth and progress. have been blocked. decoupling growth from emissions  A pro- carbon capture and storage (CCS)  A pro- cess that culminates in economic growth cess in which a relatively pure stream of no longer strongly associated with carbon carbon dioxide from industrial sources is emissions. separated, treated, and transported to a long-term storage location. disciplining incumbents  A process in which policies or actions are aimed at lim- carbon capture, utilization, and s ­torage iting the power of incumbents to capture (CCUS)  An advanced iteration of the tra- institutions or block competitors. ditional carbon capture and storage (CCS) technology. CCS focuses mainly on the Economic Complexity Index (ECI)  A rank- capture and sequestration of carbon diox- ing of countries based on the diversity and ide to mitigate emissions, and CCUS takes complexity of their export basket. High-­ it one step further by finding practical complexity countries are home to a range applications for the captured carbon. of sophisticated, specialized capabilities and xxv are therefore able to produce a highly diver- such as education, training, and better sified set of complex products. health. energy intensity  A measure of the energy leapfrogging  The process by which econ- use of an economy, calculated as units of omies attempt to become “knowledge energy per unit of gross domestic product economies” before putting in place the (GDP) or another measure of economic institutional infrastructure and develop- output. ing requisite capabilities. entrant  An entity that enters an industry Long Term Growth Model (LTGM) A with a capacity to produce goods or ser- spreadsheet-based tool to analyze future vices that can compete with those of exist- long-term growth scenarios in develop- ing entities in order to earn profits. ing countries, building on the celebrated feed-in tariff  A policy designed to ­ support Solow-Swan growth model. The LTGM the development of renewable energy sources aggregates assumptions about growth by providing a guaranteed, above-market fundamentals—such as investment, edu- price for producers. cation, and productivity—to produce a trajectory for future growth. The drivers incumbent  An established entity in soci- of growth are savings, investment, and ety, public office, or the market. This term productivity, but the model also ana- is often used to describe the existing firms lyzes human capital, demographics, the in the market, typically the leading firms, external sector (external debt, foreign as well as the prevailing technology, social direct investment, and current account elites, or technologically advanced nations balance), and labor force participation by with an established presence in the pro- gender. duction of certain goods or services. low-carbon technologies  Technologies industrial policy A policy that directs or applications intended to counter the state support toward specific technologies, effects of climate change. sectors, industries, or firms. infusion  A process in which countries merit  A person’s possession of required focus on imitating and diffusing modern skills or qualifications. technologies and business models from merit order  The sequence followed by grid more advanced economies and applying operators selling power to the market. The this knowledge at scale in their domestic starting point is the cheapest offer, made economy, thereby enabling home indus- by the power station with the lowest oper- tries to become global suppliers of goods ating costs, which determines the whole- and services. sale market prices. Any provider that can innovation  A process in which countries offer renewable energy at zero marginal focus on building home country capabil- cost—that is, with insignificant operating ities to add value to global technologies costs—should have priority in meeting so that domestic firms can become global demand. knowledge creators. middle-income trap  A situation in which investment  A process in which countries a middle-income country experiences sys- focus on increasing physical capital, such tematic growth slowdowns as it is unable as machinery, equipment, and infrastruc- to take on the new economic structures ture, as well as improving human capital, needed to sustain high-income levels. xxvi | Glossary Successful middle-income countries will resource curse  The phenomenon of coun- have to engineer two successive transi- tries with an abundance of natural resources tions to develop such economic structures. (such as fossil fuels and certain minerals) The  first transition is from a 1i strategy having lower economic growth, less democ- for accelerating investment to a 2i strategy racy, or worse development outcomes than focusing on both investment and infusion. countries with fewer natural resources. In the latter, a country brings technologies rewarding merit  The act of policies, insti- from abroad and diffuses them domesti- tutions, and other government structures cally. Once a country has succeeded in the aiding in the efficient utilization of talent, first transition, the second transition con- capital, and energy. sists of switching to a 3i strategy, which entails paying more attention to innovation. size-dependent policies Policies that, by design, stipulate different treatment of net zero  The balance between the amount firms of different sizes. of greenhouse gas produced and the amount removed from the atmosphere. social immobility  A feature of a society It can be achieved through a combina- with fixed social norms or a rigid class tion of emissions reduction and emissions system so that movement from one social removal measures. class, social or economic status, or social role to another is constrained. power purchase agreement (PPA)  A long- term agreement to purchase energy from social mobility A change in a person’s a specific asset at a predetermined price socioeconomic situation either in relation between an electricity generator and a con- to their parents (intergenerational mobil- sumer—generally a utility—or between a ity) or throughout their lifetime (intragen- developer and a supplier, which then resells erational mobility). the energy. state-owned enterprise (SOE)  A legal productivity-dependent distortion  A pol- entity created by a government to par- icy distortion related to firm size that can take in commercial activities on the discourage growth, innovation, and tech- ­government’s behalf. nology adoption. stranded assets  Assets that lose value or proximity to the frontier  A measure used turn into liabilities before the end of their in this Report to clarify the distribution expected economic life. In the context of of growth slowdowns along the national fossil fuels, this term refers to those fuels income spectrum around the world, that will not be burned and thus remain in defined as the ratio of a country’s GDP the ground. per capita to that of the frontier country each year (not adjusted for differences in total factor productivity (TFP)  A ­measure purchasing power parity). The frontier of the efficiency with which all inputs represents the growth leader—the coun- (labor, capital, and so forth) are used in the try with the most advanced combination production process. It represents the por- of economic production, innovation, and tion of output not explained by the amount workforce—which is proxied by the United of inputs used in production. States in this Report. Glossary | xxvii Abbreviations AA Account Aggregator BECCS bioenergy with carbon capture and storage BOS business of the state CCAS centralized choice and admission system CCS carbon capture and storage CCUS carbon capture, utilization, and storage CO2 carbon dioxide COVID-19 coronavirus disease 2019 DACCS direct air capture with carbon storage ECI Economic Complexity Index EMDEs emerging market and developing economies ETS emissions trading system EU European Union EV electric vehicle FAT Firm-level Adoption of Technology FDI foreign direct investment FiT feed-in tariff G20 Group of Twenty GATT General Agreement on Tariffs and Trade GDP gross domestic product GHG greenhouse gas GNI gross national income IBRD International Bank for Reconstruction and Development ICE internal combus­ tion engine ICT information and communication technology IDA International Development Association IEA International Energy Agency IITs Indian Institutes of Technology IRA Inflation Reduction Act IRENA International Renewable Energy Agency LEED Leadership in Energy and Environmental Design LTGM Long Term Growth Model MIT Massachusetts Institute of Technology MNC multinational corporation NIPO New Industrial Policy Observer NTM nontariff measure xxix OECD Organisation for Economic Co-operation and Development PMR product market regulation PPA power purchase agreement PPI Private Participation in Infrastructure PPP purchasing power parity PV photovoltaic R&D research and development RISE Regulatory Indicators for Sustainable Energy SAR special administrative region SDGs Sustainable Development Goals SMEs small and medium enterprises SOE state-owned enterprise STEM science, technology, engineering, and mathematics TCP bon price total car­ TCS Tata Consultancy Services TFP total factor productivity TVET technical and vocational education and training UNESCO United Nations Educational, Scientific, and Cultural Organization UPI Unified Payments Interface WDR World Development Report WTO World Trade Organization xxx | Abbreviations Overview: Making a Miracle In brief Middle-income countries are in a race against time. in lower-middle-income countries must add to Since the 1990s, many of them have done well investment-​ driven strategies measures to infuse mod- enough to escape low-income levels and eradicate ern technologies and successful business processes extreme poverty, leading to the general perception from around the world into their national economies. that the last three decades have been great for devel- This requires reshaping large swaths of domestic opment. But this is because of abysmally low expec- industry as global suppliers of goods and services. tations—remnants from a period when more than Once a country has succeeded in doing this, it can two-thirds of the world lived on less than a dollar a switch to a 3i strategy where it increases attention day. The ambition of the 108 middle-income coun- to innovation. Upper-middle-income countries that tries with incomes per capita of between US$1,136 have mastered infusion can complement investment and US$13,845 is to reach high-​ income status within and infusion with innovation—beginning not just the next two or three decades. When assessed against to borrow ideas from the global frontiers of tech- this goal, the record is dismal: the total population nology but also to push the frontiers outward. This of the 34 middle-income economies that transi- requires restructuring enterprise, work, and energy tioned to high-income status since 1990 is less than use once again, with an even greater emphasis on 250 ­ million, the population of Pakistan. economic freedom, social mobility, and political During the last decade their prospects have contestability. ­ worsened. With rising debt and aging populations Transitions across growth strategies are not auto- at home, growing protectionism in advanced econo- matic. Success depends on how well societies juggle mies, and escalating pressures to speed up the energy the forces of creation, preservation, and destruction. transition, today’s middle-income economies are They can do this by disciplining incumbency, reward- growing into ever-tighter spaces. The odds that the ing merit, and capitalizing on crises. Incumbents— 6  billion people in today’s middle-income countries large corporations, state-owned enterprises, and will see their countries grow to high-income status powerful citizens—can add immense value, but they within a generation or two were never that good. can just as easily reduce it. Governments must devise Now they are decidedly daunting. mechanisms to discipline incumbents through com- Drawing upon the development experience since petition regimes that encourage new entrants without the 1950s and advances in economic analysis by either coddling small and medium-size enterprises or Schumpeterian economists, World Development vilifying big corporations. Middle-income countries Report 2024 (WDR 2024) identifies pathways for have smaller reservoirs of skilled talent than advanced emerging market economies to avoid what has become economies and are less efficient in utilizing them, so known and feared as the “middle-income trap.” The they will have to become better at both accumulating Report points to the need for not one but two tran- and allocating talent. Cheap and reliable energy has sitions during middle-income. The first is to transi- been a cornerstone of rapid economic development, tion from a 1i strategy for accelerating investment to but prospering while keeping the planet livable will a 2i strategy that emphasizes both investment and now require much more attention to energy efficiency infusion in which a country brings technologies from and emissions intensity. Exigencies such as the rise of abroad and diffuses them domestically. Governments populism and climate change provide opportunities 1 to dismantle outdated arrangements and make room In implementing these strategies, the Report rec- for new ones; crises are painful, but in democracies ommends against using relatively superficial mea- they can help forge the consensus needed for tough sures like firm size, income inequality, and energy policy reforms. sources to make policy, relying instead on uncon- The handful of economies that have made speedy ditionally reliable measures such as value added, transitions from middle- to high-income have encour- socioeconomic mobility, and emissions intensity. The aged enterprise by disciplining powerful incumbents, latter are more realistic metrics for policy making, developed talent by rewarding merit, and capital- but they are also more demanding. Policy makers ized on crises to alter policies and institutions that will have to be more willing to make public sensitive no longer suit the purposes they were designed to data, to openly debate policy, and take any opportu- serve. Today’s middle-income countries will have to nity to destroy outdated arrangements. This requires do the same. The question is how. Given the com- information that is harder to get, but it is essential. plex problems they will have to deal with to prosper, Without it, middle-income countries will be sailing the imperative for today’s middle-income econo- blind into ever-stormier seas. mies is surprisingly simple: they will have to become Since the 1970s, income per capita in the median ­efficient—in the use of capital, labor, and energy. This middle-income country has stayed below a tenth of is easier said than done, but advances in economic the US level. Growing geopolitical, demographic, analysis during the last three decades provide useful and environmental complications will make eco- pointers. nomic growth harder in the years ahead. To become Readers might immediately recognize the prob- advanced economies despite these headwinds, middle-​ lem with equating a country’s development with its income countries will have to make miracles. income per capita. In fact, development practitioners have been using a raft of similarly superficial indica- tors to assess the structural strength of an economy ‘To get rich is glorious’ and its disaggregates such as industry, society, and You are a policy maker in one of the world’s 108 ecology. We have become accustomed to using the middle-income countries. You have learned the size distribution of firms in an industry to measure importance of creating a credible, solid macro- its productive efficiency, household income distribu- economic foundation for private investment, tions to assess social durability, and the distribution domestic and foreign, supported by strong insti- of energy sources to approximate ecological sustain- tutions and clean governance. And, like Deng ability. But as economic structures become more Xiaoping nearly 50 years ago, quoted here, you complex, these measures have become increasingly have big plans. inaccurate and progressively poorer guides for mak- If your country is China, your 14th Five-Year ing policy. Plan envisions reaching the median gross domes- WDR 2024 is premised on the conjecture that, tic product (GDP) per capita of developed nations relative to the complexity of their economic struc- by 2035, thereby greatly expanding your middle tures, middle-income countries have more serious class. If it is India, your prime minister’s vision information deficits than either low-income coun- is to turn the nation into a developed economy tries or advanced economies. As a result, they suffer by 2047, the centennial of independence. If it is more than the others the consequences of policies Viet Nam, your Socio-Economic Development predicated on superficial measures of economic effi- Strategy 2021–2030 outlines a strategy for sus- ciency, making them especially prone to premature tained GDP per capita growth of 7 percent slowdowns in development. This pathology was through this decade, with a transition to high-­ nicknamed the “middle-income trap” by World Bank income status by 2045. And if it is South Africa, economists, and strategies to avoid it are the subject your 2030 National Development Plan sets a goal of this Report. of raising the income per capita from US$2,800 in 2 | WORLD DEVELOPMENT REPORT 2024 2010 to US$7,000 by 2030. Other middle-income considered the world’s economic leader; people countries have similar aspirations. living in countries with incomes higher than those If these plans succeed, your country will reach of Americans add up to fewer than 25  million. high-income status in less than one generation, Since 1970 the mean income per capita of middle- or in one or two. Your firms will be earning like income countries has never risen above one-tenth never before. Your people will be consuming like that of the United States (figure O.1). never before. Far fewer people will be poor, with Compared with the United States, middle-­ none desperately poor. In the halls of government, income countries seem trapped at modest income these plans generate tremendous optimism. levels. But there is a problem. The observed rates of economic growth in According to widely used measures such as middle-income countries do not exceed those in the World Bank’s World Development Indicators, high-income countries by the margins needed you see that economic growth in middle-income to catch up in one generation—or even two or countries—including your own—is not acceler- three. Estimates using the World Bank’s Long ating. If anything, it is slowing down as incomes Term Growth Model, which is based on the cel- increase—and even more so every decade. ebrated Solow-Swan growth model, suggest that Moreover, your country is not catching up if the drivers of economic growth—investments with the income levels in advanced economies. in human capital, total factor productivity, labor Among those economies, the United States is still force participation, and the shares of economic Figure O.1  Income per capita of middle-income countries relative to that of the United States has been stagnant for decades 60 55 Korea, Rep. Share of US GNI per capita (%) 50 45 40 35 30 Poland 25 20 Chile 15 10 5 0 20 12 14 18 08 10 16 98 04 06 88 92 02 94 96 00 70 22 72 74 76 78 80 82 84 86 90 20 20 20 20 20 20 20 20 20 20 20 20 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 Middle-income countries Middle-income countries, excluding China Source: WDR 2024 team using data from WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712. Note: The plotted lines indicate the trend of average income per capita in middle-income countries and in middle-income countries, excluding China, relative to income per capita of the United States (considered the economic frontier country). Country definitions are based on the first World Development Report (World Bank 1978), in which low-income countries have gross national income (GNI) per capita of US$250 or less; middle-income countries have GNI per capita of more than US$250; and industrialized (high-income) countries consist of member countries of the Organisation for Economic Co-operation and Development, except for Greece, Portugal, Spain, and Türkiye, which are classified as middle-income countries. OVERVIEW: MAKING A MIR ACLE | 3 output devoted to public and private invest- which countries are firmly considered upper- ment—follow recent and historic trends, most middle-income. A systematic slowdown in growth middle-income countries are likely to experience then occurs. Development strategies relying cant slowdowns between 2024 and 2100. signifi­ largely on capital accumulation that served these Countries such as Brazil and Mexico are likely to countries well in their low-income phase, for many be even further behind the United States in 2100 even during their lower-middle-income phase than they are today. between US$1,136 and US$4,465—begin to yield diminishing returns. Strategies based on factor accumulation alone are likely to steadily worsen One trap or two? results—a natural occurrence as the marginal The World Bank presently classifies 108 countries productivity of capital declines. as “middle-income”—that is, those with annual To see why, consider this: if capital endowments income per capita ranging from US$1,136 to were the only economically relevant difference US$13,845.1 These countries are critical to long- between ­middle-income and high-income countries term global prosperity. They account for nearly today, the gross national income per capita of a 40 percent of global economic activity, more than typical middle-income country would have been 60 percent of people living in extreme poverty, nearly three-quarters of that of the United States in and more than 60 percent of global carbon diox- 2019 (figure O.2). In fact, it is about one-fifth that of ide (CO2) emissions (table O.1). the United States. Its growth prospects now depend Developing economies change in structure as increasingly on its ability to boost the sophistication they increase in size, which means that changes of its production methods. in the pace of growth stem from factors that are Since 2007, the World Bank has called this new to them. Although these imperatives can vary dependence the “middle-income trap.”2 And over across countries, economic expansion, on average, the last 34 years, only 34 economies have suc- begins to decelerate and often reaches a plateau ceeded in breaking out of it. in income per capita growth, typically at about To achieve high-income status, a middle-​ 11 percent of US GDP per capita. Today, this figure income country needs to ramp up the sophis- would be about US$8,000, or around the level at tication of its economic structure. Using the Table O.1  World Bank country classifications and selected global indicators, 2022 SHARE OF SHARE OF SHARE OF PEOPLE IN SHARE OF GLOBAL INCOME GLOBAL GLOBAL EXTREME POVERTY CARBON DIOXIDE (CO2) CLASSIFICATION POPULATION (%) GDP (%) GLOBALLY (%) EMISSIONS (%) Low-income 8.9 0.6 36.5 0.5 Lower-middle-income 40.3 8.3 55.4 15.7 Upper-middle-income 35.1 30.3 7.1 48.6 High-income 15.7 60.8 1.0 35.2 Sources: Population shares and global GDP shares computed from WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712; extreme poverty shares from PIP (Poverty and Inequality Platform) (dashboard), World Bank, Washington, DC, https://pip.worldbank.org/home; carbon dioxide emissions data (2022) from Climate Watch (dashboard), World Resources Institute, Washington, DC, https://www​ .climatewatchdata.org/. Note: The World Bank currently recognizes 26 economies as low-income (GNI per capita, calculated using the World Bank Atlas method, of US$1,135 or less in 2022); 54 as lower-middle-income (GNI per capita of between US$1,136 and US$4,465); 54 as upper-middle-income (GNI per capita of between US$4,466 and US$13,845); and 83 as high-income (GNI per capita of US$13,846 or more). GDP = gross domestic product; GNI = gross national income. 4 | WORLD DEVELOPMENT REPORT 2024 Figure O.2  If capital accumulation were with rising geopolitical tensions and protection- enough, work in middle-income countries ism that can slow the diffusion of knowledge to would be nearly three-quarters as rewarding middle-income countries,4 difficulties in servic- as in the United States, not just one-fifth ing debt obligations, and the additional economic and financial costs of climate change and climate 80 action. 71 70 60 50 Investment, infusion, and Percent 40 innovation—additively and 30 progressively 21 20 To achieve more sophisticated economies, middle-income countries need two successive 10 transitions, not one. In the first, investment is 0 complemented with infusion, so that countries Middle-income countries (primarily lower-middle-income countries) focus GDP per worker relative to US GDP per worker on imitating and diffusing modern technologies. Physical capital and human capital relative In the second, innovation is added to the to that in the United States investment and infusion mix, so that countries (primarily upper-middle-income countries) focus Source: WDR 2024 team using data from PWT (Penn World Table) (database version 10.1), Groningen Growth and on building domestic capabilities to add value Development Centre, Faculty of Economics and Business, to global technologies, ultimately becoming University of Groningen, Groningen, the Netherlands, innovators themselves. In general, middle-income https://www.rug.nl/ggdc/productivity/pwt/. countries need to recalibrate the mix of the three Note: The bars show the simple average for middle-income countries in 2019. The data are calculated using the drivers of economic growth—investment, infusion, methodology outlined in Jones (2016). Following Jones and innovation—as they move through middle- (2016), the figure is based on Hicks-neutral and a constant income status (table O.2). labor share of two-thirds. GDP = gross domestic product. What makes the move from middle-income status to high-income status so difficult? One economic complexity of a country’s export reason is that as  they move through middle- basket—a measure of sophistication—there is a income status, countries cannot leap all at once rising relationship between sophistication and from investment-driven growth to innovation- income for all economies that transitioned from driven growth. Infusion of technology comes a GDP per capita of less than US$13,000 to more first and then innovation. than US$31,000, regardless of whether their export baskets became more or less diversified (figure O.3). Infusion first However, the pace of progress in middle-­ Economic success in lower-income countries income countries is slowing. Average annual stems largely from accelerating investment. As income growth in these countries slipped by these economies move to middle-income status, nearly one-third in the first two decades of this continued progress requires complementing a century—from 5 percent in the 2000s to 3.5 per- good investment climate with measures delib- cent in the 2010s.3 A turnaround is not likely erately designed to bring new ideas from abroad soon because middle-income countries are facing and diffuse them across the economy—so-called ever-stronger headwinds. They are contending infusion. OVERVIEW: MAKING A MIR ACLE | 5 Figure O.3  Economies become more sophisticated as they transition from middle-income to high-income status 1.00 JPN HUN KOR DEU AUT 0.50 POL ITA FRA FIN BEL Economic Complexity Index HKG ESP EST CYP ISR PRT LVA 0 LTU GRC −0.50 −1.00 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 50,000 55,000 60,000 GDP per capita (PWT chained PPP 2017 US$) Diversified economies Concentrated economies Nonparametric trend Source: Bahar, Bustos, and Yıldırım (2024) using PWT (Penn World Table) (database version 10.1), Groningen Growth and Development Centre, Faculty of Economics and Business, University of Groningen, Groningen, the Netherlands, https:// www.rug.nl/ggdc/productivity/pwt/. Note: The figure plots for each economy that transitioned from GDP per capita of less than US$13,000 to more than US$31,000 (50th and 75th percentile, respectively, in 2019) the relationship between GDP per capita and sophistication of its exports. Sophistication is measured as the weighted average of the Economic Complexity Index. The figure shows the sample of economies that diversified (orange solid line)—that is, an economy’s final trend is more diversified than its starting point—and those whose production became more concentrated (dark blue dashed line). For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. GDP = gross domestic product; PPP = purchasing power parity. To intentionally import state-of-the-art sufficient supply of engineers, scientists, man- technology, knowledge of market potential, agers, and other highly skilled professionals. and business practices from abroad, as well as Countries that are relatively open to economic expedite their diffusion at home (figure O.4), ideas from abroad and have instituted strong newly minted middle-income economies have to secondary education and vocational training change tack. Policy makers must support firms programs at home tend to perform better than that are ready and able to incorporate global those that have not. technologies into production. For firms to make The experiences of three economies that have the most of new technologies, they need tech- grown quickly from the lower-middle-income nically skilled workers in large numbers and a to high-income levels in recent decades—Chile, 6 | WORLD DEVELOPMENT REPORT 2024 Table O.2  To achieve high-income status, countries will need to recalibrate their mix of investment, infusion, and innovation ­ INCOME CLASSIFICATION INVESTMENT INFUSION INNOVATION Low-income Higher priority Lower priority Lower priority Lower-middle-income Higher priority Higher priority Lower priority Upper-middle-income Higher priority Higher priority Higher priority Source: WDR 2024 team. Note: The orange dials indicate a strategy that is a priority for that particular income group. The blue dials indicate a strategy that is less of a priority for that particular income group until the priority strategy is successfully achieved. Figure O.4  Middle-income countries must engineer two successive transitions to move to high-income status 1i 2i 3i Investment Investment + Infusion Investment + Infusion + Innovation Relative contribution to growth Capital Productivity Proximity to the frontier Source: WDR 2024 team. Note: The curves illustrate the relative contributions of capital and productivity to economic growth (y-axis), according to countries’ proximity to the frontier (represented by the leading economies). Countries farther out on the x-axis are closer to the frontier. the  Republic of Korea, and Poland—illustrate innovation policies. Korea was among the least these ideas (figure O.5). developed countries globally in the early 1960s, Korea’s success may be the best support for the with income per capita of less than US$1,200 in argument that sustaining high growth requires 1960. By 2023, after an unparalleled five-decade adding infusion to accelerations of investment, run of high output growth, Korea’s income per and then again augmenting the 2i mix with capita had risen to about US$33,000. OVERVIEW: MAKING A MIR ACLE | 7 Figure O.5  In the Republic of Korea, Poland, and Chile, the rapid growth from middle- to high-income status has been interspersed with economic crises 35,000 Korea, Rep. GNI per capita (constant 2015 US$) 30,000 25,000 20,000 Poland 15,000 HIC threshold 10,000 Chile UMIC threshold 5,000 LMIC threshold 0 19 0 19 2 19 4 19 6 19 8 19 0 19 2 19 4 19 6 19 8 80 19 2 19 4 19 6 19 8 90 19 2 19 4 19 6 20 8 20 0 20 2 20 4 20 6 08 20 0 20 2 20 4 16 20 8 20 0 22 6 6 6 6 6 7 7 7 7 7 8 8 8 8 9 9 9 9 0 0 0 0 1 1 1 1 2 19 19 19 20 20 Source: WDR 2024 team using WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https:// datacatalog.worldbank.org/search/dataset/0037712. Note: GNI = gross national income; HIC = high-income country; LMIC = lower-middle-income country; UMIC = upper-middle- income country. In the 1960s, a combination of measures to of private institutions, did its part, setting tar- increase public investment and encourage pri- gets, increasing budgets, and monitoring the vate investment kick-started growth.5 In the development of these skills. These firms also 1970s and 1980s, Korea’s growth was powered required more specialized capital: for a growing by a potent mix of high investment rates and middle-income economy, investment remained infusion, aided by an industrial policy that important. encouraged firms to adopt foreign technolo- Poland’s case is different because of both its gies (figure O.6). Firms received tax credits for socialist past and its membership in the European royalty payments, and family-owned conglom- Union (EU), the most powerful association of erates, or chaebols, took the lead in copying economies ever assembled. But its rapid increase technologies from abroad—primarily Japan. As in income is well known, and a Korea-like 1i to 2i Korean conglomerates caught up with foreign to 3i transition is still discernible. firms and encountered resistance from their In the early 1990s, Poland underwent a erstwhile benefactors, industrial policy shifted transition from a planned economy to a market toward a 3i strategy supporting innovation. economy. It has since boosted its income per capita Then, as Korean firms became more sophis- from 20 percent of the average for the European ticated in what they produced, they needed Union to 50 percent. What is Poland’s winning workers with specialized engineering and man- strategy? It began by disciplining the large state- agement skills. The Ministry of Education, owned enterprises (SOEs). It hardened their through public universities and the regulation budget constraints by cutting subsidies, tightening 8 | WORLD DEVELOPMENT REPORT 2024 Figure O.6  From infusion to innovation in the Republic of Korea a. An agreement between companies b. Government incentives to collaborate on technology (subsidies) Figure A.1: Example of Adoption Contract 40 TECHNICAL COLLABORATION AGREEMENT BY AND BETWEEN NIPPON ELECTROIC CO., LTD. 30 Subsidy rate (%) AND SAMSUNG ELECTRON DEVICES CO., LTD. 20 Section 4 Supply of written Technical Information (a) During the term of this Agreement NEC will upon reasoable 10 request furnish SED with one transparent copy of each drawing, specification and other technical document as well asprograms as andrelated programsand rolntoddocumentation documentation within within the the scope scope specified in Section 1 (d) hereof. The time, manner and other details of furnishing such written NEC Technical 0 Information shall be separately determined by the parties 1970 1980 1990 2000 2010 2020 upon mutual consultation. Adoption subsidy Innovation subsidy Sources: Panel a: National Archives of Korea, https://www.archives.go.kr/english/index.jsp. Panel b: Choi and Shim 2024. Note: Panel b shows the adoption subsidy rate alongside the innovation (R&D) subsidy rate, calculated using the tax credit rate and the corporate tax rate. For example, a 30 percent subsidy rate indicates that firms can receive a reimbursement equivalent to 30 percent of their expenditures on adoption fees or R&D. R&D = research and development. bank loans, and liberalizing import competition— Chile’s success has similar features. In 2012, including at the iconic Stocznia Gdańsk shipyard, Chile became the first Latin American country where the Solidarność (Solidarity) movement to reach high-income status, just two years began. This discipline paved the way for after ­joining the Organisation for Economic comprehensive reform. In Polish SOEs, managers Co-operation and Development (OECD). Chile shifted their focus from production targets to has grown and diversified its exports since the profitability and market share, and they upgraded 1960s, when mining made up about four-fifths firms’ capabilities to prepare for privatization.6 of its exports. This share is now about half. Poland then built on this foundation to attract Knowledge transfers from advanced economies investment, focus on infusion next, and turn  to have been supported by both public and private innovation last. It followed this sequence largely institutions. The public Chilean Agency for by raising productivity with technologies infused Exports Promotion (ProChile) has bolstered from Western  Europe—a process accelerated small and medium enterprises (SMEs), which over in the 2000s  by its entry into the EU common 2013–16 contributed one-third of export value market, which spurred foreign direct investment. added—the other two-thirds being contributed by Poland also increased tertiary education rates large domestic exporters.7 And Fundación Chile, from  15 percent in 2000 to 42  percent in 2012. a private nonprofit created in 1976, promotes Educated  Poles put their skills to work across technology transfer for domestic ventures. One the European  Union, opening another channel example is the adaptation of Norwegian salmon to infusing global  knowledge into the Polish farming technologies to local conditions, making economy. Chile a leading world exporter of salmon. OVERVIEW: MAKING A MIR ACLE | 9 Innovation next patent office, but the patents turned out to be of low quality and lacked any relevance to Once a middle-income country has begun to global markets. Moreover, as the share of firms exhaust the potential of infusion in the most that applied for patents within the economy promising parts of its economy—running out of increased, the wage premium for skilled workers technologies to learn and adopt—it should expand declined, as did the value added.10 its efforts to become an innovation economy. But While Brazil was stumbling at home, Korea this transition is as or more daunting than the was racing around the world, making the infusion preceding one.8 Infusion is powered mainly by the of foreign technology the cornerstone of domes- technology transfers embodied in flows of physi- tic innovation. In 1980, the average productivity cal and financial capital. of a worker in Korea was just 20 percent that of Although innovation requires both of these the average US worker. By 2019, it had tripled to flows, it also needs increasingly vigorous more than 60 percent (figure O.7). By contrast, exchanges of human capital—often triggered Brazilian workers, who had been 40 percent as by a reengagement with the emigrant diaspora, productive as their US counterparts in 1980, were but also creating the conditions cherished by just 25 percent as productive by 2018. innovators such as freer economies, human There are no shortcuts to innovation. It rights, and livable cities. Moreover, to enable is unlikely that industrial policy will enable firms to innovate, governments must have done countries to leapfrog from an investment- and a lot during the infusion phase to reform and manufacturing export–driven model to an strengthen institutions. Weak institutions are innovation-oriented model or services-led model as debilitating as premature attempts to leapfrog of economic growth. The development literature from investment to innovation. In some cases, is littered with reports recommending a leap from ignoring the imperative of infusion to quicken investment to innovation, skipping the stage of innovation can even worsen the investment painful reforms to attract foreign investment climate, setting middle-income economies back and ideas. However, middle-income governments years if not decades. Latin America, ground that have tried to spare their citizenry the pains zero for the middle-income trap, provides a associated with reforms and openness have cautionary example. also kept from them the gains that come from After reaching middle-income status in the sustained growth. 1970s, Brazil veered in the wrong direction. Its policy makers attempted to encourage firms to innovate by bypassing the infusion of foreign The economics of technologies. In 2001, the government imple- mented an innovation-driven economic growth creative destruction strategy, driven in part by fears that foreign tech- The shifts from 1i to 2i to 3i strategies are nei- nology would exacerbate domestic inequality and ther smooth nor linear. They require a mix of lead to dependence on the more advanced econ- economic, social, and political change that Karl omies in the North Atlantic. Notably, it imposed Marx and other philosophers considered impossi- a 10 percent marginal tax rate on payments for ble under capitalism. They reasoned instead that international intellectual property. These tax rev- market-based economies would be riddled with enues were used to subsidize innovation in tar- a growing concentration of wealth and wracked geted sectors, including biotechnology, aviation, by crises until capitalism was replaced by com- health, and agriculture.9 munism. Joseph Schumpeter changed this debate One study found that the subsidies stimulated with his 1942 treatise Capitalism, Socialism and a rapid rise in applications at the Brazilian Democracy and the phenomenon of “creative 10 | WORLD DEVELOPMENT REPORT 2024 Figure O.7  Over the last four decades, as the Republic of Korea’s labor productivity relative to that of the United States continued to climb, Brazil’s peaked—and then sagged 70 Korea, Rep. employed relative to a US worker, %) Labor productivity (GDP per person 60 50 40 30 Brazil 20 10 0 80 82 84 86 88 90 92 94 96 98 00 04 06 08 10 12 14 16 18 20 02 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 Source: WDR 2024 team using data from PWT (Penn World Table) (database version 10.1), Groningen Growth and Development Centre, Faculty of Economics and Business, University of Groningen, Groningen, the Netherlands, https://www.rug.nl/ggdc​ /productivity/pwt/. Note: GDP = gross domestic product. destruction.”11 For Schumpeter, the crises in capi- Energy: Incumbents can collude talist economies could be simultaneously painful (Schumpeter’s view) and restorative. Joseph Schumpeter (1942) wrote that society Nearly a century later, many of Schumpeter’s benefits when entrepreneurs with talent and insights appear to have been confirmed. Indeed, vision introduce new products and technologies, his admonitions and insights have been used by displacing old products and business models and modern Schumpeterian theorists—most notably, generating ever-higher productivity and growth. Aghion and Howitt (1992) and Akcigit and Kerr Often, however, incumbents collude to preserve (2018)—to construct formal frameworks. These the status quo (figure O.8, panel a). In today’s advances in growth theory are useful in helping environment, Schumpeter’s view is perhaps solve the hardest problem facing the global econ- best reflected in the contest between high- and omy today: how should the 108 middle-income low-carbon energy. High-carbon energy, particu- countries with 75 percent of the world’s people, larly coal, has been an incumbent technology for percent 60 percent of global emissions, but just 40 ­ over 300 years (box O.1). Technical progress has of global output correct these imbalances while followed a path over which the efficiency with converging toward the living standards of which fossil fuels are extracted and burned has advanced economies? increased, urban infrastructure has been built Schumpeterian ideas provide helpful clues. around the private motor vehicle, social atti- Success seems to come most quickly to societ- tudes and personal preferences are supportive ies that balance the economic forces of creation, of high carbon consumption, and political pres- preservation, and destruction. sure groups represent carbon-intensive interests. OVERVIEW: MAKING A MIR ACLE | 11 Figure O.8  Three views of creative destruction a. Energy b. Talent c. Enterprise (Schumpeter) TECHNOLOGY FRONTIER (Aghion and Howitt) (Akcigit and Kerr) TECHNOLOGY FRONTIER TECHNOLOGY FRONTIER Entrant Incumbents Entrant Incumbent Incumbent Entrant Incumbent Economic progress Economic progress Economic progress Schumpeter’s view: incumbents Aghion and Howitt’s view: entrants Akcigit and Kerr’s view: incumbents can collude create value and displace incumbents and entrants add value Source: WDR 2024 team based on Schumpeter (1942); Aghion and Howitt (1992); Akcigit and Kerr (2018). Box O.1  Who and what are incumbents? Leading firms, technologies, nations, elites—and men Incumbents are firms that usually have well-established brand names recognized and trusted by consumers. They often have better access to financial resources, such as capi- tal for investment and technological infusion, and human resources, such as experienced employees. They also may have established relationships with suppliers and distributors, which can be leveraged to maintain a competitive edge. And they have resources to invest in research and development and to invent products and processes they can protect with patents. Incumbents are well-established energy sources such as fossil fuels. Since 1709 when Abraham Darby, a British ironmaster, first smelted iron ore with coke, coal has been the fuel of choice around the world. In the more than 300 years since Darby’s innovation, coal has become the largest source of electricity generation worldwide, producing more than one-third of global electricity in 2022.a Cities and economies have been built on cheap coal-powered energy, fueling their prosperity. However, the widespread use of coal gen- erates the highest energy-related carbon dioxide emissions—15.5 gigatons—representing 42 percent of total emissions in 2022. Incumbents are technologically advanced nations. They can share technologies with emerging economies through investing in, licensing, training, and hosting foreign stu- dents. For decades, they were instrumental in supporting the growth of emerging econo- mies. But today, they are erecting walls to subsidize their domestic firms, blocking others from joining their value chains. (Box continues next page) 12 | WORLD DEVELOPMENT REPORT 2024 Box O.1  Who and what are incumbents? Leading firms, technologies, nations, elites—and men (continued) Finally, incumbents are elites in society. They are always powerful, generally wealthy, and—in middle-income economies—mostly men. But they are not all against prog- ress. Elites can have the education and resources to accelerate growth by infusing their ­ economies with global technologies. For a middle-income country seeking to infuse and innovate, elites may serve as the go-to pool of trained professionals, managers, entrepre- neurs, and innovators. Men are also incumbents, for centuries enjoying better education and job opportunities than women and defining laws and institutions, often to buy social, economic, and political power. Such power has given them an outsize say in deciding who studies where and what, who gets a well-paid job, and who gets to start a business. Meanwhile, misogyny may keep women out of the market or at least the most desired jobs and business opportunities. Large firms, social elites, powerful men, and advanced economies have, however, also helped new entrants. The size and ownership of enterprises and the socioeconomic status and gender of individuals are not reliable attributes on which to base policy. a. IEA (2023). The result is that the returns to investing in high-­ technologies in middle-income countries is carbon activities are large because of all the com- patchy, reflecting a landscape of legacy policies plementary high-carbon investments that have that preserve a high-carbon economy. Middle- been made. income countries have a greenhouse gas (GHG) In many middle-income countries, power intensity of GDP that is 3.5 times higher than that markets are still a monopoly: an SOE operating of high-income countries. This difference reflects under a vertically integrated utility remains in both the misallocation in the use of energy (with charge of generation, transmission, distribution, the energy intensity of GDP also 2.5 times higher and the retail supply. This arrangement hinders than in high-income countries) and the lower competition and results in the inefficient use of diffusion of low-carbon energy technologies resources. In addition, in many middle-income (figure O.9, panel a). countries the first generators dispatched are often not those with the lower marginal prices (that is, power dispatch often does not follow Talent: Entrants create value and merit order), serving as a barrier to the expan- displace incumbents (Aghion and sion of renewables with rapidly declining costs. Howitt’s view) In countries that include Pakistan, Poland, South Schumpeter’s ideas on creative destruction served Africa, and Türkiye, SOEs account for 84 percent as the inspiration for one of the most influen- of total installed capacity. By contrast, the private tial papers in economics that emerged from a sector owns about an equal share (80 percent) of fortuitous collaboration between two econo- the installed capacity of renewable energy.12 mists. In the summer of 1987, Philippe Aghion, a Although advances in low-carbon energy new professor at the Massachusetts Institute of can help to decouple economic growth from Technology (MIT), and Peter Howitt, a Canadian carbon emissions, the diffusion of low-carbon economist, formalized a theory of creative OVERVIEW: MAKING A MIR ACLE | 13 destruction in which economies expand mainly of talent. Talent is wasted wherever that acquired through innovation by entrants.13 Entrants chal- through education, training, and work experience lenge incumbents and become the protagonists is allocated not by merit, but according to other of economic growth (figure O.8, panel b). factors outside the control of individuals. Gender, This formulation of creative destruction empha- family background, ethnic and cultural identity— sizes the importance of both creating ever-larger none of these factors should matter for school reservoirs of talent and improving the allocation enrollment or career prospects in a country aspir- of talent to tasks. Not investing in the talents of ing to grow rapidly through infusion and innova- women and minorities, keeping them out of the income tion. But for the average child in a middle-­ most rewarding activities, and adopting unfair country today, they matter all too much. compensation practices are surely the most self-­ Economically and socially mobile societies are defeating attributes of middle-income economies, better at developing skills and utilizing talent, where skills are already scarce. When these prac- but social mobility in middle-income countries tices are discouraged, the payoff can be immense. is about 40 percent lower than that in advanced In the United States between 1960 and 2010, the economies.15 Middle-income countries will need decline in gender and racial discrimination in edu- to ensure that more individuals, regardless of cation and work explains up to 40  ­ percent of the their parents’ circumstances, have better oppor- observed growth during that period.14 tunities to become skilled workers. And social As they grow, middle-income countries will mobility matters much more in middle-income need skilled workers such as engineers, technicians, countries than in low-income countries simply and managers, but they have smaller reservoirs because the former need more skilled workers of skilled talent than advanced economies. And to invest, infuse, innovate, and grow (figure O.9, yet preservation forces discourage the acquisition panel b). Figure O.9  Creation is a weak force in middle-income countries, where it is characterized by a rampant misallocation of resources a. In middle-income countries, economic growth is more carbon-intensive than in high-income countries, and middle-income countries lag in energy efficiency and adoption of low-carbon technologies 3.5 times higher 3.5 3.0 Proximity to the frontier 2.5 times 2.5 higher 2.0 1.5 HIC = 1 1.0 40% lower 42% lower 53% lower 0.5 0 GHG emissions Energy intensity Solar electricity Wind electricity Electric vehicle intensity (as share of GDP) generation (as share generation (as share deployment (as share of GDP) of total electricity of total electricity (per million population) generation) generation) (Figure continues next page) 14 | WORLD DEVELOPMENT REPORT 2024 Figure O.9  Creation is a weak force in middle-income countries, where it is characterized by a rampant misallocation of resources (continued) b. Intergenerational mobility matters more for skill c. “Flat and stay” versus “up or out”: Efficient development in middle-income countries than in firms do not expand, and inefficient firms do low-income countries not exit the market in India, Mexico, and Peru 8 to firms <5 years of age HICs 7 Employment relative 6 UMICs 5 4 LMICs 3 2 LICs 1 0 −0.2 0 0.2 0.4 0.6 <5 10−14 20−24 30−34 ≥40 Regression coefficient of skilled workers’ Age of firms (years) intergenerational mobility India Mexico Peru United States Sources: Panel a: Chepeliev and Corong 2022; Energy Institute 2023; Statistics Data (portal), International Renewable Energy Agency, Abu Dhabi, United Arab Emirates, https://www.irena.org/Data; WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712. Panel b: WDR 2024 team estimates based on GDIM (Global Database on Intergenerational Mobility) (dashboard), Data Catalog, World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0050771/global-database-on-intergenerational​ -mobility. Panel c: India, Mexico, and the United States: Hsieh and Klenow 2014; Peru: World Bank 2015. Note: Panel a displays for middle-income countries compared with an index of 1, representing the high-income country (HIC) frontier, the greenhouse gas (GHG) emissions intensity of the gross domestic product (GDP), the energy intensity of GDP, the share of solar and wind energy in total electricity generation, as well as the battery electric vehicles per million population. Panel b plots regression coefficients of intergenerational mobility (which is equal to 1 minus the intergenerational relative mobility) for different country groups at the 95 percent confidence interval. The dependent variable in the regression is the share of skilled workers (“Legislators, sr. officials, managers”; “Professionals”; “Technicians and associate professionals”). The regression controls for the log of GDP per capita when the 1980s birth cohort was growing up. Intergenerational mobility estimates are for educational mobility of the 1980s cohort from the World Bank’s GDIM. HICs = high-income countries; LICs = low-income countries; LMICs = lower-middle-income countries; UMICs = upper-middle-income countries. Panel c illustrates the average employment across a cohort of firms of different ages in the cross-section of firms. The number of employees serves as a proxy for firm size. The y-axis axis reports the average employment of each cohort relative to the average employment across firms under five years of age. Why do preservation forces persist in con- inequality across generations, exacerbating the straining the opportunities for so many people? inequality of opportunity. Part of the answer is that preservation insu- Three kinds of preservation forces perpetuate lates members of social elites from the forces of social immobility in middle-income countries, destruction that, in a more open society with shutting out talent from economic creation. The meritocratic institutions, might dissipate their first force is norms—biases that foreclose or limit advantages in wealth and privilege. The same opportunity for women and other members of forces ensure that, beyond elites, few children marginalized groups. Next are networks—above will get the chance to climb to a higher rung on all, family connections. And the last force is a country’s income ladder than that occupied by neighborhoods—regional and local disparities their parents. So, income inequality remains high in access to education and jobs. Although all and social mobility remains low, transmitting three factors can have positive impacts on talent OVERVIEW: MAKING A MIR ACLE | 15 ­reation—filling voids left by missing markets c it will roughly double in size. In the United States, and services—they become forces of preservation the average firm that survives that long will when they block the disadvantaged from accessing grow sevenfold (figure O.9, panel c). For firms in opportunity. middle-income countries, this implies a “flat and stay” dynamic: firms that fail to grow substan- tially can still survive for decades. By contrast, Enterprise: Incumbents and for US firms the dynamic is “up or out”: facing entrants add value (Akcigit intense competitive pressure, a few entrepreneurs and Kerr’s view) will expand their businesses rapidly, while most The original Schumpeterian premise that new others will exit quickly. Among the majority who entrants drive change and create new economic exit the market, many will become wage earners potential while incumbents are inert runs counter at the most flourishing firms. to the latest empirical evidence on enterprises. In keeping with the flat and stay dynamics, Globally, larger and more established firms are firms in India, Mexico, and Peru tend to remain infusing new knowledge into their businesses at a microenterprises: nearly nine-tenths of firms have higher rate than smaller firms. In middle-income fewer than five employees, and only a tiny minority countries, it is the large firms that are employ- have 10 or more. The longevity of undersize ing the majority of highly skilled workers.16 firms—many of them informal—points to market Throughout the twentieth century, the United distortions that keep enterprises small while States effectively transitioned its innovation keeping too many in business. For example, a high focus from individuals working in their garages regulatory cost attached to formal business growth to established firms, leveraging advantages such may inhibit an efficient firm from gaining market as risk management, market access, brand reputa- share and driving out inefficient competitors. tion, and collaboration. These firms now account Such policy-induced distortions in middle- for over 75 percent of patents filed at the United income countries result in misallocated resources, States Patent and Trademark Office.17 hampering creation and infusion at scale. A third generation of Schumpeterian econo- mists have formalized the idea that both incum- Balancing the three forces bent enterprises and entrants can create value Looked at it this way, middle-income countries (figure O.8, panel c).18 Market leaders—successful face common challenges in balancing the three incumbents—can bring scale and advance domes- forces: tic industry by investing in upgraded products and business practices, as well as technology for • Creation—the primary protagonist of eco- new markets. Scale allows for adopting modern nomic growth—is a weak force in many management practices, for hiring and rewarding middle-income countries. Large incum- skilled workers, and for making the most produc- bents are slow to develop new products tive use of large amounts of capital. In other words, and processes, and, although small firms scale gives incumbents the power to boost their are continually entering various markets, efficiency, whether in the expectation of compe- most of them do not create or disrupt. tition from other incumbents or from entrants Periods of growth are also times of cre- or in response to it. Scale also allows incumbents ation, and thus of structural change. to specialize in multiple product lines, changing • Preservation—the arch antagonist of course to parry the new offerings of competitors. ­ creation—is the strongest force in middle-­ However, the forces of creation are weak in income countries. The same market leaders middle-income countries. In India, Mexico, and who could enable middle-income coun- Peru, for example, if a firm operates for 40 years, tries to speed up the infusion of global 16 | WORLD DEVELOPMENT REPORT 2024 knowledge are too often slowing down Striking the right balance the process. Incumbent firms and elites Middle-income countries are hampered by an are often successful in keeping things as imbalance among the forces of creation, preser- they are whether through market power vation, and destruction. The forces of creation and collusion, through capture of poli- are weak, the forces of preservation are strong, cies and regulations, or through educa- and destruction is held back by the forces of tion systems and labor markets that place preservation. Middle-income countries must more importance on socioeconomic sta- therefore balance these forces (figure O.10). tus than on talent or merit. That means • Destruction—a necessary evil that clears the way for creation by freeing up misallocated • Disciplining incumbency to weaken the resources and sweeping away outdated forces of preservation institutions—is kept weak in middle-income • Rewarding merit activities—those with countries by opposition from those with positive effects on general well-being market power or government influence. and that aid in the efficient use of talent, A growing economy that requires new capital, and energy—to strengthen the arrangements in capital, labor, and energy forces of creation markets needs to release itself from less • Capitalizing on crises to aid the destruction efficient ones. To the extent that weak of outdated policies and institutions that institutions and policies preserve outdated are difficult to dislodge during boom times. arrangements, creative destruction is stifled. However, this opposition tends to These principles can help middle-income weaken during crises—whether economic, countries calibrate the mix of their three i’s— political, or ecological. When crises place investment, infusion, and innovation—as they plan intense pressure on governments to act, to accelerate economic growth. Because both a window opens for reforms. incumbent firms and entrants can add value, Figure O.10  Middle-income countries have to strike a balance among creation, preservation, and destruction Creation Creation Preservation Destruction Preservation Destruction Source: WDR 2024 team. OVERVIEW: MAKING A MIR ACLE | 17 industrial policies will need to focus on disciplining other  countries. Such contestability is central to incumbency. And because the talent of women and creative destruction. disadvantaged groups is grossly underutilized, A key part of contestability is openness to for- social policies will need to focus more on rewarding eign investors and global value chains that give their merit and advancing social and economic domestic firms access to larger markets, technol- mobility. And finally, because economic growth ogy, and know-how and allows them to add value over the last three centuries has been emissions- and grow. And they are encouraged to make use intensive, middle-income countries will need to of that access, thereby exposing domestic firms capitalize on today’s energy crisis to cut emissions to competition, but also inspiration, from inter- while balancing energy access and security. national firms that operate at or near the global technology frontier. Firms at home can seize the Enterprise, openness, and reforms opportunity to infuse technology, increase the sophistication of their operations, and scale up, Countries growing out of low-income status into or they can keep doing business as usual and be middle-income status tend to have a 1i strategy for eased out. accelerating investment. Stronger institutions are For example, in Chile imports of Chinese prod- needed to control inflation, ensure financial and ucts rose at an average pace of 27 percent a year macroeconomic stability, expand economic and from 2001 to 2007, and large Chilean incumbent political freedoms, and enforce the rule of law to firms, or market leaders, boosted their product encourage both domestic and foreign investment. innovation by 15 percent and their product quality Even if all middle-income countries enjoyed such by 22 percent.19 In Argentina, after MERCOSUR enabling conditions, a 1i strategy would not be (Southern Common Market) was established, enough to support sustained growth and move domestic firms in sectors facing export tariff reduc- these countries out of the middle-income level. tions began to invest more in computing tech- Why? The returns from capital investment alone nology and in technology transfers and patents.20 decline steadily. Growth in middle-income coun- Again, in 12 European countries over 2000–2007 tries is boosted when economies take on new more than 15 percent of the increase in patent- structures, enabled by a 2i strategy focusing on ing, information technology intensity, and pro- both investment and infusion. Institutions will ductivity was driven by import competition from need to create an environment conducive to inte- China—and successful European firms boosted grating global technologies into the domestic management quality while increasing research and economy. development (R&D) and adding new skills.21 Make markets globally contestable Contestable markets—and the institutions that Connect local firms with market leaders enable them—are vital for middle-income coun- Because local firms often do not have informa- tries that aim to become a global supplier and tion on specific technologies and the know-how sustain rapid economic growth through sophisti- to adopt them, consultants and advisory firms cation and scale. founded by experts can provide expertise and Contestability is not chaos: it does not mean advice on technology adoption and implementa- that firms in middle-income countries cannot tion. Market leaders—especially multinationals— earn comfortable market positions, becoming are often vanguards in technology and technical established and relatively difficult to displace. capabilities and can be some of the best partners However, contestability does mean that firms for local firms, working together to deploy new feel pressure to compete because their cur- technologies. The government can help make rent products and processes can be displaced by the relevant connections. For example, in Chile technologically sophisticated producers from the Supplier Development Program, which offers 18 | WORLD DEVELOPMENT REPORT 2024 large domestic buying firms an incentive to con- other firms, and misallocate resources.26 In coun- nect with suppliers that are SMEs, increased the tries that include Japan, Mexico, and Viet Nam, suppliers’ sales by 16 percent and their employ- public support for small firms—not necessarily ment by 8 percent. It also boosted the sales of large young firms—reduced productivity and increased sponsor firms by 19 percent.22 Governments can resource misallocation.27 also provide firms with information on market Even where tax codes do not create explicit pro- opportunities, enabling them to access finance visions based on firm size, middle-income coun- and strengthen their capabilities, as well as to rec- tries may be creating a practical subsidy to SMEs ognize opportunity and mobilize themselves to through size-dependent tax enforcement—that take advantage of it.23 is, governments with weak tax collection capacity may concentrate enforcement on larger firms.28 Reduce factor and product market regulations In Mexico, eliminating distortions created by Reforms that roll back protection for specific size-dependent taxation policies favoring small activities, enterprises, families, or industries rein- firms could boost output by 9 percent.29 In Chile, force the gains from openness. However, today China, and India, reductions in distortions helped middle-income countries are slow to combine these economies close the gap between actual and investment with infusion and innovation, sty- potential productivity by 10 percent. mied by the powerful institutional and regula- tory forces of preservation. Especially binding are Let go of unproductive firms product market regulations. Besides imposing Letting inefficient firms and business models fail constraints on international trade and invest- is a core principle of creative destruction. Studies ment, these regulations prop up state control of of firm exit—stemming from seminal work by business and impose legal and administrative Hopenhayn (1992)—have revealed that the exit of barriers to entrepreneurship, thereby hobbling less productive firms contributes substantially to investment and infusion at scale. raising aggregate productivity. In many countries, during periods of trade liberalization the exit of Move away from coddling small firms or the least productive firms has boosted growth.30 In ­ vilifying large firms middle-income countries, however, bureaucratic Small and medium-size enterprises are wide- frictions prolong the survival of zombie firms— spread in middle-income countries. Ideally, sub- inefficient, debt-ridden companies that crowd sidies would help SMEs grow into larger, more out investment by productive firms.31 Reforms of productive companies that pay higher wages bankruptcy laws should focus on enabling failed and adapt knowledge. But the same support also businesses to exit swiftly and predictably and on strengthens the forces of preservation by reduc- allowing viable businesses to restructure. ing incentives for a productive firm to expand, deterring it from scaling up production. Many Strengthen competition agencies firms in middle-income countries remain small As segments of an economy master infusion, even when long established; they simply do not they will need to adopt a 3i strategy. Institutions aspire to grow.24 The abundance of small firms can foster the development of new technologies in middle-income countries does not solely mir- and ensure that entrants—new entrepreneurs— ror the challenges they face. Instead, it indicates a are not blocked by established incumbents, deficiency in competition, originating from larger regulatory barriers, and entrenched industry firms that would have displaced them in the mar- practices. Antitrust laws can help prevent abuse ket if they had expanded.25 Blanket support for of dominance by established incumbents. As small firms can curtail the exit of unproductive economies (or sectors) move closer to the tech- small businesses, perpetuate smallness, crowd out nology frontier, competition agencies will need OVERVIEW: MAKING A MIR ACLE | 19 to consider a possible trade-off between inno- social mobility.34 Such conditions can provide vation incentives and market power.32 Although both social stability and economic dynamism, market power enables investment in R&D to which are equally necessary for middle-income bring new ideas to market, firms may resort to countries to grow to high-income status. In fact, anticompetitive behavior. Thus, competition barriers to social mobility can derail a country’s and innovation policies need coordination, plans for moving beyond a 1i strategy. alongside developing independent, capable com- petition authorities. Discipline, not vilify, elites For upper-middle-income countries shifting Social and economic elites can be either creative to a 3i strategy, a special concern is the con- or inimical to creation. For a middle-income tainment of killer acquisitions—that is, when country seeking rapidly to enrich its talent incumbents acquire innovative firms specifically pool, it would be self-defeating to lower elites’ to kill future competing products and technolo- ambitions. Elites are most able to invest in their gies.33 But not all acquisitions are deadly: many children’s education—and larger investments, young entrepreneurs make a deliberate effort to and better investment choices, yield increasing be acquired by an incumbent, producing com- returns to parental background.35 Elites are also plementary innovations that an incumbent can best connected for job searches and placements. scale up. And elite women can most readily become role models for other women through education Deepen capital markets and professional work. However, elites—like Switching from a 2i to a 3i strategy also has large incumbent firms—need to be disciplined implications for how firms access finance. Equity because of their power to capture institutions. If markets can be instrumental in supporting inno- elites hog education, jobs, capital, and assets for vative activities, especially in private firms, which themselves, thereby limiting access to outsiders, typically face larger financing gaps than publicly a middle-income country is suffering from elite listed firms. However, private markets for equity capture: by preserving privilege, it is stymying financing lack depth and access in emerging creation. economies (figure O.11). Start-up incubators and accelerators can be particularly helpful, providing Invest in talent and reward merit mentorship, resources, networking opportunities, People who are not only talented, but also— and sometimes funding to help start-ups grow crucially—educated and have access to labor and compete. markets, enterprise opportunities, and business financing are key to the 2i and 3i strategies. Policy Education, social mobility, and makers should especially consider initiatives to entrepreneurship educate women, along with other excluded and marginalized groups, and to let families become As more parts of an economy shift from 1i more socially and economically mobile with each to 2i and 3i strategies, demand increases for succeeding generation. highly skilled workers—technicians, managers, From the successes of former middle-income scientists, and other professionals. This demand countries that have attained high income, three can increase income inequality. But, if it is simple lessons emerge for education reform: accompanied by policies that expand access to higher education and reduce barriers for women • Broaden access to foundational skills. and other disadvantaged groups so that they Graduate more students from high are now rewarded for their skills and able to school, broadening and deepening the create new businesses, it also generates greater talent pool. 20 | WORLD DEVELOPMENT REPORT 2024 Figure O.11  In emerging market and developing economies, few companies are funded by venture capital or private equity 100 35 90 30 ventural capital per 1 million persons private equity per 1 million persons 80 Number of companies funded by Number of companies funded by 70 25 60 20 50 15 40 30 10 20 5 10 0 0 20,000 40,000 60,000 80,000 100,000 120,000 GDP per capita (US$), 2020 High-income economies funded by venture capital EMDEs funded by venture capital High-income economies funded by private equity EMDEs funded by private equity Source: Didier and Chelva 2023.  Note: The figure displays the number of companies funded by venture capital (left axis) and private equity (right axis) from deals concluded 2018–19. Economies are classified according to the World Bank’s income classification as of June 2020 (Serajuddin and Hamadeh 2020). EMDEs = emerging market and developing economies; GDP = gross domestic product. • Monitor learning outcomes using stu- Although many middle-income countries have dent assessments. Gauge progress toward expanded tertiary education, a critical difference explicit policy goals. between those that graduated to high-income sta- • Embed educational reforms in a national tus and those that did not is that the former never economic growth strategy. For example, wavered in their commitment to foundational in the early 1970s, as Finland’s economy skills, thereby developing a large pipeline of talent. became less resource-dependent and Missing the opportunity to learn while in school agrarian and more urban and industrial, is largely irreversible for children; they may not the country reformed education to meet have a chance to study later in life. Strengthening the demands of firms and of a growing foundational skills requires efficient and effective middle class. spending on education because spending by itself is not a guarantee of better learning outcomes.36 Growing the talent pool takes time, and past Countries may consider adopting the “progres- mistakes can impede countries for decades. sive universalism” principle: add incrementally OVERVIEW: MAKING A MIR ACLE | 21 to higher education investments as the quality Leverage digital technologies at lower schooling levels rises to include more Digital technologies—such as the internet, students. mobile phones, social media, and web-based Meanwhile, middle-income countries are information systems—can promote both social not only talent-scarce relative to advanced mobility and talent development. When Nandan economies, but also not nearly as effective Nilekani, one of India’s leading technology in allocating the existing talent to tasks. For entrepreneurs, was tasked with developing example, these countries do not fully reward the Aadhar (the country’s digital identification talents of women and people from less privileged system) in 2009, he paved the way for Indians families, while simultaneously protecting to accumulate digital capital (digital footprints less able people from privileged families from of online activity and digital payments). Digital competition in education. footprints become digital capital, which Policies to ensure equal opportunities for individuals own and can choose to make women, minorities, and other disadvantaged available to lenders when getting access to credit. groups whose talents have been undeveloped Digital data on payments, receipts, taxes, and or unrewarded are likely to increase both eco- loan repayments all make it possible to assess nomic efficiency and equity. However, in many financial credibility. According to a recent study, countries patriarchal gender norms are part digital capital has increased entrepreneurship of a deeply entrenched system of preservation, and business income in India and has favored limiting women’s earning power and social small-scale vendors and economically lagging and economic mobility across occupations and districts.39 By delivering instructional material, generations. Where economic and social rights digital technologies also provide students from favor men, ­m iddle-income countries that aspire disadvantaged backgrounds with opportunities to grow quickly must work hard to grant the to learn. same opportunities to women. Institutions and policies are needed to counter the exclusion Reward innovators and scientists to match of women—among others—from education, brain drain with brain gain employment, enterprise financing, and con- Investing in advanced skills is costly. Individuals tracting and to provide policies such as child- invest in these skills with the expectation that care support or flexible work for both men and their talent and acquired ability will be rewarded.40 women. However, these rewards are often found on foreign In education, policies that support girls shores. World Development Report 2023 reported who stay in school longer by offering them that in middle-income countries, 10 percent of scholarships or conditional cash transfers highly skilled workers emigrate, with high-level can improve outcomes for women.37 To boost skills in greater demand in Western Europe and female students’ interest in science, technol- North America.41 To counter the brain drain, the ogy, engineering, and mathematics (STEM), report recommended that origin countries expand mentoring and information interventions have their capacity for training highly skilled workers proven to be among the most effective meth- because greater capacity increases the likelihood ods.38 However, because women face social, that a sufficient number of highly skilled workers family, and logistical constraints—including will stay even when others migrate. household and childcare responsibilities— As countries adopt a 3i strategy, they will educating women is most effective when com- need to tap into the knowledge and know-how plemented by other interventions to address of a country’s diaspora. The emigration of highly these constraints. skilled individuals can serve as an opportunity 22 | WORLD DEVELOPMENT REPORT 2024 for the origin country if emigrants remain country’s policies address emigration. The most connected to the origin country—or even return. likely migrant to be exposed to modern pro- This is particularly relevant in conflict-affected duction processes and technologies and to countries such as Ukraine that have experienced transmit valuable knowledge back to the origin a large outflow of highly skilled individuals. country is highly skilled, moves to an advanced When the demand for advanced skills increases, economy, and works there in a leading occupa- the diaspora becomes an important talent pool to tion as a manager, professional, or technician germinate innovation at home. ­(figure O.12). As migrants acquire skills abroad, migration Building and expanding high-quality may drive a brain gain in the sending c ­ ountry. universities—institutions that can train top Whether the sending country experiences talent and contribute to innovation—require an brain drain or brain gain varies across coun- efficient system of public funding for research, tries, depending largely on how the sending along with fluid university-industry connections Figure O.12  Countries with large, successful diasporas have the highest potential for knowledge transfers Share of highly skilled migrants in “good” occupations (%) 70 LSO ZMB NZL AUS SUR ZAF 60 ZWE ISR FRA PNG MWI KEN BHS SGP VUT MLT MYS BELESP CAN ISL TZA SLE MUS UGA SWE LBN USA LUX NLD DEU IND NAM BRN FIN 50 SLB SWZ BWA BEN LBR CYP DNK GRC IRL JAM VNM BHR KWT MDGKHM AUT CHE TTO NGA ITA GBR QAT CMRGUY IRN CHN MOZ ARE TGO BLZ FJI CRI NOR SYR PAN GHA ARG JPN TJK BFA VCT BRB LAO TUN PRT VEN TUR DZA NER DJI CPV COGRWA LBY MMR ETH HRV LKAHTI PHL 40 WSM TCD CAF LCA URY CIV BLR BIH CHL BRA MAR MKD JORSEN IDN MDV GAB COL BDI SVN SVK CZE HUN EGY MRT AGO ERI NPL CUB PAK OMN GMB UZB SDN NIC BGR DOM RUS POL YEM ARM SLV THA PER UKR ROM KGZ AZE AFG GTM MEX 30 SAU ALB ECU MLI SOM BGD TKM FSM PSE PRY COD COM BOL HND IRQ GIN EST TON LVA MDA GNB GEO KAZ STP GNQ LTU 20 TMP MNG BTN 10 0 5 10 15 Number of migrants to Western Europe and North America (log scale) Low-income Middle-income High-income China and India Source: WDR 2024 team. Note: Data on migration flows by skill and current occupation are from OECD DIOC 2010–11, which covers migration flows from 200 origins to 34 OECD (Organisation for Economic Co-operation and Development) country destinations. Each scatter point represents an origin (or birth) country. For each birth country, the x-axis shows the number of migrants over 15 years of age who had completed tertiary education and were living in destination countries in Western Europe or North America (AUT, BEL, FRA, DEU, NLD, CHE, USA, GBR, IRL, CAN, ESP, ITA, DNK, NOR, SWE; in logs), and the y-axis the share of these tertiary- educated migrants working as managers, professionals, or technicians (“good” occupations) in the destination country. For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. OVERVIEW: MAKING A MIR ACLE | 23 to promote exchange of knowledge. Most efficient address the crisis. Today, two crises—the climate for middle-income countries is to focus public crisis and the global energy crisis—are combining funding on a few strategic research areas, such as to drive rapid progress in low-carbon technolo- STEM, health, and energy transition, with funds gies, defined as technologies or applications that allocated competitively and based on performance. counter the effects of climate change. Partnerships with world-class universities can be a strategy for developing a research base. Discipline advanced economies to reduce the To encourage knowledge exchanges between cost of global decarbonization universities and industries, countries could As middle-income countries move to a 2i strategy, grant R&D funding for such partnerships. they will have opportunities to join globalized Governments could also offer firms tax incentives supply chains for low-carbon products and reduce for collaborating with universities. Establishing the cost of decarbonization worldwide. However, a regulatory framework for knowledge exchange their success will depend on advanced econo- is key—especially to define the government’s mies easing up on protectionism in trade policy. intellectual property rights to knowledge Protectionist measures in advanced economies produced by universities with public resources. could prove to be the bane of the global energy One desirable outcome of university-industry transition. collaboration is venture creation by university Previous waves of middle-income countries faculty, staff, students, and postdoctoral fellows, have transitioned to high-income status with the with private investors serving as venture help of coordinated trade policies in a globally capitalists. Universities can also form partnerships integrated economy. By contrast, today’s m ­ iddle- to provide services to local companies. income countries are navigating a hazier trade landscape. Countries have not yet agreed on the key rules for low-carbon energy product supply Energy, emissions, and crisis chains. And “make local” subsidies will likely management do a lot to relocate production—to the United The destruction of outdated arrangements— States, to the European Union, and to a growing enterprises, jobs, technologies, private contracts, number of other economies that are embracing policies, and public institutions—is essential for “reshoring” efforts and enacting local content an economy to ensure that it has the appropriate requirements. For example, initial modeling balance of investment, infusion, and innovation. suggests that the US Inflation Reduction Act will But in many countries the forces of destruction substantially attract industry toward the United are weak during boom times, whereas crises States, Mexico, and Canada and away from other often play an outsize role in weakening the forces major producers.44 In effect, these subsidies and of preservation, making way for the forces of protectionist measures in high-income countries creation. threaten to lock middle-income countries out of In the context of energy, the oil price shocks low-carbon value chains. in the 1980s increased the relative cost of fos- To be clear, subsidies have a role to play in a sil fuels and played a major role in accelerating global transition to low-carbon energy sources investments in energy efficiency and the develop- in view of the positive externalities of such a ment of cleaner energy technologies.42 The global transition and the extent of today’s market financial crisis of 2007–09 coincided with a sig- failures. But they should not distinguish between nificant increase in the uptake of renewables.43 domestic and foreign suppliers. Each segment of Renewable energy use grew rapidly in the United the value chain should be located where a product States, China, and Germany in part because of can be made at the lowest cost, averting a risk of the stimulus programs governments enacted to protectionist retaliation and a race to the bottom 24 | WORLD DEVELOPMENT REPORT 2024 (the most distorted and least efficient market demand for energy will rise dramatically. In structure). But such globally rational thinking is fact, the International Energy Agency (IEA) has rarely favored by leaders with domestic politics on predicted that the electricity demand by global their minds. They are unlikely to enact subsidies data centers will more than double from 2022 to consistent with a globally integrated economy 2026, with artificial intelligence playing a major because such subsidies would allow gains from role in that increase.45 supply chain reallocation to accrue to firms based Middle-income countries will need to decide in other countries. how best to reduce the carbon emissions of their Faced with this conundrum, policy makers growing economies—a combination of energy in advanced economies should consider that the intensity (energy consumed per US dollar of energy transition to low-carbon energy sources GDP) and carbon intensity (carbon emissions has many benefits, not just through its effects per unit of energy). Today, emissions from a on the climate, but also through its implications growing economy outweigh the reductions in for the economic development of middle-income emissions from lowering energy intensity and countries. To lock middle-income countries out of carbon intensity. To decouple emissions from global value chains with protectionist measures is economic growth, governments will need to to deny firms and industries in those countries discipline incumbency, reward merit, and derisk the benefits of learning-by-doing spillovers. investments in low-carbon energy: To accommodate middle-income countries and support a global transition to low-carbon • Disciplining incumbency. Disciplining energy, policy makers in advanced economies the incumbency advantage is especially will need to update trade policy rules by limit- important for increasing energy efficiency ing green subsidies, export controls, and import and decoupling emissions from economic controls and using clear language to define their growth. Market contestability, as well as appropriate use. One option is to modify existing opportunities for value-adding firms to agreements with supplementary clauses, much grow, spurs the adoption of energy-saving in the same way that Articles 20  and  21 of the technologies. In Georgia, for example, General Agreement on Tariffs and Trade (GATT) markets with a higher concentration have were used to carve out exceptions. Such clauses lower energy efficiency. In Argentina, can transparently acknowledge that all countries firms with a higher share of skilled need to nurture emerging domestic industries if workers are better able to adopt advanced they are to achieve a just transition with energy green technologies.46 Exporters also security. But the use of subsidies should also be tend to have lower emissions intensity restricted to specific circumstances, such as the than nonexporters.47 If incumbents are need for public support to develop and commer- disciplined, energy price increases hold cialize innovative low-carbon technologies. considerable potential for firms to reduce energy intensity. In the longer term, Decouple emissions from economic growth increases in energy prices tend to be fully Rising incomes increase the demand for energy— compensated for by higher efficiency.48 even as they tend to intensify public concern A major challenge is that energy prices do about the environment and awareness that carbon not reflect costs—economic or ecological. emissions drive climate change. Furthermore, Estimates suggest that middle-income as middle-income countries ramp up the countries account for 93 percent of sophistication of their economies by switching explicit fossil fuel subsidies.49 A promising to 2i and 3i strategies and expand their use of approach is to consider the concept of artificial intelligence and machine learning, their total carbon price (TCP) to assess the OVERVIEW: MAKING A MIR ACLE | 25 price signal from a combination of direct Figure O.13  In low- and middle-income and indirect carbon pricing instruments, countries, the cost of capital for renewables including energy excise taxes and fuel is high subsidies.50 The incumbency advantage also should 9 be disciplined in the electricity industry, 8 where incumbent SOEs dominate fossil 7 Cost of capital (%) fuel power generation and block the entry 6 of new players. 5 • Rewarding merit. The most efficient way 4 to scale up the efficient provision of low- 3 carbon energy is to respect the merit order: the sequence followed by grid operators 2 selling power to the market. The starting 1 point is set by the cheapest offer, made by 0 the power station with the lowest running e e e e m m m m co co co co costs, which determines wholesale in in -in in e- e- h- market prices. Any provider who can w dl dl ig Lo id id H -m r-m offer renewable energy at zero marginal er pe w Up cost—that is, with insignificant operating Lo costs—should have priority in meeting Source: IRENA 2023. demand. When the merit order functions Note: Data are for 2021 and 2022. as designed, it shifts prices along the supply curve, which energy economics calls the “merit order curve.”51 The road ahead • Derisking investment. The cost of capital Three decades ago, Professor Robert Lucas, Jr., likened for low-carbon energy such as solar the development strategies that led to spectacular photovoltaic and wind in middle-income economic growth in Korea to the making of a countries is twice that in high-income “miracle.” 54 Given the changes in the global economy countries, averaging 3.8 percent in high- since the time that Korea was a middle-income income countries, but 7.2 percent in economy, it would be fair to conclude that it would upper-middle-income countries and more be a miracle if today’s middle-income economies than 8.5  percent in lower-middle-income manage to do in 50 years what Korea did in just 25. countries (figure O.13).52 Addressing It might even be miraculous if they replicated the technology risk, development risk, and impressive achievements of other successful countries pricing risk can help incentivize investors— such as Chile and Poland. But that is exactly what utilities, banks, or other institutions—to governments in Bangladesh, Brazil, China, India, invest in low-carbon energy. Derisking Indonesia, Mexico, Morocco, South Africa, Türkiye, requires a whole-of-economy approach. It Viet Nam, among others, hope to accomplish. depends on licensing, policy stability, and To do this, these countries will have to become social acceptance, along with reducing more disciplined. They will have to time the shift technical, market, and regulatory risks.53 from simpler investment-led growth strategies (1i) Derisking will make renewable energy that worked well in the early stages of development projects less expensive, as well as reduce to augmenting investment accelerations with the public finance needed to support these intentional policies that aid the infusion of know- projects. how from abroad (2i), and only then expend sizable 26 | WORLD DEVELOPMENT REPORT 2024 resources on innovation (3i). Put another way, they more sophisticated mix of policies (table O.3). will have to become more efficient in their use of Low-income countries can focus solely on policies capital—both financial and human—and labor designed to increase investment—the 1i approach. and energy. Once they attain lower-middle-income status, they To do this, they will have to shed long-held will need to shift gears and expand the policy mix prejudices about enterprise, talent, and energy. They to 2i , investment + infusion. At the upper-middle- will have to appreciate the importance of reliable income level, countries will have to shift gears information to shape and quicken the structural again to 3i : investment + infusion + innovation. transformations that must accompany any Middle-income countries will need progressively durable increase in incomes and living standards. greater economic freedom, more open and informed Depending on their special circumstance and debates, and—frequently—the political courage to the stage of development they have reached, they change stubborn institutions and long-standing will need to adopt a sequenced and progressively arrangements. Table O.3  The 3i strategy: What countries should do at different stages of development UPPER-MIDDLE-INCOME LOW-INCOME COUNTRIES COUNTRIES LOWER-MIDDLE-INCOME COUNTRIES 3i: Investment + Infusion + 1i: Investment 2i: Investment + Infusion Innovation Enterprise • Improve the • Discipline market leaders through integration into • Deepen capital markets investment climate globally contestable markets. and expand equity to increase • Diffuse global technologies with fluid factor and financing. domestic and product markets. • Strengthen antitrust foreign investment. regulation and • Reward value-adding firms to stimulate business dynamism. competition agencies. • Protect intellectual property rights. Talent • Invest in human • Discipline elites by providing equal opportunities • Strengthen industry- capital by for women, minorities, and disadvantaged groups. academia links broadening • Improve allocation of talent to task. domestically. foundational skills • Expand programs to • Develop links among local and globally leading and improving connect with diaspora in universities. learning outcomes. advanced economies. • Allow emigration of educated workers whose skills are not valued in domestic markets. • Enhance economic and political freedoms. Energy • Increase investment • Discipline SOEs by hardening budget constraints. • Lower the cost of capital in expanding access • Use international coalitions to encourage for low-carbon energy by and grid networks. advanced economies to ease protection of reducing risks involving • Reform regulatory domestic incumbents. technology, markets, and frameworks to policy. • Aid adoption of energy-efficient practices. attract private • Increase multilateral • Enhance economic efficiency by reflecting investment and finance for very environmental costs in energy prices. ensure fair long-term investments. competition. Source: WDR 2024 team. Note: SOEs = state-owned enterprises. OVERVIEW: MAKING A MIR ACLE | 27 Notes  1. Throughout this Report, data on GDP and income per move away from their parents’ position relative to gen - capita are as of July 1, 2023. erational peers.  2. Gill and Kharas (2007). 35. Becker et al. (2018).  3. Kose and Ohnsorge (2024). 36. Angrist et al. (2023); World Bank (2018).  4. Melitz and Redding (2021). 37. Chaudhury and Parajuli (2010).  5. Soh, Koh, and Aridi (2023). 38. Muñoz-Boudet et al. (2021).  6. Pinto (2014). 39. Dubey and Purnanandam (2023).  7. Marcel and Vivanco (2021). 40. Akcigit, Baslandze, and Stantcheva (2016).  8. Lucas (1988); Romer (1990). 41. World Bank (2023).  9. de Souza (2022). 42. Peters et al. (2012). 10. de Souza (2022, 2023). 43. UNEP (2009). 11. Schumpeter (1942). 44. Baqaee and Farhi (2023). 12. Vagliasindi (2023). 45. https://time.com/6987773/ai-data-centers-energy​ 13. Aghion and Howitt (1992). -usage-climate-change/. 14. Hsieh et al. (2019). 46. Albornoz et al. (2009). 15. van der Weide et al. (2021). 47. Holladay (2016); Richter and Schiersch (2017). 16. Gottlieb, Poschke, and Tueting (2024). 48. Bashmakov (2007); Bashmakov et al. (2023). 17. Akcigit, Grigsby, and Nicholas (2017). 49. Black et al. (2023). 18. Akcigit and Kerr (2018). 50. Agnolucci, Gencer, and Heine (2024). TCP components 19. Cusolito, Garcia-Marin, and Maloney (2023). labeled “energy taxes” and “energy subsidies” are 20. Bustos (2011). based on “net” computed values (as proxies for the 21. Bloom, Draca, and Van Reenen (2016). actual values of energy taxes and subsidies) due to 22. Arráiz, Henríquez, and Stucchi (2011). data limitations. Energy taxes and subsidies are esti- 23. Cirera and Maloney (2017). mated based on the “price gap” between retail prices 24. Eslava and Haltiwanger (2020); Hsieh and Olken and supply costs for a particular energy carrier used in (2014). a specific sector in a jurisdiction in a given year. The net 25. Akcigit, Alp, and Peters (2021). energy taxes and subsidies are then aggregated across 26. Bertoni, Colombo, and Quas (2023); Kersten et al. (2017). sectors, fuels, and countries to yield a global value. 27. Aivazian and Santor (2008); López and Torres (2020); More details on this methodology are provided in Tsuruta (2020); Vu and Tran (2021). Agnolucci, Gencer, and Heine (2024). 28. Bachas, Fattal Jaef, and Jensen (2019). 51. Acemoglu, Kakhbod, and Ozdaglar (2017). 29. López and Torres (2020). 52. Estimates of the cost of capital are based on the cost 30. Melitz (2003). of debt and the cost of equity. The cost of debt is the 31. Didier and Cusolito (2024). cost to finance a loan for a renewable energy asset. 32. Cheng (2021); Gal et al. (2019). The cost of equity is the return on equity required by 33. Cunningham, Ederer, and Ma (2021). the project developer (IRENA 2023). 34. Social mobility is intergenerational movement up or 53. Noothout et al. (2016). down a country’s income ladder, allowing children to 54. Lucas (1988). References Acemoglu, Daron, Ali Kakhbod, and Asuman Ozdaglar. Akcigit, Ufuk, Harun Alp, and Michael Peters. 2021. “Lack 2017. “Competition in Electricity Markets with of Selection and Limits to Delegation: Firm Dynamics Renewable Energy Sources.” Energy Journal 38 (1): in Developing Countries.” American Economic Review 137–55. 111 (1): 231–75. Aghion, Philippe, and Peter Howitt. 1992. “A Model of Akcigit, Ufuk, Salomé Baslandze, and Stefanie Stantcheva. Growth through Creative Destruction.” Econometrica 2016. “Taxation and the International Mobility of 60 (2): 323–51. Inventors.” American Economic Review 106 (10): 2930–81. Agnolucci, Paolo, Defne Gencer, and Dirk Heine. 2024. Akcigit, Ufuk, John Grigsby, and Tom Nicholas. 2017. “The “Total Carbon Pricing for Energy Consumption: The Rise of American Ingenuity: Innovation and Inventors of Importance of Energy Taxes and Subsidies.” ESMAP the Golden Age.” NBER Working Paper 23047 (January), Technical Report, Energy Subsidy Reform in Action National Bureau of Economic Research, Cambridge, MA. Series, World Bank, Washington, DC. Akcigit, Ufuk, and William R. Kerr. 2018. “Growth through Aivazian, Varouj Aram, and Eric Santor. 2008. “Financial Heterogeneous Innovations.” Journal of Political Constraints and Investment: Assessing the Impact of Economy 126 (4): 1374–443. a World Bank Credit Program on Small and Medium Albornoz, Facundo, Matthew A. Cole, Robert J. R. Elliott, and Enterprises in Sri Lanka.” Canadian Journal of Marco G. Ercolani. 2009. “In Search of Environmental Economics 41 (2): 475–500. Spillovers.” World Economy 32 (1): 136–63. 28 | WORLD DEVELOPMENT REPORT 2024 Angrist, Noam, Elisabetta Aurino, Harry A. Patrinos, George Chepeliev, Maksym, and Erwin Corong. 2022. “Revisiting Psacharopoulos, Emiliana Vegas, Ralph Nordjo, and the Environmental Bias of Trade Policies Based on an Brad Wong. 2023. “Improving Learning in Low- and Environmentally Extended GTAP MRIO Data Base.” Lower-Middle-Income Countries.” Journal of Benefit- Paper presented at the virtual 25th Annual Conference, Cost Analysis 14 (S1): 55–80. Global Economic Analysis, Center for Global Trade Arráiz, Irani, Francisca Henríquez, and Rodolfo Stucchi. Analysis, Department of Agricultural Economics, 2011. “Impact of the Chilean Supplier Development Purdue University, West Lafayette, IN, June 8–10, 2022. Program on the Performance of SMEs and Their Large https://www.gtap.agecon.purdue.edu/resources/res​ Firm Customers.” OVE Working Paper OVE/WP–04/11 _display.asp?RecordID=6548. (May), Office of Evaluation and Oversight, Inter- Choi, Jaedo, and Younghun Shim. 2024. “From Adoption American Development Bank, Washington, DC. to Innovation: State-Dependent Technology Policy in Bachas, Pierre, Roberto N. Fattal Jaef, and Anders Developing Countries.” STEG Working Paper WP091 Jensen. 2019. “Size-Dependent Tax Enforcement (March), Structural Transformation and Economic and Compliance: Global Evidence and Aggregate Growth, Centre for Economic Policy Research, Implications.” Journal of Development Economics 140 London. (September): 203–22. Cirera, Xavier, and William F. Maloney. 2017. The Innovation Bahar, Dany, Sebastian Bustos, and Muhammed A. Yıldırım. Paradox: Developing-Country Capabilities and the 2024. “Stages of Diversification Revisited.” Background Unrealized Promise of Technological Catch-Up. paper prepared for World Development Report 2024, Washington, DC: World Bank. World Bank, Washington, DC. Cunningham, Colleen, Florian Ederer, and Song Ma. 2021. Baqaee, David Rezza, and Emmanuel Farhi. 2023. “Killer Acquisitions.” Journal of Political Economy “Networks, Barriers, and Trade.” NBER Working Paper 129 (3): 649–702. 26108 rev. (February), National Bureau of Economic Cusolito, Ana Paula, Alvaro Garcia-Marin, and William F. Research, Cambridge, MA. Maloney. 2023. “Proximity to the Frontier, Markups, and Bashmakov, Igor. 2007. “Three Laws of Energy Transitions.” the Response of Innovation to Foreign Competition: Energy Policy 35 (7): 3583–94. Evidence from Matched Production-Innovation Surveys Bashmakov, Igor, Michael Grubb, Paul Drummond, in Chile.” American Economic Review: Insights 5 (1): Robert Lowe, Anna Myshak, and Ben Hinder. 2023. 35–54. “‘Minus  1’ and Energy Costs Constants: Empirical de Souza, Gustavo. 2022. “The Labor Market Consequences Evidence, Theory, and Policy Implications.” SSRN of Appropriate Technology.” Working Paper WP 2022- Preprint (March 30), Social Science Research Network, 53 (September 6), Federal Reserve Bank of Chicago, Rochester, NY. https://papers.ssrn.com/sol3/papers​ Chicago. .cfm?abstract_id=4401851. de Souza, Gustavo. 2023. “R&D Subsidy and Import Becker, Gary Stanley, Scott Duke Kominers, Kevin M. Substitution: Growing in the Shadow of Protection.” Murphy, and Jörg L. Spenkuch. 2018. “A Theory of Working Paper WP 2023-37 rev. (October 5), Federal Intergenerational Mobility.” Journal of Political Economy Reserve Bank of Chicago, Chicago. 126 (S1): S7–S25. Didier, Tatiana, and Beulah Chelva. 2023. “Private Equity Bertoni, Fabio, Massimo G. Colombo, and Anita Quas. 2023. Markets in EMDEs.” World Bank, Washington, DC. “The Long-Term Effects of Loan Guarantees on SME Didier, Tatiana, and Ana Paula Cusolito. 2024. Unleashing Performance.” Journal of Corporate Finance 80 (June): Productivity through Firm Financing. Washington, DC: 102408. World Bank. Black, Simon, Antung A. Liu, Ian W. H. Parry, and Nate Dubey, Tamanna Singh, and Amiyatosh Purnanandam. Vernon. 2023. “IMF Fossil Fuel Subsidies Data: 2023 2023. “Can Cashless Payments Spur Economic Update.” IMF Working Paper WP/23/169 (August), Growth?” Presentation at the National Bureau of International Monetary Fund, Washington, DC. Economic Research’s Summer Institute 2023, “Macro, Bloom, Nicholas, Mirko Draca, and John Van Reenen. Money, and Financial Frictions.” Cambridge, MA, July 2016. “Trade Induced Technical Change? The Impact 12–13, 2023. of Chinese Imports on Innovation, IT and Productivity.” Energy Institute. 2023. “Statistical Review of World Energy Review of Economic Studies 83 (January): 87–117. 2023.” June, Energy Institute, London. Bustos, Paula. 2011. “Trade Liberalization, Exports, and Eslava, Marcela, and John C. Haltiwanger. 2020. “The Technology Upgrading: Evidence on the Impact of Life-Cycle Growth of Plants: The Role of Productivity, MERCOSUR on Argentinian Firms.” American Economic Demand and Wedges.” NBER Working Paper 27184 Review 101 (1): 304–40. (May), National Bureau of Economic Research, Chaudhury, Nazmul, and Dilip Parajuli. 2010. “Conditional Cambridge, MA. Cash Transfers and Female Schooling: The Impact Gal, Peter, Giuseppe Nicoletti, Theodore Renault, Stéphane of the Female School Stipend Programme on Public Sorbe, and Christina Timilioti. 2019. “Digitalisation School Enrolments in Punjab, Pakistan.” Applied and Productivity: In Search of the Holy Grail; Firm- Economics 42 (28): 3565–83. Level Empirical Evidence from EU Countries.” Cheng, Thomas K. 2021. The Patent-Competition Interface OECD Economics Department Working Paper 1533 in Developing Countries. Oxford, UK: Oxford University (February  6), Organisation for Economic Co-operation Press. and Development, Paris. OVERVIEW: MAKING A MIR ACLE | 29 Gill, Indermit Singh, and Homi Kharas. 2007. An East Noothout, Paul, David de Jager, Lucie Tesnière, Sascha van Asian Renaissance: Ideas for Economic Growth. With Rooijen, Nikolaos Karypidis, Robert Brückmann, Filip Deepak Bhattasali, Milan Brahmbhatt, Gaurav Datt, Jirouš, et al. 2016. The Impact of Risks in Renewable Mona Haddad, Edward Mountfield, Radu Tatucu, and Energy Investments and the Role of Smart Policies: Final Ekaterina Vostroknutova. Washington, DC: World Bank. Report. Karlsruhe, Germany: Fraunhofer Institute for Gottlieb, Charles, Markus Poschke, and Michael Tueting. Systems and Innovation Research. 2024. “Skill Supply, Firm Size, and Economic Peters, Glen P., Gregg Marland, Corinne Le Quéré, Thomas Development.” Background paper prepared for World Boden, Josep G. Canadell, and Michael R. Raupach. Development Report 2024, World Bank, Washington, DC. 2012. “Rapid Growth in CO2 Emissions after the 2008– Holladay, J. Scott. 2016. “Exporters and the Environment.” 2009 Global Financial Crisis.” Nature Climate Change Canadian Journal of Economics 49 (1): 147–72. 2: 2–4. Hopenhayn, Hugo A. 1992. “Entry, Exit, and Firm Dynamics Pinto, Brian. 2014. “Why Poland Beat the Odds.” Chapter 4 in in Long-Run Equilibrium.” Econometrica 60 (5): 1127–50. How Does My Country Grow? Economic Advice through Hsieh, Chang-Tai, Erik Hurst, Charles I. Jones, and Peter Story-Telling, 55–76. New York: Oxford University Press. J. Klenow. 2019. “The Allocation of Talent and U.S. Richter, Philipp M., and Alexander Schiersch. 2017. “CO2 Economic Growth.” Econometrica 87 (5): 1439–74. Emission Intensity and Exporting: Evidence from Firm- Hsieh, Chang-Tai, and Peter J. Klenow. 2014. “The Life Level Data.” European Economic Review 98 (September): Cycle of Plants in India and Mexico.” Quarterly Journal 373–91. of Economics 129 (3): 1035–84. Romer, Paul Michael. 1990. “Endogenous Technological Hsieh, Chang-Tai, and Benjamin A. Olken. 2014. “The Change.” Journal of Political Economy 98 (5, Part 2): Missing ‘Missing Middle’.” Journal of Economic S71–S102. Perspectives 28 (3): 89–108. Schumpeter, Joseph Alois. 1942. Capitalism, Socialism and IEA (International Energy Agency). 2023. World Energy Democracy. New York: Harper and Brothers. Outlook 2023. Paris: IEA. Serajuddin, Umar, and Nada Hamadeh. 2020. “New World IRENA (International Renewable Energy Agency). 2023. Bank Country Classifications by Income Level: 2020– “The Cost of Financing for Renewable Power.” IRENA, 2021.” Data Blog (blog), July 1, 2020. https://blogs​ Abu Dhabi, United Arab Emirates. .worldbank.org/opendata/new-world-bank- ­c ountry​ Jones, Charles I. 2016. “The Facts of Economic Growth.” In -classifications-income-level-2020-2021. The Handbook of Macroeconomics, vol. 2A, edited by John Soh, Hoon Sahib, Youngsun Koh, and Anwar Aridi, eds. 2023. B. Taylor and Harald Uhlig, 3–69. Amsterdam: Elsevier. Innovative Korea: Leveraging Innovation and Technology Kersten, Renate, Job Harms, Kellie Liket, and Karen Maas. for Development. Washington, DC: World Bank. 2017. “Small Firms, Large Impact? A Systematic Review Tsuruta, Daisuke. 2020. “SME Policies as a Barrier to Growth of the SME Finance Literature.” World Development 97 of SMEs.” Small Business Economics 54 (4): 1067–106. (September): 330–48. UNEP (United Nations Environment Programme). 2009. Kose, M. Ayhan, and Franziska Ohnsorge, eds. 2024. Falling “The Global Financial Crisis and Its Impact on Long-Term Growth Prospects: Trends, Expectations, and Renewable Energy Finance.” Division of Technology, Policies. Washington, DC: World Bank. Industry, and Economics, UNEP, Nairobi, Kenya. López, José Joaquín, and Jesica Torres. 2020. “Size- Vagliasindi, Maria. 2023. “The Role of SOEs in Climate Dependent Policies, Talent Misallocation, and the Change.” Background paper prepared for The Business Return to Skill.” Review of Economic Dynamics 38 of the State, World Bank, Washington, DC. (October): 59–93. van der Weide, Roy, Christoph Lakner, Daniel Gerszon Lucas, Robert B., Jr. 1988. “On the Mechanics of Economic Mahler, Ambar Narayan, and Rakesh Ramasubbaiah. Development.” Journal of Monetary Economics 22 (1): 2021. “Intergenerational Mobility around the World.” 3–42. Policy Research Working Paper 9707, World Bank, Marcel, Mario, and Diego Vivanco. 2021. “Measuring Small Washington, DC. and Medium-Size Enterprises’ Contribution to Trade in Vu, Quang, and Tuyen Quang Tran. 2021. “Government Value Added: The Case of Chile 2013–2016.” Working Financial Support and Firm Productivity in Vietnam.” Paper of the Central Bank of Chile 914 (April), Central Finance Research Letters 40 (May): 101667. Bank of Chile, Santiago. World Bank. 1978. World Development Report 1978. Melitz, Marc J. 2003. “The Impact of Trade on Intra-Industry Washington, DC: World Bank. Reallocations and Aggregate Industry Productivity.” World Bank. 2015. Peru: Building on Success: Boosting Econometrica 71 (6): 1695–725. Productivity for Faster Growth. Washington, DC: Melitz, Marc J., and Stephen J. Redding. 2021. “Trade World Bank. and Innovation.” NBER Working Paper 28945 (June), World Bank. 2018. World Development Report 2018: National Bureau of Economic Research, Cambridge, MA. Learning to Realize Education’s Promise. Washington, Muñoz-Boudet, Ana María, Lourdes Rodríguez-Chamussy, DC: World Bank. Cristina Chiarella, and Isil Oral Savonitto. 2021. World Bank. 2023. World Development Report 2023: “Women and STEM in Europe and Central Asia.” Report Migrants, Refugees, and Societies. Washington, DC: AUS0002179, World Bank, Washington, DC. World Bank. 30 | WORLD DEVELOPMENT REPORT 2024 Part 1 Middle-Income Transitions T he share of middle-income countries in Notably, the progress of the middle-income the global economy is increasing—over countries has slowed in recent decades. The half of countries today are middle-income. median middle-income economy has income per As classified by the World Bank, 108 coun- capita that is less than one-tenth that of the United tries qualify as middle-income. With about States. More surprising, this figure has remained three-fourths of the world’s population, these almost unchanged for 50 years. Meanwhile, the 108 middle-income countries account for nearly prospects for middle-income countries are not 40  percent of global economic activity. Of every improving in view of the direction the global five people in extreme poverty globally, more than economy is going, from healthy to wobbling. three live in middle-income countries. And they Against these headwinds, today’s middle-income generate well over 60 percent of all carbon dioxide countries need to make miracles to develop at emissions (table P1.1). Not surprisingly, they will the pace of the 34 economies that reached high-​ play a central role in global development, and the income status between 1990 and 2021. And even difficulties they face should be of global concern. if these headwinds were not getting stronger, So where are these economies headed? middle-​income countries would still face long Table P1.1 World Bank country classifications and selected global indicators, 2022 SHARE OF PEOPLE IN SHARE OF GLOBAL INCOME SHARE OF GLOBAL SHARE OF EXTREME POVERTY CARBON DIOXIDE CLASSIFICATION POPULATION (%) GLOBAL GDP (%) GLOBALLY (%) (CO2) EMISSIONS (%) Low-income 8.9 0.6 36.5 0.5 Lower-middle-income 40.3 8.3 55.4 15.7 Upper-middle-income 35.1 30.3 7.1 48.6 High-income 15.7 60.8 1.0 35.2 Sources: Population shares and global GDP shares computed from WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712; extreme poverty shares from PIP (Poverty and Inequality Platform) (dashboard), World Bank, Washington, DC, https://pip.worldbank.org/home; carbon dioxide emissions data (2022) from Climate Watch (dashboard), World Resources Institute, Washington, DC, https://www​ .climatewatchdata.org/. Note: The World Bank currently recognizes 26 economies as low-income (GNI per capita, calculated using the World Bank Atlas method, of US$1,135 or less in 2022); 54 as lower-middle-income (GNI per capita of between US$1,136 and US$4,465); 54 as upper-middle-income (GNI per capita of between US$4,466 and US$13,845); and 83 as high-income (GNI per capita of US$13,846 or more). GDP = gross domestic product; GNI = gross national income.  | 31 odds of achieving high-income ­ status because Pathways to high-income status can differ among of growth trajectories suggestive of a “middle-​ countries sectorally and spatially. But they gen- income trap.” erally involve transitioning first from relying Part 1 of this Report examines the evidence for largely on investment in physical and human the middle-income trap and asks three questions. capital—the mainstay of successful growth at First, is growth in middle-income countries slower, low-income levels of development—to combin- with investment-led growth running out of steam ing investment with the infusion of global tech- (chapter 1, Slowing Growth)? Second, is growth nologies and know-how, which applies largely in middle-income countries different, requiring a to lower-middle-income countries. This tran- ­ significant change in growth s chapter 2, ­ trategies (­ sition is necessary, but it is not enough to move Structural Stasis)? And, third, is growth in to the high-income level. The second transition middle-​ income countries now harder (­ chapter  3, involves adding an emphasis on innovation to the Shrinking Spaces)? mix, which is more applicable to upper-middle-​ Chapter 1 summarizes the evidence on growth income countries. This mapping of transitions slowdowns and highlights that in middle-​ income to income levels should be considered indicative. countries a majority of growth slowdowns take Strictly speaking, it is the structure of an econ- place as the returns from capital investment omy that determines the timing of the shift, not diminish sharply. The median growth slowdown its gross national income per capita. episode occurs when a country reaches a little Chapter 3 examines the forces that today are more than 11 percent of the gross domestic prod- making growth harder to achieve. Foreign trade uct (GDP) per capita level of the United States. and investment channels are in danger of becom- Policy and institutional deficiencies exacerbate ing constricted by geopolitical tensions. The growth slowdowns; countries with weaker polit- room for governments to act has shrunk because ical institutions experience a growth slowdown of multiple crises and populist pressures. And much earlier, and at lower incomes, than coun- in many middle-income countries, government tries with stronger institutions. debt—which is more expensive for this income Chapter 2 identifies the two successive group than for any other—is at an all-time high. transitions that middle-income countries Further complicating matters, fragility, con- must undergo  to achieve high-income status. flict, and violence hamper development in some Specifically, countries need to recalibrate their middle-​ income countries. And in every country, mix of investment, infusion, and innovation as climate change is putting pressure on the govern- they move through the middle-income status. ment to rethink its growth strategies. 32 | WORLD DEVELOPMENT REPORT 2024 1 Slowing Growth Key messages • Today’s 108 middle-income countries represent about 40 percent of the global economy, are home to about 75 percent of the world’s population and more than 60 percent of the world’s poor, and contribute nearly two-thirds of global carbon dioxide emissions. • Middle-income countries are prone to systematic growth slowdowns—a concept termed the “middle-income trap.” The median growth slowdown episode occurs when a country reaches about 11 percent of the gross domestic product per capita of the United States. • Although income per capita is the metric most commonly used to measure the pace of economic development, measures of average income can differ greatly, depending on the measure. • Countries with weaker institutions—and especially those with lower levels of economic and political freedom—are more susceptible to slowdowns at even lower levels of income. Introduction • They had more resources available to raise the living standards of the poor. The problem of economic growth in middle-­income economies has been a concern of development pol- The 1978 Report emphasized the central role of icy practitioners for at least five decades. In the cultivating engineering talent to design products first World Development Report, published in 1978, that change continually and rapidly, alongside “middle-income” was an omnibus term applied to better organizing workshops and other produc- countries with diverse economic characteristics tion facilities so they are made efficiently.3 at various stages of development.1 Middle-income Interest in the economic growth of middle-­ countries were defined as those with annual income countries rose over the last two decades, income per capita of over US$250.2 By that defini- especially after a 2007 World Bank regional report tion, 58 countries, home to about 900 million peo- on East Asia introduced the term “­ middle-income ple, were designated middle-income. Despite  the trap.”  The term encapsulated the concern that 4 ­diversity, World Development Report 1978 identified middle-income countries are prone to systematic two characteristics that distinguished middle-­ slowdowns in growth demonstrated, for example, income from low-­ income countries: by the economic stagnation in Latin America and the Middle East since the mid-1970s. This chapter • Their growth prospects were more sen- assesses whether the experience of the developing sitive to economic conditions in the world is consistent with this concern. It finds that industrialized (high-income) countries, the majority of growth slowdowns do take place particularly the environment for trade and in middle-income countries. The median growth commercial capital flows. slowdown episode occurs when a country reaches 33 about 11 percent of the gross domestic product model features vigorous trade and capital flows, (GDP) per capita of the United States. This chapter freer enterprise, free worker mobility, stronger also documents that countries with weaker insti- institutions, and social inclusion—at a time of tutions—and especially those with lower levels of ­ relatively rapid economic growth in Western economic and political f ­reedom—are susceptible Europe. They benefited greatly from institutional to slowdowns at even lower levels of income. and regulatory reforms that enabled transitions Chapter 2 explores a related but relatively to a market economy, incentivized emerging qualitative question: Is economic growth during economies to attract foreign direct investment the middle-income stage systematically different and infuse new technologies into their produc- from growth in low- and high-income countries? tion structures while pushing advanced econo- Chapter 3 examines the growing concern that rap- mies to innovate, and fostered an environment idly changing economic conditions and policies for developing a skilled workforce. in the advanced economies of the North Atlantic Among the other newcomers to high-income will make development in middle-­ income coun- status, resource-rich economies such as Chile and tries even more difficult. Saudi Arabia benefited when they timed policy reforms to coincide with high commodity prices. East Asian economies such as the Republic of Growth in middle-income Korea and Taiwan, China,6 stand out for following countries a path of high savings and investment, enlight- The share of middle-income countries in the ened education policies, expansion of trade with global economy has increased rapidly since export-oriented policies and technology adoption the 1990s, suggesting that it is easier to enter the from more advanced economies, and a transition ­ middle-income stage than to exit it. According to local innovation well after closing the gaps to the World Bank’s 2023 income classifications, with the global technology frontier. the 108 current middle-income countries are For countries that are not fortunate enough split evenly between lower- and upper-middle-­ to  be in the European Union, are not endowed income countries. Representing about 40 percent with a ­bundant resources, or are not fiercely of the global economy, middle-income countries focused, progress through the m ­ iddle-income are home to about 75 percent of the world’s pop- stage has been slower. The average middle-​ ulation and more than 60 percent of the world’s income  economy still has an income per capita poor. In other words, more than 400 million of less than one-tenth that of the United States the extreme poor globally live in middle-income (figure 1.2). countries, a statistic that should concern wealth- It is understandable why middle-income ier countries. They also contribute nearly two- countries are not satisfied with the status quo thirds of global carbon dioxide (CO2) emissions and why most have plans for faster growth in (see table P1.1), a statistic that is of global concern. living standards. China’s 14th Five-Year Plan Over the last three decades, the world’s two outlines a vision of achieving the median GDP most populous countries, China and India, joined per capita of developed nations by 2035, with the club of middle-income countries, in 1997 and a large increase in the middle class. The vision 2007, respectively. It is not surprising, then, that document also highlights that China’s growth growth in middle-income countries will play a will be driven by major breakthroughs in key pivotal role in international development. technologies, making it one of the most inno- Since 1990, 34 middle-income economies have vative nations in the world, buttressed by a transitioned to high-income status (figure 1.1).5 modern economic system with digitalization, Thirteen benefited from deep integration with thriving cities, and modern agriculture. In the European Union (EU)—whose economic India, the prime minister’s vision is to transform 34 | WORLD DEVELOPMENT REPORT 2024 Figure 1.1 A handful of economies have transitioned from middle-income to high-income status over the last three decades Lower- Low- middle- Upper-middle- income income income High-income 45,000 MAC 40,000 KOR 35,000 MLT SVN GNI per capita (US$), 2022 High-income 30,000 CZE BHR EST SAU PRI PRT 25,000 LTU SVK LVA GRC KNA OMN 20,000 HRV HUN BRB POL URY ATG GUY ROM PAN TTO SYC 15,000 CHL 13,845 middle- middle- income income Lower- Upper- 10,000 5,000 4,465 1,135 0 5 0 8, 0 0 0 1, 0 0 0 0 00 0 0 0 00 00 00 00 00 00 00 00 00 61 46 62 ,0 9, 2, 3, 4, 5, 6, 7, 2, 7, 10 GNI per capita (US$), 1990 Source: WDR 2024 team using WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https:// datacatalog.worldbank.org/search/dataset/0037712. Note: Each scatter point indicates an economy’s 1990 and 2022 gross national income (GNI) per capita in current US dollar terms (using the World Bank Atlas method). The blue vertical lines show thresholds to transition to lower-middle-income status, upper-middle-income status, and high-income status in 1990 (US$610, US$2,465, and US$7,620, respectively), while the blue horizontal lines show these thresholds in 2022 (US$1,135, US$4,465, and US$13,845, respectively) based on the World Bank income classifications (https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country​ -and-lending-groups). The figure includes only economies at middle-income levels in 1990. For legibility, only economies that have transitioned to high-income status since 1990 are labeled (1990 data were unavailable for six economies). For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. the nation into a developed economy by 2047— Other middle-income countries have similar the hundredth year of independence. In Viet aspirations. Nam, its Socio-Economic Development Strategy But the growth prospects of middle-income 2021–2030 outlines a plan to sustain GDP per countries are not improving. Over the last decade, capita growth of 7 percent through this decade, the global economy has gone from healthy to with a transition to high-income status by 2045. hobbling and from largely integrated to increas- ­ In South Africa, the 2030 National Development ingly fragmented.7 Foreign trade and investment Plan has prioritized raising its income per cap- channels are also becoming more constricted—or ita from US$2,800 in 2010 to US$7,000 by 2030. at least encumbered—by geopolitical tensions. Slowing Grow th | 35 Figure 1.2 Income per capita of middle-income countries relative to that of the United States has been stagnant for decades 60 55 Share of US GNI per capita (%) 50 45 40 35 30 25 20 15 10 5 0 20 22 14 18 08 10 12 16 98 04 00 02 06 88 92 94 96 70 72 74 76 78 80 82 84 86 90 20 20 20 20 20 20 20 20 20 19 19 19 20 20 20 19 19 19 19 19 19 19 19 19 19 19 19 Middle-income countries Middle-income countries, excluding China Source: WDR 2024 team using data from WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712. Note: The plotted lines indicate the trend of average income per capita in middle-income countries and in middle-income countries, excluding China, relative to income per capita in the United States (considered as the economic frontier country). Country definitions are based on the first World Development Report (World Bank 1978), in which low-income countries have gross national income (GNI) per capita of US$250 or less, and middle-income countries have GNI per capita of more than US$250; industrialized (high-income) countries consist of member countries of the Organisation for Economic Co-operation and Development (OECD), except for Greece, Portugal, Spain, and Türkiye, which are classified as middle-income countries. Meanwhile, the room for governments to act is of sovereign debt distress heightens policy uncer- shrinking due to rapidly changing demographic tainty and dampens economic activity. trends (more rapid than countries had planned for), These difficulties are compounded by others. multiple crises, and populist pressures. In many In some middle-income countries, fragility, con- ­ middle-income c ­ountries, government debt— flict, and violence are hampering development. which is more ­ expensive for this income group And in almost every country, climate change is than for any other—is at an all-time high. And putting pressure on the government to rethink its the belated efforts of advanced economies’ central development strategy. banks to normalize monetary policy and control Given these headwinds, an economy at the inflation by raising interest rates has increased sov- middle-income stage will have to “make a mir- ereign spreads (the difference between bond yields acle” to develop at the pace of the 34 economies issued on international markets by the country that reached high-income status between 1990 in question versus those offered by governments and 2021.8 That would require having a business with AAA ratings) and raised borrowing costs for sector that facilitates a radical transformation of emerging markets, in some cases to prohibitive lev- enterprises, having a government that assuages els. As a consequence, middle-­ income economies the growing expectations of an increasingly rest- are being squeezed from several sides: tighter fiscal less middle class, and having a country transition space reduces public investment and the cushion sooner to less emissions-intensive ways of pro- for structural reforms; higher public debt service ducing and consuming than those engineered by crowds out private borrowing; and a higher risk the middle-income economies of the 1990s. 36 | WORLD DEVELOPMENT REPORT 2024 Even without these headwinds, today’s m ­ iddle- These tasks have proved to be surprisingly dif- income countries would still face long odds of ficult, and they are likely to become even harder. achieving high-income status because of what the World Bank has called—since 2007—a “­ middle-​ income trap.”  9 Although the term connotes inevi- Measuring progress through tability, the original proposition was that getting the middle stages of mired in the middle stages of development is a development possibility, not an inevitability. It would be inevi- What indicates that an economy is developing? table only if countries did not adapt their policies Income per capita is the most commonly used and institutions to changing economic and struc- metric to measure the pace of economic develop- tural needs. The three priorities for middle-income ment. But measures of average income can differ countries to evade the trap and maintain a growth greatly, depending on the measure. For exam- momentum10 could be summarized as: ple, the World Bank, other international organi- • Increasing the sophistication of processes zations, and bilateral aid agencies use GDP per and products through integration into capita at market exchange rates for analysis and world markets, generally accompanied by lending (box 1.1). On the other hand, the use of the growing specialization of production GDP per capita based on adjustments for purchas- • Keeping up with changing education ing power parity (PPP, which reflects the purchas- system priorities to help workers acquire ing power of a consumer for goods and services) skills that enable them to adjust to new can yield different results.11 technologies and shape new products and A comparison of these two sets of mea- processes sures for Türkiye and Chile illustrates the • Quickening the pace of innovation by problem. According to World Bank estimates both fostering entrepreneurial activity that use market exchange rates, Türkiye and keeping markets open to competition. is a middle-­ i ncome country, and Chile is a Box 1.1 Misunderstanding through misclassification The World Bank’s income classification method for grouping countries was first presented in the 1978 World Development Report.a It introduced groupings of “low-​ income” and “middle-income” countries using a threshold of US$250 gross national income (GNI) per capita between the groups. The low-income threshold was set in keeping with the guidelines for procurement of goods and services for civil works projects for countries eligible for assistance from the International Development Association (IDA), the organization in the World Bank Group that supports the world’s least developed countries. Specifically, the threshold was based on the “civil works preference” operational guideline for IDA countries. The process of setting thresholds for income per capita began with finding a “stable relationship” between a summary measure of well-being such as poverty incidence and infant mortality, on the one hand, and economic variables, including GNI per capita esti- mated using the World Bank’s Atlas method, on the other.b Based on such a relationship and the annual availability of the World Bank’s resources, the original income per capita thresholds were established.c They were last updated in 1989, using GNI per capita valued (Box continues next page) Slowing Grow th | 37 Box 1.1 Misunderstanding through misclassification (continued) annually in US dollars based on a three-year average exchange rate and were expanded to four categories: • Low-income. The low-income threshold was officially set in 1988, still based on the value of the IDA’s “civil works preference” and updated for inflation.d • Lower-middle-income. The lower-middle-income threshold is based on the oper- ational guidelines cutoff for determining access to 17-year repayment terms for loans through the World Bank Group’s International Bank for Reconstruction and Development (IBRD), although these terms are no longer available. It appears to have first been introduced in the 1983 edition of the World Development Report.e • Upper-middle-income. The upper-middle-income threshold is the range between lower-​middle-income and high-income. • High-income. The high-income threshold does not relate to a cutoff derived from the operational guidelines, but was set at GNI per capita of US$6,000 in 1987 prices in a paper presented to the World Bank’s Board of Executive Directors in January 1989, which also reconfirmed the low- and lower-­ middle-income threshold levels.f The US$6,000 level has been updated over time for what is called “international infla- tion,” defined as the average inflation rates of Japan, the United Kingdom, the United States, and the euro area. The choice of the high-income threshold was made to address anomalies in the classification of high-income and industrialized economies used in the World Bank’s World Development Indicators prior to that point. Under this current classification method, Zambia (with income per capita of US$1,170) and Bulgaria (with income per capita of US$13,250) are both middle-income economies. But few people would disagree with the observation that these countries have had vastly different development experiences and face vastly different growth challenges and tra- jectories. Yet these income classifications continue to be used widely in the development discourse g and in analyses of economic growth.h Moreover, although the World Bank may not use the income classifications for operational or lending purposes,i other international organizations and bilateral aid agencies do. Given its widespread use, many economists have called for a revision of the current income classification system.j The proposals include: • Reclassifying levels based on fiscal capacity. Ravallion (2009) argues that levels of devel- opment should be assessed based on countries’ internal capacities for redistribution (through taxes) in favor of their poorest citizens. Similarly, Ceriani and Verme (2014) propose a measure of a country’s capacity to reduce its own poverty levels and show how these tools can be used to guide budget or aid allocations. • Reflecting the multidimensional nature of development. Sumner and Vázquez (2012) use a set of indicators covering definitions of development from four conceptual frameworks (development as structural transformation; development as human development; development as democratic participation and good governance; and development as sustainability) to identify five types of developing countries. Similarly, (Box continues next page) 38 | WORLD DEVELOPMENT REPORT 2024 Box 1.1 Misunderstanding through misclassification (continued) Nielsen (2011) suggests the need to introduce a development taxonomy to classify countries based on a variety of existing development proxies (the Human Develop- ment Index, lifetime income measure, and so on) rather than income levels. Even the World Bank Group Strategy adopted in 2013 recognizes the need for an approach that pays more attention to the multiple facets and fragility across the development spectrum.k a. World Bank (1978). b. For more on the Atlas method, see World Bank Atlas Method: Detailed Methodology (Data Help Desk), Data, World Bank, Washington, DC, https://datahelpdesk.worldbank.org/knowledgebase​ articles/378 /­ 832-what-is-the-world-bank-atlas-method. c. Ravallion (2013). d. Ravallion (2012). e. World Bank (1983). f. World Bank (1989). g. Dolan (2016). h. Summers and Pritchett (2014). i. Operationally, borrower countries are distinguished by their lending category within the World Bank: IDA-only; blend (both IDA and IBRD); and IBRD-only. IDA provides countries with the most difficulty borrowing externally with grants and concessional loans. IBRD offers nonconcessional loans to coun- tries that it finds creditworthy. Because IBRD terms are “harder” (more market-based) than those of IDA, IBRD borrowers tend to be perceived as more developed than IDA borrowers. But graduating from low-income status is not the same as graduating from IDA. Eligibility for IDA benchmarks a country’s income against different thresholds, while graduation from IDA takes into account factors other than income. See Dolan (2016) for more information. j. Fantom and Serajuddin (2016). k. World Bank (2015). high-income  country. Yet  when adjusted for countries in the production of tradable as well as purchasing power, Türkiye’s GDP per capita nontradable goods and services, including infra- is higher than Chile’s. The country for which structure, health care, and education. both measures are defined as identical is the However, recent assessments of PPPs appear United States because purchasing power in any to be highly correlated with economic activity. country is measured relative to what a dollar A ­comparison of countries using PPP adjust- can buy in the United States. Türkiye’s GDP ments could produce a better understanding of per capita relative to that of the United States the distance to the technology frontier than gross is nearly 50 percent when adjusted for PPP but national income. However, income or GDP per less than 15 percent using market exchange capita does  not reflect the wide array of growth rates. For Chile, the numbers are, respectively, ­ challenges that countries face. Two high-income 40 percent and 20 percent. countries provide an example. In 2022, Qatar’s PPP adjustments have been criticized for their GDP  per capita was US$88,046, driven mostly by inability to reflect the complexity and diversity exports of hydrocarbons from its abundant reserves of economic production and capabilities in indi- of oil and natural gas. The same year, Denmark’s vidual economies.12 The adjustments do not con- GDP per capita was US$66,983, and its services sec- sider quality or productivity differences among tor employs about 80 percent of labor. Meanwhile, Slowing Grow th | 39 countries vary ­ the ­ significantly in their levels of short-lived, even in countries that historically have technical sophistication; in 2021, Denmark was enjoyed high growth rates (figure 1.3). ranked twenty-fourth in the Economic Complexity The standout economy—the growth superstar Index and Qatar eighty-second.13 even—is the Republic of Korea, and this Report prominently features its experiences. What Growth in middle-income was behind its success? As subsequent chapters explain, Korea’s remarkable transformation from countries is slower a postconflict country in the 1950s to an economy Economic growth is not a smooth process, and the- powered by the infusion of ideas from abroad to ory does not stipulate that it should be smooth.14 one that is transitioning to innovation at the Instead, economic growth tends to be highly vol- global frontiers of technology makes its eco- atile, and long-run growth averages tend to mask nomic history required reading for policy makers periods of success, struggle, and failure.15 Growth in in any middle-income country hoping to achieve low- and middle-income countries is an “episodic” high-income levels of living within their lifetimes phenomenon, with countries experiencing distinct (box 1.2; see chapter 2 for more information). patterns of economic growth.16 In fact, a key char- Growth slowdowns occur more frequently in acteristic of economic growth in middle-income middle-income countries than in low- or high-­ countries is its lack of persistence. The volatility income countries (box 1.3). Research conducted of growth rates is even higher in  those countries, for this Report uses a measure, proximity to the with sustained growth p ­ eriods that are typically economic frontier (leading economies), to clarify Figure 1.3 Sustained growth periods are short-lived, even in rapidly growing economies 8.5 8.0 Average annual growth rate (%) 7.5 7.0 6.5 6.0 5.5 5.0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 Number of years during sustained growth periods Bangladesh China India Korea, Rep. Malaysia Source: WDR 2024 team using data from Maddison Project Database 2023, Groningen Growth and Development Centre, Faculty of Economics and Business, University of Groningen, Groningen, the Netherlands, https://www.rug.nl/ggdc​ /­historicaldevelopment/maddison/releases/maddison-project-database-2023. Note: The figure illustrates the growth paths of countries whose average gross domestic product (GDP) growth rate per capita is higher than 5 percent per year for at least eight years. 40 | WORLD DEVELOPMENT REPORT 2024 Box 1.2 A growth superstar: How the Republic of Korea leveraged foreign ideas and innovation Over the last seven decades, Korea has engineered the most remarkable transformation in recorded economic history. It went from a war-torn, desperately poor country in the 1950s to one of the most prosperous, healthiest, and best-educated countries in the world today. With fewer than 50 million people, it is a global leader in innovation and technology and the tenth-largest economy. What was behind Korea’s success? First, Korea prioritized openness through export promotion and leveraged the interna- tional markets to expose domestic firms to competition. Over time, it reduced tariff barri- ers and loosened restrictions on foreign investors to open the domestic market to foreign competition. It also promoted private enterprises through policies that first favored the growth and expansion of large conglomerates (efficiency-driven) and then shifted to favor smaller firms and entrepreneurs (equity-driven). Investments in infrastructure helped its rapid economic growth, with physical capital accounting for about 60 percent of growth in the gross domestic product (GDP) between 1990 and 1997. Early investments in infor- mation and communication technology infrastructure, amounting to US$32.5 billion between 1995 and 2005 and an additional US$2.6 billion between 2005 and 2014, enabled Korea to leverage new sources of growth driven by digital and technology adoption. Second, Korea devised public policies to ensure contestability. The government rewarded firms for investments in research and development (R&D) and exports through R&D; it promoted science, technology, and innovation policies; it offered tax incentives; and it adopted export facilitation measures. Korea’s spending on R&D jumped from 0.5 percent of GDP in 1980 to 1.6 percent in 1990 when Korea was still an upper-middle- income country. Private R&D expenditures increased by an unprecedented 26 times from 1980 to 1990 and exceeded 80 percent of total R&D spending by the end of the 1990s. In parallel, Korea invested heavily in human capital and ensured that job creation was matched with the needed supply of skills at the different stages of development— vocational and technical secondary, STEM (science, technology, engineering, and mathematics) education, and R&D accelerators—which was carried out more effectively than by much richer countries. Third, Korea got better at regulating the relationships among large, medium, and small firms. Initially, large firms were favored as the instrument for infusing new technologies into the economy. But by the mid-1990s the limits of this approach had become obvious, and yet powerful incumbents stymied the efforts of policy makers to change course. The 1997 Asian financial crisis changed the balance of power, and Korea established a new state–market relationship by adopting reforms to strengthen (1) financial market institu- tions with greater oversight and rules that diminished distortions; (2) competition policies that ended tacit government support for market collusion and concentration of market power; and (3) a pro-entrepreneurship policy regime with improved financing mechanisms for domestic technology ventures. The effectiveness of all of these policies was enhanced by stronger bureaucratic capacities, anticorruption initiatives, more transparent legal frameworks, better coordination mechanisms, and monitoring and evaluation systems. Sources: Kim 2006; Soh, Koh, and Aridi 2023. Slowing Grow th | 41 Box 1.3 Identifying growth slowdowns Several studies have examined the reasons for growth slowdowns. Eichengreen, Park, and Shin (2011) identify frequent slowdowns in middle-income countries. They define a slowdown as a decline in the seven-year average growth rate of gross domestic product (GDP) per capita by at least 2 percentage points, with growth being higher than 3.5 per- cent in the preceding years. In addition, they limit slowdowns to cases in which GDP per capita is greater than US$10,000 in 2005 constant international prices adjusted for pur- chasing power parity (PPP). They discover slowdowns when GDP per capita reaches about US$16,540 (in 2005 constant international PPP prices). Extensions of the analysis indicate that growth in middle-income countries slows even at points early in the middle-income stage: specifically, in the range of US$10,000–US$11,000 GDP per capita (in 2005 con- stant prices) and in the range of US$15,000–US$16,000 (in 2005 constant prices). Eichengreen, Park, and Shin (2011) find that slowdowns are driven largely by low pro- ductivity growth. They also find that the probability of middle-income traps is higher in countries with high investment rates, high old-age dependency ratios,a and undervalued real exchange rates that translate into a barrier to move up the technology ladder. In addi- tion, they find that the level and structure of human capital, the level and structure of exports (specifically, the relative importance of low- and high-tech exports), financial and political stability, and external shocks are among the significant correlates of slowdowns. Aiyar et al. (2013) define the middle-income trap as a special case of growth slowdowns. They distinguish between natural slowdowns in growth and unusually severe slowdowns. Although economies in all income groups experience growth slowdowns, based on their analysis covering 1960–2005 middle-income countries are especially vulnerable to growth slowdowns. They point to steep drops in the growth of total factor productivity (TFP)b as a key driver of such slowdowns. Spence (2011) also finds slowdowns clustering in a narrow band of countries with income per capita of between US$5,000 and US$10,000. Im and Rosenblatt (2013) focus on the probability of a country transitioning to the next income category. They find that the transition from upper-middle- to high-income status is just as likely as the transition from lower-middle- to upper-middle-income status. They argue that income per capita relative to the frontier stagnates after reaching middle-­ income status (for both lower-middle- and upper-middle-income countries). Their analysis suggests that it will take a century or more for middle-income countries to catch up to high-income countries if middle-income economies grow by 3–4 percent in per capita terms, assuming that the growth rate of high-income countries proceeds at the world average, which is 1.8 percent. Robertson and Ye (2013) identify the middle-income trap as an ailment in which a country’s GDP per capita is time-invariantc and stays in the middle-income range, defined as between 8 percent and 36 ­ percent of GDP per capita of the United States. a. The old-age dependency ratio is the ratio of older dependents (age 65 and over) to the working-age population (ages 15–64). b. Total factor productivity, a concept created by Robert Solow, is an equation used in economics to measure the impact of technological advancements and changes in worker knowledge. It attempts to measure the effects that these changes have on the long-term output of an economic system. c. A time-invariant variable refers to a variable whose value does not change across time. 42 | WORLD DEVELOPMENT REPORT 2024 the distribution of growth slowdowns along the Figure 1.4 Growth slowdowns are most national income spectrum around the world. The frequent when countries’ GDP per capita is frontier represents the growth leader: the country less than one-fourth of the United States’ with the most advanced combination of economic 45 production, innovation, and workforce. For this Likelihood of growth slowdowns (%) analysis, the United States is used as a proxy for 40 the frontier. Technically, a growth slowdown is 35 defined as a break in the time series of the growth rate of GDP per capita, whereby the growth follow- 30 ing the break is distinctly lower than the growth 25 preceding it.17 Proximity to the frontier is the ratio of a country’s GDP per capita to that of the frontier 20 country (the United States) each year (not adjusted 15 for differences in PPP). 10 Measured by their proximity to the frontier, the types of countries that experience growth slow- 5 downs vary widely. When they enter a slowdown, 0 their proximity to the frontier can range anywhere 0–5 5–25 25–50 50–100 from just above 0 percent all the way to 150 percent. GDP per capita relative to the The median growth slowdown episode occurs in a frontier (United States) (%) country-year with just 11 percent proximity to the Source: WDR 2024 team. frontier, and the mean episode occurs at 21 per- Note: GDP = gross domestic product. cent proximity to the frontier—approximately the 75th percentile in the distribution. Together, these productivity, human capital, and demographics— median and mean measures imply that a majority are already running out of steam. of growth slowdowns take place in middle-income Although researchers continue to debate the countries (figure 1.4). In fact, a middle-income coun- existence of a middle-income trap along the lines try is three times as likely to experience a growth of that first flagged by the World Bank in the mid- slowdown compared with a high-income country. 2000s, policy makers in middle-income econo- Using the World Bank’s Long Term Growth mies generally consider it a serious possibility. Model (LTGM),18 figure 1.5 sheds more light on Their concerns are the motivation for this Report. growth slowdowns in low- and middle-income Developing countries should also seriously countries. Assuming a “business as usual” base- consider the close correlation between the qual- line, where the growth drivers (ratios of public ity of institutions and the probability of falling investment to GDP and private investment to GDP, into the trap. Economists have conjectured that total factor productivity, human capital, and labor poor institutional quality discourages invest- force participation rates) follow their historical ment and innovation, distorts allocation, and or recent trends,19 most low- and ­ middle-income lowers returns to entrepreneurship.20 And policy countries are forecasted to experience significant and institutional deficiencies can put the brakes slowdowns as they approach the economic fron- on and even derail development.21 Research con- tier country (the United States) over 2023–2100. ducted for this Report reveals that countries with In addition, middle-income countries whose weaker political institutions—measured in many growth has already significantly slowed, such as ways—experience growth slowdowns at lower Argentina, Bulgaria, and Mexico, are expected to levels of development than countries with stron- diverge from the economic frontier over the next ger ones (figure 1.6). Panel a of figure 1.6 suggests 70 years. This is an unfortunate outcome because that civil liberties may influence the overall con- the key drivers of growth—savings, investment, ditions for investment, innovation, and growth. Slowing Grow th | 43 Figure 1.5 Growth is expected to slow down as countries approach the economic frontier (United States) Annual growth in GDP per capita (%) 7 6 5 4 3 2 1 0 10 20 30 40 50 60 70 80 90 100 GDP per capita relative to the frontier (United States) (%) Argentina Bangladesh Bulgaria Chile China India Indonesia Mexico Poland Türkiye Source: WDR 2024 team. Note: The dashed lines represent countries that will experience a divergence from the frontier over 2023–2100, although they were closer to the frontier in 2022. The solid lines indicate a convergence to the frontier over 2023–2100. These projections are based on extrapolation of recent historical trends using the World Bank’s Long Term Growth Model. GDP = gross domestic product. Figure 1.6 Weak institutions hasten and worsen growth slowdowns a. Civil liberties and political rights b. Economic freedom 100 100 Cumulative probability Cumulative probability of slowdown (%) of slowdown (%) 75 75 50 50 25 25 0 50 100 150 0 50 100 150 Proximity to the frontier (%) Proximity to the frontier (%) Not free Free Less free More free Source: Chikis 2024. Note: The empirical distribution function for growth decelerations is as defined in Kar et al. (2013). Gross domestic product (GDP) is measured using market exchange rates. Proximity to the frontier indicates a country’s GDP per capita relative to US GDP per capita. In panel a, political institutions are based on scores for “civil liberties” and “political rights” from Freedom House. In panel  b, the plots of economic freedom use data from the Heritage Foundation Index of Economic Freedom. Countries with scores above the median are freer. That methodology identifies slowdowns in 69 countries between 1972 and 2010. See Countries and Territories (dashboard), Freedom House, Washington, DC, https://freedomhouse.org/countries​ /freedom-world/scores; Index of Economic Freedom, 30th Edition (dashboard), Heritage Foundation, Washington, DC, https:// www.heritage.org/index/. 44 | WORLD DEVELOPMENT REPORT 2024 Countries with weaker economic freedoms also that tighter economic restrictions may mean experience growth slowdowns while remaining forgoing opportunities to close the gaps in liv- far from the global frontier (figure 1.6, panel b). ing standards between their own economies and In other words, policy makers in middle-income the more advanced economies in North America, countries should be mindful of the possibility Northeast Asia, Western Europe, and Oceania. Notes  1. World Bank (1978).  8. Lucas (1993).  2. Four member countries of the Organisation for  9. Gill and Kharas (2007). Economic Co-operation and Development (OECD)— 10. Gill and Kharas (2007, 2015); Kharas and Gill (2020). Greece, Portugal, Spain, and Türkiye—were included 11. Hamadeh et al. (2022); World Bank (2020). among the middle-income countries. The other OECD 12. Almås (2012). member countries were placed in the industrialized 13. The Economic Complexity Index is a measure of a (high-income) group. country’s productive capabilities. It is defined as “the  3. Engineering talent and organizational structure are cen- composition of a country’s productive output, reflect- tral themes of this Report, which underpins an expansive ing the structures able to hold and combine knowl- economics literature on creative destruction (Aghion and edge” (Hausmann et al. 2013, 18; see also Balland et al. Howitt 1992; Aghion et al. 2019; Grossman and Helpman 2022; Hidalgo and Hausmann 2009). 1991; Segerstrom, Anant, and Dinopoulos 1990). The cen- 14. See, for example, Solow (1956); Mankiw, Romer, and tral feature of the Industrial Revolution and its aftermath Weil (1992). was the gradual shift from tacit knowledge (as embodied 15. Easterly et al. (1993); Jones and Olken (2008). in craftsmanship and simple production techniques) to 16. Pritchett et al. (2016). more formal knowledge created by mathematicians, 17. Kar et al. (2013). physicists, chemists, medical doctors, and people 18. The Long Term Growth Model (LTGM) is a spread- schooled in engineering science (Mokyr 2023). sheet-based tool used to analyze future long-term  4. Gill and Kharas (2007). growth scenarios in developing countries building on  5. The 34 economies that transitioned to high-income the celebrated Solow-Swan growth model. The LTGM status since 1990 are American Samoa (United States); aggregates assumptions about growth fundamen- Antigua and Barbuda; Bahrain; Barbados; Chile; Croatia; tals—such as investment, education, and productivity— Czechia; Estonia; Gibraltar; Greece; Guam; Guyana; to produce a trajectory for future growth. The drivers of Hungary; Isle of Man; the Republic of Korea; Latvia; growth are savings, investment, and productivity, but Lithuania; Macao SAR, China; Malta; New Caledonia; the model also analyzes human capital, demographics, Northern Mariana Islands; Oman; Panama; Poland; the external sector (external debt, foreign direct invest- Portugal; Puerto Rico (United States); Romania; Saudi ment, current account balance), and labor force partic- Arabia; the Seychelles; the Slovak Republic; Slovenia; ipation by gender. St. Kitts and Nevis; Trinidad and Tobago; and Uruguay. 19. The recent trend spans 2000–2019.  6. Taiwan, China, transitioned to high-income status 20. Aiyar et al. (2013). before 1990. 21. Kharas and Gill (2020).  7. Kose and Ohnsorge (2024). References Aghion, Philippe, Ufuk Akcigit, Antonin Bergeaud, Richard David Rigby. 2022. “The New Paradigm of Economic Blundell, and David Hémous. 2019. “Innovation and Top Complexity.” Research Policy 51 (3): 104450. Income Inequality.” Review of Economic Studies 86 (1): Ceriani, Lidia, and Paolo Verme. 2014. “The Income Lever 1–45. and the Allocation of Aid.” Journal of Development Aghion, Philippe, and Peter Howitt. 1992. “A Model of Studies 50 (11): 1510–22. Growth through Creative Destruction.” Econometrica Chikis, Craig. 2024. “The Developmental Trinity: Institutions, 60 (2): 323–51. Infrastructure, and Technology.” Background paper pre - Aiyar, Shekhar S., Romain A. Duval, Damien Puy, Yiqun pared for World Development Report 2024, World Bank, Wu, and Longmei Zhang. 2013. “Growth Slowdowns Washington, DC. and the Middle-Income Trap.” IMF Working Paper Dolan, Lindsay. 2016. “What’s in a World Bank Income WP/13/71 (March), International Monetary Fund, Classification?” CGD Blog Post (blog), July 11, 2016. Washington, DC. https://www.cgdev.org/blog/whats-world-bank-income​ Almås, Ingvild. 2012. “International Income Inequality: -classification. Measuring PPP Bias by Estimating Engel Curves Easterly, William Russell, Michael R. Kremer, Lant H. Pritchett, for Food.” American Economic Review 102 (2): 1093–117. and Lawrence H. Summers. 1993. “Good Policy or Good Balland, Pierre-Alexandre, Tom Broekel, Dario Diodato, Luck? Country Growth Performance and Temporary Elisa Giuliani, Ricardo Hausmann, Neave O’Clery, and Shocks.” Journal of Monetary Economics 32 (3): 459–83. Slowing Grow th | 45 Eichengreen, Barry Julian, Donghyun Park, and Kwanho Mankiw, N. Gregory, David Romer, and David N. Weil. 1992. Shin. 2011. “When Fast Growing Economies Slow Down: “A Contribution to the Empirics of Economic Growth.” International Evidence and Implications for China.” Quarterly Journal of Economics 107 (2): 407–37. NBER Working Paper 16919 (March), National Bureau of Mokyr, Joel. 2023. “The Industrial Revolution and the Economic Research, Cambridge, MA. Origins of Modern Economic Growth: A New Look.” Fantom, Neil James, and Umar Serajuddin. 2016. “The World Presentation at Tel Aviv University, Tel Aviv, Israel, May 1, Bank’s Classification of Countries by Income.” 2023. https://en-social-sciences.tau.ac.il/sites/­socsci​ Policy Research Working Paper 7528, World  Bank, -english.tau.ac.il/files/media_server/social/2023/Tel​ Washington, DC. -Aviv-May’23.pdf. Gill, Indermit Singh, and Homi Kharas. 2007. An East Nielsen, Lynge. 2011. “Classifications of Countries Based on Asian Renaissance: Ideas for Economic Growth. With Their Level of Development: How It Is Done and How It Deepak Bhattasali, Milan Brahmbhatt, Gaurav Datt, Could Be Done.” IMF Working Paper WP/11/31 (February), Mona Haddad, Edward Mountfield, Radu Tatucu, International Monetary Fund, Washington, DC. and Ekaterina Vostroknutova. Washington, DC: Pritchett, Lant H., Kunal Sen, Sabyasachi Kar, and Selim World Bank. Raihan. 2016. “Trillions Gained and Lost: Estimating the Gill, Indermit Singh, and Homi Kharas. 2015. “The Middle- Magnitude of Growth Episodes.” Economic Modelling Income Trap Turns Ten.” Policy Research Working 55 (June): 279–91. Paper 7403, World Bank, Washington, DC. Ravallion, Martin. 2009. “Do Poorer Countries Have Less Grossman, Gene M., and Elhanan Helpman. 1991. Innovation Capacity for Redistribution?” Policy Research Working and Growth in the Global Economy. Cambridge, MA: MIT Paper 5046, World Bank, Washington, DC. Press. Ravallion, Martin. 2012. “Should We Care Equally about Hamadeh, Nada, Aart C. Kraay, Eric Metreau, Marko Poor People Wherever They May Live?” Let’s Talk Rissanen, Giovanni Tonutti, Catherine Van Rompaey, Development (blog), November 8, 2012. https://blogs​ Mizuku Yamanaka, and Kathryn Young. 2022. “Using .worldbank.org/en/developmenttalk/should-we-care​ Purchasing Power Parities in the World Bank’s -equally-about-poor-people-wherever-they-may-live. Classifications of Countries by Income Level.” Ravallion, Martin. 2013. “Why $12,616?” CGD Blog Post (blog), World Bank, Washington, DC. July 8, 2013. https://www.cgdev.org/blog/why-12616. Hausmann, Ricardo, César A. Hidalgo, Sebastián Bustos, Robertson, Peter E., and Longfeng Ye. 2013. “On the Existence Michele Coscia, Alexander James Gaspar Simoes, and of a Middle Income Trap.” Economics Discussion Paper Muhammed A. Yildirim. 2013. The Atlas of Economic 13-12 (February), Department of Economics, University Complexity: Mapping Paths to Prosperity. Cambridge, of Western Australia, Perth. MA: MIT Press. Segerstrom, Paul S., T. C. A. Anant, and Elias Dinopoulos. Hidalgo, César A., and Ricardo Hausmann. 2009. “The 1990. “A Schumpeterian Model of the Product Life Building Blocks of Economic Complexity.” Proceedings Cycle.” American Economic Review 80 (5): 1077–91. of the National Academy of Sciences 106 (26): Soh, Hoon Sahib, Youngsun Koh, and Anwar Aridi, eds. 2023. 10570–75. Innovative Korea: Leveraging Innovation and Technology Im, Fernando Gabriel, and David Rosenblatt. 2013. “Middle- for Development. Washington, DC: World Bank. Income Traps: A Conceptual and Empirical Survey.” Solow, Robert M. 1956. “A Contribution to the Theory of Policy Research Working Paper 6594, World Bank, Economic Growth.” Quarterly Journal of Economics Washington, DC. 70 (1): 65–94. Jones, Benjamin F., and Benjamin A. Olken. 2008. “The Spence, Michael. 2011. The Next Convergence: The Future Anatomy of Start-Stop Growth.” Review of Economics of Economic Growth in a Multispeed World. New York: and Statistics 90 (3): 582–87. Farrar, Straus and Giroux. Kar, Sabyasachi, Lant H. Pritchett, Selim Raihan, and Kunal Summers, Lawrence H., and Lant H. Pritchett. 2014. “Growth Sen. 2013. “Looking for a Break: Identifying Transitions Slowdowns: Middle-Income Trap vs. Regression to the in Growth Regimes.” Journal of Macroeconomics 38 Mean.” VoxEU Column: Development, December 11, (Part B): 151–66. 2014. https://cepr.org/voxeu/columns/growth-­slow​ Kharas, Homi, and Indermit Singh Gill. 2020. “Growth downs-middle-income-trap-vs-regression-mean#/. Strategies to Avoid the Middle-Income Trap.” In Sumner, Andy, and Sergio Tezanos Vázquez. 2012. Trapped in the Middle? Developmental Challenges “Beyond Low- and Middle-Income Countries: What If for Middle-Income Countries, edited by José Antonio There Were Five Clusters of Developing Countries?” Alonso and José Antonio Ocampo, 24–47. Oxford, UK: IDS Working Paper 404, Institute of Development Oxford University Press. Studies, Brighton, UK. Kim, Kihwan. 2006. “The 1997–98 Korean Financial World Bank. 1978. World Development Report 1978. Crisis: Causes, Policy Response, and Lessons.” Paper Washington, DC: World Bank. presented at the International Monetary Fund and World Bank. 1983. World Development Report 1983. Singapore government’s High-Level Seminar on Crisis Washington, DC: World Bank; New York: Oxford Prevention in Emerging Markets, Singapore, July 10–11, University Press. 2006. https://www.imf.org/external/np/seminars/eng​ World Bank. 1989. “Per Capita Income.” Board document /2006/cpem/pdf/kihwan.pdf. SecM89-73 (January 17), World Bank, Washington, DC. Kose, M. Ayhan, and Franziska Ohnsorge, eds. 2024. Falling World Bank. 2015. The World Bank Group: A to Z, 2015. Long-Term Growth Prospects: Trends, Expectations, and Washington, DC: World Bank. Policies. Washington, DC: World Bank. World Bank. 2020. Purchasing Power Parities and the Size Lucas, Robert E. B., Jr. 1993. “Making a Miracle.” of World Economies: Results from the 2017 International Econometrica 61 (2): 251–72. Comparison Program. Washington, DC: World Bank. 46 | WORLD DEVELOPMENT REPORT 2024 Spotlight From X-rays to MRIs: The Enterprise need for a clear picture of Middle-income growth requires a shift from economic structure investment in physical capital to infusion of tech- nology and innovation. At this stage, countries What indicates that a middle-income economy need to improve their capabilities to produce a is advancing in its efforts to catch up with high- range of sophisticated products. The Economic income economies? Complexity Index (ECI) provides a measure of a The World Bank’s income classification country’s productive capabilities based on the aims to reflect a country’s level of development, diversity and complexity of its export basket. drawing on Atlas gross national income (GNI) At this stage of growth, countries also need per capita as a broadly available indicator of to improve the allocation of resources to their economic capacity.1 GNI per capita is a useful most productive firms. In fact, efficient allocation indicator that is often closely correlated with of the factors of production accounts for about other, nonmonetary measures of quality of 25 percent of productivity growth in developing life. However, it does not directly measure a countries.4 Efficiency in allocating resources country’s level of development or the welfare of boosts job and output growth as well as creates its residents.2 Moreover, measures of income per positive spillovers for other businesses along the capita can reflect high levels of investment and value chain.5 consumption expenditure by governments— By reducing productivity-constraining as well as good fortune in the area of natural distortions—defined as policy measures that curb resources.3 the expansion and growth of productive firms— Because of these limitations, better indicators countries can encourage productive firms to are needed to provide a clear picture of the infuse new technologies and grow. Finally, when underlying structure of an economy, much like countries focus on innovation, there is a rise in an MRI provides three-dimensional images of the number of their patents and the importance the human body, going beyond the less-detailed of these patents in the global production of view provided by an X-ray. To that end, the first knowledge. step is to examine the dynamism of a country’s enterprises, its talent, and its changing carbon emissions profile. The second step is to examine Talent specific factors that drive progress and identify Countries need a more skilled workforce as their those that hold it back. production processes become more complex with 47 infusion and innovation. Although low-skilled supply disrupts production and creates additional workers can fuel demand from firms at early stages, pressures on firms to invest in alternate backup infusion and innovation require a shift toward sources. workers with technical and professional skills. Further, growing energy use often exacer- Enrollment in formal schooling can signal a bates emissions. The ultimate impact of rising country’s commitment to education—even if it energy demand on carbon emissions will depend may not reflect the actual skills and ­competencies on the carbon intensity of production, reflecting being developed. The first step for countries, both the energy mix and the energy intensity of therefore, is to provide primary and secondary production. Middle-income countries will need education to their youth. to enhance energy access and reliability while The second step requires countries to improve “decoupling” from their economic growth and social mobility—that is, to create better incen- carbon emissions. tives for individuals to invest in their own skills The indicators suggested in table S.1 development. With improved social mobility, they complement the use of GNI per capita to enjoy better opportunities to take advantage of understand a country’s economic structure. their potential. The selected countries in table S.1 account for As countries improve their talent pools, 62 percent of the global population and 72 percent innovation and infusion require countries to of global GDP. The proposed indicators can help invest in their researchers, who contribute to the policy makers gain a clearer and more complete expansion of knowledge in various fields. picture of their countries’ economic health. In  the  twenty-first century, business dynamism, the talent pool, and energy and carbon intensity Energy are much more reliable indicators of the progress A reliable energy supply is vital for a country’s an economy is making toward reaching high- economic prosperity and security. Unreliable ­ income status. 48 | WORLD DEVELOPMENT REPORT 2024 Table S.1 Suggested indicators provide a clear picture of the underlying structure of an economy Income per capita Enterprise Talent Energy Upper- Energy stability Carbon intensity GNI per secondary (% of firms Energy of energy capita, Atlas completion identifying intensity (energy consumption (CO2 method Economic Productivity- rate (% of Researchers electricity as a consumption/ emissions/energy Income (current Complexity constraining Technology relevant Social per million major GDP , exajoules consumption, Country group US$, 2023) Index distortions index age group) mobility inhabitants constraint) per trillion US$) Gt/exajoules) Pakistan LMIC 1,500 –0.6 0.47 0.0 23.25 0.23 394.9 24.20 9.57 0.12 India LMIC 2,540 0.5 0.36 1.0 54.40 0.23 255.7 12.60 10.77 0.09 Viet Nam LMIC 4,180 0.0 — 0.0 63.40 0.45 772.4 3.70 11.23 0.10 Indonesia UMIC 4,870 –0.1 — 0.0 65.25 0.53 336.2 11.40 7.41 0.10 South Africa UMIC 6,750 –0.1 — 0.2 53.20 0.77 491.7 19.40 11.87 0.10 Colombia UMIC 6,870 –0.1 0.37 0.0 72.80 0.45 — 50.10 6.38 0.09 Brazil UMIC 9,070 –0.2 — 0.3 69.00 0.57 — 46.00 6.98 0.08 Kazakhstan UMIC 10,940 –0.5 — 0.0 98.00 0.65 658.2 21.90 14.15 0.09 Türkiye UMIC 11,650 0.6 — 0.2 63.00 0.38 1,736.7 17.00 7.74 0.07 Malaysia UMIC 11,970 0.9 0.25 0.4 — 0.69 1,433.0 22.20 11.90 0.06 Mexico UMIC 12,100 1.1 — 0.2 58.25 0.55 348.7 46.70 6.17 0.07 Argentina UMIC 12,520 –0.2 — 0.1 63.50 — 1,249.7 47.20 5.69 0.10 China UMIC 13,400 1.3 — 13.1 73.50 0.44 1,523.5 1.80 8.87 0.08 Costa Rica UMIC 13,850 0.4 — 0.1 62.20 — 369.6 63.20 — — Russian UMIC 14,250 0.2 — 4.6 90.40 0.58 2,727.0 7.70 12.90 0.08 Federation Bulgaria UMIC 14,460 0.6 0.31 0.2 86.00 0.35 2,354.6 12.70 9.37 0.06 Chile HIC 15,820 –0.3 0.31 0.1 84.50 0.60 520.1 30.10 5.94 0.06 Romania HIC 16,670 1.2 0.30 0.2 83.60 0.28 922.2 36.10 4.31 0.07 Hungary HIC 19,820 1.5 0.25 0.5 85.60 0.36 4,156.8 12.40 5.35 0.06 Portugal HIC 26,270 0.7 0.50 0.4 83.20 0.38 5,177.3 43.80 3.68 0.06 (Table continues next page) Spotlight | 49 Table S.1 Suggested indicators provide a clear picture of the underlying structure of an economy (continued) Income per capita Enterprise Talent Energy Upper- Energy stability Carbon intensity GNI per secondary (% of firms Energy of energy capita, Atlas completion identifying intensity (energy consumption (CO2 method Economic Productivity- rate (% of Researchers electricity as a consumption/ emissions/energy Income (current Complexity constraining Technology relevant Social per million major GDP , exajoules consumption, Country group US$, 2023) Index distortions index age group) mobility inhabitants constraint) per trillion US$) Gt/exajoules) Korea, Rep. HIC 35,490 2.0 — 92.5 98.75 0.79 8,483.2 — 7.63 0.05 Japan HIC 39,030 2.3 — 50.1 95.50 0.65 5,476.3 — 4.22 0.06 France HIC 45,070 1.3 0.11 11.5 88.40 0.72 4,947.1 40.20 3.01 0.04 United HIC 47,800 1.6 — 9.9 83.33 0.79 4,491.3 — 2.38 0.05 Kingdom 50 | WORLD DEVELOPMENT REPORT 2024 United Arab HIC 53,290 0.1 — 0.2 — — 2,582.8 — 9.95 0.05 Emirates Belgium HIC 54,530 1.2 0.28 4.7 86.20 0.65 5,576.4 29.60 4.23 0.04 Qatar HIC 70,070 –0.4 — — 84.00 — 783.5 — 7.94 0.06 United HIC 80,300 1.4 0.00 100.0 94.00 0.66 4,340.9 — 3.77 0.06 States Source: WDR 2024 team. Note: Countries are listed within income groups from lowest to highest GNI per capita. The Economic Complexity Index is a ranking of countries based on the diversity and complexity of their export basket. Productivity-constraining distortions are estimates of a regression coefficient between the logarithm of idiosyncratic distortions and the logarithm of idiosyncratic physical productivity across firms, both computed as in Hsieh and Klenow (2009). The productivity-constraining distortions are reported relative to the level observed in the United States. The technology index is a composite index of patents per capita and the network centrality of the patents created by a country. Completion rate is the percentage of a cohort of children or young people three to five years older than the intended age for the last grade of each level of education (primary, lower secondary, or upper secondary) who have completed that level of education. Social mobility shows the intergenerational mobility between children (in the 1980s birth cohort) and parents’ years of schooling. Social mobility is measured by 1 minus the intergenerational correlation in schooling. Energy stability is measured by the percentage of firms identifying electricity as a major constraint. Energy intensity is defined as the ratio of energy consumption to GDP (in exajoules per trillion US dollars). Carbon intensity of energy consumption is measured by carbon dioxide (CO2) emissions per energy consumption (gigatons per exajoules). GDP = gross domestic product; GNI = gross national income; Gt = gigatons; HIC = high-income country; LMIC = lower-middle-income country; UMIC = upper-middle-income country; — = not available. Notes 1. Hamadeh, Van Rompaey, and Metreau (2023). 4. Cusolito and Maloney (2018). 2. Fantom and Serajuddin (2016). 5. Grover Goswami, Medvedev, and Olafsen (2019). 3. World Bank (2020). References Cusolito, Ana Paula, and William F. Maloney. 2018. Hamadeh, Nada, Catherine Van Rompaey, and Eric Productivity Revisited: Shifting Paradigms in Analysis Metreau. 2023. “World Bank Group Country and Policy. Washington, DC: World Bank. Classifications by Income Level for FY24 (July 1, Fantom, Neil James, and Umar Serajuddin. 2016. “The 2023–June 30, 2024).” Data Blog (blog), June 30, 2023. World Bank’s Classification of Countries by Income.” https://blogs.worldbank.org/en/opendata/new-world​ Policy Research Working Paper 7528, World Bank, -bank-group-country-classifications-income-level-fy24. Washington, DC. Hsieh, Chang-Tai, and Peter J. Klenow. 2009. “Misallocation Grover Goswami, Arti, Denis Medvedev, and Ellen Olafsen. and Manufacturing TFP in China and India.” Quarterly 2019. High-Growth Firms: Facts, Fiction, and Policy Journal of Economics 124 (November): 1403–48. Options for Emerging Economies. Washington, DC: World Bank. 2020. Purchasing Power Parities and the Size World Bank. of World Economies: Results from the 2017 International Comparison Program. Washington, DC: World Bank. Spotlight | 51 2 Structural Stasis Key messages • Economic growth in middle-income countries is different than that for countries at other income levels. Capital returns diminish at later development stages, and therefore countries, to achieve sustained growth, need to also focus on technological progress and improved efficiency in converting capital and labor into goods and services. • Successful middle-income countries will have to engineer two successive transitions to develop economic structures that can eventually sustain high-income levels. • The first transition is from a 1i strategy for accelerating investment to a 2i strategy focus- ing on both investment and infusion in which a country brings technologies from abroad and diffuses them domestically. Policy makers in lower-middle-income countries will need to add to investment strategies to infuse modern technologies and business prac- tices from global leaders into their own economies. • Once a country has succeeded in the first transition, the second transition is to switch to a 3i strategy, which entails paying more attention to innovation. Upper-middle-​income countries that have mastered infusion can complement investment and infusion with innovation, thereby developing industrial structures and technical competencies to add value to and advance the global technology frontier. Meanwhile, in Northeast Asia something Introduction completely different was happening in the ­ In Brazil in the early 1970s, after several Republic of Korea and Taiwan, China. In the decades of impressive output growth, the aver- 1970s, the productivity of Korea’s manufactur- age worker in the manufacturing sector was ing workers was less than one-tenth of their more than 40  ­ percent as productive as his American counterparts. By 2008, their produc- American c ­ounterpart. By 2008, this ratio had tivity was greater than 70 percent that of the fallen to 17 percent. Up until about 1980, Brazil ­ average American worker in the same sector.2 implemented protectionist policies from for- ­ Their enterprises became well known globally eign competition and provided incentives to because their economies reinvented themselves substitute imports by domestic manufacturers. after they had reached middle-income status— Although all these policies were intended to make not once, but twice. Both economies grew rap- Brazil more competitive, they led to a decline in idly to the upper-middle-income level, followed the productivity of Brazilian workers, and enter- by the high-income level, and subsequently to prises became less competitive than those in the levels of income and standards of living simi- United States.1 lar to those of advanced economies such as 53 Germany, Japan, the United Kingdom, and the Once a country has succeeded in infusing United States. global technologies and know-how in specific What changed between the 1950s and 1960s sectors or industries, it can switch to a 3i strategy (when Brazil was a low-income country), the by paying more attention to innovation. Upper- 1970s (when Brazil achieved rapid growth), and middle-income countries that have mastered the 1980s and 1990s (when Brazil—and its neigh- infusion can complement investment and infu- bors such as Argentina and Colombia—became sion with innovation, thereby developing indus- both a middle-income economy and an also-ran)? trial structures and technical competencies to How did the trajectory of economic develop- add value to and advance the global technology ment differ in Korea and its neighbors Japan and frontier. Taiwan, China? Infusion is powered mainly by technology This chapter explores whether economic transfers embodied in flows of physical and growth in middle-income countries is different financial capital, while innovation requires both than that at other income levels. The simple logic of these flows, as well as increasingly vigorous is that if it is different, then these countries’ devel- exchanges of human capital through engagement opment strategies cannot remain the same. The with the diaspora and the emigration of talented chapter points to evidence consistent with the workers. However, these are not hard-and- hypothesis that successful middle-­ income coun- fast rules. Some countries have succeeded in tries have to engineer two successive transitions to attaining high income levels without instituting develop economic structures that can eventually the structural prerequisites needed to sustain sustain high-income levels. The first is to transi- them. They did so by getting rid of obsolete tion from a 1i strategy for accelerating investment economic arrangements, by weakening the forces to a 2i strategy focusing on both investment and of preservation, and by creating the necessary infusion in which a country brings technologies new ones. However, it appears that these from abroad and diffuses them domestically countries—such as Argentina and República (table 2.1). Policy makers in l ­ower-middle-income Bolivariana de Venezuela—also find it difficult to countries will need to add to investment strat- ensure that their income gains are durable, and egies to infuse modern technologies and busi- even more difficult to continue to close the gaps ness practices from global leaders into their own in living standards with economies at the global economies. economic frontier. Table 2.1 Middle-income countries will need to engineer two successive transitions to develop economic structures that can sustain high-income status INCOME CLASSIFICATION INVESTMENT INFUSION INNOVATION Low-income Higher priority Lower priority Lower priority Lower-middle-income Higher priority Higher priority Lower priority Upper-middle-income Higher priority Higher priority Higher priority Source: WDR 2024 team. Note: The orange dials indicate a strategy that is a priority for that particular income group. The blue dials indicate a strategy that is less of a priority for that particular income group until the priority strategy is successfully achieved. 54 | WORLD DEVELOPMENT REPORT 2024 Economic development = that, although there has been considerable con- vergence in capital to output ratios between low- structural change and middle-income countries and high-income As described in chapter 1, the term “middle-­ income countries, income levels have not converged. trap” refers to the risk of an economic slowdown And so other factors are clearly at work. A simple or stagnation if a country fails to adapt its policies decomposition of factor endowments and total and institutions to changing economic and struc- factor productivity (TFP) reveals that the contri- tural needs.3 Strategies based on factor accumula- bution of physical capital per worker diminishes tion alone are likely to steadily worsen results—a at later development stages (figure  2.1). What natural occurrence as the marginal productivity of really matters for growth is TFP growth, which is ­ iddle-income countries capital declines. Even if all m clumsy longhand for the effects of technological enjoyed the enabling conditions of peace, freedom, progress and improved efficiency in converting factor mobility, and rule of law, the returns from capital and labor into goods and services. In fact, capital investment alone would decline too sharply much of the growth in the United States between to support the countries’ sustained and ongoing 1909 and 1949—when it was a middle-income economic growth.4 country—stemmed from technical change, not If capital endowments were the only economi- an increase in capital per worker.5 cally relevant difference between middle-­ income If much of growth everywhere is the result of and high-income countries today, the gross technical change, then conventional thinking national income (GNI) of a typical middle-income would follow that every middle-income coun- country would be about 75 percent that of the try needs to figure out how to quickly institute United States. In China, for example, its invest- arrangements that foster technical progress, ment to gross domestic product (GDP) ratios not only (or primarily) the accumulation of cap- have been stratospherically high for decades, but ital. But this interpretation is not helpful for its GNI is less than 25 percent that of the United policy making because of the ongoing impor- States. Another way to understand the problem is tance of capital deepening across all categories Figure 2.1 As economies develop, capital accumulation brings diminishing returns 60 (% of GDP per capita growth) 50 Contribution to growth 40 30 20 10 0 Low-income Lower-middle-income Upper-middle-income High-income Physical capital per worker Total factor productivity Source: WDR 2024 team using data from Lange, Wodon, and Carey (2018); PWT (Penn World Table) (database version 10.1), Groningen Growth and Development Centre, Faculty of Economics and Business, University of Groningen, Groningen, the Netherlands, https://www.rug.nl/ggdc/productivity/pwt/. Note: GDP = gross domestic product. Struc tur al Stasis | 55 of countries, but especially in middle-income and diffusing this knowledge at scale in their economies. domestic economy. These technologies and mod- In fact, the growth of middle-income countries els have, often in short order, enabled enterprises depends on both capital accumulation and tech- to become regional and global suppliers of goods nical change, making the growth challenge twice and services. as complex as it is for either low-income countries The second phase of structural change, com- that primarily must focus on accumulation or monly called innovation, occurs mainly in suc- high-income countries that must rely largely on cessful upper-middle-income economies. This technical change, even if a large part of it is in the transition involves a deliberate shift from imitat- form of new investment. ing and adapting technologies used in advanced economies to building home country capabilities to change leading global technologies and prod- Infuse first, then innovate ucts. An increasing number of domestic firms How have the most successful middle-income can become global knowledge creators and— countries engineered progress? Modern economic eventually—leading innovators themselves. history provides one valuable lesson. Countries The term infusion has been carefully chosen to that have made technological advances and connote both deliberately imitating technology achieved high-income status did so through two and business practices from abroad and expedit- successive transitions. ing their diffusion at home (figure 2.2). Not sur- The first set of changes, described as infusion prisingly, countries that have been relatively open in this Report, dominates development strat- to economic developments abroad and have been egies in rapidly growing lower-middle-income successful at instituting general secondary edu- countries. Policy makers in these countries have cation and technical and vocational training pro- emphasized importing modern technologies and grams at home have done better than those that business models from more advanced economies failed to do one or both. Figure 2.2 A middle-income country will need to engineer two successive transitions to achieve high-income status: Infusion, followed by innovation 1i 2i 3i Investment Investment + Infusion Investment + Infusion + Innovation Relative contribution to growth Capital Productivity Proximity to the frontier Source: WDR 2024 team. Note: The curves illustrate the relative contributions of capital and productivity to economic growth (y-axis) according to countries’ proximity to the frontier (represented by leading economies). Countries farther out on the x-axis are closer to the frontier. 56 | WORLD DEVELOPMENT REPORT 2024 The mechanics of the innovation stage of eco- technologies and innovation through tax nomic development are more difficult, and so this credits. Specifically, firms received tax stage has received much attention.6 Like others, credits for royalty payments or research this Report warns developing countries against and development (R&D) expenditures. The attempts to “leapfrog” (that is, to prematurely policy first subsidized technology adop- attempt to transition) to the innovation stage tion when the technology gap with foreign (generally through the use of industrial policy firms was large. But as Korean conglom- interventions). What is more novel in this Report erates caught up with foreign firms, the is its emphasis on the changing nature of knowl- approach gradually shifted toward sup- edge exchanges and the successful impact of such porting innovation. Korean policy makers exchanges on fostering development by benefit- ensured that public support was moni- ing from the international mobility, not only of tored and evaluated, and data on innova- capital and know-how but also of highly skilled tion grants were made publicly available. ­ people. The prerequisites needed to capitalize • Malaysia. Malaysia became a success- on the global nature of human capital—such as ful industrialized country through policies to attract entrepreneurs from the dias- ­ infusion-centered and export-oriented pora and ensure greater freedom of expression— growth that replaced import substitution are more difficult to institute and can stymie policies in the mid-1980s. Technology progress. embodied in foreign direct investment Successful infusion efforts have marked rever- (FDI) was important for developing and sals of fortune in several parts of the world marred structuring the country’s industrial base. by war and violence: Malaysia offered a spectrum of tax incen- tives to attract FDI through the Promo- • Postwar Europe. The onset of the European tion of Investment Act in 1986.9 Malaysia’s Golden Age was powered by infusion. The growth in the 1980s was marked by large two world wars in the first half of the twen- productivity gains from adopting and dif- tieth century destroyed much of Europe’s fusing technology. But Malaysia did not capital stock and skills. And the exodus perform as well as Singapore in attracting of talent in the interwar years meant that entrepreneurs of Malaysian origin living Europe lagged behind the United States abroad. in technology.7 The Marshall Plan was developed to transfer technologies from As these examples show, infusion, tapping the United States to Europe. European into global knowledge, and a country’s institu- managers were sent to the United States tional structure play a key role in supporting to acquire skills, and businesses could the economic growth of middle-income coun- obtain loans to purchase technologically tries beyond just increasing a country’s income advanced American capital goods under per capita. And the key to infusion at scale is the US Productivity Program. As a result, openness and exchange—through paths such as Europe rapidly adopted modern technol- competition regulation, licensing, trade, FDI, pro-­ ogies and best practices, allowing West- migration, and knowledge exchanges. A combi- ern European countries to accomplish in nation of investment and infusion can engineer 30 years what might otherwise have taken high growth through investments in physical twice as long.8 capital (infrastructure), structural change that • Korea. An important component of indus- improves the allocation of productive resources trial policy in Korea was incentives for across firms and sectors, and technological con- technological investment. In particular, vergence through the adoption and infusion of Korea subsidized the adoption of foreign foreign technologies.10 Struc tur al Stasis | 57 Figure 2.3 The demand for highly skilled workers increases in middle-income countries 60 Share of highly skilled workers in labor force (%) 50 40 30 20 10 0 20 40 60 80 GDP per capita (constant PPP 2017 US$1,000, log scale) Low-income Middle-income High-income Source: WDR 2024 team using DataBank: Jobs, World Bank, Washington, DC, https://databank.worldbank.org/source​ /­jobs; WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org​ /­search/dataset/0037712. Note: Skilled workers consist of the top three International Standard Classification of Occupations (ISCO) codes (“Legislators, sr. officials, managers”; “Professionals”; “Technicians and associate professionals”). See ISCO (International Standard Classification of Occupations), International Labour Organization, Geneva, https://ilostat.ilo.org/methods/concepts-and​ -definitions/classification-occupation/. GDP = gross domestic product; PPP = purchasing power parity. To undertake infusion at scale, however, coun- the science, technology, engineering, and math- tries need both globally competitive firms and spe- ematics (STEM) fields play a central role in gen- cialized talent. As firms adopt newer technologies, erating and spreading ideas and technologies. their need for engineers, scientists, managers, and Encouragingly, three-quarters of STEM graduates other highly skilled professionals increases. The are now in middle-income countries, and Chinese variety and skill content of work also increase and Indian STEM graduates make up about half of in middle-income countries (figure  2.3). In fact, global STEM graduates (figure 2.4). the central feature of the Industrial Revolution The example of Korea, the only country ever and its aftermath was the slow shift from tacit to sustain economic growth that averaged more knowledge that is not codified or easily expressed than 5 percent for more than 50 years, is espe- (as embodied in craftsmanship and simple pro- cially illustrative. How did Korean industry do duction techniques) to more formal knowledge it? Domestic firms such as Samsung embarked created by mathematicians, physicists, chem- on a journey that began with infusion. By licens- ists, medical doctors, and people schooled in ing technologies from Sanyo and NEC in Japan, “engineering science.”11 Today, graduates from Samsung transitioned from making noodles 58 | WORLD DEVELOPMENT REPORT 2024 Figure 2.4 STEM graduates are increasingly through a combination of investment and imita- concentrated in middle-income countries, tion (see box 2.1 for a description of how Japan thereby increasing opportunities for connected with global knowledge). But eventu- technology infusion ally the gains from imitation begin to subside. As an economy approaches the global technology 50 frontier, policies that supported growth even- Share of global STEM graduates (%) 43.8 tually become a burden for sustaining further 40 development.14 When Samsung reached that point, it moved toward innovation—nudged by 30.2 support from the Korean government, which 30 had calibrated its incentives to encourage imi- 23.5 tation first and support innovation much later 20 (figure 2.5, panel b). To nurture innovation, insti- tutions must give inventors and entrepreneurs 10 incentives and ensure that they can acquire the technical and financial resources to carry out 2.5 their designs.15 0 Once a middle-income country has infused its e e e e economy with technology from global innova- m m m m co co co co tors and is sustaining rapid growth, it can aspire -in in in in e- e- h- w dl dl ig Lo to converge to the global technology frontier by id id H -m r-m preparing to join those innovators—that is, to er pe w Up Lo become an innovation economy. To reach this Source: WDR 2024 team. point, however, governments must have done Note: The primary data for the calculation of the everything possible in the infusion phase to not percentage share of science, technology, engineering, and mathematics (STEM) graduates in tertiary education are just prepare the economic structure for the dif- from UIS.Stat (dashboard), Institute for Statistics, United ferent next stage, but also reform and strengthen Nations Educational, Scientific, and Cultural Organization, supporting institutions. Those who falter in infu- Montreal, http://data.uis.unesco.org/. The STEM data set covers 144 countries. sion or try to leapfrog will find it much more chal- lenging to transition toward innovation. to manufacturing televisions for domestic and A comparison of Estonia, Poland, and Bulgaria regional markets (figure 2.5, panel a).12 This tran- is illustrative. Transitioning from central plan- sition created a higher demand for engineers, ning, Estonia reached 80 percent of Western managers, and skilled professionals that was European income by 2021, Poland 75 percent, and monitored, and the targets were met by the Bulgaria 50 percent (box 2.2). Estonian indepen- Ministry of Education through both providing dence in 1991 catalyzed a swift transition to high the needed education in public universities and levels of innovation. By contrast, Bulgaria pro- regulating private institutions. Korea also gen- tected many incumbent state-owned enterprises erated a demand for more specialized capital: (SOEs) from competition and stymied efficient for economies at the infusion stage, investment resource allocation, preventing the contrac- remains important. tion of low-­ productivity sectors. As for Poland, The first World Development Report (1978) it privatized many of its SOEs and ­ championed highlighted the need to differentiate between competition. strategies of imitation and innovation in driv- Middle-income countries lag noticeably behind ing growth in middle-income countries.13 A large high-income countries in terms of the “novelty” part of middle-income country growth happens of their knowledge, as well as in producing new Struc tur al Stasis | 59 Figure 2.5 Calibrating policies to a country’s stage of development: From imitation to innovation in the Republic of Korea a. An agreement between companies b. Government incentives to collaborate on technology (subsidies) Figure A.1: Example of Adoption Contract 40 TECHNICAL COLLABORATION AGREEMENT BY AND BETWEEN NIPPON ELECTROIC CO., LTD. 30 Subsidy rate (%) AND SAMSUNG ELECTRON DEVICES CO., LTD. 20 Section 4 Supply of written Technical Information (a) During the term of this Agreement NEC will upon reasoable 10 request furnish SED with one transparent copy of each drawing, specification and other technical document as well as programs and related documentation within the scope specified in Section 1 (d) hereof. The time, manner and other details of furnishing such written NEC Technical 0 Information shall be separately determined by the parties 1970 1980 1990 2000 2010 2020 upon mutual consultation. Adoption subsidy Innovation subsidy Sources: Panel a: National Archives of Korea, https://www.archives.go.kr/english/index.jsp. Panel b: Choi and Shim 2024. Note: Panel b shows the adoption subsidy rate alongside the innovation (R&D) subsidy rate, calculated using the tax credit rate and the corporate tax rate. For example, a 30 percent subsidy rate indicates that firms can receive a reimbursement equivalent to 30 percent of their expenditures on adoption fees or R&D. R&D = research and development. Box 2.1 The Meiji Restoration reconnected Japan with global knowledge The turning point for Japan’s rapid industrialization was the Meiji Restoration in the late 1800s. The g ­ overnment embarked on a project, Shokusan Kogyo (industrial development and promotion of industries).a Under it, Japan began investing in modernizing infrastruc- ture (such as telegraph, rail, and electricity) and deploying demonstration factories to facilitate private sector learning and technological diffusion and to assume the first-mover risks of deploying technology. Trade flows were a critical factor in launching Japan toward the technological frontier. Government-sponsored trips to the United States and Europe were also instrumental in acquiring technical expertise from frontier countries. The 1871–73 Iwakura Mission to both areas proved critical to facilitating the knowledge transfers needed to push Japan to the technology frontier. The statesmen and students who participated set out to study Western institutions, economic structures, educational systems, and industrial capabil- ities. Although German and English industry impressed senior diplomats,b Mission staff took an interest in US applications of new technologies and the productivity gains reaped from such innovations.c (Box continues next page) 60 | WORLD DEVELOPMENT REPORT 2024 Box 2.1 The Meiji Restoration reconnected Japan with global knowledge (continued) In 1888, engineers in the newly formed Ministry of Communications were dispatched to the United States, among other Western countries, to collect information on the state- of-the-art technology needed to construct the country’s telecommunications network.d The ministry officials who attended this trip visited US telecommunications firms such as Western Electric, brought back equipment such as switchboards, and urged the Japanese government to implement Western Electric’s systems for the country’s network.e Foreign expertise was transmitted to Japanese students through foreign practitioners who taught in domestic technical schools (usually in English, indicating that some of the professors must have been American).f The government also sent engineering students to Western countries. By the end of the 1880s, one-quarter of the students who had traveled abroad had visited the United States. a. Genther (2020); Odagiri and Goto (1996). b. Beasley (1972). c. Swale (2008). d. Ohno (2019). e. Mason (1992). f. Ohno (2019). Box 2.2 Three ways to evade the middle-income trap: Swiftly (Estonia), steadily (Poland), or slowly (Bulgaria) Over the last three decades, Bulgaria, Estonia, and Poland have transitioned simultane- ously from c ­ entral planning to market economies, and Estonia and Poland have moved from middle-income to high-income status.a By 2021, Estonia had reached 80 percent of Western European income, Poland 75 percent, and Bulgaria 50 percent.b This remarkable leap from income per capita of between 20 and 30 percent of Western European levels in the early 1990s occurred at different speeds, which provides valuable policy insights. Policy insight 1: Comprehensive reforms unlock productivity and prosperity Structural reforms, specifically privatization and market liberalization, have played a crucial role in speeding up income convergence among Central and Eastern European economies, as documented in an extensive cross-country study by Matkowski, Prochniak, and Rapacki (2016). They find that the rate of convergence has significantly bene- fited from enhancements of institutions, economic freedom, and governance  quality. (Box continues next page) Struc tur al Stasis | 61 Box 2.2 Three ways to evade the middle-income trap: Swiftly (Estonia), steadily (Poland), or slowly (Bulgaria) (continued) Higher investment rates, a skilled labor force, low budget deficits, and lower tax burdens have been associated with accelerated economic growth. Estonian independence in 1991 catalyzed a swift transition to high levels of productiv- ity growth driven by a strategic divestment of public sector assets, trade liberalization, and a flat tax system. These policies opened up the market for new entrants, attracted a surge in foreign direct investment (FDI), and boosted private sector productivity. The magni- tude of FDI inflows to Estonia during the 1990s was seven times that to Bulgaria and three times that to Poland.c Estonia maintained its status as a leading innovator in the region, with the highest research and development (R&D) intensity—1.75 percent of its gross domestic product (GDP) in 2021—and a high ranking in the Global Innovation Index (GII).d As for Poland, early “big bang” reforms, trade competition, and hard budget constraints for state-owned enterprises (SOEs) systematically activated a cycle of creative destruc- tion, closing the gap with advanced countries, as evidenced by an increase in R&D expen- diture to 1.3 percent of GDP in 2021, as well as higher GII scores. Policy insight 2: Incentives for incumbents to drive innovation are crucial Productivity growth is a dominant driver of income convergence. Between 1996 and 2021, Estonia displayed robust annual productivity growth of 3.8 percent, followed by Poland at 2.6 percent and Bulgaria at 1.4 percent. Lowering barriers to entry and streamlining regulation—enhanced by accession to the European Union (EU)—have initiated a virtuous cycle that has sustained productivity growth. In this cycle, competition spurs innovation, which then fuels further competition, ultimately raising societal well-being. During Bulgaria’s initial phase of its transition, many incumbent SOEs were shielded by regulatory safeguards and anticompetitive practices. This environment hampered productivity growth by delaying the shift of resources toward more productive sectors. Throughout much of the period preceding EU accession, growth among Bulgaria’s high-productivity industries was limited. At the same time, low-productivity sectors that relied on low-skilled labor—such as construction, retail, and ground transportation—­ continued to expand. Unlike Estonia and Poland, Bulgaria took nearly a decade longer to free up resources for more productive uses. Even now, some industries benefit from regulation that inhibits healthy competition. Poland’s mass privatization program is a compelling example of how to catalyze a virtu- ous cycle of competition and innovation, despite some challenges. By redistributing equity from more than 500 SOEs—which constituted up to 5 percent of the nation’s wealth—to 27 million citizens, Poland created incentives for both new and established companies to champion competition. In Poland’s early transition, the implementation of hard bud- get constraints, established so that SOEs would avoid reliance on unlimited government support, led to a monumental shift in attitude toward market competition. Even iconic SOEs such as the Stocznia Gdańsk shipyard, where Lech Walesa’s Solidarność (Solidarity) movement began, triggering the collapse of communism in Central and Eastern Europe, (Box continues next page) 62 | WORLD DEVELOPMENT REPORT 2024 Box 2.2 Three ways to evade the middle-income trap: Swiftly (Estonia), steadily (Poland), or slowly (Bulgaria) (continued) were not bailed out. The alignment of managerial incentives—both explicit and implicit— propelled creative destruction. Managers at Polish SOEs shifted their attention from pro- duction targets to profitability and market share.e These managers became the primary agents of restructuring and innovation, facing both financial constraints and competition. Policy insight 3: Skilled labor is essential for moving from infusion to innovation As economies mature, the returns from capital become dependent on a supply of sophis- ticated workers. Continual enhancements in firm efficiency and workforce skills are vital to staying competitive in the rapidly changing productivity landscape. The success stories of Estonia and Poland illustrate strategic ways to close the tech- nology gap—­ initially by adopting existing technology and subsequently by developing innovative capacity. Poland narrowed its productivity gap primarily by adopting older technologies from more advanced economies rather than through the transfer of new, cutting-edge technologies.f Domestic investments in innovation further accelerated Poland’s move toward the technology frontier, enhancing firms’ ability to contribute to global technological progress. Labor also played a critical role in the transformation in Poland, according to a breakdown of the growth in total factor productivity.g A notable 20 percent of the contribution came from labor alone, a figure driven not by a reduction in workforce numbers but by the enhanced quality and diversity of the labor force. A sig- nificant rise in the proportion of individuals between the ages of 25 and 34 with a tertiary education, which increased from 15 percent to 42 percent between 2000 and 2012, had a significant impact on output growth.h Bulgaria’s experience illustrates how shortages of skilled labor can impede new invest- ment and the growth of high-productivity ventures. The educational landscape in Bulgaria has lagged in quality, participation, equity, and intergenerational mobility. Certain seg- ments of the population still struggle with the acquisition of basic skills, undermining their potential contribution to productivity. Prioritizing human capital development to take advantage of adoption and innovation is essential, especially in view of the demo- graphic changes in many middle-income countries, which are at risk of “growing old before becoming rich,” as the saying goes. a. Gross national income (GNI) per capita and gross domestic product (GDP) per capita data in this report are as of July 2023. As of July 2023, Bulgaria was classified by the World Bank as an upper- middle-income country, with a GNI per capita (Atlas method, current US dollars) of US$13,350, short of crossing the high-income country threshold of US$13,846. b. These figures are measured using purchasing power parity for the four largest EU economies. c. These figures capture net FDI per capita accumulated by the end of 1999. d. See GII (Global Innovation Index) (dashboard), World Intellectual Property Organization, Geneva, https://www.wipo.int​/­global_innovation_index/en/. e. Pinto, Belka, and Krajewski (1993). f. Kolasa (2008). g. Gradzewicz et al. (2018). h. Bukowski et al. (2006). Struc tur al Stasis | 63 knowledge (figure 2.6). Innovation is concen- Narayana Murthy, cofounder of India-based trated in a handful of high-income economies: Infosys, have grown their enterprises to scale Germany; Japan; the Republic of Korea; Taiwan, and created thousands of jobs by successfully China; and the United States. The magnitude and competing in global markets and even pushing impact of new research in other countries is quite the technology frontier outward. Still, too many limited (figure 2.7). markets are hobbled by excessive business regula- To support innovation, countries will have to tions, government patronage, and limited inter- find ways to make existing firms (incumbents)— national competition. In such an environment, both industrial conglomerates and economic powerful owners and unproductive large firms elites—innovative and more productive and to can stifle growth, lobbying to protect their pref- make way for newcomers. To be sure, the emerg- erential access and monopoly rents when they ing markets are already hotbeds of entrepre- could instead be investing in productivity-­ neurial activity. Novel products are being used enhancing technology. by millions, and new production methods are As policy makers shift their emphasis toward increasing consumer choices and lowering prices. innovation, they should first combine a lot Jack Ma, cofounder of China-based Alibaba, and of investment with a lot of infusion (box 2.3). Figure 2.6 The innovation gap between Figure 2.7 Middle-income countries high-income countries and others is significantly lag behind high-income substantial countries in research capacity 10 4,500 4,149 Researchers per million inhabitants 9 4,000 8.4 8 3,500 Technology index 7 3,000 6 2,500 5 2,000 4 1,500 3 994 1,000 2 490 1 0.6 500 0.0 0.0 48 0 0 e e e e e e e e m m om m m m om om co co co co co c c c in in -in -in -in in -in in e- h- w le e- h- w le dl ig Lo d dl ig Lo d id id H id id H -m -m -m r-m er r pe er pe w w Up Lo Up Lo Source: WDR 2024 team. Source: WDR 2024 team. Note: The technology index is constructed using the number Note: The primary data for number of researchers per of granted patents per capita and network centrality. million inhabitants are from UIS.Stat (dashboard), Institute Network centrality is calculated to measure a country’s for Statistics, United Nations Educational, Scientific, and “frontierness” in technology and is defined by the citations Cultural Organization, Montreal, http://data.uis.unesco.org. in a country’s patents of other countries’ patents in a given period. Therefore, the technology index embeds both patent importance and scale. 64 | WORLD DEVELOPMENT REPORT 2024 Box 2.3 The magic of investment accelerations Investment is a fundamental pillar of economic progress. Not only does investment growth allow countries to enhance their stocks of physical capital such as factories, offices, roads, bridges, schools, and clinics, but it is also a necessary condition for infusing global technol- ogies in domestic production possibilities. Because technology is embodied in capital, a country will find it challenging to advance technologically without scaling up investment. Investors look for macroeconomic stability and ease of doing business in deciding where, in what, and how much to invest. The experiences of Colombia, Türkiye, and the Republic of Korea are examples. In 2001, Colombia implemented a comprehensive reform package to stabilize its economy by restraining public spending, increasing cen- tral bank independence, and introducing a floating exchange rate. Similarly, in the early 2000s Türkiye implemented a primary surplus target, the central bank became indepen- dent, and reforms to improve the business climate and liberalize the banking sector were adopted. Earlier, Korea implemented two rounds of reform packages. In the mid-1980s, Korea adopted a balanced budget, improved the business climate by promoting compe- tition, and liberalized trade. A second round of reforms in the late 1990s improved the independence of the central bank, consolidated government finances, strengthened the financial sector, and liberalized the capital account. Following these reform efforts, all three countries experienced investment accelera- tions: Colombia from 2001 to 2007, Türkiye from 2003 to 2008, and Korea in 1985–96 and from 1999 to 2007. Investment accelerations are periods with a sustained increase in investment growth. During these periods, investment as well as productivity grew much faster than in nonacceleration years (figure B2.3.1). More broadly, across a sample of 104 economies, including 69 emerging market and developing economies (EMDEs) and 35  advanced economies covering the years 1950 to 2022, 192 episodes of investment acceleration occurred in which per capita investment growth averaged at least 4 percent per year over at least six years.a On average, an EMDE has experienced about 1.7 invest- ment accelerations. During these accelerations, investment growth more than tripled to 10 percent a year over that of nonacceleration years; output growth increased by 2 per- centage points; and productivity growth quadrupled to 1.7 percent per year. In the sample of 104 economies, 82 percent of the transitions from middle-income status to high-income status that occurred over the last three decades happened during or shortly after investment accelerations. Sectoral shifts gained momentum during invest- ment accelerations because output grew substantially faster in the manufacturing and services sectors than before the acceleration. Accelerations were also often periods during which more progress was made in reducing poverty, living standards improved, and the pace of convergence to advanced economy income per capita levels increased. (Box continues next page) Struc tur al Stasis | 65 Box 2.3 The magic of investment accelerations (continued) Figure B2.3.1 Investment growth accelerations: Colombia, Republic of Korea, and Türkiye 16 4 Investment growth (%) 12 3 TFP growth (%) 8 2 4 1 0 0 –4 –1 Investment TFP Investment TFP Investment TFP (right axis) (right axis) (right axis) Colombia Korea, Rep. Türkiye Acceleration years Nonacceleration years Sources: Dieppe 2021; Feenstra, Inklaar, and Timmer 2015; World Bank 2024; Database Profiles (dashboard), Our Data, Haver Analytics, New York, https://www.haver.com/our-data; WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog​ .worldbank​.org/search/dataset/0037712. Note: Investment refers to gross fixed capital formation. The sample period is 1980–2022. Bars are simple averages of growth in investment and total factor productivity (TFP). Acceleration years cover the full duration of the episodes. Nonacceleration years exclude acceleration years that are not included in this box (see World Bank 2024, box 3.1). a. World Bank 2024. And  they should not bet the farm—or even a human capital, limited venture capital, field—on leapfrogging. Infusion is imperative. A ­ barriers and incomplete rule regulatory ­ review of the literature on middle-income traps of law present significant barriers to highlights that countries often try to make pre- becoming an innovation-driven economy. mature leaps from investment to innovation. Middle-income countries that invest According to Gill and Kharas (2015, 28–29), heavily and prematurely in trying to become “knowledge economies” can find [ Some countries tried] to leapfrog low returns to such investments. The prematurely into “knowledge economies,” combination of wasted fiscal spending with none of the institutional and a faulty growth diagnostic can lead infrastructure in place to accomplish this. to substandard performance—another Poor quality universities, low levels of example of the middle-income trap. 66 | WORLD DEVELOPMENT REPORT 2024 Notes 1. Nassif, Feijó, and Araújo (2015). procure modern machinery from the United States 2. Branstetter and Kwon (2018). (technology transfer). Specifically, firms that engaged 3. Gill and Kharas (2007). in both management and technology transfers wit- 4. Mokyr (2018). nessed the most substantial long-term productivity 5. Solow (1957). Technical change is an economic term growth, highlighting the important role of infusion meaning a change in the amount of output produced (Giorcelli 2019). from the same amount of inputs.  9. Chuah, Loayza, and Nguyen (2018). 6. Lucas (1988); Romer (1990). 10. Acemoglu, Aghion, and Zilibotti (2006); Gerschenkron 7. Toniolo (1998). (1962); König et al. (2022); Zilibotti (2017). 8. Fernández-Villaverde and Ohanian (2018). In the 1950s, 11. Mokyr (2023). for example, Italian firms benefited from sponsored 12. Choi and Shim (2023). training trips for their managers, enabling them to 13. World Bank (1978). acquire modern management practices from firms in 14. Zilibotti (2017). the United States. Some firms also received loans to 15. Mokyr (2023). References Acemoglu, Daron, Philippe Aghion, and Fabrizio Zilibotti. Genther, Phyllis A. 2020. A History of Japan’s Government- 2006. “Distance to Frontier, Selection, and Economic Business Relationship: The Passenger Car Industry. Growth.” Journal of the European Economic Association Michigan Papers in Japanese Studies Series. Ann 4 (1): 37–74. Arbor: University of Michigan Press. Beasley, William G. 1972. The Meiji Restoration. Stanford, Gerschenkron, Alexander. 1962. Economic Backwardness in CA: Stanford University Press. Historical Perspective. Cambridge, MA: Belknap Press. Branstetter, Lee G., and Namho Kwon. 2018. “South Korea’s Gill, Indermit Singh, and Homi Kharas. 2007. An East Transition from Imitator to Innovator: The Role of Asian Renaissance: Ideas for Economic Growth. With External Demand Shocks.” Journal of the Japanese and Deepak Bhattasali, Milan Brahmbhatt, Gaurav Datt, International Economies 49 (September): 28–42. Mona Haddad, Edward Mountfield, Radu Tatucu, Bukowski, Maciej, Iga Magda, Łukasz Marć, and Julian and  Ekaterina Vostroknutova. Washington, DC: Zawistowski. 2006. “Źródła i perspektyw wzrostu pro - World Bank. duktywności pracy w Polsce” [Sources and prospects Gill, Indermit Singh, and Homi Kharas. 2015. “The Middle- of productivity growth in Poland]. December, Instytut Income Trap Turns Ten.” Policy Research Working Badań Strukturalnych, Warsaw. Paper 7403, World Bank, Washington, DC. Choi, Jaedo, and Younghun Shim. 2023. “Technology Giorcelli, Michela. 2019. “The Long-Term Effects of Adoption and Late Industrialization.” STEG Management and Technology Transfers.” American Working Paper WP033 rev. (September), Structural Economic Review 109 (1): 121–52. Transformation and Economic Growth, Centre for Gradzewicz, Michał, Jakub Growiec, Marcin Kolasa, Economic Policy Research, London. Łukasz Postek, and Paweł Strzelecki. 2018. “Poland’s Choi, Jaedo, and Younghun Shim. 2024. “From Adoption Uninterrupted Growth Performance: New Growth to Innovation: State-Dependent Technology Policy in Accounting Evidence.” Post-Communist Economies Developing Countries.” STEG Working Paper WP091 30 (2): 238–72. (March), Structural Transformation and Economic Kolasa, Marcin. 2008. “How Does FDI Inflow Affect Growth, Centre for Economic Policy Research, Productivity of Domestic Firms? The Role of Horizontal London. and Vertical Spillovers, Absorptive Capacity and Chuah, Lay Lian, Norman V. Loayza, and Ha Nguyen. Competition.” Journal of International Trade and 2018. “Resource Misallocation and Productivity Gaps Economic Development 17 (1): 155–73. in Malaysia.” Policy Research Working Paper 8368, König, Michael, Kjetil Storesletten, Zheng Song, and World Bank, Washington, DC. Fabrizio Zilibotti. 2022. “From Imitation to Innovation: Dieppe, Alistair, ed. 2021. Global Productivity: Trends, Where Is All That Chinese R&D Going?” Econometrica Drivers, and Policies. Washington, DC: World Bank. 90 (4): 1615–54. Feenstra, Robert C., Robert Inklaar, and Marcel Peter Lange, Glenn-Marie, Quentin T. Wodon, and Kevin Carey, eds. Timmer. 2015. “The Next Generation of the Penn 2018. The Changing Wealth of Nations 2018: Building a World Table.” American Economic Review 105 (10): Sustainable Future. Washington, DC: World Bank. 3150–82. Lucas, Robert E., Jr. 1988. “On the Mechanics of Economic Fernández-Villaverde, Jesús, and Lee Edward Ohanian. 2018. Development.” Journal of Monetary Economics 22 (1): “The Lack of European Productivity Growth: Causes 3–42. and Lessons for the U.S.” PIER Working Paper 18-024, Mason, Mark. 1992. American Multinationals and Japan: Penn Institute for Economic Research, Department of The Political Economy of Japanese Capital Controls, Economics, University of Pennsylvania, Philadelphia. 1899–1980. Harvard East Asian Monographs Series, Struc tur al Stasis | 67 vol.  154. Cambridge, MA: Asia Center, Harvard Technological Learning, Industrial Policy, and Catch-Up, University. edited by Arkebe Oqubay and Kenichi Ohno, 85–106. Matkowski, Zbigniew, Mariusz Prochniak, and Ryszard New York: Oxford University Press. Rapacki. 2016. “Real Income Convergence between Pinto, Brian, Marek Belka, and Stefan Krajewski. Central Eastern and Western Europe: Past, Present, 1993.  “Transforming State Enterprises in Poland: and Prospects.” Paper presented at the 33rd Centre for Evidence on Adjustment by Manufacturing Firms.” International Research on Economic Tendency Surveys’ Brookings Papers on Economic Activity BPEA 1993 (1): Conference on Economic Tendency Surveys and 213–70. Economic Policy, Copenhagen, September 14–17, 2016. Romer, Paul Michael. 1990. “Endogenous Technological https://www.econstor.eu/handle/10419/146992. Change.” Journal of Political Economy 98 (5, Part 2): Mokyr, Joel. 2018. “The Past and the Future of Innovation: S71–S102. Some Lessons from Economic History.” Explorations in Solow, Robert M. 1957. “Technical Change and the Economic History 69 (July): 13–26. Aggregate Production Function.” Review of Economics Mokyr, Joel. 2023. “Creative Destruction or Destructive and Statistics 39 (3): 312–20. Creation? A Prelude to the Industrial Revolution.” In The Swale, Alistair. 2008 “America, 15 January to 6 August Economics of Creative Destruction: New Research on 1872: The First Stage in the Quest for Enlightenment.” Themes from Aghion and Howitt, edited by Ufuk Akcigit In The Iwakura Mission to America and Europe: A New and John Michael Van Reenen, 714–35. Cambridge, Assessment, edited by Ian Nish, 7–23. Meiji Japan MA: Harvard University Press. Series, vol. 6. London: Routledge. Nassif, André, Carmem Feijó, and Eliane Araújo. 2015. Toniolo, Gianni. 1998. “Europe’s Golden Age, 1950–1973: “Structural Change and Economic Development: Is Speculations from a Long-Run Perspective.” Economic Brazil Catching Up or Falling Behind?” Cambridge History Review 51 (2): 252–67. Journal of Economics 39 (5): 1307–32. World Bank. 1978. World Development Report 1978. Odagiri, Hiroyuki, and Akira Goto. 1996. Technology and Washington, DC: World Bank. Industrial Development in Japan: Building Capabilities World Bank. 2024. Global Economic Prospects, January by Learning, Innovation, and Public Policy. New York: 2024. Washington, DC: World Bank. Oxford University Press. Zilibotti, Fabrizio. 2017. “Growing and Slowing Down Like Ohno, Kenichi. 2019. “Meiji Japan: Progressive Learning China.” Journal of the European Economic Association of Western Technology.” In How Nations Learn: 15 (5): 943–88. 68 | WORLD DEVELOPMENT REPORT 2024 3 Shrinking Spaces Key messages • The growth prospects of middle-income countries are becoming more problematic because of an increasingly fragmented global economy, rapidly changing demographic trends, multiple crises, populist pressures, rising government debt, and climate change pressures. • Previous episodes of growth acceleration have been accompanied by trade integration, but rising geopolitical tensions have affected trade policy, and further protectionism can potentially worsen the diffusion of knowledge to low- and middle-income countries. • Many middle-income countries are severely indebted in the aftermath of the COVID-19 pandemic, and monetary tightening in high-income countries risks compounding the burden of high debt. ­ • Middle-income countries will need significant resources to scale up reliable low-­carbon energy. But many face the rising cost of borrowing, high up-front infrastructure requirements, and high capital costs, all of which could increase the cost of low-carbon technologies and delay the diffusion of low-carbon energy. Introduction with the retrenchment of globalization (a force Growing past middle-income status has never that has spurred infusion and innovation), dif- been easy—slowing growth (chapter 1) and ficulties in servicing debt obligations, and the structural stasis (chapter 2) dampen growth additional economic and financial costs of cli- prospects. This chapter asks whether growth in mate change and climate action. On the lat- middle-income countries is becoming harder. ­ iddle-income countries will need to build ter, m The answer: it is. resilience to the shocks arising from a changing Two decades into the twenty-first century, the climate, as well as accelerate their energy transi- world is at a historic crossroads. Foreign trade tions. Such challenges will have to be confronted and investment are in danger of becoming con- against the backdrop of a rapidly aging world stricted by geopolitical tensions, and the room for population and therefore waning demographic governments to act is shrinking because of rising dividends (box 3.1). At the same time, middle-­ populism and public debt. As a result, economic income countries need to ensure macroeconomic growth in the remainder of this decade will likely stability since it is essential for high and sustain- be weaker than it was in the last two.1 able rates of growth. Macroeconomic instability For middle-income countries, this prospect will increase the cost of borrowing and keep both means they are growing into shrinking spaces domestic and foreign investors away. 69 Box 3.1 Graying growth Historically, demographic dividends have fostered economic growth in many countries. An increase in the working-age population and a consequent decrease in dependency ratios free up resources to be spent on education, health care, employment, and social protection schemes.a However, the world is aging rapidly and fertility rates are declin- ing rapidly,b which will especially affect middle-income countries projected to face labor crunches in the coming decades. In fact, today’s middle-income countries are aging more rapidly than high-income countries did in the past. Transitioning from an aging society to an aged society took about 61 years in today’s high-income countries and as long as 69 years in the United States and 115 years in France.c By contrast, transitioning from an aging society to an aged society is estimated to take about 26 years for today’s middle-­ income countries (figure B3.1.1). Figure B3.1.1 Today’s middle-income countries are aging more rapidly than high-income countries did in the past Korea, Rep. 18 Bangladesh 20 Türkiye 20 Brazil 22 Malaysia 23 Indonesia 24 India 25 China 25 Japan 25 Middle-income 26 Upper-middle-income 27 Lower-middle-income 31 United Kingdom 45 Argentina 60 High-income 61 United States 69 France 115 1860 1880 1900 1920 1940 1960 1980 2000 2020 2040 2060 Sources: WDR 2024 team using data of World Population Prospects (dashboard), Population Division, Department of Economic and Social Affairs, United Nations, New York, https://population​ .un.org/wpp/; World Bank (2016). Note: The bars indicate beginning and end years for transitioning from 7 percent (aging) to 14  percent (aged) of the population age 65 and older. The number corresponding to each bar shows total number of years to transition from aging to aged. Aging and aged thresholds are based on definitions by the Population Division of the United Nations Department of Economic and Social Affairs (UN DESA). Bars for income groups represent the averages of countries included in the corresponding income group. The dashed vertical line indicates the year 2023. (Box continues next page) 70 | WORLD DEVELOPMENT REPORT 2024 Box 3.1 Graying growth (continued) Working-age populations are already declining in middle-income countries in East Asia.d Latin America will retain a large workforce until the early 2040s, while the Middle East and North Africa and South and Central Asia regions will begin to lose working-age persons by 2045. As a result, middle-income countries can no longer rely on increasing employment as a primary source of output growth. This prospect of “growing old before becoming rich” makes escaping the middle-income trap more challenging and more urgent, includ- ing among the largest middle-income countries such as Brazil, China, Mexico, and Türkiye. The demographic challenge has implications for public policy. Investments will be needed in infrastructure and technologies to support the elderly, in addition to expanding fiscal transfers for older persons. These measures will, in turn, necessitate policies that support savings among the working-age population. Meanwhile, countries will need to invest in upgrading skills of their workforce and ensuring that everyone’s talent is used appropriately and efficiently. Furthermore, automation and other productivity-enhancing technologies to compensate for a shrinking workforce will have to be deployed and have been prioritized by aging countries such as Japan.e The path from imitative to innova- tive growth is the key for middle-income countries to address the needs of their aging populations. Finally, the booming populations in low-income countries and lower-middle-income countries in Sub-Saharan Africa and elsewhere will need to be integrated into graying societies’ labor markets (especially those in middle-income countries).f Canada has suc- cessfully relied on immigration as a strategy to manage its low fertility rates and labor shortages.g World Development Report 2023: Migrants, Refugees, and Societies argues that if managed effectively, migration could unlock further growth for destination countries.h And this migration need not be a brain drain imposed on lower-income countries with high population growth rates. For example, professional links established between expat scientists and those they leave behind may generate knowledge spillovers in their coun- tries of origin.i a. UN DESA (2022). Dependency ratios refer to the number of children (ages 0–14) and older persons (age 65 and over) to the working-age population (ages 15–64). b. Lee and Mason (2006); UN DESA (2022). c. Aging society and aged society thresholds are based on the definition of the Population Division of the United Nations Department of Economic and Social Affairs (UN DESA). According to the definition, the share of people age 65 and older exceeds 7 percent of the population in an aging society, and a country qualifies as an aged society if this share exceeds 14 percent. d. UN DESA (2022). e. Moss (2017). f. Wakeman-Linn et al. (2015). g. Green (2023). h. World Bank (2023b). i. Prato (2023). Shrinking Spaces | 71 Fragmenting geopolitical tensions have affected trade pol- icy,6 and further protectionism could worsen international trade the ­ m iddle-​ d iffusion of knowledge to low- and ­ Episodes of growth acceleration have been accom- income countries. Although there is limited 7 panied by trade integration.2 The Republic of evidence of widespread “friend-shoring” or Korea’s economic metamorphosis was under- realignment of trade along geopolitical lines pinned by a licensing regime to import frontier beyond the United States–China decoupling, technologies, followed by a shift toward promo- this situation could change in the near future if tion of research and development.3 The Korean trade tensions continue to rise.8 Many industries story suggests that catching up and eventually that use technology and innovation intensively leading the pack requires trade liberalization to will face extensive costs and risks if critical buy foreign technology and to sell a country’s global links are reduced because these links are own inventions. More broadly, globalization has at the heart of technology-intensive industries increased incentives to adopt foreign technologies such as information and communication tech- by enhancing international competition.4 nology (ICT) and smartphones.9 Globally, since But the international trade architecture that 2019 the number of harmful trade policies has allowed countries such as Korea to ascend is exceeded the number of helpful trade policies, facing challenges. First, declining potential ­ and the pace of their enactment is increasing: global output threatens trade flows, compound- the number of new harmful trade measures ing the blows dealt by the global financial crisis enacted per year doubled over the last decade and the COVID-19 pandemic.5 Second, rising (figure 3.1). Figure 3.1 Globally, harmful trade policies outnumber helpful trade policies 100,000 Number of new global trade policies 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Helpful trade policies Harmful trade policies Source: WDR 2024 team using the GTA (Global Trade Alert) database, St. Gallen Endowment for Prosperity through Trade, University of St. Gallen, St. Gallen, Switzerland, https://www.globaltradealert.org/data_extraction. Note: Measuring trade policies is fraught with challenges, and the GTA database may overrepresent countries that issue a relatively high quantity of legislative documents or those with greater regulatory transparency. The plotted lines indicate global trade policies that were introduced in the corresponding year. Helpful trade policies include interventions that liberalize on a nondiscriminatory basis or improve the transparency of a relevant policy. Harmful trade policies include interventions that discriminate against foreign commercial interests. See Harmonized System (dashboard), World Customs Organization, Brussels, https://www.wcotradetools.org/en/harmonized-system. 72 | WORLD DEVELOPMENT REPORT 2024 Figure 3.2 Harmful interventions in the global semiconductor trade have skyrocketed since 2019 a. New helpful trade interventions b. New harmful trade interventions in critical technologies per year in critical technologies per year 350 350 Number of new interventions Number of new interventions 300 300 250 250 200 200 150 150 100 100 50 50 0 0 23 20 0 20 1 20 2 20 1 13 20 4 20 6 20 7 20 8 20 9 20 0 20 2 20 5 20 0 20 1 20 9 20 2 20 3 20 4 20 5 20 6 20 7 20 8 20 0 20 1 20 2 23 2 2 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 20 20 20 8541 semiconductors 8542 semiconductors Source: WDR 2024 team using the GTA (Global Trade Alert) database, St. Gallen Endowment for Prosperity through Trade, University of St. Gallen, St. Gallen, Switzerland, https://www.globaltradealert.org/data_extraction. Note: Measuring trade policies is fraught with challenges, and the GTA database may overrepresent countries that issue a relatively high quantity of legislative documents or those with greater regulatory transparency. The bars indicate global trade policies introduced in the corresponding year. Helpful trade policies include interventions that liberalize on a nondiscriminatory basis or improve the transparency of a relevant policy. Harmful trade policies include interventions that discriminate against foreign commercial interests. “8541” and “8542” refer to the four-digit Harmonized System codes for semiconductors. See Harmonized System (dashboard), World Customs Organization, Brussels, https://www.wcotradetools​ .org/en/harmonized-system. Moreover, most high-income and large ­middle-​ unable to function, or it would take decades for income countries are now increasingly resorting those industries to catch up their current levels of to industrial policy and trade-related measures functionality because of their massively modular that may result in restrictions to trade flows.10 For industrial organization.13 Unfortunately, harm- example, some of the largest trading nations have ful interventions to block the trade of critical adopted a “friend-shoring” approach that concen- technologies have more than tripled since 2019 trates their supply chain network among allies (figure 3.2). and friendly countries in response to rising ten- sions with the Russian Federation and China.11 In addition, the costs associated with technolog- Elevated debt ical fragmentation are estimated to be as high as Global debt relative to GDP has remained stub- 5 percent of the gross domestic product (GDP) bornly high in the aftermath of the COVID-19 for many economies.12 Compounding that are pandemic (figure 3.3). Indeed, many middle-­ the high risks associated with disrupting critical income countries are more severely indebted industries and global supply chains. In a deglo- than ever before.14 Postpandemic fiscal deficits balized scenario, some critical industries may be in response to spikes in food and energy costs Shrinking Spaces | 73 Figure 3.3 Most developing economies are more severely indebted than ever 200 Debt as share of GDP (%) 175 COVID-19 onset 150 125 100 75 50 25 0 1950 1960 1970 1980 1990 2000 2010 2019 2020 2021 2022 EMEs’ public debt EMEs’ private debt Source: WDR 2024 team based on IMF (2023b); Kose and Ohnsorge (2024). Note: The figure shows the total debt of emerging market economies (EMEs) in terms of weighted averages broken down into public debt and private debt. GDP = gross domestic product. have kept public debt levels e­ levated. Meanwhile, the green transition, education, the capacity to monetary tightening in high-​ income countries innovate, and infrastructure—are likely to be put risks compounding the burden of high  debt-to- on hold. GDP ratios for middle-income countries. About Encouraging growth in low- and middle-­income three-quarters of countries tightened both fiscal countries under the current circumstances will and monetary policy in 2022.15 As interest rates not be easy. Reducing debt vulnerabilities and have surged, government budgets have been reversing long-term debt trends should be at the squeezed. This squeeze has increased sovereign core of these countries’ agendas. In particular, spreads and borrowing costs for many emerging the governments of middle-income countries and frontier markets. The resulting loss of mar- urgently need to take steps to alleviate debt dis- ket access for some of these countries and the tress in the medium term. More broadly, reforms lack of access to these avenues of relief are likely to improve debt transparency and strengthen to result in a wave of uncoordinated defaults debt management policies and frameworks are in middle-income countries over the medium key to reducing the risks of debt distress. These term. Unfortunately, there is no sign of relief reforms could also foster economic growth if because high-income countries are likely to sus- they are coupled with reforms in labor markets tain high interest rates for the foreseeable future and product markets. In addition, reducing debt (figure 3.4). Many low- and middle-income coun- burdens is crucial to creating the fiscal space for tries now spend more on debt service payments investments in health, education, and infrastruc- than they do on health, education, and infra- ture. Middle-income countries can make quick structure. Because fiscal space is eroding due to progress in transitioning to higher-income status costlier borrowing in low- and middle-income only if they achieve strong and sustained invest- countries, new investment needs—including in ment growth. 74 | WORLD DEVELOPMENT REPORT 2024 Figure 3.4 Debt service payments in emerging markets and middle-income countries may skyrocket as the cost of borrowing soars 9.6 9.7 10 9.1 9.4 8.9 9 8.3 8 7.0 6.6 6.9 7 6.5 6.3 6.5 6.2 6.1 6.5 6.4 6.5 5.7 5.9 Projected 6 5.4 5.5 5.6 Percent 5 4 2.5 2.6 2.6 3 2.2 2.4 2.5 1.8 1.7 1.7 1.8 1.7 1.6 1.6 1.6 1.7 1.7 1.8 1.8 1.8 1.8 1.8 1.8 2 1 0 13 15 16 19 24 25 07 08 09 10 11 12 14 17 18 20 21 22 23 26 27 28 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Interest expense (% of GDP) Interest expense (% of general government revenues) Source: WDR 2024 team using data from Fiscal Monitor (dashboard), International Monetary Fund, Washington, DC, https:// data.imf.org/?sk=4be0c9cb-272a-4667-8892-34b582b21ba6. Note: GDP = gross domestic product. Climate action countries (figure 3.5). Energy intensity (panel a) Fossil fuels—coal, oil, and natural gas—are the in middle-income countries is about double that major contributors to global climate change, in high-income countries, and the carbon inten- accounting for over three-fourths of total green- sity of their energy consumption (panel b) is also house gas emissions and about 90 percent of car- ­relatively inefficient. bon dioxide (CO2) emissions.16 The International At the same time, reliable cheap energy is Energy Agency has announced that achieving ­ critical to supporting industrial activities and net zero emissions by 2050 will require reduc- powering businesses. It is also correlated with ing between 2020 and 2050 the global demand higher living standards. But over 300 million for coal by 90 percent, oil by 75 percent, and ­ people in middle-income countries still lack natural gas by 55 percent.17 Therefore, climate access to ­ electricity. Moreover, blackouts and action—measures to mitigate climate change— brownouts are common in middle-­ income requires changing the way energy is produced countries. As energy demand rises with income, and consumed. expanding access to reliable energy will be Because middle-income countries pres- as critical as ­ transitioning to lower-carbon ently account for about two-thirds of global energy. 18 CO2 emissions, their decarbonization is mate- Today, energy transition technologies con- rial to the world’s ability to meet its climate stitute the largest investment gap. Significant goals.  Today,  both energy intensity and car- resources are needed to scale up reliable, low-­ bon intensity are quite high in middle-income carbon energy. Low-carbon energy technologies Shrinking Spaces | 75 Figure 3.5 In middle-income countries, the energy intensity and carbon intensity of energy consumption are quite high a. Energy intensity b. Carbon intensity of energy consumption 15 0.120 (GtCO2 equivalent/exajoule) (exajoules/trillion US$) of energy consumption 12.1 0.100 0.098 0.086 Carbon intensity Energy intensity 10 9.0 0.080 0.060 0.056 5.1 5 0.040 0.020 0 0 LMICs UMICs HICs LMICs UMICs HICs Source: WDR 2024 team using data from Energy Institute (2023); WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712. Note: Data for low-income countries are not available. GtCO2 = gigatons of carbon dioxide; HICs = high-income countries; LMICs = lower-middle-income countries; UMICs = upper-middle-income countries. require considerable up-front capital despite lower low-carbon energy. In addition, d ­eveloping operating expenses.19 Thus middle-income coun- ­ countries are currently spending more on ­ interest tries need access to low-cost capital to increase payments than on climate i ­nvestments even the affordability of low-carbon energy transitions though they need to more than triple ­ climate because considerable improvement of capital-­ investments to meet the Paris Agreement intensive clean energy assets—such as wind targets.22 power, solar photovoltaic (PV), electric vehicles, Meanwhile, middle-income countries are and hydrogen electrolyzers—is required. The cost more exposed to risks from climate change than of capital is key because the weighted average high-income countries and have less capacity to cost of capital can account for 20–50 percent of adapt to such risks (figure 3.7). They also face the levelized cost of electricity from utility-scale overwhelmingly high costs to finance climate solar PV projects.20 And yet the cost of capital adaptation. In 2030, annual adaptation costs for utility-scale solar power projects in middle-​ are estimated to exceed 1 percent of GDP per income countries is significantly higher than that year for some middle-income countries, while in the United States and Europe (figure 3.6).21 annual global adaptation needs will reach about High up-front infrastructure requirements and percent of global GDP.23 0.25 ­ capital costs could potentially increase the cost The room for governments to act in middle-­ of low-carbon technologies. income countries is shrinking because of rising High borrowing costs make it more chal- debt and interest rates, and their feeble growth lenging for middle-income countries to finance prospects stem from a weakening foreign trade green energy investments and to diffuse and investment outlook. In addition, international 76 | WORLD DEVELOPMENT REPORT 2024 Figure 3.6 In middle-income countries, the weighted average cost of capital for utility-scale solar power projects is substantially higher than the cost in high-income countries 70,000 21 Weighted average cost of capital (%) GDP per capita (constant 2015 US$) 60,000 18 50,000 15 40,000 12 30,000 9 20,000 6 10,000 3 0 0 vak Polaary Kaz Mexi n Ind Rep. . Cze blic Uni Austr m Eth aso tes Sen ep. Gha sh Col Africa Ma stan th sia Sri am ka Bra ia ia Tür ina Ban Ken n ede lgaria ted alia Uga pia en, nda Pak egal Kin ny Yem wa da , Isl ab R a akh co Spaia na kiye Rep nd gla ya zil ted Germace Arg China Fra an Vie India Hun hile in Jap ly am ep o Iran , Ar Algeri ista Ita b lays ch gdo Lan de n rati n g one ent R om io Sta tN u C aF ic u R kin B Sou nF Bur p t sia Slo Egy Uni Rus Low- income Lower-middle-income Upper-middle-income High-income GDP per capita Weighted average cost of capital (right axis) Low-income WACC average Lower-middle-income WACC average Upper-middle-income WACC average High-income WACC average Source: WDR 2024 team using IRENA (2023); WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org/search/dataset/0037712. Note: GDP = gross domestic product; WACC = weighted average cost of capital. pressures and domestic pledges for climate action The constraints facing middle-income coun- will require these countries to allocate ­substantial tries and the extensive and frenetic pace of change investment to adaptation and mitigation. With required for them to transition to high-income many middle-income economies facing high debt status make policy making more challenging and higher borrowing costs, a policy dilemma than at the low- or high-income stage. The chap- arises between achieving ambitious ­ climate ters that follow focus on feasible steps that could actions that pay off in the longer term and avert- aid this transition. ing short-term macroeconomic imbalances. Shrinking Spaces | 77 Figure 3.7 Low- and middle-income countries are exposed to similar levels of risk from climate change, and they have less adaptive capacity 1.0 (higher scores indicate lower capacity) 0.9 0.8 Adaptive capacity index 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Exposure index (higher scores indicate greater risk) Low-income Middle-income High-income Source: WDR 2024 team using data from ND-GAIN Country Index (University of Notre Dame Global Adaptation Index) (dashboard), University of Notre Dame, Notre Dame, IN, https://gain.nd.edu/our-work/country-index/. Note: The solid lines show the linear relationship for each country group. Notes  1. World Bank (2023a). 13. Thun et al. (2022).  2. Hausmann, Pritchett, and Rodrik (2005). 14. Kose and Ohnsorge (2024).  3. Choi and Shim (2024). 15. IMF (2023a).  4. IMF (2018). 16. UNEP (2023).  5. Goldberg and Reed (2023); Kose and Ohnsorge (2024). 17. IEA (2021b).  6. Góes and Bekkers (2022); Goldberg and Reed (2023). 18. Kahn and Lall (2022).  7. Melitz and Redding (2021). 19. Erdogan and Arboleya Sarazola (2023).  8. Pierce and Yu (2023). 20. IEA (2021a).  9. Thun et al. (2022). 21. Steffen (2020); UNEP (2023). 10. Barattieri, Mattoo, and Taglioni (2024). 22. UNCTAD (2024). 11. Gill (2023). 23. Aligishiev, Bellon, and Massetti (2022). 12. Cerdeiro et al. (2021). References Aligishiev, Zamid, Matthieu Bellon, and Emanuele Massetti. Policy Research Working Paper 10806, World Bank, 2022. “Macro-Fiscal Implications of Adaptation to Washington, DC. Climate Change.” IMF Staff Climate Note 2022/002 Cerdeiro, Diego A., Johannes Eugster, Rui C. Mano, Dirk (March), International Monetary Fund, Washington, DC. Muir, and Shanaka J. Peiris. 2021. “Sizing Up the Barattieri, Alessandro Eugenio Maria, Aaditya Mattoo, Effects of Technological Decoupling.” IMF Working and Daria Taglioni. 2024. “Trade Effects of Industrial Paper WP/21/69 (March), International Monetary Fund, Policies: Are Preferential Agreements a Shield?” Washington, DC. 78 | WORLD DEVELOPMENT REPORT 2024 Choi, Jaedo, and Younghun Shim. 2024. “From Adoption Lee, Ronald Demos, and Andrew Mason. 2006. “What Is to Innovation: State-Dependent Technology Policy in the Demographic Dividend?” Finance and Development Developing Countries.” STEG Working Paper WP091 43 (3): 16–17. (March), Structural Transformation and Economic Melitz, Marc J., and Stephen J. Redding. 2021. “Trade Growth, Centre for Economic Policy Research, London. and Innovation.” NBER Working Paper 28945 (June), Energy Institute. 2023. “Statistical Review of World Energy National Bureau of Economic Research, Cambridge, 2023.” June, Energy Institute, London. MA. Erdogan, Musa, and Lucila Arboleya Sarazola. 2023. “Cost Moss, Daniel. 2017. “Aging Japan Wants Automation, Not of Capital Survey Shows Investments in Solar PV Can Be Immigration.” Bloomberg: Opinion, August 22, 2017. Less Risky than Gas Power in Emerging and Developing https://www.bloomberg ​.com ​/ view/articles/2017​- 08​ Economies, though Values Remain High.” Commentary, -22​/aging-japan​-wants-automation​- not​- immigration. November 30, 2023. https://www.iea.org​ /commentaries​ Pierce, Justin R., and David Yu. 2023. “Assessing the Extent of /cost-of-capital-survey-shows​ - investments​ - in-solar​ Trade Fragmentation.” FEDS Notes. Board of Governors -pv-can-be-less-risky-than​- gas-power​- in-emerging-and​ of the Federal Reserve System, Washington,  DC. -developing-economies​-though​-values-remain-high. https://www.federalreserve.gov/econres/notes/feds​ Gill, Indermit Singh. 2023. “The New Threat to -notes/assessing-the-extent-of-trade-fragmentation​ Prosperity Everywhere.” Paradigm Shifts (blog), -20231103.html. March 13, 2023. https://www.project-syndicate.org​ Prato, Marta. 2023. “The Global Race for Talent: Brain Drain, /magazine ​ / deglobalization - new-fiscal - monetar y​ Knowledge Transfer, and Economic Growth.” Working -policy ​ - norms-bad-for​­- economic- development-by​ Paper, Ettore Bocconi Department of Economics, -indermit-gill-2023-03. Bocconi University, Milan. Góes, Carlos, and Eddy Bekkers. 2022. “The Impact Steffen, Bjarne. 2020. “Estimating the Cost of Capital for of Geopolitical Conflicts on Trade, Growth, and Renewable Energy Projects.” Energy Economics 88 Innovation.” WTO Staff Working Paper ERSD-2022-09 (May): 104783. (July 4), Economic Research and Statistics Division, Thun, Eric, Daria Taglioni, Timothy J. Sturgeon, and Mark World Trade Organization, Geneva. P. Dallas. 2022. “Massive Modularity: Understanding Goldberg, Pinelopi Koujianou, and Tristan Reed. 2023. “Is Industry Organization in the Digital Age: The Case of the Global Economy Deglobalizing? And If So, Why? And Mobile Phone Handsets.” Policy Research Working What Is Next?” Brookings Papers on Economic Activity Paper 10164, World Bank, Washington, DC. BPEA 2023 (Spring): 347–96. UN DESA (United Nations Department of Economic and Green, Mark A. 2023. “Canada: Ageing Workforce, Boosting Social Affairs). 2022. “World Population Prospects Immigration.” Stubborn Things (blog), April 18, 2023. 2022: Summary of Results.” Document UN DESA/ https://www.wilsoncenter.org/blog - post/canada​ POP/2021/TR/NO. 3, United Nations, New York. https:// -ageing-workforce-boosting-immigration. www.un.org/development/desa/pd/sites/www.un.org​ Hausmann, Ricardo, Lant H. Pritchett, and Dani Rodrik. .development.desa.pd/files/wpp2022_summary_of​ 2005. “Growth Accelerations.” Journal of Economic _results.pdf. Growth 10 (4): 303–29. UNCTAD (UN Trade and Development). 2024. A World of IEA (International Energy Agency). 2021a. “The Cost of Debt: A Growing Burden to Global Prosperity. New York: Capital in Clean Energy Transitions.” Report, December United Nations. 17, IEA, Paris. UNEP (United Nations Environment Programme). 2023. IEA (International Energy Agency). 2021b. Net Zero by Broken Record: Temperatures Hit New Highs, Yet World 2050: A Roadmap for the Global Energy Sector. 4th rev. Fails to Cut Emissions (Again). Emissions Gap Report (October). Paris: IEA. 2023. Nairobi, Kenya: UNEP. IMF (International Monetary Fund). 2018. World Economic Wakeman-Linn, John, Rahul Anand, Paulo Drummond, Outlook, April 2018: Cyclical Upswing, Structural Change. Richard Erlebach, Francisco Roch, Vimal Thakoor, Washington, DC: IMF. and Juan Treviño. 2015. “How Can Africa Harness IMF (International Monetary Fund). 2023a. Fiscal Monitor: On the Demographic Dividend?” In Sub-Saharan Africa: the Path to Policy Normalization. April. Washington, DC: IMF. Navigating Headwinds, 25–45. With research assis- IMF (International Monetary Fund). 2023b. “2023 Global tance of Idan Elmelech, Cleary Haines, and George Debt Monitor.” September, Fiscal Affairs Department, Rooney. Regional Economic Outlook, April. World IMF, Washington, DC. Economic and Financial Surveys Series. Washington, IRENA (International Renewable Energy Agency). 2023. DC: International Monetary Fund. Renewable Power Generation Costs in 2022. Abu Dhabi, World Bank. 2016. Live Long and Prosper: Aging in East Asia United Arab Emirates: IRENA. and Pacific. World Bank East Asia and Pacific Regional Kahn, Matthew E., and Somik Vinay Lall. 2022. “Will the Report. Washington, DC: World Bank. Developing World’s Growing Middle Class Support Low World Bank. 2023a. Global Economic Prospects, January Carbon Policies?” NBER Working Paper 30238 (July), 2023. Washington, DC: World Bank. National Bureau of Economic Research, Cambridge, MA. World Bank. 2023b. World Development Report 2023: Kose, M. Ayhan, and Franziska Ohnsorge, eds. 2024. Falling Migrants, Refugees, and Societies. Washington, DC: Long-Term Growth Prospects: Trends, Expectations, and World Bank. Policies. Washington, DC: World Bank. Shrinking Spaces | 79 Part 2 Creative Destruction P art 1 of this Report highlights the need weakened, and the forces of destruction man- for middle-income countries to undergo aged. Policy makers thus need to understand not one but two transitions. In the first, and account for these countries transition from a “1i” strategy for accelerating investment to a “2i” strategy focus- • Creation. Incumbents and entrants com- ing on both investment and infusion in which a pete by creating economic value (see country brings technologies from abroad and chapter 4). The forces of creation are diffuses them domestically. Once a country has strengthened by fostering openness in successfully undertaken this effort, it can switch the exchange of goods, services, and ideas. to a “3i” strategy in which it pays greater atten- The same forces are weakened by excessive tion to innovation—beginning to not just borrow government regulation and macroeco- ideas from the global frontiers of technology but nomic uncertainty. also to push the frontiers outward. Part 2 of this • Preservation. Economic, social, and polit- Report provides a diagnostic framework based on ical structures are susceptible to being advances in Schumpeterian economic theory to captured by powerful incumbents (see examine what hinders and what helps countries chapter 5). The forces of preservation must engineer shifts in growth strategies (box  P2.1). be reined in and prevented from smoth- Economic analysis using a Schumpeterian cre- ering creation. Where institutions are ative destruction framework provides more weak, incumbents often capture them to reliable diagnoses because it draws on the micro- preserve the status quo. economic foundations of growth. But like most • Destruction. The forces of preservation are new technologies, it requires greater expertise weakened only when crises arise (see chap- in assessing the evidence, higher-quality data for ter 6). The vital forces of destruction place informing the analysis, and greater care in inter- tremendous pressure on governments to preting the results. act because a growing economy needs to Underpinning the Schumpeterian frame- shed outdated arrangements (in capital, work are three sets of forces: creation, preserva- labor, and energy markets) as much as tion, and destruction (figure P2.1). These unruly it needs to invent new ones. Where the forces have to be domesticated. To advance forces of destruction are constrained by technological progress, the forces of creation misguided policies, creation struggles and need to be amplified, the forces of preservation advances slowly.  | 81 Box P2.1 Joseph Schumpeter and creative destruction The concept of creative destruction was featured in the early writings of many influen- tial political economists such as Werner Sombart (1863–1941) and Friedrich von Wieser (1851–1926).a However, it became the trademark of Joseph A. Schumpeter (1883–1950), an Austrian economist, who widely popularized the term and made it accessible through his book Capitalism, Socialism and Democracy. Schumpeter’s view was that creative talents with vision would create new products and technologies to change the way we live. According to Schumpeter, “Railroads did not emerge because some consumers took the initiative in displaying an effective demand for their service in preference to the services of mail coaches. Nor did consumers exhibit the wish to have electric lamps or rayon stock- ings, or to travel by motorcar or airplane, or to listen to radios, or to chew gum.”b What leads to economic growth is the “change” that is ultimately generated by an entrepre- neur’s desire and leadership. Without change, a society is doomed. Schumpeterian growth theory provides sharper instruments for diagnosing the eco- nomic health of economies because it features • Heterogeneous agents. The theory is premised on differentiation among firms (such as small and large; new and old) and workers (such as unskilled and skilled; rural and urban). It can usefully be extended to distinguish between energy sources (such as renewable and carbon-intensive; reliable and risky). • Continuous dynamics. The theory recognizes the importance of both creation (start- ups, investment, new skills, and innovation) and destruction (firm closures, skill obso- lescence, and stranded assets) in the process of structural change and economic development. • Institutional inertia. The theory provides insights into the forces of preservation of societal arrangements and status quo biases in industrial organization and public policy. • Constructive crises. Related to inertia, the theory recognizes that economic and envi- ronmental crises present opportunities for improvements; with the right policies, creation can emerge from destruction. a. Campagnolo and Vivel (2012); Reinert and Reinert (2006). b. Schumpeter (1942, 73). The forces of creation, preservation, and to illustrate the forces of preservation, and chap- destruction appear in each transition of middle-​ ter 6 on energy to highlight how new technolo- income economic growth—first, in adding infusion gies and climate crises generate both the need and to investment, and, second, in adding innovation to the potential to destroy current arrangements in the mix. energy markets. Firms, talent, and energy are Chapter 4 focuses on enterprise dynamics to closely connected, and their interactions shape illustrate the forces of creation, chapter 5 on talent the forces of creative destruction. 82 | WORLD DEVELOPMENT REPORT 2024 Figure P2.1 Rebalancing the forces of creation, preservation, and destruction to advance infusion and innovation Creation Creation Preservation Destruction Preservation Destruction Source: WDR 2024 team. References Campagnolo, Gilles, and Christel Vivel. 2012. “Before Reinert, Hugo, and Erik S. Reinert. 2006. “Creative Schumpeter: Forerunners of the Theory of the Destruction in Economics: Nietzsche, Sombart, Entrepreneur in 1900s German Political Economy— Schumpeter.” In Friedrich Nietzsche 1844–1900: Werner Sombart, Friedrich von Wieser.” European Economy and Society, edited by Jürgen G. Backhaus Journal of the History of Economic Thought 19 (6): and Wolfgang Drechsler, 55–85. New York: Springer. 908–43. Schumpeter, Joseph Alois. 1942. Capitalism, Socialism and Democracy. New York: Harper and Brothers. CREATIVE DESTRUC TION  | 83 4 Creation Key messages • In an economy, both entrants and incumbents can add value. Entrants bring change in the form of enterprises with new products or production processes, workers with new skills and ideas, or energy sources such as renewables that embody new technologies. Incumbents bring scale—and can compete with entrants in the market to jointly expand a country’s technological capabilities—moving the country closer to the global technology frontier. • In most middle-income countries, too few entrants disrupt because the majority of entrepreneurs start businesses driven by necessity, not business opportunity. And too few ­ incumbents infuse global technologies and know-how. Instead, they abuse their dominance to block entry, resist innovation, and capture political institutions. • Government policies in middle-income countries are often based on the size of firms rather than their ability to create value. As a result, policies tend to favor small firms and reduce the incentives of firms to grow. • Governments will need more reliable diagnostics to implement effective policies, ­ including microlevel data on firms, a more comprehensive approach to examining business d ­ ynamism, and greater analytical expertise. Creation: The protagonist Services (TCS). Tapping into talent from its of economic growth, where various businesses and India’s elite technical incumbents create value institutions, TCS became the first Indian exporter of computing services to the United States. As part alongside entrants of a large conglomerate, TCS could have stifled In the 1960s, the Tata Business Group, a major creativity and misused its import allowance. In conglomerate based in India, needed computers practice, however, TCS opened the world’s eyes for its various businesses, but they were difficult to to the high-quality, l­ower-cost talent that India’s purchase and import because the government had software development had to offer. Many entrants enacted import restrictions to protect state-owned took advantage of this opportunity. Among enterprises (SOEs). To allow foreign hardware them, Infosys, a company founded by former IBM to be imported, the government required Indian engineers and financed by their own savings, companies to commit to exporting products or dramatically increasing the value it created, has services. This challenge inspired the conglomerate become a global leader in information technology to create a new business, Tata  Consultancy consulting (figure 4.1). | 85 Figure 4.1 Both entrants and incumbents create value and reinforce one another’s growth through competition in India’s computing services industry 300,000 Number of employees (log scale) 100,000 30,000 10,000 3,000 2000 2005 2010 2015 Industry average Infosys Tata Consultancy Services Source: Worldscope (database), Baker Library, Bloomberg Center, Harvard Business School, Boston, https://www.library​ .hbs.edu/find/databases/worldscope. Note: The figure reports the number of employees in Infosys and Tata Consultancy Services and the average across firms listed in India’s stock exchange in the same two-digit industry (Mean-Industry-India). “Two-digit” refers to NAICS (North American Industry Classification System) (database), US Census Bureau, Suitland, MD, https://www.census.gov/naics/. The gales of creative destruction carry new ideas, figure  4.2, panel a) in the form of enter- (­ products, processes, and practices to the shores prises with new products or production of middle-income economies. The term creative processes, workers with new skills and destruction was popularized by Austrian econo- ideas, or energy sources, such as renew- mist Joseph A. Schumpeter in his book Capitalism, ables, that embody new technologies. In the Socialism and Democracy (1942). Schumpeter was process, they expand a country’s technol- concerned by the growing concentration of wealth ogy frontier. Interactions among firms, in market economies where dominant incum- workers, and energy/technology underpin bents colluded to preserve the status quo. He creative destruction (see box P2.1). In the argued that economic and social improvements current environment, Schumpeter’s view is arise from new products and technologies intro- reflected in the competition between high- duced by entrepreneurs with talent and vision. and low-carbon energy. Fossil fuels have The forces of creation, preservation, and been the dominant incumbent t ­echnology destruction interact to shape the growth of for over 300 years. Technical progress, com- nations through technological advancement. plementary public and private ­ investment, This chapter examines how the forces of creation individual preferences, and powerful play out in middle-income countries to create ­ interest groups interact to preserve a value. Chapters 5 and 6 examine the forces of high-carbon status quo. preservation and destruction. To shed light on the Schumpeter’s ideas inspired the for- potency of creation in middle-income countries, mal models of creative destruction devel- this chapter examines three questions: oped by Philippe Aghion and Peter Howitt, 1. Who creates value? Schumpeter highlighted who make a strong case that the entirety that entrants bring change, and incum- of growth stems from entrants and that dominance bents collude to preserve their ­ incumbents assume the primary task of 86 | WORLD DEVELOPMENT REPORT 2024 producing goods until they are replaced by more workers, and displace enterprises these new players (figure 4.2, panel b). Their that fail to generate additional value. This approach is particularly useful in high- approach ensures that capital and labor lighting the value added by investing in the are not held captive by unproductive firms talents of women and enabling women to but are utilized more efficiently by grow- seek the most rewarding opportunities. ing enterprises. In fact, 50–70 percent of But must a new entrant always drive productivity growth stems from success- creation and displace incumbents? ­ ful resource reallocation among firms— Although Schumpeter and his immediate whether incumbents or entrants.1 followers celebrated entrants and down- 2. Are entrants and incumbents creating played the role of incumbents, Tata Con- value in middle-income countries? Too few sultancy Services’ journey reveals that entrants disrupt, and too few incumbents incumbents can also create value (figure infuse global technologies and know-how. 4.2, panel c). Incumbents bring scale, which The majority of entrepreneurs start busi- enables them to invest in upgrading prod- nesses driven by necessity, not business ucts, hire and reward skilled workers, and opportunity. Incumbents that create value effectively use large amounts of capital. are unable to expand, limiting their poten- They can compete with entrants in the tial for infusing technologies. Dominant market jointly to expand a country’s tech- incumbents can be vanguards for infusing nological capabilities, advancing the coun- global technologies and exporting glob- try closer to the global frontier. ally. But, too often, they abuse their domi- The symbiotic view is most applicable nance to capture political institutions. to the modern enterprise. Regardless of whether they are entrants or established 3. Are middle-income country governments companies, firms that create value must strengthening the forces of creation? Gov- be able to expand their operations, hire ernments often use outdated rules of Figure 4.2 The interactions between entrants and incumbents set the pace of creative destruction a. Schumpeter’s view: b. Aghion and Howitt’s view: c. Akcigit and Kerr’s view: Dominant incumbents can Entrants create value and Incumbents and collude to block competition displace incumbents entrants add value TECHNOLOGY FRONTIER TECHNOLOGY FRONTIER TECHNOLOGY FRONTIER Entrant Incumbents Entrant Incumbent Incumbent Entrant Incumbent Economic progress Economic progress Economic progress Schumpeter’s view: incumbents Aghion and Howitt’s view: entrants Akcigit and Kerr’s view: incumbents can collude create value and displace incumbents and entrants add value Source: WDR 2024 team based on Schumpeter (1942); Aghion and Howitt (1992); Akcigit and Kerr (2018). Creation | 87 thumb to regulate and lack the informa- Figure 4.3 Entrants drive growth: Insights tion and capabilities needed to identify from Aghion and Howitt’s seminal paper on key constraints impeding the growth of creative destruction firms. Over the last three decades, econo- mists have developed modern techniques to provide calibrated assessments of what is enabling or suppressing the growth of IN firms. These techniques include better Productivity or quality of production Firm 5 measurement of business dynamics that includes market concentration, markups, productivity dispersion, firm entry and Firm 4 exit rates, and job reallocation. OUT This chapter examines the forces of creation Firm 3 from the perspective of firms with an emphasis on the process in middle-income countries. A broader treatment that includes talent and energy Firm 2 sources appears in part 3 of this Report. Creative destruction: Three Firm 1 decades of increasingly refined analysis Source: WDR 2024 team. A first look: Entrants create value Note: The figure depicts a quality ladder based on the Aghion-Howitt model of creative destruction. IN represents When Canadian economist Peter Howitt vis- a new firm entering the market and OUT represents a firm ited the Massachusetts Institute of Technology exiting the market. Each innovation adds a new rung to the (MIT) in the summer of 1987, he crossed paths ladder and elevates the prospects for all future producers. This hierarchy suggests that there may be underinvestment with Philippe Aghion, an assistant professor at in innovation because innovators often fail to consider MIT. Their collaboration gave birth to the for- the height of each rung, which determines the long-term mal theory of creative destruction, and their benefits of their creative endeavors for society. On the other hand, there may be an overinvestment in entry. paper “A Model of Growth through Creative Entrants primarily focus on displacing incumbents, so more Destruction,” published in Econometrica in entry is associated with shorter durations of monopoly for 1992, became one of the most influential papers incumbents, discouraging innovations by future entrants. in economics.2 Their main insight was that economies expand organically through innova- determines the long-term benefits of their cre- tion driven by new entrants. Aghion and Howitt ative endeavors for society. On the other hand, attributed innovation and growth exclusively to there may be an overinvestment in entry. Entrants newcomers (figure 4.3). primarily focus on displacing incumbents, and Such models of creative destruction, that is, so more entry is associated with shorter dura- quality ladder models, depict a hierarchical struc- tions of monopoly for incumbents, discouraging ture with multiple rungs in which each innova- innovations by future entrants. This motive— tion adds a new rung to the ladder and elevates called “business ­stealing”—­carries the risk that the prospects for all future producers—a phe- a society could end up overinvesting in research nomenon known as intertemporal spillovers. This and development (R&D) and creating too much hierarchy suggests that there may be underin- entry. Whether equilibrium results in an excess or vestment in innovation because innovators often deficiency of R&D depends on the varying mag- fail to consider the height of each rung, which nitudes of spillovers and business stealing, which 88 | WORLD DEVELOPMENT REPORT 2024 can differ significantly from one sector to another their existing products or services. In essence, this and from one country to another. This  novel framework indicated that, within each narrow framework has opened the door to a more nuanced sector, all innovations originate from external and realistic discourse on innovation policy. sources and never from incumbents within the production unit. In 2018, Ufuk Akcigit and William Kerr fur- A recent view: Incumbents also create ther advanced the framework to enable a more value accurate description of innovation and growth.4 Schumpeter’s ideas, as well as the framework In their framework, incumbents not only expand presented in Aghion and Howitt’s seminal 1992 into new markets through external R&D, but also paper, focused predominantly on entrants and enhance their existing products through internal small businesses, downplaying the role of incum- R&D without directly displacing other producers. bents (often larger firms) in fostering infusion and In fact, upon venturing into new markets, many innovation. In fact, the theory of creative destruc- established incumbents allocate a significant por- tion highlights an important tension between the tion of their resources to refining and elevating incumbent currently in production and the new their current products and technologies. The entrant that endeavors to replace the incumbent significant insight is that creative destruction with a superior product or technology. can be instigated not only by entrants but also In 2004, Tor Klette and Sam Kortum refined by incumbents. Moreover, not every innovation the framework by reimagining each firm as a con- introduced by incumbents inevitably results in glomerate of production units in different sectors.3 creative destruction through the dismantling of Within this framework, firms are able to expand “ecosystem” (figure 4.4). others in their ­ their products or services into new markets and The United States and Germany are contrast- compete with other incumbents, but not scale up ing cases of creative destruction. In Germany, Figure 4.4 Entrants and incumbents drive growth through turnover and upgrading: Insights from Akcigit and Kerr’s refined approach to creative destruction New entrant Level of product quality (contribution of the Existing quality latest innovation) Internal R&D External R&D Product lines Firm 1 Firm 2 Source: Akcigit and Kerr 2018. Note: R&D = research and development. Creation | 89 companies established before 1950 constitute a manufacturing enterprise in the United States. substantial 70 percent of publicly traded compa- Characterized until the early 1900s by nies—a figure that diverges markedly from that ­ single-unit firms with one or a small number for the United States, where the equivalent pro- of owners, the modern multiunit enterprise, portion is approximately 50 percent. Furthermore, administered by salaried middle and top man- although the annual rate of new firm creation agers overseeing complex layers of production, is about 8 percent in the United States, it is became the typical business model by World ­ 3 percent in Germany. Despite these pronounced War I. Recordkeeping, accounting, and inven- disparities in terms of incumbent longevity and tory control were all tasks required by the the pace of firm entry, there is a striking symme- ­ “visible hand” managing a firm.5 try in the number of patents per capita between This transformation of business proved to be the two economies. This convergence in patent fertile ground for the emergence of what became output underscores a noteworthy distinction: a powerhouse of corporate innovation, the innovation in Germany is primarily championed Computing-Tabulating-Recording-Company, by long-standing incumbents, whereas in the later IBM. In 1939, when William Hewlett United States it is shouldered to a greater extent and David Packard created their company in a by emerging young entrants (figure 4.5). rented garage in Palo Alto, California, IBM was already a well-established company with more than 10,000 employees. Herman Hollerith, A more nuanced view of incumbents: founder of one of the companies consolidated Incumbents’ power can drive or into IBM in 1911, could not anticipate that his suppress creation invention, the electromechanical tabulation of The turn of the twentieth century wit- punched card data, would become the backbone nessed a rapid transformation in the typical of a computer hardware and software company. Figure 4.5 Contrasting examples of innovation: Growth is driven by entrants in the United States and by incumbents in Germany a. Firm entry rate b. Patents per capita 12 60 Patents per 100,000 inhabitants 10 50 Firm entry rate (%) 8 40 6 30 4 20 2 10 0 0 80 85 90 95 00 05 10 00 05 10 15 19 19 19 19 20 20 20 20 20 20 20 Germany United States Source: Akcigit et al. 2024. 90 | WORLD DEVELOPMENT REPORT 2024 Similarly, Hewlett and Packard did not envi- Figure 4.6 Entrants and incumbents can sion that their frequency oscillator would pave reinforce one another’s growth: The case of the way for their small firm to become a leading the US business services industry manufacturer of personal computers that would compete with a giant like IBM. Yet both compa- 400,000 Number of employees (log scale) nies found themselves competing against each 100,000 other for decades. It was only later that many others, such as Dell and Microsoft, entered the 10,000 market to compete. On the surface, the forces of creation are 1,000 more apparent than the forces of destruction in the Schumpeterian view of forces of innovation. 100 In the US business services industry, as HP, Dell, Microsoft, and Oracle entered various branches of 10 the hardware, software, and computing ­ services markets, the pace of growth of IBM (as the 0 ­ incumbent) slowed—but not to the level of shut- 1910 1930 1950 1970 1990 2010 ting down the company (figure 4.6). Destruction Dell HP IBM still occurs within firms as they replace older Microsoft Oracle products and tap into newer markets. But dominant firms can also use their Source: Luttmer 2011. incumbency advantage to block entry and resist Note: The figure reports the number of employees for a selected group of US business services firms: Dell, Hewlett innovation. They can collude, with the most Packard (HP), IBM, Microsoft, and Oracle. significant players coordinating price-setting to outbid smaller competitors, while avoiding price wars against one another. For example, in In middle-income countries, 2013 the US Department of Justice uncovered a too few small entrants decade-long cartel of Japanese auto part makers disrupt, and too few large conspiring against the United States car indus- incumbents innovate or try (figure 4.7). Mitsuba Corporation, a promi- nent cartel member, increased its sales relative infuse global technologies to the average performance of publicly listed Although large incumbent firms can spur as well companies in the same industry. However, this as block the forces of creation, large incumbents growth was not fueled by innovation but rather do not capture the broader economic landscape by suppressing competition, partly reflected in in middle-income countries. Most establishments the decline in the number of public firms in the are microenterprises that neither grow nor exit, same industry. and therefore they do not contribute to growth When incumbents respond to competition by in productivity. Furthermore, many medium and upgrading technologies and business practices large firms that have the potential to leverage the either to defend existing market positions or to incumbency advantage to lead infusion and inno- tap into newer ones, they drive progress, as in the vation often achieve scale by benefiting from dis- United States (box 4.1). When incumbents resist tortions rather than from merit (see next section competition by enacting barriers to entry, it not for that discussion). only stalls infusion and innovation but also can “Flat and stay” and “up or out” describe entirely prevent them. entrepreneurial dynamics. One reason for the Creation | 91 Figure 4.7 A cartelized industry suppresses innovation and dynamism: Evidence from the Japanese auto parts sector a. Log sales b. R&D intensity 1.0 45 Absolute difference from 1990 35 Absolute difference from 1990 (%) 0.5 25 15 0 5 0 –5 –0.5 –15 90 95 00 05 10 15 90 95 00 05 10 15 19 19 20 20 20 20 19 19 20 20 20 20 Mitsuba Industry mean c. Number of public firms in industry 700 650 Number of firms 600 550 500 450 90 95 00 05 10 15 19 19 20 20 20 20 Source: WDR 2024 team based on Worldscope (database), Baker Library, Bloomberg Center, Harvard Business School, Boston, https://www.library.hbs.edu/find/databases/worldscope. Note: The figure reports a series of sales (panel a) and shares of research and development (R&D) expenditures relative to total investment (panel b) for Mitsuba Corporation and the average across listed firms in Japan’s stock exchange within the same two-digit industry. Panel c shows the number of listed companies in Japan’s stock exchange in the same two- digit industry. Sales are presented in logarithms and reported as a difference from their value in 1990, the first year in the sample. The share of R&D expenditures to total investment, R&D intensity, is also reported as a difference from the share in 1990. “Two-digit” refers to NAICS (North American Industry Classification System) (database), US Census Bureau, Suitland, MD, https://www.census.gov/naics/. 92 | WORLD DEVELOPMENT REPORT 2024 Box 4.1 Vibrant corporate R&D, connected places, mobile people, and successful markets for patents: How the United States nurtured an ­ ­innovation ecosystem The US economy has fostered innovation for centuries, inspiring numerous inventors to generate brilliant ideas that have, subsequently, been harnessed by businesses to create consumer products or production technologies that have had a widespread impact. Four dimensions have been important: • Who “creates” has shifted from small, independent inventors to corporate research and development (R&D) and innovation. Over the last 150 years, there has been a remark- able transition from the realm of “garage inventors” to the corporate R&D facilities of modern corporations (figure B4.1.1). The transition has coincided with the growing sophistication of the US economy. • Connected places are more inventive. Connectivity enhances the market size for inno- vation and the flow of knowledge. Inventive activity in the early nineteenth century accelerated in US locations that were near navigable waterways and had developed railroads.a • The migration of people is critical to innovation. Technology sectors in the United States with a higher concentration of immigrant inventors between 1880 and 1940 experienced accelerated growth from 1940 to 2000.b Furthermore, immigrant inven- tors exhibited greater productivity throughout their careers than their native-born counterparts. Unfortunately, despite their heightened productivity, immigrant inno- vators found that their earnings were much lower. Figure B4.1.1 The number of patents filed by corporations with the US Patent and Trademark Office has skyrocketed since 1880 80 70 Share of patents filed by corporations (%) 60 50 40 30 20 10 0 1880 19 0 00 19 0 00 20 19 0 19 0 19 0 19 0 19 0 19 0 20 0 9 1 3 4 5 6 7 8 9 18 19 19 Source: Akcigit, Grigsby, and Nicholas 2017b. (Box continues next page) Creation | 93 Box 4.1 Vibrant corporate R&D, connected places, mobile people, and successful markets for patents: How the United States nurtured an ­ ­innovation ecosystem (continued) • The secondary market for innovations (such as patent resale or licensing) is as important as innovation itself. This secondary market has played an outsize role in the United States. Between 1870 and 1910, inventors increasingly sought the services of the more than 500 specialized registered patent agents nationwide to navigate the intri- cacies of patent-related transactions. For example, Edward Van Winkle, a mechanical engineer who pursued a law degree via correspondence courses, ­ possessed the ideal skill set to provide clients on both sides of the market with expert advice on the legal and technical aspects of inventions. He established a network of businessmen, inven- tors, and fellow lawyers to broker patent deals between buyers and sellers.c a. For waterways, see Sokoloff (1988); for railroads, see Donaldson and Hornbeck (2016) and Perlman (2015). b. Akcigit, Grigsby, and Nicholas (2017a). c. For historical markets for technologies, see Lamoreaux and Sokoloff (2002). For more contemporary markets, see Akcigit, Celik, and Greenwood (2016). lack of business dynamism in the economies of at most four workers declines by 60  percent by ­ middle-income countries can be traced to the the age of 25 in the United States, the decrease is typical life cycle of a firm’s growth (figure 4.8, only about 10 percent in India (figure 4.8, panel b). panel a). In the United States, a celebrated feature Consequently, the Indian economy lacks the of the economy is the selectivity of its markets. mechanism for effective selection among firms, Start-ups and young businesses are confronted hindering the reallocation of resources to more by pressure to move up or out. The average productive users. young US firm grows by a factor of 7 by age 40, The life cycle dynamics of firms exhibits a assuming it is still in business. Failing entrepre- similar flatness when expanding the sample to neurs either move up to newer ventures, or out include a few East European and East Asian econ- to find wage-earning opportunities by means of omies (figure 4.8, panel a). Serbia and Malaysia the rising labor demand in flourishing firms. By seem to have stabilized at a notably lower level contrast, in middle-­ income countries flat and of life cycle growth than the United States, com- stay is a more accurate description of entrepre- parable with the dynamics of the other middle-­ neurial dynamics. The growth rates of firms in income economies in the figure. Viet Nam seems India, Mexico, and Peru are far lower than those to be at a critical point at which it either contin- of firms in the United States, with firms expand- ues its promising trend and accompanies a simi- (figure 4.8, panel a). ing by less than a factor of 3 ­ lar degree of dynamism as observed in the United Conversely, when firms with growth potential States, or it succumbs to the lackluster perfor- lack dynamism, they fall short of displacing mance characteristic of the middle-income trap. unproductive small firms from the market. This The weakness of firm dynamics in middle-­ absence of creative destruction results in a stark income countries translates into stark differences contrast: although the share of small firms with in the distribution of firms across sizes (figure 4.9). 94 | WORLD DEVELOPMENT REPORT 2024 Figure 4.8 In middle-income countries, the growth rate of firms across their life cycles is much lower than in the United States a. Firm size over time b. Share of small firms over time, Employment relative to firms <5 years of age India and the United States 8 100 7 90 Share of small firms (%) 80 6 70 5 60 4 50 3 40 30 2 20 1 10 0 0 <5 10−14 20−24 30−34 ≥40 0–5 6–10 11–15 16–20 21–25 26+ Age of firms (years) Age of firms (years) India Malaysia Mexico Peru India United States Serbia United States Viet Nam Sources: Panel a: India, Mexico, and the United States: Hsieh and Klenow 2014; Peru: World Bank 2015; Malaysia: de Nicola and Timmis, forthcoming; Serbia and Viet Nam: World Bank 2024b. Panel b: Akcigit, Alp, and Peters 2021. Note: Panel a illustrates the average employment across a cohort of firms of different ages in the cross-section of firms. The number of employees serves as a proxy for firm size. The vertical axis reports the average employment of each cohort relative to the average employment across firms under five years of age. In the United States, microenterprises are the middle-income countries reflect a misallocation dominant form of production in terms of share in jobs whereby distortions in firm growth reduce of firms, but firms are distributed more uniformly wages and wage-earning opportunities in the for- across the size spectrum. Start-ups enter small and mal sector. Such distortions misallocate resources proceed to grow up or get out, which explains the from their best use and discourage incentives to prominence of medium and large firms. In India, adopt and innovate technologies. This vicious Mexico, and Peru, the flat and stay behavior trans- cycle, in turn, reduces higher-paying formal wage lates into a market in which more than 80 percent jobs and further fuels informality. of firms employ fewer than five workers each. Middle-income countries are not benefiting Informal establishments account for most of from technological disruption by their smaller the microenterprises in the firm size distribu- firms. This outcome challenges the commonly tion in India, Mexico, and many other countries.6 perceived notion that micro- and small enterprises Although informal entrepreneurs may have had are the drivers of job creation and economic dyna- greater success at growing and becoming formal mism. That misconception is compounded by the if the costs of business entry were lower, attempts fact that start-ups are primarily small. However, to reduce the costs of business entry for firms in age, not size, should be the measure used to assess developing countries have had modest results at dynamism. In fact, in the United States small firms best.7 Instead, the large informal sectors in many are the net destroyers of jobs, except for start-ups Creation | 95 Figure 4.9 Microenterprises dominate firm of creative destruction. Governments often use size distributions in India, Mexico, and Peru outdated rules of thumb to identify who creates value, and they lack the information and capa- bilities to identify key constraints to the growth Share of establishments (%) 80 of productive firms. When size is used to target larger firms to advance social objectives, the incen- tives to grow are dampened. As a consequence, an 50 economy produces far below its potential. Adopting and developing technologies are costly endeavors, and so entrepreneurs will incur 25 these costs only if rewarded with a sufficient rate 15 of return. When governments effectively lower 0 the rate of return to adopting and developing technologies, they undermine the whole econo- <5 9 9 9 0+ 9 5− −1 −9 49 my’s production possibilities relative to the poten- 50 0− 10 20 tial frontier, locally and globally. 10 Number of employees Screening firms by size—not by their ability India Mexico Peru United States to create value—often ends up being misguided. Talented entrepreneurs are not able to attract Sources: India: Sixth Economic Census 2013–14 more capital and labor, jobs shift from high-­ (dashboard), National Data Archive, Ministry of Statistics and Programme Implementation, New Delhi, https:// productivity firms to low-productivity ones, and microdata.gov.in/nada43/index.php/catalog/47; Mexico: value creation declines. Compounding this effect Iacovone et al. (2022), based on tabulations from is the more consequential one of reduced upgrad- Economic Censuses 2019 (dashboard), Instituto Nacional de Estadistica y Geografia (National Institute of Statistics ing, which lowers the number of firms that have and Geography Mexico), Aguascalientes, Mexico, https:// the ability to grow. en.www.inegi.org.mx/programas/ce/2019/; Peru: IV Censo Nacional Económico 2008 (dashboard), Instituto Nacional de Estadística e Informática, Lima, Peru, http://censos1​ Outdated rules of thumb stifle .inei.gob.pe/cenec2008/redatam_inei/; United States: 2019 progress data from BDS (Business Dynamics Statistics) (dashboard), US Census Bureau, Suitland, MD, https://www.census.gov​ The goals of protecting small businesses and pre- /programs-surveys/bds.html. venting the concentration of economic power Note: The figure reports the share of establishments of various size classes. Shares for India and the United States have motivated many countries to implement pol- are for the manufacturing sector only. icies based on firm size (size-dependent policies). By using size to screen which firms should be pro- tected and which firms should be penalized, pol- younger than a year—they are not the net creators icy makers end up taxing firms that create value (figure 4.10). (box 4.2). A firm that creates value attracts capital and workers—it grows in size. Thus a policy that How governments stifle screens by size is effectively a policy that curtails firms’ incentives to grow, productivity and value creation. Such policies infuse global technologies, keep productive firms smaller and less productive firms larger than they should be based on their and innovate technical capabilities. Most firms in middle-income countries are small, An example is the Small-Scale Reservation Laws do not grow, and do not exit the economy. This implemented in India between the 1960s and early lackluster performance is due, in large part, to 2000s. The laws stipulated that most manufac- government policies that interfere with the forces tured goods could be produced only by small-scale 96 | WORLD DEVELOPMENT REPORT 2024 Figure 4.10 Young firms—not small firms—create the most jobs (net) in the United States Number of jobs lost or created 750,000 500,000 250,000 0 –250,000 1–4 5–9 10–19 20–49 50–99 100–499 500+ Number of employees Age of firms (years): <1 1–4 5–9 ≥10 Source: 2019 data from Business Employment Dynamics (dashboard), Bureau of Labor Statistics, United States Department of Labor, Washington, DC, https://www.bls.gov/bdm/business-employment-dynamics-data-by-age-and-size.htm. Note: The figure reports the net number of jobs created by firms of different sizes and ages in the United States. Box 4.2 Examples of size-dependent policies In some countries, smaller firms below a specific size are exempt from regulations or tax- ation. In other countries, once the firm exceeds a specified size threshold, it faces higher taxes and more regulations. In extreme versions, such policies restrict the production of certain items to firms of a given size. Many other policies and market frictions may also result in a size-dependent distortion based on the abilities of firms of different sizes, productivity levels, and internal wealth to circumvent such policies. For example, stringent collateral requirements for business loans disproportionately affect young, productive firms with strong borrowing needs but little history of retaining enough earnings to accumulate internal wealth. Some examples of explicit size-dependent policies in various middle- and high-income countries follow. Although individual policies like these, in isolation, may have only a minor effect on the full extent of misallocation and lack of innovation in middle-income economies, explicit size-dependent policies may interact with implicit size-dependent distortions and reinforce one another. For example, a regulation that increases the fixed and variable costs of labor for firms with 50 or more workers (an explicit size-dependent distortion) may coexist with a financial market that imposes tighter credit standards on young businesses and for loans aimed at financing intangible capital investments such as research and development (an implicit size-dependent distortion). The overall degree (Box continues next page) Creation | 97 Box 4.2 Examples of size-dependent policies (continued) of resource misallocation and underinvestment in innovation is the outcome of multiple distortions interacting and reinforcing one another. Mexico. Mexico’s income tax law created REPECO (Régimen de Pequeños Contribuyentes), a special provision for small businesses based on their level of sales. For ordinary firms whose sales exceed the threshold, the tax on capital income amounts to 38 percent (28 percent for the government and 10 percent for profit-sharing with employ- ees). Firms with annual sales below US$163,000 (Mex$2 million) do not have to pay the capital tax, but instead pay a 2 percent sales tax, with 7.5 percent of the 2 percent sales tax directed to the profit-sharing scheme. Once the sales of a REPECO firm exceed the threshold, it cannot become a REPECO firm again.a Türkiye. Türkiye’s labor and safety laws impose several regulations that apply to firms above certain size thresholds. Firms employing more than 50 workers must establish a health and safety board; set up a health unit and hire physicians and other health staff; and employ disabled workers and formerly convicted workers.b India. Small-Scale Reservation Laws in India are the poster child of a size-dependent policy that fosters misallocation and discourages innovation.c Since 1960, the govern- ment of India has “reserved” the production of many manufactured goods for small-scale industries only. The definition of a small-scale industry is based on an industry’s cumulative investment in plant and machinery. Effectively, all establishments with plant and machin- ery below a specific limit have been considered small and allowed to produce the reserved goods. Although the number of products falling into the reserved category increased steadily between 1960 and 2002, the policy has been progressively dismantled, with only 20 products remaining under the reservation law in 2010. Peru. Various labor laws in Peru are size-dependent. For example, a profit-sharing agreement applies to firms with more than 21 salaried workers. The fraction of profits ­ to be distributed ranges from 10 percent in manufacturing to 8 percent in mining, retail, wholesale, and restaurants. Workers can join a firm-specific union of at least 20 members, exposing firms beyond this size to the unionization of their workforce. And firms exceed- ing 20 employees must set up a health and safety committee.d Portugal. The average size of firms in Portugal fell sharply between the mid-1980s and mid-2000s. Size-dependent policies may have contributed to such a dramatic shift in the production structure.e Notably, firms with less than 50 workers are allowed to pay up to 50 percent less than the minimum wage, are subsidized in hiring young workers, and receive support for worker hiring. Firms with more than 50 workers must maintain a worker health protection system. Italy. Italy’s Employment Protection Legislation imposes a disproportionately higher cost on firms employing more than 15 workers to fire workers. Firms with fewer than 15 workers must pay a dismissal cost equivalent to between 2.5 and 6 months of salary in cases of unjustified firing. Firms with more than 15 workers must compensate workers for forgone wages between the time of dismissal and the court’s sentence. Dismissal trials can last up to five years, and there is no upper limit on the amount of forgone wages.f (Box continues next page) 98 | WORLD DEVELOPMENT REPORT 2024 Box 4.2 Examples of size-dependent policies (continued) France. Labor laws in France apply special provisions to firms with more than 10, 11, 20, or 50 employees.g In particular, as a firm reaches 50 employees, a committee for hygiene, safety, and work conditions must be formed and trained; a works council must be formed and meet at least every other month; and a higher payroll tax rate, which goes from 0.9 percent to 1.5 percent, subsidizes worker training. In addition, if a firm fires more than nine workers for “economic reasons,” it must follow a special legal process, which increases dismissal costs and creates legal uncertainty for the firm. Republic of Korea. When a firm is classified as a small or medium enterprise (SME), it can receive about 160 benefits from the SME support policy. Benefits include differential corporate tax rates, tax relief benefits, and government-guaranteed loans for SMEs.h a. Sánchez-Vela and Valero-Gil (2011). b. Akcigit et al. (2023). c. García-Santana and Pijoan-Mas (2014). d. Dabla-Norris et al. (2018). e. Braguinsky, Branstetter, and Regateiro (2011). f. Schivardi and Torrini (2004). g. Gourio and Roys (2014). h. Jung and Jung (2022). plants, defined as those whose capital stock did enforcement in low- and middle-income coun- not exceed a government-set threshold.8 This pol- tries, whereas this practice is almost absent in icy was complemented by several additional bene- advanced economies.12 fits for small-scale enterprises, such as the Private In terms of credit, when contract enforce- Sector Lending Program, in effect until the 1990s, ment is weak and secured transactions and which stipulated that commercial banks had to bankruptcy laws are poorly designed and allocate up to 40 percent of their private credit at enforced, banks may demand higher interest subsidized rates to small firms. In Peru, the con- rates, impose more stringent lending stan- stitution and legislation require firms with more dards, and tighten their collateral require- than 21 workers to distribute a portion of their ments.13 Some firms may have more collateral before-tax profits to their formal labor force.9 And, to pledge and thus may not be affected by the in France, firms with more than 50 workers pay financial requirements. Others may have accu- a higher payroll tax rate and must comply with mulated enough internal liquid funds to cope additional regulations.10 Such programs create with the pressing needs of working and phys- perverse incentives in terms of resource allocation ical capital without financial intermediation. and innovation. They also contribute to increased However, for many other firms credit is nec- informality in developing countries.11 essary. Tighter collateral constraints hamper Turnover taxes that tax intermediate and these firms’ abilities to scale up to the level capital goods and corporate taxes, even when merited by their capabilities. It is usually young set at uniform rates, are also examples of size-­ entrepreneurs with innovative business ideas dependent policies in the way in which they are but without collateral they can pledge who are enforced. Larger firms are more likely to face tax most negatively affected (box 4.3).14 Creation | 99 Box 4.3 The productivity effects of credit misallocation and capital market underdevelopment A concerning feature of credit markets is the misallocation of credit away from the most productive businesses in middle-income economies. More productive firms in these countries receive lower amounts of debt and equity financing than merited by their technical efficiency, a property that is less prevalent in high-income countries (figure  B4.3.1). In Mexico, for example, access to finance across firms favors the least productive over the most productive firms.a Another feature of developing economies is the underdevelopment of capital markets. Unlike bank-provided credit, debt and equity financing are associated with productive investments by firms, leading to expansions in infusion, innovation, and physical capital.b By excessively relying on banks for financial intermediation, middle-income countries are limiting their capable firms to access finance. Figure B4.3.1 Productivity-dependent financial distortions, by GDP per capita 0.35 Ukraine Productivity-dependent financial distortions 0.30 0.25 Montenegro Bulgaria Slovak Republic Bosnia and Herzegovina Serbia Czechia Luxembourg 0.20 Spain Romania Portugal Hungary Estonia Italy Belgium Croatia 0.15 Poland Slovenia Austria North Macedonia Latvia Germany Finland 0.10 Norway France 0 7 8 9 10 11 12 GDP per capita (log scale) Source: Cusolito et al., forthcoming. Note: The figure reports the estimate of a regression coefficient between the logarithm of idiosyncratic financial distortions and the logarithm of idiosyncratic physical productivity across firms, both computed as in Whited and Zhao (2021). A high value of the coefficient means that more productive firms face higher distortions in financial markets than less productive ones. Thus, there is a weaker relationship between a firm’s productivity and its debt and equity financing compared with efficient allocation. GDP = gross domestic product. (Box continues next page) 100 | WORLD DEVELOPMENT REPORT 2024 Box 4.3 The productivity effects of credit misallocation and capital market underdevelopment (continued) Episodes of capital account liberalization, opening middle-income countries’ econo- mies to foreign capital financing, can reduce the misallocation of capital.c For example, India’s capital market liberalization contributed to a more efficient allocation of resources across firms over the last 20 years. a. Iacovone et al. (2022). b. Didier et al. (2020). c. Bau and Matray (2023). Figure 4.11 Productivity-dependent distortions are more severe in low- and middle-income countries 0.4 Productivity-dependent distortions Kenya Portugal Bangladesh El Salvador Pakistan 0.3 Ethiopia Ghana China India Colombia 0.2 Peru Bulgaria Chile Belgium Romania 0.1 Hungary Malaysia Latvia Spain Italy 0 Finland France –1.0 6 7 8 9 10 11 GDP per capita (log scale) Low-income Middle-income High-income Sources: Productivity-dependent distortions: WDR 2024 team calculations and Fattal Jaef (2022). Gross domestic product (GDP) per capita: PWT (Penn World Table) (database version 9.0), Groningen Growth and Development Centre, Faculty of Economics and Business, University of Groningen, Groningen, the Netherlands, https://www.rug.nl/ggdc/productivity/pwt​ /­pwt-releases/pwt9.0?lang=en . Note: The figure reports the estimates of a regression coefficient between the logarithm of idiosyncratic distortions and the logarithm of idiosyncratic physical productivity across firms, both computed as in Hsieh and Klenow (2009). The corresponding estimate for the United States is subtracted from each country’s estimate so that productivity-dependent distortions are reported relative to the level observed in the United States. A high value for this elasticity means that a productive firm confronts higher distortions than less productive ones, and thus there is a weaker relationship between firm productivity and firm size than the output-maximizing allocation would require. More broadly, these rules of thumb impose a also discourage the adoption of modern manage- hefty tax on productive firms, a practice known ment practices, a necessary condition for improv- as “taxing the good” (figure 4.11). By discouraging ing efficiency and adopting technologies.15 Alfred the expansion and growth of firms, these policies D. Chandler, Jr., a renowned economic historian, Creation | 101 Figure 4.12 Management practices are worse in economies with more policy distortions 0.4 Productivity-dependent distortions Kenya Portugal Ethiopia 0.3 Ghana China Colombia 0.2 India Chile 0.1 Spain Italy 0 France −0.1 2.0 2.2 2.4 2.6 2.8 3.0 3.2 Management indicator Sources: Productivity-dependent distortions: Fattal Jaef 2022; management indicator: WMS (World Management Survey) (dashboard), London School of Economics, London, https://worldmanagementsurvey.org (see Bloom and Van Reenen 2007). Note: Productivity-dependent distortions refer to policy distortions related to firm size that can discourage innovation and technology adoption. The figure reports, on the y-axis, the estimates of a regression coefficient between the logarithm of idiosyncratic distortions and the logarithm of idiosyncratic productivity of the firms. The idiosyncratic distortion and the physical productivity are computed as in Hsieh and Klenow (2009). On the x-axis, the figure reports the management indicator across firms, as provided in the World Management Survey. suggested that one of the critical transforma- ventures in the market. However, these prefer- tions in the US economy between the mid-1880s ences hurt private enterprise. Although SOEs and World War I was the replacement of the benefit from lower costs of labor and capital, this traditional family enterprise with the modern ­ benefit increases the cost of production for their multiunit enterprise managed by a complex layer private counterparts, which are often more pro- of top- and middle-management structures.16 ductive. Benefits extended to SOEs not only mis- Countries with a higher quality of man- allocate resources away from more productive agement practices—as measured by the World firms but also discourage the private sector from Management Surveys—have lower ­ productivity-​ investing in costly initiatives (such as R&D) to dependent distortions (figure 4.12).17 In the become more innovative or produce more inno- United States, a firm’s managerial capabilities vative goods or services.19 SOEs in manufacturing are central to its ability to grow and innovate. and contestable service industries have become a However, in most middle-income countries, source of resource misallocation.20 managerial c ­ ­ apabilities of firms are underdevel- In the critical area of energy supply, electricity oped.18 The exceptions are a handful of countries continues to be delivered primarily by SOEs. In that have enjoyed sustained growth accelerations. 2020, 60 percent of 125 low- and middle-income countries still relied on a public distribution utility.21 The role of SOEs in fossil fuel power investment Benefits extended to state-owned increased from 43 percent to 50 percent from 2015 enterprises hurt private enterprise to 2019, notably due to the expansion of coal plants When countries are early in their development in India and South Africa and gas plants in the journey, the state favors its SOEs for undertaking Middle East and North Africa. 102 | WORLD DEVELOPMENT REPORT 2024 Preferences for incumbents block • Market power is sustained over time by government interventions that protect a firm ­ progress or provide it with specific advantages. In Large enterprises are important drivers of inno- ­ Mexico, for example, although the tele- vation and are able to imitate at scale. However, communications sector was open to com- weak institutions may encourage these firms to petition in 1995 after 20 years of public protect their profitability by deterring competi- monopoly, followed by five years of private tion rather than innovating their way out of com- monopoly,25 it was not until 2014 that the petition. A firm’s market power and incentives to government imposed the first obligations abuse its dominance are shaped by features of the on Telmex and Telcel as operators with market, government interventions, and its strate- significant market power.26 The inability gic behavior. of the sector regulator and the competi- The problem is not the market power of tion authority to impose such conditions some firms per se, but the exercise of domi- in previous years made its owner, Carlos nance that undermines effective competition, Slim, the richest man in the world between dampening innovation, harming consumers, 2011 and 2013, according to Forbes mag- and driving equally efficient competitors out azine. Similarly, overly long concession of the market (table 4.1). Market leaders can contracts for limestone and restrictive also acquire smaller competitors to wind down standards for cement helped Aliko Dan- the operations of potential competitors, such gote become the richest man in Africa.27 as in the pharmaceutical industry.22 Market • Market power is sustained over time by illegal power becomes problematic under the follow- strategic behavior, such as when dominant ing conditions:23 firms exclude their rivals from the market or when a few firms agree to collude. Anti- • Market power is not the result of innovation. competitive practices that close entry and For example, politically connected firms in place competitors at a disadvantage are Italy’s local markets changed their behav- common in digital platform markets.28 ior as they grew. Rather than innovating and competing with other firms, they If large incumbent firms are not disciplined by brought local politicians onto their pay- means of regulations that promote competition roll.24 At the same time, their innovation or antitrust rules, firms with market power are (measured in terms of numbers of patents) more prone to abusive practices—such as preda- began to decline. tory pricing, price squeezing, or denial to supply Table 4.1 Examples of possible effects of market power on development outcomes DEVELOPMENT OUTCOMES PROS CONS Productivity Mergers and acquisitions can lead to Conglomerates can cross-subsidize unproductive more efficiency. firms, crowding out more productive competitors. Consumer welfare More efficient structures can lead to It is easier to form cartels when fewer players are lower costs and therefore lower prices. in the market. Innovation Some firms acquire market power There are fewer incentives to innovate, and firm because they innovate and push the entry is low. frontier. Jobs The demand for highly skilled workers Wages for lower-skilled workers are depressed, grows. while the most talented workers are captured. Source: World Bank 2024a. Creation | 103 other competitors—contributing to a decrease in grow from “within” by enhancing their existing competition. Although many competition author- product portfolio, or do they expand “across” ities have the mandate to sanction anticompeti- the market by challenging other incumbents tive practices, prevent anticompetitive mergers, in different markets? If growth is from within, and provide policy advice to eliminate regulatory are gatekeepers hindering firms from develop- restrictions to competition, not all competition ing new products for new markets? Examining authorities in middle-income countries have the the health of an economy and measuring its resources needed to enforce their mandate. For vital signs require a comprehensive approach example, although Brazil, China, India, Mexico, to weighing its business dynamism, including and South Africa have a wide range of resources responsiveness to productivity shocks, job cre- such as enforcement tools, budget, and staff, ation and destruction rates, turnover among authorities in Bangladesh, Malaysia, Morocco, market leadership, entry and exit rates, life and Nigeria have limited resources.29 cycle (age and size) dynamics, R&D investment, and spending on technology licensing, among other things. Modernizing data and New sources of data and new opportunities for diagnostic tools to data-driven policy analysis. The creative destruc- understand and regulate tion framework offers a distinctive advantage in creative destruction—from understanding aggregate economies by address- ing microlevel intricacies and frictions that X-rays to MRIs combine to form the macroeconomy. Although Modern Schumpeterian thinking on creative certain underperforming countries may exhibit destruction has the potential to provide much analogous macroeconomic trends in terms of more reliable diagnostics than what analysts, investment dynamics and sluggish productivity researchers, and regulators have been using so growth, their microlevel challenges can vary sig- far. No longer is it enough to decompose aggre- nificantly. For example, one economy may grap- gate output growth into the growth of measur- ple with financial frictions affecting the “cost” of able production factors—physical and human investment for firms, while another may contend capital and the residual, total factor productivity. with high market concentration hindering invest- Nor is it enough to rely solely on measures such as ment incentives by lowering “returns.” Similarly, firm employment and size distribution to gauge entrants may face obstacles in one economy, while an industry’s health. These diagnostics are much incumbent firms encounter distinct challenges in like using two-dimensional images from X-ray others. A microlevel investigation, informed by machines to analyze economic structures. They reliable data on firms and individuals, is needed. are necessary but not sufficient. The recent digital revolution facilitates not only A nuanced understanding of the process of the collection and organization of extensive firm- creative destruction emphasizes the tensions level data sets but also their processing through that arise between entrants and incumbents, as innovative techniques for data-driven policy well as among incumbents competing for mar- analysis. ket leadership. Diagnostics that examine these Several “must-have” data sets exist. One note- links need the latest instruments—akin to three-­ worthy example is the census of firms, a thor- dimensional images from MRIs. As with MRIs, ough and complete collection of information using these instruments also requires more about all businesses or firms within a specific skilled practitioners, more data, and more care. geographic area or industry. Many countries now New questions to answer. Moving beyond grant experts access to their census of firms for firm employment size, diagnostics will need research purposes. Unlike a sample survey, which to examine the following questions: Do firms gathers data from a subset of the population, 104 | WORLD DEVELOPMENT REPORT 2024 a  census aims to encompass every entity within security, and flexibility in managing their finan- the defined scope. Typically, census of firms data cial processes. As a result, platform data can pro- include details such as the number of firms, their vide information, even on the smallest businesses, size, location, ownership structure, industry, almost in real time, even when government statis- number of employees, revenue, and other perti- tics are unavailable. nent information. This data collection method Several additional “good-to-have” data sources provides a detailed, accurate representation of are available for conducting in-depth firm-level the entire population of firms, enabling a com- investigations, including Orbis Europe,32 Orbis prehensive analysis of many measures of business M&A,33 PitchBook,34 and PATSTAT.35 These data dynamism. sets offer valuable insights into firm performance The World Bank’s Enterprise Surveys are at the micro level. the go-to resource for cross-country analysis Greater analytical expertise. As the sources of firm-level information about the business and volume of data soar, avoiding misdirection environment. More recently, the surveys have and framing the appropriate questions become been expanded to include the informal sector. challenging. A starting point is to carefully The World Bank also conducts the Firm-level investigate whether promising new firms can Adoption of Technology (FAT) survey, collect- easily enter the economy or they face obstacles, ing information on the adoption of general be it from direct means such as licensing or business functions and sector-specific technol- indirectly from a significant presence of SOEs. ogies in 11  developing economies. Some com- Furthermore, a better understanding of the mercial databases have also been widely used productivity dynamics among incumbent firms in the ­literature. Salient examples are the Orbis is needed. As firms expand, there is typically database, produced by Bureau van Dijk, and an upsurge in the concentration of economic Worldscope, accessible via online platform.30 activities. Market concentration can either Social Security records are also widely utilized. signify “productive behavior” or be linked to rent- These records often contain employment-related seeking and “strategic behavior.” The challenge information, including earnings, work history, lies in distinguishing between the two. Analysts and contributions to social security programs. must therefore scrutinize whether growing firms Experts can scrutinize labor market dynamics, are expanding due to enhanced productivity employment patterns, and strategic hiring, even or are exhibiting strategic behavior without an investigating “killer acquisitions” by examining actual improvement in productivity. Improving which firms hire workers from their rivals to stifle the analytic capabilities of analysis is just as competition. Emerging online platforms present important as improving the quality of data. invaluable opportunities for real-time tracking of In conclusion, the forces of creation will need businesses.31 In middle-income countries, espe- strengthening in middle-income countries. Both cially among small firms with limited internal entrants and incumbents can create value—by resources, susceptibility to macroeconomic con- infusing global technologies and by innovating. ditions has increased, underscoring the need for Outdated rules of thumb, a preference for SOEs, frequent real-time monitoring of the health of the and political capture by large incumbents stymie small business sector. Meeting this need was pre- the forces of creation. By modernizing data and viously challenging because of a lack of the appro- diagnostics, governments can help in the mod- priate data. However, the emergence of online ernization of productive firms, leading to better cloud-based accounting software has bridged managers, better professionals, and better tech- this gap. Many businesses now leverage these nologies. Governments can also discipline dom- platforms for improved efficiency, accessibility, inant incumbents. Creation | 105 Notes  1. Foster, Haltiwanger, and Krizan (2001); Lentz and 23. World Bank (2024a). Mortensen (2016). 24. Akcigit, Baslandze, and Lotti (2023).  2. Aghion and Howitt (1992). 25. OECD (2012).  3. Klette and Kortum (2004). 26. OECD (2017).  4. Akcigit and Kerr (2018). 27. Economist (2014).  5. Chandler (1977). 28. World Bank (2021).  6. See, for example, Abreha et al. (2022) for a discussion 29. World Bank (2024a). of informality and the firm size distribution in Sub- 30. See Orbis (database), Baker Library, Bloomberg Center, Saharan Africa. Harvard Business School, Boston, https://www.library​  7. Bruhn and McKenzie (2014); Ulyssea (2020). .hbs.edu/find/databases/orbis#:~:text=ORBIS%2C%20  8. García-Santana and Pijoan-Mas (2014). a%20global%20company%20database,company%20  9. Dabla-Norris et al. (2018). data% 20 inc luding% 20 su bsidiar y % 20 loc ations; 10. Gourio and Roys (2014). Worldscope (database), Baker Library, Bloomberg 11. Akcigit et al. (2023). Center, Harvard Business School, Boston, https:// 12. Bachas, Fattal Jaef, and Jensen (2019). www.library.hbs.edu/find/databases/worldscope. 13. Rodano, Serrano-Velarde, and Tarantino (2016). 31. Akcigit et al. (2024). 14. Buera, Kaboski, and Shin (2011); Midrigan and Xu 32. See https://www.eui.eu/Research/Library/Research​ (2014). Guides/Economics/S tatistics/DataPor tal/Orbis​ 15. Cirera and Maloney (2017). Europe. 16. Chandler (1977, 123–26). 33. See https://libguides.eur.nl/az.php?q=ZEPHYR. 17. Bloom and Van Reenen (2007). 34. PitchBook (portal), PitchBook, Seattle, https://­ 18. Bloom and Van Reenen (2010). pitchbook.com/. 19. Akcigit and Cilasun (2023). 35. PATSTAT (Patent Statistical Database), European 20. Brandt, Kambourov, and Storesletten (2020); Hsieh and Patent Office, Munich, https://www.epo.org/en​ Klenow (2009); Whited and Zhao (2021); World Bank /searching-for-patents/business/patstat#:~:text​ (2023). = PAT S TAT % 20 c o n tains% 20 bibliog r ap hic al% 20 21. Küfeoğlu, Pollitt, and Anaya (2018). and%20legal,or %20can%20be%20consulted%20 22. Cunningham, Ederer, and Ma (2021). online. References Abreha, Kaleb Girma, Xavier Cirera, Elwyn Adriaan Robin Akcigit, Ufuk, Murat Alp Celik, and Jeremy Greenwood. Davies, Roberto N. Fattal Jaef, and Hibret Belete 2016. “Buy, Keep or Sell: Economic Growth and the Maemir. 2022. “Deconstructing the Missing Middle: Market for Ideas.” Econometrica 84 (3): 943–84. Informality and Growth of Firms in Sub-Saharan Africa.” Akcigit, Ufuk, and Seyit Mumin Cilasun. 2023. “A Brief Policy Research Working Paper 10233, World Bank, Overview of State-Owned Enterprises in Türkiye.” Washington, DC. Background paper prepared for The Business of the Aghion, Philippe, and Peter Howitt. 1992. “A Model of State report, World Bank, Washington, DC. Growth through Creative Destruction.” Econometrica Akcigit, Ufuk, John Grigsby, and Tom Nicholas. 2017a. 60 (2): 323–51. “Immigration and the Rise of American Ingenuity: Akcigit, Ufuk, Yusuf Emre Akgündüz, Harun Alp, Seyit Innovation and Inventors of the Golden Age.” American Mumin Cilasun, and Jose M. Quintero. 2023. “Good Economic Review 107 (5): 327–31. Firms, Bad Policies: The Dynamics of Informality Akcigit, Ufuk, John Grigsby, and Tom Nicholas. 2017b. “The and Industry Policy in Shaping Economic Growth.” Rise of American Ingenuity: Innovation and Inventors University of Chicago Working Paper, University of of the Golden Age.” NBER Working Paper 23047 Chicago, Chicago. (January), National Bureau of Economic Research, Akcigit, Ufuk, Harun Alp, and Michael Peters. 2021. “Lack Cambridge, MA. of Selection and Limits to Delegation: Firm Dynamics Akcigit, Ufuk, and William R. Kerr. 2018. “Growth through in Developing Countries.” American Economic Review Heterogeneous Innovations.” Journal of Political 111 (1): 231–75. Economy 126 (4): 1374–443.  Akcigit, Ufuk, Sina T. Ates, André Diegmann, Furkan Kilic, Bachas, Pierre, Roberto N. Fattal Jaef, and Anders and Steffen Müller. 2024. “Steadfast Germany, Dynamic Jensen. 2019. “Size-Dependent Tax Enforcement USA: Incumbents and Entrants Shaping Economic and Compliance: Global Evidence and Aggregate Growth.” Working paper, University of Chicago, Chicago; Implications.” Journal of Development Economics 140 Halle Institute for Economic Research, Halle, Germany. (September): 203–22. Akcigit, Ufuk, Salomé Baslandze, and Francesca Lotti. 2023. Bau, Natalie, and Adrien Matray. 2023. “Misallocation “Connecting to Power: Political Connections, Innovation, and Capital Market Integration: Evidence from India.” and Firm Dynamics.” Econometrica 91 (2): 529–64. Econometrica 91 (1): 67–106. 106 | WORLD DEVELOPMENT REPORT 2024 Bloom, Nicholas, and John Van Reenen. 2007. “Measuring Fattal Jaef, Roberto N. 2022. “Entry Barriers, Idiosyncratic and Explaining Management Practices across Firms Distortions, and the Firm Size Distribution.” American and Countries.” Quarterly Journal of Economics 122 (4): Economic Journal: Macroeconomics 14 (2): 416–68. 1351–408. Foster, Lucia, John C. Haltiwanger, and C. J. Krizan. Bloom, Nicholas, and John Van Reenen. 2010. “Why 2001. “Aggregate Productivity Growth: Lessons from Do Management Practices Differ across Firms and Microeconomic Evidence.” In New Developments in Countries?” Journal of Economic Perspectives 24 (1): Productivity Analysis, edited by Charles R. Hulten, Edwin 203–24. R. Dean, and Michael J. Harper, 303–72. National Bureau Braguinsky, Serguey, Lee G. Branstetter, and Andre of  Economic Research Studies in Income and Wealth Regateiro. 2011. “The Incredible Shrinking Portuguese Series, vol. 63. Cambridge, MA: National Bureau of Firm.” NBER Working Paper 17265 (July), National Economic Research; Chicago: University of Chicago Press. Bureau of Economic Research, Cambridge, MA. García-Santana, Manuel, and Josep Pijoan-Mas. 2014. Brandt, Loren, Gueorgui Kambourov, and Kjetil “The Reservation Laws in India and the Misallocation Storesletten. 2020. “Barriers to Entry and Regional of Production Factors.” Journal of Monetary Economics Economic Growth in China.” Working Paper 652 66 (September): 193–209. (January 5), Department of Economics, University of Gourio, François, and Nicholas A. Roys. 2014. “Size- Toronto, Toronto. Dependent Regulations, Firm Size Distribution, and Bruhn, Miriam, and David J. McKenzie. 2014. “Entry Reallocation.” Quantitative Economics 5 (2): 377–416. Regulation and the Formalization of Microenterprises in Hsieh, Chang-Tai, and Peter J. Klenow. 2009. “Misallocation Developing Countries.” World Bank Research Observer and Manufacturing TFP in China and India.” Quarterly 29 (2): 186–201. Journal of Economics 124 (4): 1403–48. Buera, Francisco Javier, Joseph P. Kaboski, and Yongseok Hsieh, Chang-Tai, and Peter J. Klenow. 2014. “The Life Shin. 2011. “Finance and Development: A Tale of Two Cycle of Plants in India and Mexico.” Quarterly Journal Sectors.” American Economic Review 101 (5): 1964–2002. of Economics 129 (3): 1035–84. Chandler, Alfred D., Jr. 1977. The Visible Hand: The Iacovone, Leonardo, Rafael Muñoz Moreno, Eduardo Managerial Revolution in American Business. Olaberria, and Mariana De La Paz Pereira López. 2022. Cambridge, MA: Harvard University Press. Productivity Growth in Mexico: Understanding Main Cirera, Xavier, and William F. Maloney. 2017. The Innovation Dynamics and Key Drivers. Washington, DC: World Bank. Paradox: Developing-Country Capabilities and the Jung, A‐Reum, and Do‐Jin Jung. 2022. “The Effects of Size‐ Unrealized Promise of Technological Catch-Up. Dependent Policy on the Sales Distortion Reporting: Washington, DC: World Bank. Focusing on the Discretionary Sales Management of Cunningham, Colleen, Florian Ederer, and Song Ma. 2021. Korean SMEs.” Managerial and Decision Economics “Killer Acquisitions.” Journal of Political Economy 43 (2): 301–20. 129 (3): 649–702. Klette, Tor Jakob, and Samuel Kortum. 2004. “Innovating Cusolito, Ana Paula, Roberto Fattal Jaef, Davide Salvatore Firms and Aggregate Innovation.” Journal of Political Mare, and Akshat Singh. Forthcoming. “From Financial Economy 112 (5): 986–1018.  to Real Misallocation: Evidence from a Global Küfeoğlu, Sinan, Michael Gerald Pollitt, and Karim L. Sample.” Policy Research Working Paper, World Bank, Anaya. 2018. “Electric Power Distribution in the World: Washington, DC. Today and Tomorrow.” Cambridge Working Paper in Dabla-Norris, Era, Laura Jaramillo Mayor, Frederico Lima, Economics 1846 (August 16), University of Cambridge, and Alexandre Sollaci. 2018. “Size-Dependent Policies, Cambridge, UK. Informality, and Misallocation.” IMF Working Paper Lamoreaux, Naomi R., and Kenneth L. Sokoloff. 2002. WP/18/179 (August), International Monetary Fund, “Intermediaries in the U.S. Market for Technology, Washington, DC. 1870–1920.” NBER Working Paper 9017 (June), National de Nicola, Francesca, and Jonathan D. Timmis. Bureau of Economic Research, Cambridge, MA. Forthcoming. “Firm Foundations: Digital Technologies Lentz, Rasmus, and Dale T. Mortensen. 2016. “Optimal and Productivity Growth in East Asia and Pacific.” Growth through Product Innovation.” Review of World Bank, Washington, DC. Economic Dynamics 19 (January): 4–19. Didier, Tatiana, Ross Eric Levine, Ruth Llovet Montanes, and Luttmer, Erzo G. J. 2011. “On the Mechanics of Firm Growth.” Sergio Schmukler. 2020. “Capital Market Financing and Review of Economic Studies 78 (3): 1042–68. Firm Growth.” Policy Research Working Paper 9337, Midrigan, Virgiliu, and Daniel Yi Xu. 2014. “Finance and World Bank, Washington, DC. http://dx.doi.org/10.2139​ Misallocation: Evidence from Plant-Level Data.” /ssrn.3666203. American Economic Review 104 (2): 422–58. Donaldson, Dave, and Richard Hornbeck. 2016. “Railroads OECD (Organisation for Economic Co-operation and American Economic Growth: A ‘Market Access’ and Development). 2012. OECD Review of Approach.” Quarterly Journal of Economics 131 (2): Telecommunication Policy and Regulation in Mexico. 799–858. Paris: OECD. Economist. 2014. “Building on Concrete Foundations.” OECD (Organisation for Economic Co-operation and Business, April 12, 2014. https://www.economist.com​ Development). 2017. OECD Telecommunication and /­business/2014/04/12/building-on-concrete-foundations. Broadcasting Review of Mexico 2017. Paris: OECD. Creation | 107 Perlman, Elisabeth Ruth. 2015. “Dense Enough to Be Records, 1790–1846.” NBER Working Paper 2707 Brilliant: Patents, Urbanization, and Transportation (September), National Bureau of Economic Research, in Nineteenth Century America.” CEH Discussion Cambridge, MA. Paper 2015–06 (March), Centre for Economic History, Ulyssea, Gabriel. 2020. “Informality: Causes and Research School of Economics, Australian National Consequences for Development.” Annual Review of University, Canberra. Economics 12 (1): 525–46. Rodano, Giacomo, Nicolas Serrano-Velarde, and Emanuele Whited, Toni M., and Jake Zhao. 2021. “The Misallocation of Tarantino. 2016. “Bankruptcy Law and Bank Financing.” Finance.” Journal of Finance 76 (5): 2359–407. Journal of Financial Economics 120 (2): 363–82. World Bank. 2015. Peru: Building on Success, Boosting Sánchez-Vela, Claudia, and Jorge N. Valero-Gil. 2011. Productivity for Faster Growth. Washington, DC: World “The Effect of Firm-Size Dependent Policies on the Bank. Economy: The Case of the Repecos Law in Mexico.” IDB World Bank. 2021. “Antitrust and Digital Platforms: An Working Paper IDB-WP-216 (February), Inter-American Analysis of Global Patterns and Approaches by Development Bank, Washington, DC. Competition Authorities.” Equitable Growth, Finance, Schivardi, Fabiano, and Roberto Torrini. 2004. “Firm Size and Institutions Insight. Trade, Investment, and Distribution and Employment Protection Legislation in Competitiveness, Levelling Up Series, World Bank, Italy.” Temi di discussione (Economic Working Paper) Washington, DC. 504, Economic Research and International Relations World Bank. 2023. The Business of the State. Washington, Area, Bank of Italy, Rome. DC: World Bank. Schumpeter, Joseph Alois. 1942. Capitalism, Socialism and World Bank. 2024a. From Market Power to Powering Democracy. New York: Harper and Brothers. Markets for Development. Washington, DC: World Bank. Sokoloff, Kenneth L. 1988. “Inventive Activity in World Bank. 2024b. Investment Climate Assessment 2.0 Early Industrial America: Evidence from Patent Diagnostic Manual. Washington, DC: World Bank. 108 | WORLD DEVELOPMENT REPORT 2024 5 Preservation Key messages • Many middle-income countries have a shortage of talent, a key ingredient for bolstering the forces of creation. Yet the forces of preservation prevent talent from developing as well as waste the existing talent. • Talent matters more in middle-income countries than in low-income countries because in middle-income countries skilled workers have become increasingly key to their economic, structural, and technological transformation. • The opportunities for social mobility in middle-income countries are few due to three main forces of preservation: networks, neighborhoods, and norms. • Networks, such as elite pacts, facilitate the formation of groups that can determine access to jobs, services, and policy making. As a result, they tend to keep in-groups in and out- groups out, devalue talent and merit, perpetuate inequities of opportunity, and depress expectations for upward mobility in middle-income countries. • Neighborhoods shape the access of those who were born, grew up, and lived there to oppor- tunities and aspirations. In many middle-income countries, neighborhoods keep people stuck in place, impede migration, hinder productive agglomeration, slow knowledge diffu- sion, and block the pathways to scale that efficient small enterprises need to become large productive and innovative firms. • Norms—particularly patriarchal social norms—can hold back women and other marginal- ized communities from entering the labor force, as well as from benefiting from educational opportunities. In this way, norms prevent the development and growth of talent among half the population of middle-income countries. Preservation is an antagonist secondary schools. One such reform included of creation because it is also a rigorous teacher certification requirement in exchange for doubling teachers’ salaries. an antagonist of destruction Although the certification process was supposed In 2005, the government of Indonesia attempted to include high-quality external assessments of to implement reforms that would heighten teachers’ subject knowledge and pedagogical ­ students’ achievement in elementary and lower practice, the issue became highly politicized. | 109 Consequently the intended competency tests of education, out of skills, out of luck, and were replaced with weak certification require- out of hope. ments, and the reform amounted to a 100 percent salary hike for all teachers. Despite its very high The first two forces of preservation emerge fiscal price tag, the reform yielded no improve- from elite pacts, keeping out-groups out. The ment in student achievement,1 although such a third is a consequence of patriarchal social systemic improvement of teacher requirement norms and a system of attitudes and beliefs. may take a longer time to have a systemwide The result is that women have unequal access impact on learning. to resources and power, thereby holding back a As enterprises in middle-income countries large proportion of the population. To be sure, infuse global technologies, they will need tech- all three factors can be helpful in filling the nical workers and specialized professionals to gaps left by missing markets and missing ser- adopt and use technology as well as managers vices, but they also become sources of preserva- to run modern firms. Middle-income countries tion when they restrict and ration access based have a shortage of such talent, which is a key on social status. ingredient of efforts to bolster the forces of cre- This chapter highlights how social immobil- ation (see chapter 4). Yet the forces of preserva- ity preserves the status quo, exacerbating the tion keep talent from developing as well as waste stunting and waste of talent in middle-income the existing talent, therefore reducing the incen- countries. Although the forces of preservation tives for many individuals to invest in developing discussed here focus on talent for a growing their human capital. In many ­ middle-income middle-income economy, the principles broadly countries, access to opportunities is not based on apply to enterprise and energy as well. For merit, resulting in high income inequality and example, obstructions posed by incumbents, few opportunities for social mobility, thereby including state-owned enterprises, hobble perpetuating “social immobility.” The roots of the drive for a quick lower-carbon transition, social immobility can be traced to n ­ etworks, mainly led by young firms in the private s­ ector. neighborhoods, and norms: Incumbents can cause significant delays by refusing network connection on shared assets (“deep connection”). Incumbents can also • Networks keep in-groups in and out- “curtail” the distribution of power generated groups out, devalue talent and merit, by wind and solar energy—a persistent prob- perpetuate inequities of opportunity, lem for renewable energy developers in several and depress expectations for upward middle-income countries, despite a “must-run” mobility, all for the sake of preserving assurance in regulations to deliver reliable social elites—and, with them, inefficient energy supplies. incumbent firms. This chapter examines three questions: • Neighborhoods keep people stuck in place, impeding migration, hindering produc- • How harmful are the forces of preservation in tive agglomeration, slowing knowledge middle-income countries? As countries tran- diffusion, and blocking the pathways to sition from low-income to middle-income scale that efficient small enterprises need status, the demand for skilled workers such to become large productive firms—and to as technicians, professionals, and man- become the innovative incumbents of the agers increases substantially. As income future. per capita increases, the share of workers • Norms keep women and marginalized employed in small firms declines, and the communities out of the labor force, out share of those employed in medium and 110 | WORLD DEVELOPMENT REPORT 2024 large firms steadily increases. ­ Countries Talent drives economic with greater social mobility have more skills to draw on in the workforce. The progress, but social forces of preservation, holding back immobility holds back the social mobility, are much more harmful development of talent middle-income countries than in low-­ in ­ Fostering talent is a priority for income countries. middle-income countries • How do the economic and social “elite” pre- serve the status quo? Although economic As economies grow, their production processes and social elites have the resources and become more complex; they rely on a growing the education to help accelerate growth division of labor, and the need for new talents through the infusion of global technol- emerges. Agrarian societies need few skills, whereas ogies, their dominance can also buy high-­income countries need many more for their economic, social, and political power. high-end, sophisticated services. The share of By capturing political and social insti- skilled workers among the workforce is very low tutions, they have an outsize say in who in low-­income countries, but it increases steadily studies where and what, who gets a as countries move from lower-middle-income to sought-after job and what they are paid, upper-­middle-income to high-income status (see and who gets to start a business. They chapter 2, figure 2.3). Pakistan would need to double also influence housing markets to deter- its share of skilled workers to reach the level of Chile, mine who lives where and whether new- and China would also need to increase its share sub- comers to a city or country are welcome. stantially. In general, as middle-income countries The status quo is preserved by keeping grow—particularly as they approach high-income “others” out. status and must innovate rather than simply adopt • How does gender inequality hold back the technologies—they require increasingly sophisti- potential of women? In many middle-­ cated talent. Such transformations in the economy income countries, patriarchal gender make the development and efficient allocation of norms hold back women from taking talent particularly important for m ­ iddle-income advantage of attractive opportunities in countries and place social mobility and equitable the labor market and for entrepreneur- access to opportunities at the forefront of policies to ship. Discrimination, sexism, and misog- promote growth and social welfare. yny occur in all walks of life, including the businesses women own, the jobs Skilled workers are key to economic, they get and the pay they receive; how much their families spend on educating structural, and technological them and for how long; and their abil- transformation ity to operate bank accounts. Unequal Firms grow as skilled workers, such as manag- social norms and beliefs and the institu- ers and professionals, become more abundant. tions that reinforce men’s status advan- In lower-income countries with gross domestic tage and access to more resources and product (GDP) per capita under US$3,000, most power hold back nearly half a country’s workers are unskilled and employed in small people, curtailing an economy’s growth. firms with fewer than 10 employees (figure 5.1). It is of particular concern in graying As GDP per capita increases, the share of work- middle-income countries projected to ers employed in small firms declines, and the face labor crunches. share of those employed in medium and large Preservation | 111 Figure 5.1 The share of skilled workers in talent from being nurtured? A better-educated large firms increases with GDP per capita and wealthier parent has a greater capacity to finance investments and make better investment 100 decisions, has better connections for job searches and placements, and can serve as a role model in Share of workers (%) 80 terms of education and professional work. In a self-­ reinforcing cycle, greater investments—and better 60 investment choices—yield increasing benefits to 40 parental background: that is, they create increas- ing returns.2 In particular, greater investments by 20 parents in early childhood increase the returns on later investments.3 This approach can lead to higher 0 inequality and the development of a “human capital Country Country Country Country income income income income elite,” where there is considerable mobility within a group 1 group 2 group 3 group 4 class boundary but not across classes.4 A striking example is the intergenerational persistence among Small firms, low skills Small firms, high skills Medium firms, low skills Medium firms, high skills political elites (where the social and economic sta- Large firms, low skills Large firms, high skills tus of family members between generations stays the same). In many low-income and middle-income Source: Gottlieb, Poschke, and Tueting 2024. countries, the descendants of political elites also Note: The figure shows the share of low-skilled and highly skilled workers who are employed in small, medium, and tend to be involved in and consolidate power and large firms across four country income categories, which resources through politics.5 correspond to gross domestic product (GDP) per capita Countries characterized by higher income levels of US$0–US$3,000 (group 1), US$3,000–US$10,000 (group 2), US$10,000–US$30,000 (group 3), and more than inequality are often those in which a significant US$30,000 (group 4). Shares are computed from nationally portion of economic advantage is transmitted representative labor force surveys from 76 countries from parents to their children. The association covering 805 country-years. Small firms have fewer than between income inequality and intergenerational 10 employees. Medium firms have 10–50 employees. Large firms have more than 50 employees. Workers are classified immobility—often referred to as the “Great Gatsby as low-skilled if they have less than nine years of formal Curve”—is positive as more unequal countries are education, and as highly skilled otherwise. For countries in more socially immobile (figure 5.2).6 This associ- each income group, the bars add up to 100 percent. ation is much stronger for middle-income coun- firms increases steadily. As economies transition tries than for high-income countries. to higher-income status, the demand for skilled At one end of the income equality scale is workers such as technicians, professionals, and Finland, where schooling is largely free at all lev- managers increases substantially. In turn, more els and is of very high quality by international educated managers are more likely to adopt tech- standards. There, 80 percent of children attain nology for general and sector-specific business a level of social mobility that is not dependent functions, thereby raising not only the produc- on their parents’ social status.7 By contrast, in tivity of their firm but also contributing to cre- ­ middle-income countries on average, intergener- ation and economic progress for the economy as ational persistence is much higher (40 percent). a whole (box 5.1). Specifically, the share of individuals whose social and economic status is the same as that of the Improving social mobility is a priority previous generation is more than twice as large for middle-income countries as in Finland. Closing even a fraction of this gap How do societies select who has access to educa- means a great boost to acquisition and allocation tion, employment, and finance? What prevents of talent and to growth. 112 | WORLD DEVELOPMENT REPORT 2024 Box 5.1 Firms with better-educated managers adopt more technology If technology is available, why don’t more firms adopt it? Data from the World Bank’s Firm-level Adoption of Technology (FAT) survey of 12,000 firms in 11 developing coun- tries reveal that the average firm has adopted an intermediate level of technology, scoring 2.4 on the technology adoption index out of a possible 5.a The education level of managers is an important factor in the adoption of technology because better-educated managers are more likely to adopt technology (figure B5.1.1). A manager’s education is particularly important to the adoption of advanced technologies such as enterprise resource planning for business administration, software-based statistical control or automated systems for quality control, and robots or additive manufacturing for advanced manufacturing. The process may also work two ways: firms that adopt more technology hire more educated managers. But a considerable share of firms (more than 30 percent in Georgia, Ghana, India, Kenya, and Senegal) view the lack of capabilities—including managers’ and workers’ skills—as an important barrier to technology adoption, suggesting that having better-educated managers may be a prerequisite for greater technology adoption. Figure B5.1.1 Better-educated managers are more likely to adopt technology in middle-income countries 0.5 Regression coefficient of technology adoption 0.4 0.3 0.2 0.1 0 Top manager Top manager with BA+ studied abroad General business function Sector-specific business function Source: WDR 2024 team based on data from the Firm-level Adoption of Technology (FAT) survey. Note: The figure shows coefficients from regressing a technology adoption index (extensive index, 1–5) on the independent variables shown in the figure, controlling for fixed effects by country, sector, and firm size (small/medium/large). The figure covers the following middle-income countries: Bangladesh, Brazil, Cambodia, Georgia, Ghana, India, Kenya, Senegal, and Viet Nam. BA = bachelor’s degree. a. The World Bank–administered FAT survey provides firm-level data on the adoption of more than 300 technologies across approximately 50 business functions. These include general business functions that are common to all firms and sectors, such as business administration, operations planning, sales, and quality control. They also include sector-specific business functions that vary across sectors. For food processing, for example, functions include input testing; mixing, blending, and cooking; antibacterial procedures; packaging; and storage. For further information, see Cirera, Comin, and Cruz (2022). Preservation | 113 Figure 5.2 Higher inequality is associated with higher intergenerational immobility 1.0 Intergenerational relative immobility in schooling 0.8 0.6 0.4 0.2 0 0.2 0.4 0.6 0.8 1.0 Schooling inequality in parents’ generation (Gini coefficient) Low-income countries Middle-income countries High-income countries Source: WDR 204 team based on 2018 data from GDIM (Global Database on Intergenerational Mobility) (dashboard), Data Catalog, World Bank, Washington, DC,  https://datacatalog.worldbank.org/search/dataset/0050771/global​ - database​ -on-intergenerational-mobility. Note: The y-axis depicts the intergenerational relative immobility, which is the slope estimate from a regression of children’s schooling on parents’ schooling. The estimate shows the extent to which children can pull away from the shadow of their parents. The x-axis measures the Gini coefficient of years of schooling for the parents’ generation. Social mobility is key to enabling growth, social mobility arises because a country with particularly more equitable and inclusive higher mobility is able not only to produce more growth. Countries with greater social mobil- talent but also to generate more skilled jobs. ity have more skilled workforces. Data reveal More advanced economies are also better able that if two countries have the same income to ensure that more individuals, regardless of per capita but one has more social mobility their parents’ circumstances, have better oppor- than the other from one generation to the next tunities to become skilled workers. And social (intergenerational mobility), the country with ­ mobility matters much more in middle-income greater intergenerational mobility has a higher countries than in low-income countries simply share of skilled workers. This positive associa- because the former need more skilled workers tion between the share of skilled workers and to invest, infuse, innovate, and grow (figure 5.3). 114 | WORLD DEVELOPMENT REPORT 2024 Figure 5.3 Intergenerational mobility of on innovation has moved toward even more spe- skilled workers matters more for middle- cialized education, exemplified by the pursuit of income countries than for low-income advanced degrees such as PhDs.9 When deciding countries how much to invest in their own human c ­ apital— or that of their children—individuals weigh the expected returns of those investments against HICs their costs, largely based on societal norms and UMICs rules that determine an individual’s access to the relevant markets and services. LMICs When these norms and rules are biased—for example, in favor of the wealthy and the elite— LICs they restrict access to opportunity. Inequitable access to opportunity exacerbates social immobil- −0.2 0 0.2 0.4 0.6 ity, which preserves the existing social hierarchy, Regression coefficient of skilled workers’ perpetuating inequality. Social immobility, in the intergenerational mobility aggregate, holds back the energies that drive the Source: WDR 2024 team estimates based on GDIM (Global forces of creation. Database on Intergenerational Mobility) (dashboard), Data Catalog, World Bank, Washington, DC, https://datacatalog​ .worldbank.org/search/dataset/0050771/global-database-on​ Elite pacts hinder learning, -intergenerational-mobility; WDI (World Development Indicators) (DataBank), World Bank, Washington, DC, https://databank​ employment, and entrepreneurship .worldbank.org​/­source/world-development-indicators. Networks facilitate the formation of groups Note: The figure plots regression coefficients of that can determine access to jobs, services, and intergenerational mobility (which is equal to 1 minus the intergenerational relative mobility) for different country policy making. Better-educated and wealthier groups at the 95 percent confidence interval. The dependent parents have broader social networks to assist variable in the regression is the share of skilled workers in their child’s job search.10 Networks also (“Legislators, sr. officials, managers”; “Professionals”; “Technicians and associate professionals”). The regression matter for entrepreneurship. Because better-­ controls for the log of gross domestic product (GDP) per educated and wealthier entrepreneurs have capita when the 1980s birth cohort was growing up. access to wider and better social networks than Intergenerational mobility estimates are for educational mobility of the 1980s cohort from the World Bank GDIM. others, they have greater access to opportuni- HICs = high-income countries; LICs = low-income countries; ties and credit. LMICs = lower-middle-income countries; UMICs = upper- middle-income countries. Networks keep outsiders out Networks and group memberships based on parental and family ties can secure access to jobs, Elite pacts perpetuate social public services, and political power for individu- immobility and preserve the als from wealthy and close-knit groups. Such tight social groups are common in many countries. status quo Although these social relationships are often Creative destruction requires talent; individuals, instrumental in building trust and facilitating in turn, need opportunities to develop their tal- business transactions, they also create unequal ents and an expectation that investing in such playing fields, limiting opportunities for those talent will improve their lives. In the early twen- outside the network. And they keep outsiders out. tieth century, a high school or college degree In fact, the majority of people in many countries held the greatest potential for fostering inven- believe that social connections, mostly through tions.8 However, in today’s world marked by ever-­ family, are a key to success, as opposed to personal advancing and complex technologies, the focus effort, grit, or talent.11 Preservation | 115 In the Middle East, an implicit social con- When official credentials (such as degrees) convey tract known as wasta obliges those within the little information about a job candidate’s skills group—typically a tribal group—to aid others and personal traits, recruitment through social from the group. The use of wasta is common when networks helps employers lower the risk of choos- searching for a job, procuring a driver’s license or ing unqualified candidates. But because social business license, gaining admission to a university, ­ networks are mostly defined by a person’s socio- and performing many other day-to-day transac- economic background and where they grew up, tions. Compared with other individuals, those they can worsen existing inequalities. Research who have access to a wasta obtain more favorable shows that new technologies, particularly when rulings from agencies and courts, are more likely introduced by entrants, can foster social mobil- to obtain government contracts, and benefit more ity, although this effect diminishes in economies from government rules that limit competition.12 where incumbents spend more time and money Other examples of such social contracts include on lobbying activities.16 guanxixue in China, blat in the Russian Federation, In lower- and middle-income countries, compadrazgo in Latin America, and the “old boys between 40 and 80 percent of workers find jobs network” in Western countries. through social networks.17 Hiring through social A situation in which only well-connected indi- networks can lead to discrimination against indi- viduals obtain rewarding jobs, irrespective of skills viduals without access to high-quality networks. and talent, can have profound negative effects on the One reason is people’s tendency to associate with incentives for outsiders to attend school and even to others of similar backgrounds or characteristics.18 perform well in school. Cronyism and corruption in In Malawi, for example, men systematically refer education lead to lower academic achievement for a fewer women than men to jobs.19 given level of public spending on education, thereby Social connections matter to entrepreneurial lowering the efficacy of such expenditures.13 success—and the connections that matter most Networks also facilitate the formation of inter- are often parents and extended families. Social est groups, which block the entry of new actors. For and ethnic networks help the next generation of example, in public school systems there is often a entrepreneurs by facilitating access to credit.20 symbiotic relationship between teachers’ unions They also help with enforcing contracts, provid- and political leaders, which can hamper children’s ing operational support, and developing further achievement. When in office, political leaders pro- connections. Because many entrepreneurs in vide teachers with benefits such as higher salaries. developing countries lack a legal and institutional Teachers then pay union fees from their salaries, framework within which to operate, they rely and unions contribute a portion of this revenue instead on their kinship networks. to politicians’ campaigns. Although replacing the Developing social networks from scratch is lowest-performing teachers with more effective costly for entrants to any circle, which strengthens ones would improve children’s achievement, the the status quo—the forces of preservation—and often politically powerful teachers’ unions typi- leads to persistent social inequality. Furthermore, cally block efforts aimed at improving education the persistence of social networks is resistant to quality.14 This occurs in many middle-income shocks. countries, including Indonesia and Mexico.15 An inability to trust people and institutions beyond one’s own family and social network can Connections in job recruitment and starting a limit firm growth and productivity. In develop- business can worsen existing inequalities ing countries, firm owners generally make major Job-seekers and firms alike rely on social networks management decisions themselves because they for recruitment. Social networks provide job-­ fear the consequences of delegating to their seekers with information about job opportuni- managers. But because their time and talent are ties, access to hiring managers, and other support. limited, owners are compelled to manage firms 116 | WORLD DEVELOPMENT REPORT 2024 through their children. Consequently, the num- urban neighborhoods is high in Cape Town, and ber of male children emerges as one of the best overall in South Africa, and mobility between predictors of firm size, thereby impeding the generations is quite low. Cities where consump- growth and profitability of their firm.21 In India, tion is more unequal across neighborhoods—due, for example, this factor underlies firms’ inabil- for example, to income differences across neigh- ity to grow. Its effect is sizable; poor delegation borhoods—are located in countries with fewer of managerial responsibilities could account for opportunities for social mobility. 11 percent of the difference in income per capita By shaping the laws, regulations, and rules between India and the United States.22 determining who lives where, the social elite keep outsiders out—relegated to rural areas or ­d isadvantaged neighborhoods. A dominant Insiders keep outsiders out of mechanism is setting urban planning standards sight in distant and disadvantaged that are unaffordable for outsiders.29 When the neighborhoods US city of Philadelphia was settled, for exam- Where a person is born, grows up, and lives shapes ple, city authorities set a minimum lot size of that person’s access to opportunities and aspira- about 30 square meters. By contrast, minimum tions. In the United States, children of similar lot sizes in Ethiopia range from 75 to 300 square family incomes raised in nearby neighborhoods meters. with different postal codes may have vastly dif- Even in accessing finance for housing, there ferent chances of succeeding as adults.23 This is is a long history of discrimination. In the United equally true in middle-income countries, where States, the Home Owners Loan Corporation social mobility is typically limited to only some (HOLC) drew maps in the 1930s for more than geographic areas. Individuals who live in areas 200 cities as part of its City Survey Program to with high levels of poverty fare much worse than document the relative riskiness of lending in others on a wide range of economic, health, and neighborhoods. Risk factors included race, eth- educational outcomes.24 nicity, and immigration status. The lowest-rated In middle-income countries, income inequal- neighborhoods, most of whose residents were ity is typically higher in urban areas than in rural African American, were drawn in red. Borrowers areas because cities attract both highly skilled from these “redlined” neighborhoods were and unskilled workers. But cities also have higher denied access to credit due to the demographic social mobility because they offer more oppor- composition of their neighborhoods. For more tunities to develop talent than rural areas.25 For than two decades, the redlining in effect barred example, since 1950 social mobility in Latin African Americans from buying homes in attrac- American cities has been higher for larger cities,26 tive neighborhoods, even when they could afford although not for all cities. In Brazil, dynamic cit- them, and kept their home values low. ies such as São Paulo and Rio de Janeiro offer scant Although this discriminatory practice was opportunities to poor children.27 In India, social banned by the Fair Housing Act in 1968, the mobility has improved greatly in urban areas in effects still linger today. Neighborhoods that were recent decades,28 and has delivered larger gains for formerly redlined fare worse in terms of housing the disadvantaged than in rural areas during the value, homeownership rates, racial composition, period following India’s economic liberalization. and exposure to pollutants. The inability to own Cities can be formidable engines of social a home prevented those discriminated against mobility, but cities in middle-income countries from generating home equity, the main source are less socially mobile than cities in high-income of wealth for most American households and countries. Further, they have greater inequal- the major source of inherited (intergenerational) ity (figure 5.4). For example, inequality between wealth.30 Preservation | 117 Figure 5.4 High inequality within cities is associated with low social mobility from one generation to the next 0.9 0.8 Country-level mobility index Almaty 0.7 Dar es Salaam 0.6 Bangkok Tehran Mexico City Madrid 0.5 Rio de Janeiro Santiago Bogotá 0.4 Cape Town 0.3 0.2 0.2 0.3 0.4 0.5 0.6 Intraurban inequality index Sources: GDIM (Global Database on Intergenerational Mobility) (dashboard), Development Research Group, World Bank, Washington, DC, https://www.worldbank.org/en/topic/poverty/brief/what-is-the-global-database-on-intergenerational​ -mobility-gdim; Park et al. 2022. Note: The mobility index is defined as 1 minus the correlation between parents’ and children’s years of schooling. The intraurban inequality index is defined as the Gini coefficient based on consumption data from the Global Monitor Database (GMD) surveys, matched with administrative area and GHS-Urban Center Database (GHS-UCDB) data. See Aron et al. 2023; GHSL (Global Human Settlement Layer) (dashboard), Joint Research Center, European Commission, Ispra, Italy, https:// ghsl.jrc.ec.europa.eu/index.php. Distant and disadvantaged neighborhoods outcomes stem from the neighborhood itself (that restrict upward social mobility is, living in certain neighborhoods).33 The bene- Living in a disadvantaged neighborhood affects fits of moving to a better neighborhood are larger adults’ outcomes and children’s trajectories for younger children because they are exposed to because residential segregation impairs school- beneficial effects for a longer period. ing, health outcomes, intergenerational mobility, Neighborhoods matter for children’s life and the formation of social capital.31 outcomes. First, richer neighborhoods tend to Children who grow up in better neighborhoods have higher school quality, a major determi- have improved outcomes in their education and in nant of upward social mobility.34 Differences in the labor market.32 These outcomes could result ­ neighborhood-level school quality can arise when from being born into a family that would choose schools are financed locally, which is more common to live in these better neighborhoods (sorting) or in higher-income countries. In ­ middle-income could be attributable to the neighborhood itself. countries, central governments are responsible In developing countries, about one-third of these for the large part of the public expenditure on 118 | WORLD DEVELOPMENT REPORT 2024 education, although in many countries provin- In addition, violence may disrupt school routines, cial governments also contribute. In South Africa, increase teacher and student absenteeism, and for example, more school funding is allocated to cause major psychological distress that can lower poorer neighborhoods. Yet the quality of schools test scores for students exposed to violence. In is lower in poorer neighborhoods in cities and in the United States, children who move to a county more remote rural areas, where schools struggle with lower crime rates, lower concentration of to attract and retain high-quality teachers, doc- poverty, less income inequality, stronger schools, tors, and other service providers because of poorer and a greater share of two-parent households infrastructure, services, and amenities. Having experience better outcomes. For example, moving high-quality teachers not only improves test a child out of public housing in the United States scores but also can influence important noncog- to an area with a low poverty level when the child nitive and behavioral attributes in positive ways. is young using a subsidized voucher has been esti- The absence of high-quality teachers contributes mated to increase the child’s total lifetime earn- to the poor performance of schools in disadvan- ings by about US$302,000.40 taged neighborhoods.35 Second, children find their peers and role mod- In disadvantaged neighborhoods, occupational els in their neighborhoods and form the social choices are limited networks that can help them in their future job Nearly one in six people around the world lives search. Better neighborhoods help in all these in urban slums, areas characterized by inade- aspects.36 For example, the probability of drop- quate infrastructure and property status. The ping out of school or committing crimes is similar largest slums—Khayelitsha in Cape Town, Kibera among children who attend the same school or in Nairobi, Dharavi in Mumbai, Ciudad Neza in grow up in the same neighborhood (and are pre- Mexico City, and Orangi Town in Pakistan—are sumably peers). Similarly, children in poor neigh- located in some of the largest cities in middle-­ borhoods may have strong ties with friends and income countries and form their own towns. neighbors, but these are of little use in searching Many migrants settle in slums in search of better for a job because they do not include contacts economic opportunities and intend to stay there with people outside the community.37 temporarily, yet often remain there for decades.41 Third, poor neighborhoods typically have poor For some individuals, slums are a “social eleva- infrastructure and services. Lack of sanitation tor”—a temporary stop before finding regular and greater exposure to pollution are common in housing.42 For others, slums are a poverty trap urban slums. As a result, children growing up in they cannot escape. Slum dwellers face risks from slums are more susceptible to diseases. All these criminal gangs, contagious diseases, and pollu- factors have negative impacts on early-life health, tion, and often struggle with long commutes and human capital, and labor market outcomes.38 relatively high housing costs. Fourth, poorer neighborhoods tend to have a In urban slums in middle-income countries, high incidence of crime and violence. For example, children can often access education opportuni- young children growing up in one of the numer- ties, but still have limited job opportunities. In the ous slums ( favelas) in Rio de Janeiro affected by slums in Bangalore, India, parents’ top priority has conflicts between drug gangs perform signifi- been investing in their children’s education, which cantly worse at school.39 Between 2003 and 2009, has led the children to have higher education lev- at least one favela was in a drug-related conflict els than their parents.43 Although most families in Rio de Janeiro during four of those six years. experience gains in income and assets over time, Living in such poor neighborhoods in middle-­ longer-term residents (extending to a fourth gen- income countries can carry a social stigma that eration) have not been able to move out. In the affects life outcomes for the residents (box 5.2). slums in Jakarta, Indonesia, intergenerational Preservation | 119 Box 5.2 Living in favelas makes it more difficult to get a job In her book Favela: Four Decades of Living on the Edge in Rio de Janeiro, Janice Perlman documented the experiences of families living in Brazilian slums ( favelas) for more than 30 years.a The respondents were interviewed in 1969 and again in 2001, when the original interviewees and their children and grandchildren were asked for their perceptions about why they were the targets of discrimination. The reasons most frequently mentioned by the original interviewees and their children for not getting a job were their residence in favelas, followed by skin color, their appearance, and being a migrant (figure B5.2.1). Grandchildren, on the other hand, perceived less discrimination than their grandparents based on skin color or migrant status, although they perceived living in favelas and their own appearance as major impediments to obtaining jobs. Figure B5.2.1 Slum residents in Rio de Janeiro identified their residence in a favela as the largest impediment to getting a job Impediments to employment Birthplace outside Rio de Janeiro Appearance Skin color Residence in favela 0 10 20 30 40 50 60 70 80 90 Percent Original interviewees Children Grandchildren Source: Perlman 2010. Note: The figure shows the responses of residents of favelas in Rio de Janeiro surveyed in 1969 and 2001. a. Perlman (2010). mobility is higher among younger children than Most slum residents—particularly women— older children because they have benefitted from work in the slums and cannot obtain formal jobs recent improvements in educational mobility.44 because of their lack of access to job networks and Educational mobility is also relatively high in their isolation from city centers. Moreover, slum favelas in Rio de Janeiro, Brazil.45 However, edu- residents are highly vulnerable to adverse events cational mobility does not translate into higher such as spells of bad health.46 Overall, residents in cities. occupational mobility in any of these three ­ the Bangalore, Jakarta, and Rio de Janeiro slums 120 | WORLD DEVELOPMENT REPORT 2024 are neither stuck in poverty traps nor are they on a ­ arriers can include caste boundaries in India; gov- b steady trajectory toward the middle class.47 Their ernment regulations such as hukou in China and main constraint is securing a better job, particu- propiska in Central Asia; and welfare schemes tied larly outside of their own neighborhoods. to residence.49 In China, the hukou system has his- torically imposed large costs on working and living Keeping migrants out misallocates talent outside of one’s hukou location, primarily through Internal migration allows individuals to meet restricted access to social services and limited and learn from more productive people, sell their employment rights.50 In 2000, the average cost of ideas in better markets, and expand job oppor- moving from rural to urban areas within a Chinese tunities, thereby contributing to a more efficient province was equivalent to reducing one’s real allocation of workers across an economy.48 By income by a factor of nearly 3; moves between prov- contrast, barriers to internal migration are costly inces were even costlier.51 Between 2000 and 2005, for growth. The low internal migration rate in following a reform of the hukou system, migration some middle-income countries suggests the pres- costs declined by 18 percent on average and by ence of high mobility barriers, even among highly about 40 percent for moves between provinces. educated individuals (figure 5.5). Location preferences or discrimination can The lack of information or social networks in also limit migration opportunities.52 On average, the destination, as well as market or policy distor- individuals with a tertiary education (univer- tions, can limit migration opportunities. Migration sity or beyond) face much lower migration costs Figure 5.5 In many middle-income countries, movement of workers from one part of the country to another is more limited than in high-income countries such as France and the United States 45 Internal lifetime migration rate 40 35 of the employed (%) 30 25 20 15 10 5 0 p. o e a in a a ile il a a es az nc in si in ric bi ic a Re Ch at ne Ch nt ex Sp m Br a Af St Fr ab ge lo M do h Co d Ar Ar ut In ite So t, Un yp Eg Middle income: Nontertiary-educated Tertiary-educated High income: Nontertiary-educated Tertiary-educated Source: WDR 2024 team based on data of IPUMS International (Integrated Public Use Microdata Series, Harmonized International Census Data for Social Science and Health Research) (dashboard), Minnesota Population Center, University of Minnesota, Minneapolis, https://international.ipums.org/international/. Note: The sample includes China (2000), Brazil (2000), Spain (2001), the Arab Republic of Egypt (2006), Colombia (2005), Indonesia (2010), Argentina (2001), Mexico (2000), Chile (2002), South Africa (2001), France (1999), and the United States (2000). Internal lifetime migration is defined as current residence different from residence at birth within the same country. Tertiary education refers to schooling at the university level or beyond. Preservation | 121 Figure 5.6 In many middle-income countries, migration costs are higher for individuals without high levels of education 50 % wage increase needed to offset 45 40 35 moving costs 30 25 20 15 10 5 0 o e es ile a in a a a il a az nc si in in ric bi ic a Ch at ne nt ex Ch Sp m Br a Af St Fr ge lo M do h Co d Ar ut In ite So Un Middle income: Nontertiary-educated Tertiary-educated High income: Nontertiary-educated Tertiary-educated Source: Census data of IPUMS International (Integrated Public Use Microdata Series, Harmonized International Census Data for Social Science and Health Research) (dashboard), Minnesota Population Center, University of Minnesota, Minneapolis, https://international.ipums.org/international/. Note: Moving costs refer to the percentage of wage increase at the destination needed to compensate for the cost of moving. The sample includes Chile (2002), France (1999), South Africa (2001), the United States (2000), Spain (2001), Argentina (2001), Brazil (2000), Mexico (2000), Indonesia (2010), Colombia (2006), and China (2000). Tertiary education refers to schooling at the university level or beyond. than individuals with less education (figure 5.6). Learning poverty is alarmingly high in For example, on average, migrants with a tertiary middle-income countries. In the median lower-­ education in China need a 39.5 percent wage middle-income country and upper-middle-income increase to compensate for their moving costs, country, only 31 and 63 percent of children ages while migrants who lack a tertiary education need 10 or younger, respectively, are able to under- a 45.1 percent wage increase.53 stand a text relative to 94 percent in the median high-­income country.54 Among 15-year-olds, A missing opportunity: Education only half of high school students are proficient in Although norms, networks, and neighborhoods math, reading, and science in the median upper-­ contribute to preserving the status quo, policy middle-income country and 30  percent in the can disrupt them and unleash creation and social median lower-­ middle-income country, relative mobility. A critical policy is expanding quality to 80 percent in the median high-income coun- education, as it represents for many the best—and try.55 Low shares of young people are enrolled in perhaps the only—hope to climb the social ladder. higher education (18  percent and 45  percent in Education systems that promote human capital the median lower-middle-­ income country and accumulation are therefore key to disrupting the upper-middle-income country, respectively, rel- status quo, and yet middle-income countries have ative to 70 percent in the median high-income largely failed at building those systems. country).56 And even lower shares have graduated 122 | WORLD DEVELOPMENT REPORT 2024 from higher education: 12 percent in the median within  societies. Patriarchal social norms per- lower-middle-income country, 28 percent in petuate gender inequality. They hold back the median upper-­ middle-income country, and women—out of the labor force, out of education, 43 percent in the median high-income country.57 out of skills, out of luck, and out of hope. Men, The education system failures are particularly who benefit from more access to resources and acute for disadvantaged students. Gender, location, opportunities, have the most incentives to protect and wealth create large and worrisome access and the status quo. Other norms keep marginalized completion gaps in elementary and secondary edu- groups down—and are shaped by the social elite. cation.58 In higher education, the poorest students These unequal norms and beliefs can be and those in rural areas are much less likely than deeply ingrained in a nation’s social fabric and others to complete at least two years of higher edu- exert a powerful influence on individual actions. cation.59 These gaps are so large that the percentage Norms can strongly influence the behaviors and of individuals who have completed at least two years choices of caregivers and parents—often not would rise by about 30–40 percent if location gaps treating their daughters on a par with their sons. were eliminated and would double if wealth-related These norms define a child’s access to educa- gaps were eliminated.60 tion, liberty, employment, and entrepreneurship. Furthermore, higher education contributes Furthermore, parents pass on cultural norms to social mobility only if it provides skills that to their children, perpetuating and reinforcing are effectively rewarded in the labor market. inequality, whether based on gender, race, ethnic- Educational institutions can identify and fulfill ity, or religion. These norms hold back a country’s the skill needs of the economy by connecting with growth and development. enterprises and the labor market, and yet this link is often broken in middle-income countries. In a Girls: Starting to show up in school World Economic Forum executive survey, when Improving women’s educational attainment companies are asked to rank the skills of higher ensures that economies can expand their tal- education graduates relative to their needs (on ent pool. Therefore, expanding the middle class a scale of 1–7), the average score is 4.13 in the requires providing the needed skills and compe- median middle-income country, well below 5.03 tencies to all members of society. in the median high-income country. Moreover, In the area of ensuring access to basic higher education in middle-income countries pro- education, remarkable progress across the duces relatively few graduates in fields typically world during the past two decades has reduced supportive of infusion and innovation—engineer- (and in some countries, even eliminated) ing, information and communication technology, gender gaps in enrollment and educational science, and health. Similarly, a low share of higher attainment. In some countries, the higher education students (15 percent in the median educational attainment of women has resulted ­ middle-income country) are enrolled in short-­ in an improvement in mobility from one cycle programs (two or three years long), which generation to the next (intergenerational provide the technical skills needed to engage in mobility)—that is, daughters can move up midlevel knowledge-­ intensive occupations.61 the educational ladder even if their mothers or parents were lower down that ladder.62 Patriarchal gender norms Despite this improvement, in many countries hold back a large proportion higher educational attainment for successive generations is still lower for women and of the population disadvantaged groups (figure 5.7),63 although Norms are the unwritten rules and shared not in Sub-Saharan Africa and Latin America expectations that govern human behavior and the Caribbean.64 Preservation | 123 Figure 5.7 There is a substantial gap Figure 5.8 Female labor force participation between low- and high-income countries in is low in the Middle East and North Africa and female educational attainment in South Asia 80 60 least upper-secondary education (%) 69.3 Share of women age 25+ with at Female labor participation, 70 50 60 ages 15−64 (%) 52.5 40 50 40 30 30 27.0 20 20 10 10 6.4 0 0 1960 1980 2000 2020 LICs LMICs UMICs HICs East Asia and Pacific Source: WDR 2024 team. Europe and Central Asia Note: The figure illustrates educational attainment of Latin America and the Caribbean women 25+ years of age for at least upper-secondary education. The data cover 141 countries. HICs = high- Middle East and North Africa income countries; LICs = low-income countries; LMICs = South Asia Sub-Saharan Africa lower-middle-income countries; UMICs = upper-middle- income countries. Source: WDI (World Development Indicators) (DataBank), World Bank, Washington, DC, https://databank.worldbank​ Women: Missing at work .org/source/world-development-indicators. Female labor force participation is low in sev- Note: Data are averages of national estimates for middle- income countries in each region. eral middle-income countries, particularly in the Middle East and North Africa and in South Asia returns to talent, education, and experience. (­ figure 5.8). By contrast, female labor force par- Compounding the problem, fewer higher-skilled ticipation has increased in many countries over jobs are available in middle-income countries— the last decades (figure 5.9). In 1990, the Republic for men and for women—than in high-income of Korea had the same level of GDP per capita (in countries, and women in middle-income coun- terms of purchasing power parity, PPP) as India tries have relatively less access to them than in 2020. However, the female labor force par- women in high-income countries. However, in ticipation rate in Korea was about 51 percent in some middle-income countries such as Indonesia, 1990, while India’s was 30 percent in 2020 but has the share of women in professional occupations improved in recent years.65 In the Arab Republic of has grown rapidly in recent decades (figure 5.10). Egypt, India, and Türkiye, female labor force par- In some lower-­ middle-income countries, such as ticipation is well below what would be expected Egypt, the share has grown as well but remains given their levels of income per capita, whereas the low, indicating a large talent misallocation. rates are much higher in Indonesia and Malaysia. In high-income countries and middle-income Even when women are employed, they countries alike, women are less likely than men are more  likely to work in lower-paid jobs or to enroll in science, technology, engineering, and be self-employed (and thus have unpredict- mathematics (STEM) fields.66 The share of women able incomes) than men in these types of jobs. among STEM graduates is on average higher in They are much less likely than men to work in India compared with even developed countries, higher-paid jobs such as professional, manage- ­ and yet women’s representation in prestigious rial, and technical positions, which have high colleges lags behind. In 2016, for example, only 124 | WORLD DEVELOPMENT REPORT 2024 Figure 5.9 Female labor force Figure 5.10 The share of female participation has evolved differently professionals has risen in some countries but across countries not others 80 60 participation, ages 15−64 (%) 70 Share of females in highly skilled occupations (%) 50 60 Female labor force 50 40 40 30 30 20 20 10 0 10 1 5 10 20 40 60 1990 2000 2010 2019 GDP per capita (US$1,000, log scale) China Egypt, Arab Rep. Brazil, 1970–2010 Chile, 1960–2002 India Indonesia Egypt, Arab Rep., 1986–2006 India, 1983–2009 Japan Korea, Rep. Indonesia, 1971–2005 Mexico, 1960–2015 Malaysia Türkiye South Africa, 1996–2007 United States, 1960–2015 Source: WDI (World Development Indicators) (DataBank), Source: WDR 2024 team based on census data of IPUMS World Bank, Washington, DC, https://databank.worldbank​ International (Integrated Public Use Microdata Series, .org/source/world-development-indicators. Harmonized International Census Data for Social Science and Health Research) (dashboard), Minnesota Population Center, University of Minnesota, Minneapolis, https:// international.ipums.org/international/. about 8  percent of students admitted to India’s Note: Professional occupations combine the top three prestigious Indian Institutes of Technology (IITs) 1-digit International Standard Classification of Occupations were women, compared to 46 percent admitted to (ISCO) occupation codes (“Legislators, sr. officials, the Massachusetts Institute of Technology (MIT) managers”; “Professionals”; “Technicians and associate professionals”). GDP = gross domestic product. in the United States. The IITs went on to establish a female enrollment target of 20 percent, which was achieved in the 2023 cohort. Women: Missing independence in owning Some middle-income countries educate rel- property, opening financial accounts, and atively more females in STEM fields than high-­ running businesses income countries, and yet they employ relatively The gender gap extends to formal entrepreneur- fewer.67 Why would women pursue STEM fields ship. Women are more likely than men to work but not work in them? Recent evidence suggests a as subsistence microentrepreneurs and earn possible driver: higher returns to a STEM educa- lower profits than male microentrepreneurs. tion in the marriage market. In Pakistan, female Furthermore, women are less likely than men to physicians are considered “trophy brides” in the work in formal firms. Globally, only 23 percent of marriage market. More than 70 percent of gradu- businesses are female-owned, with large variation ates of medical school are women in Pakistan, and across sectors (figure 5.11). Female-owned busi- yet only 23 percent of them practice their pro- nesses are more egalitarian employers: although fession after they graduate.68 Similarly, in Egypt male-owned firms employ few female employ- returns to higher education—not just in STEM ees (25 percent) and even fewer female managers fields—are much higher in marriage markets than (6  percent), female-owned firms tend to employ in labor markets.69 males and females equally. Preservation | 125 Figure 5.11 Globally, women own a smaller share of firms than men Share of firms that are female-owned (%) 40 30 20 10 0 em ale W r p eum el nd an sm s he th M n s an rod ts an ct s al se cts od s ts s rm ls to ts rv s Te es au g an ser g IT tal ctu g on an pro ring tru er r d bli rt el rub l od ec be nt M ufa ice pr ice ho ic ile Se ent i io st in r m r nin M ufa urin Ga ica an nd eta o lit p uc uc d uc ns Oth ic W ron Ch les ct m d I du sp i ra re sh Fo xt ur v lic rv ta od od l i ry s a R tro t r Pe an u T r et T p e Co he d e o y at Ot O ne ic ot a hi st Le H ing s ac la pi M P t in os N Pr H Sources: Chiplunkar and Goldberg 2021; WBES (World Bank Enterprise Surveys) (dashboard), World Bank, Washington, DC, https://www.enterprisesurveys.org/en/enterprisesurveys. Note: The figure shows the share of female-owned firms across 25 sectors, as well as the confidence intervals for the 95th percentile. The WBES sample covers 141 countries across 13 years (2006–18). IT = information technology. The gender gap in access to financial accounts obstacle, and that there is an estimated US$1.7 (such as formal and mobile banking) is also still trillion global financing gap.72 very large (figure 5.12). Even in countries such as One barrier to women gaining expanded access Bangladesh and Nigeria, where mobile phone and to finance is the continued restrictions they face in mobile banking penetration have been impres- asset ownership. The property rights of women and sive, there are still large gender gaps in finan- disadvantaged groups—manifested, for example, cial inclusion. In Morocco, the gap is more than through property and inheritance laws—also vary. 25  percentage points, whereas the gap does not Across the world, 40 percent of economies still con- exist in Sweden. strain women’s property rights, denying them equal Indeed, a 10 percent or higher gender gap in access to essential resources for financial security account ownership persists in 41 countries, and and economic independence.73 Among middle-­ women are 37 percent less likely than men to have income economies, 14 percent do not grant women an account in fragile and conflict-affected situa- equal ownership of immovable property such as tion countries.70 Women are also less likely than real estate or land; 24 percent have unequal inheri- men to own a debit or credit card, have borrowed tance rights that favor sons over daughters; 25 per- from a formal financial institution, or have bor- cent do not grant equal inheritance rights to male rowed to support a farm or business.71 It is thus and female surviving spouses, further marginaliz- not surprising that women-led businesses are ing women’s and girls’ economic empowerment and more likely to identify access to finance as a major autonomy; and 34 percent still do not recognize 126 | WORLD DEVELOPMENT REPORT 2024 Figure 5.12 Women lag behind men in having financial accounts 40 30 Account ownership gap (percentage points) Morocco 20 Nigeria Bangladesh 10 Colombia China Chile 0 Sweden –10 50 100 GDP per capita (constant PPP 2017 US$1,000, log scale) Sources: Global Findex (Global Financial Inclusion Database), World Bank, Washington, DC, https://globalfindex.worldbank​ .org/#data_sec_focus; WDI (World Development Indicators) (DataBank), World Bank, Washington, DC, https://databank​ .worldbank.org/source/world-development-indicators. Note: The figure shows the absolute difference (in percentage points) in 157 countries between the share of men with a financial account and the share of women with a financial account. Accounts include those in financial institutions and mobile accounts. GDP = gross domestic product; PPP = purchasing power parity. nonmonetary contributions, including caring for However, there is still room for progress: if the minor children or taking care of the family home, chance of becoming an innovator in the United undermining women’s crucial roles in caregiving States today were as high for women, minorities, and domestic responsibilities.74 and children from low-income families as for men Gender inequality is a major barrier to socio- from high-income families, innovation in the US economic mobility and a growing middle class. economy would increase fourfold.76 Box 5.3 outlines a program of research to examine The forgone growth in some middle-income the economic growth lost to gendered barriers. countries is likely much larger than in the United States. For example, removing barriers to entrepre- neurship for women in India would double female The cost of social immobility labor force participation and raise real income by and preservation: Holding 40 percent.77 And globally, closing the gender gap in back the energies that drive employment and entrepreneurship could raise the global GDP by more than 20 percent.78 Eliminating creation the gender gap over the next decade would essentially Elite pacts and patriarchal gender norms maintain double the current global economic growth rate. the status quo, stunting and misallocating talent. High migration costs prevent workers from These are costly missteps. In the United States, locating where they are most productive. In for example, the reduction of gender and racial Indonesia, reducing migration costs to levels sim- barriers in educational and occupational choices ilar to those in the United States would lead to a between 1960 and 2010 explains 20–40 percent of 7.1 percent boost in productivity.79 The reform of the observed economic growth over that period.75 the Chinese hukou system in 2003 led to a 5 percent Preservation | 127 Box 5.3 Global Gender Distortions Index: Measuring economic growth lost to gendered barriers In the United States, policies that reduced labor market barriers and other forms of dis- crimination against women and African Americans contributed up to 30 percent of post– World War II economic growth. How can today’s middle-income countries evaluate the economic dividends of progress toward equal opportunity and improved talent allocation? How can policy makers identify specific barriers within their labor markets that need to be addressed and given priority? Researchers at Yale University’s Economic Growth Center, working under the Gender and Growth Gaps project, are developing a Global Gender Distortions Index (GGDI) to measure the losses in global economic growth stemming from gender gaps in the labor market. The GGDI links changes in gender gaps in the labor market to produc- tivity growth through improvements in the allocation of women’s talent. Specifically, this index measures by how much the gross domestic product (GDP) of a country (or subnational unit) has grown, or could grow, from improvements in women’s labor market opportunities. The index highlights that women often do not choose the occupation in which they have a comparative advantage because of (1) labor demand distortions that lead to a wedge between wages and marginal products and (2) differences in occupational preferences that capture factors such as social norms and other labor supply distor- tions. The GGDI is computed by using observed differences in women’s wages, labor supply, and employment across job type (formal versus informal) to derive an estimate of economywide productivity losses or gains. By quantifying growth losses stemming from gender inequality and distilling them into a single measure, the GGDI allows comparisons across time and locations that can inform policy decisions. It can also complement the World Bank’s Women, Business and the Law (WBL) index by measur- ing the aggregate consequences of de facto labor market barriers, whereas the WBL measures de jure barriers.a In a proof-of-concept exercise, the GGDI team uses a cross-sectional analysis across Indian states for 2018 and finds that labor demand distortions are negatively related to state-level economic development.b Poorer states such as Bihar gain 10 percent in state GDP from removal of labor demand distortions, whereas richer states such as Kerala gain 4 percent in GDP. By contrast, labor supply distortions are not related to state-level GDP. The GGDI, which will be computed for 30 countries over the next 24 months, can act as a dynamic barometer for countries and regions, providing researchers and policy makers with a valuable new resource. a. World Bank (2024). b. Goldberg et al. (2024). 128 | WORLD DEVELOPMENT REPORT 2024 increase in labor productivity.80 Furthermore, low destruction and growth. Creative destruction mobility costs can mitigate the shocks inherent requires the development of talent, which indi- to creative destruction. During the nineteenth viduals undertake when they expect economic century, the US city of Detroit grew into a thriv- returns and social mobility. Poor expectations ing hub of commerce and industry largely based of social mobility hamper talent development. on the auto industry. But when the industry con- Similarly, inequitable opportunities in markets tracted, low mobility costs allowed workers to relo- and education hinder talent development and cate to other production hubs in the United States. social mobility, producing a talent misalloca- In middle-income countries, in addition to high tion—a waste—with costly consequences for migration costs, the costs of ineffective education individuals and countries alike. These consider- systems are high because they fail to develop tal- ations suggest two main roles for policy in cre- ent and perpetuate existing inequalities. ative destruction: removing barriers to developing The forces of preservation are holding talent and actively promoting talent development ­ middle-income countries back from creative (see chapter 8 for further details). Notes  1. de Ree et al. (2018). 17. Beaman and Magruder (2012); Caria, Franklin, and  2. Becker et al. (2018). Witte (2022); Gatti et al. (2014); Mani and Riley (2021);  3. This is due to the dynamic complementarities studied Nicodemo and García (2015). by García, Heckman, and Ronda (2023). Also see, 18. Jackson (2021). Schady et al. (2023). 19. Beaman, Keleher, and Magruder (2018).  4. The term “human capital elite” is from Becker et al. 20. Blanchflower and Oswald (1998); Evans and Jovanovic (2018). (1989); Evans and Leighton (1989).  5. Stone (1990). 21. Bloom et al. (2010).  6. Social immobility is measured by intergenerational cor- 22. Akcigit, Alp, and Peters (2021). relation between children’s and parents’ years of 23. Chetty et al. (2014); Chyn and Katz (2021). schooling. The correlation is the slope coefficient in the 24. Chyn and Katz (2021). regression of children’s schooling on parents’ years of 25. Lall et al. (2023). schooling. This measure is also known as the intergen- 26. Neidhöfer, Ciaschi, and Gasparini (2021). erational relative mobility. Other measures of intergen- 27. Britto et al. (2022). erational mobility are available such as absolute 28. Emran, Jiang, and Shilpi (2021); Emran and mobility, which indicates whether the distribution of Shilpi (2015). children’s years of schooling moved relative to that of 29. Lall, Henderson, and Venables (2017). parents. A similar concept is upward mobility. These 30. Rothstein (2017). two measures relate to overall progress in schooling, 31. Acevedo-Garcia et al. (2003); Alexander and Currie not just whether a child’s status depends on that of par- (2017); Baum-Snow and Lutz (2011); Chetty, Hendren, ents. This Report uses the relative mobility measure and Katz (2016); Chetty et al. (2022); Granovetter instead of the absolute mobility measure. Relative (1973). mobility also provides a direct estimate of inequality of 32. Chetty and Hendren (2018); Chetty, Hendren, and Katz opportunity. Social mobility is measured by 1 minus the (2016); Chyn (2018); Deutscher (2020); Laliberté (2021); intergenerational correlation in schooling. The correla- Nakamura, Sigurdsson, and Steinsson (2022). tion in schooling is used because of the lack of data to 33. Alesina et al. (2021). estimate intergenerational income mobility in develop- 34. Britto et al. (2022); Chetty et al. (2014). ing countries where a large fraction of the population is 35. Jackson (2021). self-employed. 36. Bayer, Ross, and Topa (2008); Hellerstein, Kutzbach,  7. Narayan et al. (2018); WEF (2020). and Neumark (2019); Hellerstein, McInerney, and  8. Akcigit, Grigsby, and Nicholas (2017). Neumark (2011); Topa and Zenou (2015).  9. Akcigit, Pearce, and Prato (2020). 37. Chetty et al. (2022). 10. Becker et al. (2018). 38. Chyn and Katz (2021). 11. Coco and Lagravinese (2014). 39. Monteiro and Rocha (2017). 12. Barnett, Yandle, and Naufal (2013). 40. Chetty, Hendren, and Katz (2016), looking at the Moving 13. Coco and Lagravinese (2014); Suryadarma (2012). to Opportunity system of subsidized vouchers. 14. Zingales (2014). 41. Marx, Stoker, and Suri (2013). 15. See, for example, de Ree et al. (2018); Schneider (2022). 42. Glaeser (2011). 16. Aghion et al. (2019). 43. Krishna (2013). Preservation | 129 44. Bryan and Morten (2019). 60. WDR 2024 team calculations based on UNESCO data. 45. Perlman (2010). 61. Bianchi and Giorcelli (2020) show that greater 46. See Perlman (2006) for Brazil; Turok and Borel-Saladin access to science, technology, engineering, and math - (2018) for South Africa; Ziraba, Kyobutungi, and Zulu ematics  (STEM) and vocational and technical pro - (2011) for Kenya. grams in Italy in the 1960s led to an increase in 47. Rains and Krishna (2020). patenting. For Finland, Toivanen and Väänänen (2016) 48. Akcigit et al. (2018, 2022); Lucas (2009); Lucas and find large effects of greater access to engineering Moll (2014); Perla and Tonetti (2014). master’s programs on patenting. 49. Grover, Lall, and Maloney (2022); Selod and Shilpi 62. Asher, Novosad, and Rafkin (2024); Emran and Shilpi (2021). (2015). 50. Hukou is a system of household registration used in 63. Asher, Novosad, and Rafkin (2024); Emran and Shilpi China. Under the system, each citizen is required to (2015). register in only one place of permanent residence. An 64. Alesina et al. (2021); Neidhöfer, Ciaschi, and Gasparini individual’s hukou status defines his or her rights and (2021). eligibility for social welfare and various services, 65. Estimates on India’s female labor force participation including public education and housing, within a spe- rate  vary. See, for example, India’s  Economic cific administrative unit. Survey  for  2022–23, https://www.indiabudget.gov​ 51. Tombe and Zhu (2019). .in​/economicsurvey​/doc/eschapter/echap06.pdf. 52. D’Aoust, Galdo, and Ianchovichina (2023). 66. Hammond et al. (2020). 53. Wong (2019). 67. Hammond et al. (2020). 54. Based on the Learning Poverty Index, calculated by 68. Hammond et al. (2020). World Bank and UNESCO. 69. In Egypt, it is common for the groom to transfer signifi - 55. Based on the 2018 Programme for International cant resources to the new household at the time of Student Assessment (PISA), a study administered by marriage. This is a direct and informative measure of the Organisation for Economic Co-operation and the monetary gains women may obtain through the Development (OECD), and the 2019 Trends in marriage market. Exploiting the staggered rollout of a International Mathematic and Science Study (TIMSS) school reform in Egypt that reduced the number of administered by the International Association for the years required to complete primary education from six Evaluation of Educational Achievement (IEA). to five, Deng et al. (2023) find that the return to a bride’s 56. UIS.Stat (dashboard),  Institute for Statistics, United compulsory education is about 100 percent for the Nations Educational, Scientific, and Cultural marital transfer, about 14 percent for the husband’s Organization,  Montreal,  http://data.uis.unesco.org/. wage at the time of marriage, and about 16 percent for These figures correspond to the higher education a measure of the husband’s permanent income. These gross enrollment ratio, defined as the number of indi - returns to education in the marriage market are much viduals enrolled in higher education relative to all higher than the returns to education that Egyptian age-relevant individuals (18–24 years of age). women experience in the labor market. 57. UIS.Stat (dashboard),  Institute for Statistics, United 70. Demirgüç-Kunt et al. (2022). Nations Educational, Scientific, and Cultural 71. Demirgüç-Kunt et al. (2022). Organization,  Montreal,  http://data.uis.unesco.org/. 72. IFC (2017). These figures correspond to the gross graduation ratio, 73. World Bank (2024). defined as the ratio between the number of graduates 74. World Bank (2024). from bachelor’s or graduate degree programs by the 75. Hsieh et al. (2019). population of the theoretical graduation age of the 76. Bell et al. (2019). most common bachelor’s program. 77. Chiplunkar and Goldberg (2021). 58. World Bank (2018). 78. Pennings (2022). 59. WDR 2024 team analysis using UNESCO Education 79. Bryan and Morten (2019). Inequalities Database based on household surveys. 80. Tombe and Zhu (2019). References Acevedo-Garcia, Dolores, Kimberly A. Lochner, Theresa in Developing Countries.” American Economic Review L. Osypuk, and S. V. Subramanian. 2003. “Future 111 (1): 231–75. Directions in Residential Segregation and Health Akcigit, Ufuk, Santiago Caicedo, Ernest Miguelez, Stefanie Research: A Multilevel Approach.” American Journal of Stantcheva, and Valerio Sterzi. 2018. “Dancing with the Public Health 93 (2): 215–21. Stars: Innovation through Interactions.” NBER Working Aghion, Philippe, Ufuk Akcigit, Antonin Bergeaud, Richard Paper 24466 (March), National Bureau of Economic Blundell, and David Hemous. 2019. “Innovation and Top Research, Cambridge, MA. Income Inequality.” Review of Economic Studies 86 (1): Akcigit, Ufuk, John Grigsby, and Tom Nicholas. 2017. “The 1–45. Rise of American Ingenuity: Innovation and Inventors of Akcigit, Ufuk, Harun Alp, and Michael Peters. 2021. “Lack the Golden Age.” NBER Working Paper 23047 (January), of Selection and Limits to Delegation: Firm Dynamics National Bureau of Economic Research, Cambridge, MA. 130 | WORLD DEVELOPMENT REPORT 2024 Akcigit, Ufuk, John Grigsby, Tom Nicholas, and Stefanie Blanchflower, David G., and Andrew J. Oswald. 1998. “What Stantcheva. 2022. “Taxation and Innovation in the Makes an Entrepreneur?” Journal of Labor Economics Twentieth Century.” Quarterly Journal of Economics 16 (1): 26–60. 137 (1): 329–85. Bloom, Nicholas, Aprajit Mahajan, David J. McKenzie, and Akcigit, Ufuk, Jeremy G. Pearce, and Marta Prato. 2020. John Roberts. 2010. “Why Do Firms in Developing “Tapping into Talent: Coupling Education and Innovation Countries Have Low Productivity?” American Economic Policies for Economic Growth.” NBER Working Paper Review 100 (2): 619–23. 27862 (September), National Bureau of Economic Britto, Diogo G. C., Alexandre de Andrade Fonseca, Paolo Research, Cambridge, MA. Pinotti, Breno Sampaio, and Lucas Warwar. 2022. Alesina, Alberto Francesco, Sebastian Hohmann, Stelios “Intergenerational Mobility in the Land of Inequality.” Michalopoulos, and Elias Papaioannou. 2021. CESifo Working Paper 10004, Munich Society for the “Intergenerational Mobility in Africa.” Econometrica Promotion of Economic Research, Center for Economic 89 (1): 1–35. Studies, Ludwig Maximilian University and Ifo Institute Alexander, Diane, and Janet Currie. 2017. “Is It Who You for Economic Research, Munich. Are or Where You Live? Residential Segregation and Bryan, Gharad T., and Melanie Morten. 2019. “The Aggregate Racial Gaps in Childhood Asthma.” Journal of Health Productivity Effects of Internal Migration: Evidence Economics 55 (September): 186–200. from Indonesia.” Journal of Political Economy 127 (5): Aron, Danielle Victoria, R. Andres Castaneda Aguilar, 2229–68. Carolina Diaz-Bonilla, Maria Gabriela Farfan Betran, Caria, Stefano, Simon Franklin, and Marc Josef Witte. 2022. Elizabeth Mary Foster, Tony Henri Mathias Jany Fujs, “Searching with Friends.” Journal of Labor Economics Dean Mitchell Jolliffe, et al. 2023. “September 2023 41 (4): 887–922. Update to the Poverty and Inequality Platform (PIP): Chetty, Raj, and Nathaniel Hendren. 2018. “The Impacts What’s New.” Global Poverty Monitoring Technical Note of Neighborhoods on Intergenerational Mobility I: 32 (September), World Bank, Washington, DC. Childhood Exposure Effects.” Quarterly Journal of Asher, Sam, Paul Novosad, and Charlie Rafkin. 2024. Economics 133 (3): 1107–62. “Intergenerational Mobility in India: New Measures and Chetty, Raj, Nathaniel Hendren, and Lawrence F. Katz. 2016. Estimates across Time and Social Groups.” American “The Effects of Exposure to Better Neighborhoods on Economic Journal: Applied Economics 16 (2): 66–98. Children: New Evidence from the Moving to Opportunity Barnett, Andy, Bruce Yandle, and George Naufal. 2013. Experiment.” American Economic Review 106 (4): “Regulation, Trust, and Cronyism in Middle Eastern 855–902. Societies: The Simple Economics of ‘Wasta.’” Journal Chetty, Raj, Nathaniel Hendren, Patrick Kline, and Emmanuel of Socio-Economics 44 (June): 41–46. Saez. 2014. “Where Is the Land of Opportunity? The Baum-Snow, Nathaniel, and Byron F. Lutz. 2011. “School Geography of Intergenerational Mobility in the United Desegregation, School Choice, and Changes in States.”  Quarterly Journal of Economics  129 (4): Residential Location Patterns by Race.” American 1553–623. Economic Review 101 (7): 3019–46. Chetty, Raj, Matthew O. Jackson, Theresa Kuchler, Johannes Bayer, Patrick, Stephen  L. Ross, and Giorgio Topa. 2008. Stroebel, Nathaniel Hendren, Robert B. Fluegge, Sara “Place of Work and Place of Residence: Informal Hiring Gong, et al. 2022. “Social Capital I: Measurement and Networks and Labor Market Outcomes.” Journal of Associations with Economic Mobility.” Nature 608 Political Economy 116 (6): 1150–96. (7921): 108–21. Beaman, Lori A., Niall Keleher, and Jeremy Magruder. 2018. Chiplunkar, Gaurav, and Pinelopi Koujianou Goldberg. “Do Job Networks Disadvantage Women? Evidence 2021. “Aggregate Implications of Barriers to Female from a Recruitment Experiment in Malawi.” Journal of Entrepreneurship.” NBER Working Paper 28486 Labor Economics 36 (1): 121–57. (February), National Bureau of Economic Research, Beaman, Lori A., and Jeremy Magruder. 2012. “Who Gets Cambridge, MA. the Job Referral? Evidence from a Social Networks Chyn, Eric. 2018. “Moved to Opportunity: The Long-Run Experiment.” American Economic Review 102 (7): Effects of Public Housing Demolition on Children.” 3574–93. American Economic Review 108 (10): 3028–56. Becker, Gary Stanley, Scott Duke Kominers, Kevin M. Chyn, Eric, and Lawrence F. Katz. 2021. “Neighborhoods Murphy, and Jörg L. Spenkuch. 2018. “A Theory of Matter: Assessing the Evidence for Place Intergenerational Mobility.” Journal of Political Economy Effects.”  Journal of Economic Perspectives 35 (4): 126 (S1): S7–S25. 197–222. Bell, Alexander M., Raj Chetty, Xavier Jaravel, Neviana Cirera, Xavier, Diego  Comin, and Marcio Cruz. 2022. Petkova, and John Van Reenen. 2019. “Who Becomes Bridging the Technological Divide: Technology Adoption an Inventor in America? The Importance of Exposure by Firms in Developing Countries.  Washington, DC: to Innovation.”  Quarterly Journal of Economics 134 (2): World Bank.  647–713. Coco, Giuseppe, and Raffaele Lagravinese. 2014. “Cronyism Bianchi, Nicola, and Michela Giorcelli. 2020. “Scientific and Education Performance.” Economic Modelling Education and Innovation: From Technical Diplomas 38 (February): 443–50. to University Stem Degrees.” Journal of the European D’Aoust, Olivia Severine, Virgilio Galdo, and Elena Economic Association 18 (5): 2608–46. Ivanova Ianchovichina. 2023. “Territorial Productivity Preservation | 131 Differences and Dynamics within Latin American Grover, Arti, Somik Vinay Lall, and William F. Maloney. Countries.” Policy Research Working Paper 10480, 2022. Place, Productivity, and Prosperity: Revisiting World Bank, Washington, DC. Spatially Targeted Policies for Regional Development. Demirgüç-Kunt, Asli, Leora Klapper, Dorothe Singer, and Washington, DC: World Bank. Saniya Ansar. 2022.  The Global Findex Database 2021: Hammond, Alicia, Eliana Rubiano Matulevich, Kathleen Financial Inclusion, Digital Payments, and Resilience in Beegle, and Sai Krishna Kumaraswamy. 2020. The the Age of COVID-19. Washington, DC: World Bank. Equality Equation: Advancing the Participation of Women Deng, Jingyuan, Nelly Youssef Elmallakh, Luca Flabbi, and Girls in STEM. Washington, DC: World Bank. and Roberta V. Gatti. 2023. “Returns to Education in Hellerstein, Judith K., Mark J. Kutzbach, and David Neumark. the Marriage Market: Bride Price and School Reform 2019. “Labor Market Networks and Recovery from Mass in Egypt.” Policy Research Working Paper 10288, Layoffs: Evidence from the Great Recession Period.” World Bank, Washington, DC. Journal of Urban Economics 113 (September): 103192. de Ree, Joppe, Karthik Muralidharan, Menno Prassad Hellerstein, Judith K., Melissa McInerney, and David Pradhan, and F. Halsey Rogers. 2018. “Double for Neumark. 2011. “Neighbors and Coworkers: The Nothing? Experimental Evidence on an Unconditional Importance of Residential Labor Market Networks.” Teacher Salary Increase in Indonesia.” Quarterly Journal Journal of Labor Economics 29 (4): 659–95. of Economics 133 (2): 993–1039. Hsieh, Chang-Tai, Erik Hurst, Charles I. Jones, and Peter Deutscher, Nathan. 2020. “Place, Peers, and the Teenage J. Klenow. 2019. “The Allocation of Talent and U.S. Years: Long-Run Neighborhood Effects in Australia.” Economic Growth.” Econometrica 87 (5): 1439–74. American Economic Journal: Applied Economics 12 (2): IFC (International Finance Corporation). 2017. “MSME 220–49. Finance Gap: Assessment of the Shortfalls and Emran, M. Shahe, Hanchen Jiang, and Forhad J. Shilpi. Opportunities in Financing Micro, Small and Medium 2021. “Gender Bias and Intergenerational Educational Enterprises in Emerging Markets.” IFC, Washington, DC.  Mobility: Theory and Evidence from China and India.” Jackson, Matthew O. 2021. “Inequality’s Economic and Social Policy Research Working Paper 9250, World Bank, Roots: The Role of Social Networks and Homophily.” Washington, DC. SSRN Electronic Journal, vol. 25. https://papers.ssrn​ Emran, M. Shahe, and Forhad J. Shilpi. 2015. “Gender, .com/sol3/papers.cfm?abstract ​_ id=3795626. Geography, and Generations: Intergenerational Krishna, Anirudh. 2013. “Stuck in Place: Investigating Social Educational Mobility in Post-Reform India.” World Mobility in 14 Bangalore Slums.” Journal of Development Development 72 (August): 362–80. Studies 49 (7): 1010–28. Evans, David S., and Boyan Jovanovic. 1989. “An Estimated Laliberté, Jean-William. 2021. “Long-Term Contextual Effects Model of Entrepreneurial Choice under Liquidity in Education: Schools and Neighborhoods.” American Constraints.” Journal of Political Economy 97 (4): 808–27. Economic Journal: Economic Policy 13 (2): 336–77. Evans, David S., and Linda S. Leighton. 1989. “Some Lall, Somik Vinay, J. Vernon Henderson, and Anthony J. Empirical Aspects of Entrepreneurship.” American Venables. 2017. Africa’s Cities: Opening Doors to the Economic Review 79 (3): 519–35. World. With Juliana Aguilar, Ana Aguilera, Sarah Antos, García, Jorge Luis, James J. Heckman, and Victor Paolo Avner, Olivia Severine D’Aoust, Chyi-Yun Huang, Ronda. 2023. “The Lasting Effects of Early-Childhood Patricia Jones, Nancy Lozano Gracia, and Shohei Education on Promoting the Skills and Social Nakamura. Washington, DC: World Bank. Mobility of Disadvantaged African Americans and Lall, Somik Vinay, Jon Kher Kaw, Forhad J. Shilpi, and Sally Beth Their Children.”  Journal of Political Economy 131 (6): Murray. 2023. Vibrant Cities: On the Bedrock of Stability, 1477–506. Prosperity, and Sustainability. Washington, DC: World Bank. Gatti, Roberta V., Diego F. Angel-Urdinola, Joana Silva, Lucas, Robert E. B., Jr. 2009. “Ideas and Growth.” Economica and András Bodor. 2014. Striving for Better Jobs: The 76 (301): 1–19. Challenge of Informality in the Middle East and North Lucas, Robert E. B., Jr., and Benjamin Moll. 2014. “Knowledge Africa. Directions in Development: Human Development. Growth and the Allocation of Time.” Journal of Political Washington, DC: World Bank. Economy 122 (1): 1–51. Glaeser, Edward L. 2011. “Cities, Productivity, and Quality of Mani, Anandi, and Emma Riley. 2021. “Social Networks Life.” Science 333 (6042): 592–94. as Levers of Mobility.” In Social Mobility in Developing Goldberg, Pinelope, Somik Lall, Meet Mehta, Michael Countries: Concepts, Methods, and Determinants, Peters, and Aishwarya Ratan. 2024. “Gender edited by Vegard Iversen, Anirudh Krishna, and Kunal Growth Gaps across Indian States.” EGC Discussion Sen, 424–47. UNU-WIDER Studies in Development Paper 1108, EliScholar, Economic Growth Center, Yale Economics. Helsinki: United Nations University– University, New Haven, CT. World Institute for Development Economics Research; Gottlieb, Charles, Markus Poschke, and Michael New York: Oxford University Press. Tueting. 2024. “Skill Supply, Firm Size and Economic Marx, Benjamin, Thomas Stoker, and Tavneet Suri. 2013. Development.” Background paper prepared for World “The Economics of Slums in the Developing World.” Development Report 2024, World Bank, Washington, DC. Journal of Economic Perspectives 27 (4): 187–210. Granovetter, Mark S. 1973. “The Strength of Weak Ties.” Monteiro, Joana, and Rudi Rocha. 2017. “Drug Battles and American Journal of Sociology 78 (6): 1360–80. School Achievement: Evidence from Rio de Janeiro’s 132 | WORLD DEVELOPMENT REPORT 2024 Favelas.” Review of Economics and Statistics 99 (2): Schneider, Ben Ross. 2022. “Teacher Unions, Political 213–28. Machines, and the Thorny Politics of Education Reform Nakamura, Emi, Jósef Sigurdsson, and Jón Steinsson. 2022. in Latin America.” Politics and Society 50 (1): 84–116. “The Gift of Moving: Intergenerational Consequences of Selod, Harris, and Forhad J. Shilpi. 2021. “Rural-Urban a Mobility Shock.” Review of Economic Studies 89 (3): Migration in Developing Countries: Lessons from the 1557–92. Literature.” Regional Science and Urban Economics Narayan, Ambar, Roy van der Weide, Alexandru Cojocaru, 91 (November): 103713. Christoph Lakner, Silvia Redaelli, Daniel Gerszon Stone, Samuel Z. 1990. The Heritage of the Conquistadors: Mahler, Rakesh Gupta N. Ramasubbaiah, and Stefan Ruling Classes in Central America from the Conquest Thewissen. 2018.  Fair Progress? Economic Mobility to the Sandinistas. Foreword by Richard E. Greenleaf. across Generations around the World. Equity and Lincoln: University of Nebraska Press. Development Series. Washington, DC: World Bank. Suryadarma, Daniel. 2012. “How Corruption Diminishes Neidhöfer, Guido, Matías Ciaschi, and Leonardo Gasparini. the Effectiveness of Public Spending on Education in 2021. “Intergenerational Mobility in Education in Latin Indonesia.” Bulletin of Indonesian Economic Studies America.” CAF Working Paper 2021/14 (December 14), 48 (1): 85–100. Corporación Andina de Fomento (Development Bank Toivanen, Otto, and Lotta Väänänen. 2016. “Education of Latin America), Caracas, República Bolivariana de and Invention.” Review of Economics and Statistics Venezuela. 98 (2): 382–96. Nicodemo, Catia, and Gustavo Adolfo García. 2015. “Job Tombe, Trevor, and Xiaodong Zhu. 2019. “Trade, Migration, Search Channels, Neighborhood Effects, and Wages and Productivity: A Quantitative Analysis of China.” Inequality in Developing Countries: The Colombian American Economic Review 109 (5): 1843–72. Case.” Developing Economies 53 (2): 75–99. Topa, Giorgio, and Yves Zenou. 2015. “Neighborhood and Park, Hogeun, Nancy Lozano Gracia, Giuseppe Rossitti, Network Effects.” In Handbook of Regional and Urban and Olivia Severine D’Aoust. 2022. “A Global Review Economics, vol. 5, edited by Gilles Duranton, J. Vernon of Intra-Urban Spatial Inequality.” World Bank, Henderson, and William C. Strange, 561–624. Oxford, Washington, DC. UK: Elsevier. Pennings, Steven Michael. 2022. “A Gender Employment Turok, Ivan, and Jackie Borel-Saladin. 2018. “The Theory Gap Index (GEGI): A Simple Measure of the Economic and Reality of Urban Slums: Pathways-Out-of-Poverty Gains from Closing Gender Employment Gaps, with or Cul-de-Sacs?” Urban Studies 55 (4): 767–89. an Application to the Pacific Islands.” Policy Research WEF (World Economic Forum). 2020. The Global Social Working Paper 9942, World Bank, Washington, DC. Mobility Report 2020: Equality, Opportunity and a New Perla, Jesse, and Christopher Tonetti. 2014. “Equilibrium Economic Imperative. Insight Report (January). Geneva: Imitation and Growth.” Journal of Political Economy WEF. 122 (1): 52–76. Wong, Mark Tsun On. 2019. “Intergenerational Family Perlman, Janice E. 2006. “The Metamorphosis of Support for ‘Generation Rent’: The Family Home for Marginality: Four Generations in the Favelas of Rio de Socially Disengaged Young People.” Housing Studies Janeiro.” Annals of the American Academy of Political 34 (1): 1–23. and Social Science 606 (July): 154–77. World Bank. 2018. World Development Report 2018: Perlman, Janice E. 2010. Favela: Four Decades of Living on the Learning to Realize Education’s Promise. Washington, Edge in Rio de Janeiro. New York: Oxford University Press. DC: World Bank. Rains, Emily, and Anirudh Krishna. 2020. “Precarious Gains: World Bank. 2024. Women, Business and the Law 2024. Social Mobility and Volatility in Urban Slums.” World Washington, DC: World Bank. Development 132 (August): 105001. Zingales, Luigi. 2014. A Capitalism for the People: Rothstein, Richard. 2017. The Color of Law: A Forgotten Recapturing the Lost Genius of American Prosperity. History of How Our Government Segregated America. New York: Basic Books. New York: Liveright. Ziraba, Abdhalah Kasiira, Catherine Kyobutungi, and Eliya Schady, Norbert R., Alaka Holla, Shwetlena  Sabarwal, Msiyaphazi Zulu. 2011. “Fatal Injuries in the Slums of Joana Silva, and Andres Yi Chang. 2023.  Collapse and Nairobi and Their Risk Factors: Results from a Matched Recovery: How the COVID-19 Pandemic Eroded Human Case-Control Study.” Journal of Urban Health 88 Capital and What to Do about It. Washington, DC: World (Supplement 2): 256–65. Bank.  Preservation | 133 6 Destruction Key messages • When a crisis strikes, such as that surrounding global energy, middle-income countries should seize the opportunity to instigate the needed restructuring and reallocation. For example, the global energy and climate crises have spurred rapid progress in the develop- ment and deployment of low-carbon technologies. • In middle-income countries, the rate of adoption of key clean energy technologies is growing more rapidly than in high-income countries, but levels of adoption by firms and households remain significantly lower, particularly for solar energy, wind energy, and electric vehicles. • Incumbents in high-carbon industries, which tend to be state-owned enterprises in middle-income countries, erect barriers to the entry of low-carbon technology because they have the strongest incentive to maintain the status quo and limit competition from low-carbon energy providers. • Middle-income countries run the risk of becoming stranded nations—not because of anything inherent in the scaling up of low-carbon technologies and the winding down of fossil fuels, but because of (1) outdated policies and rules of thumb that limit the growth of value-creating enterprises and the exit of unproductive ones; (2) limited improvements in human capital and the mobility of workers; and (3) a refusal to let go of state control of productive assets that are being decommissioned ahead of time. Destruction: To be expected, tactics, suppressing technological improvements, and persuading government to restrict new entry managed, and mitigated or open market policies. Therefore, the crisis, The Republic of Korea’s financial crisis in 1997–98 although triggered by external events, was largely had paradoxical economic impacts on innovation, a product of internal problems related to a weak restructuring, and growth. Before the crisis, large system of corporate governance, a dysfunctional family-owned industrial conglomerates, chaebols, financial system, and poor labor relations.1 enjoyed almost unrestrained market power and The crisis triggered major reforms and a com- expansion thanks to excessive debt financing. prehensive restructuring of the financial sector They were then able to drive rival firms and small and the chaebols. Nearly 500 nonviable finan- industries out of business by adopting predatory cial institutions closed, including two-thirds of | 135 commercial banks, and 15 of the 30 top conglom- United States, on average nearly 16 percent of jobs erates went bankrupt. The collapse of many chae- in the private sector have been destroyed each bols made room for venture capital–backed firms, year.4 Literature on firm exit—stemming from which led to the rapid growth of the information seminal work by Hopenhayn (1992)—reveals that and communication technology (ICT) sector. The the exit of low-productivity firms contributes sub- ICT boom—spurred by new technologies such as stantially to raising aggregate productivity. For mobile phones, thin film transistor liquid crys- example, the Great Depression (1929–39) ushered tal displays, and broadband and wireless inter- in a permanent structural change toward mass net—kick-started an unusually swift recovery in production and automation in the motor vehicle 1999, with economic growth of 10.3 percent. The industry through the exit of smaller, less produc- reforms and restructuring led to long-term inno- tive plants and the need for surviving plants to vation-led growth and contributed to Korea’s tran- innovate—a process that likely would have taken sition to high-income status. Overall, then, it took much longer without the crisis.5 a crisis to lay bare the need for economic reforms Financial crises also spur readjustments of and drive the reallocation of economic activity new technologies. For example, the financial cri- toward more productive, more innovative firms. sis of 2007 accelerated skill-biased technological The destruction of outdated arrangements— change.6 More recently, during the COVID-19 enterprises, jobs, technologies, private contracts, crisis economic activity was reallocated toward policies, and public institutions—is essential to more productive firms, with reallocation effects create value through investment and reallocation, higher than before the pandemic.7 infusion, and innovation. But in many countries In the context of energy, the oil price shocks these destructive forces are weak during boom in the 1980s played a major role in accelerating times, with crises playing a disproportionate role investments in energy efficiency and the develop- in driving the process of resource reallocation.2,3 ment of cleaner energy technologies.8 One impact In some cases, downturns can serve as times of of the oil price shocks was an increase in the rel- cleansing in which older, less-productive firms ative cost of fossil fuels. Another was new policy die, making way for newer, more productive firms. support for less energy-intensive activities. More As Joseph  A.  Schumpeter (1942, 113), writing in recently, the global financial crisis of 2007–09 the aftermath of the Great Depression, argued coincided with a significant increase in the uptake about crises: of renewables.9 Renewable energy use grew rap- idly in the United States, China, and Germany in They are but temporary. They are the part because of the stimulus programs govern- means to reconstruct each time the eco- ments enacted to address the crisis. nomic system on a more efficient plan. This chapter focuses on the process of cre- But they inflict losses while they last, drive ative destruction being accelerated today by two firms into the bankruptcy court, throw crises—the climate crisis and the 2022–23 global people out of employment, before the energy crisis triggered by the Russian Federation’s ground is clear and the way paved for new invasion of Ukraine. In this context, this chapter achievement of the kind which has created explores three questions: modern civilization and made the great- • How is the energy crisis building momentum ness of this country. for change? The global energy crisis spurred rapid progress in the development and One of Schumpeter’s key observations is that deployment of low-carbon technologies. the process of creating new industries does not It also triggered reshaping of the climate go forward without sweeping away the existing policy landscape in 2022 and 2023. Devel- order. For example, over the last 30 years in the oped nations such as the United States 136 | WORLD DEVELOPMENT REPORT 2024 and European Union (EU) member states The climate and energy crises introduced a wide variety of incentives for producing and deploying low-carbon tech- could trigger restructuring nologies. In September 2023, the leaders and reallocation of the Group of Twenty (G20) agreed to Disruptions are accelerating the triple renewable energy capacity by 2030. diffusion of innovative lower-carbon Although fossil fuels have shaped the technologies world’s economy and economic geography for over a century, low-carbon technologies Today, two crises—the climate crisis and the favor new trends: urban agglomerations global energy crisis—are combining to drive rapid and new spatial clusters; the use of highly progress in low-carbon technologies (box  6.1). skilled workers to develop, modify, adapt, Four technologies—solar panels, wind turbines, apply, and maintain new technologies; and lithium-ion batteries, and electrolyzers used the entry of younger firms in the private for green hydrogen—have been shown to fol- sector. Entrants are instigators of change. low “learning curves,” as formalized by Wright’s • Who are the antagonists blocking creative Law: costs fall as a power function of cumulative destruction in energy markets? Incumbents deployment due to the positive effects of learn- may resist change. Many incumbents ing by doing or increasing returns to scale in the in the energy market have the strongest unit’s production.11 Since the first commercial incentive to maintain the status quo and use of solar panels in 1958, their costs have fallen limit competition from low-carbon energy by more than three orders of magnitude. This providers. High-carbon firms tend to lobby technology is therefore in a category that has against pro-environmental regulations.10 been characterized by exponential rather than Power purchase agreements (PPAs) with linear growth, along with computer processing long time horizons and inflexible terms power, Ford’s Model T cars in the 1900s, and DNA that create “lock-ins” impede change. sequencing. Figure 6.1 compares the cost trajec- These agreements often lock in polluting tories of solar and wind power with other tech- assets, resulting in significant inertia in nologies that have undergone rapid cost declines. energy systems. Critically, this pattern is in sharp contrast to that • Do middle-income countries run the risk of of fossil fuels, whose prices have stayed broadly becoming stranded nations? Yes, they do. constant when adjusted for inflation over the last But stranding is not driven by anything century. inherent in the scaling up of low-­ carbon Diffusion of these technologies, although slow technologies and the winding down of for many decades, recently accelerated. Analysis of fossil fuels. It is the result of outdated 700 million online job postings from 35 predomi- policies and rules of thumb that limit nantly advanced economies reveals that after very the growth of value-creating enterprises modest growth beginning in 2014, the share of jobs and the exit of unproductive ones; lim- related to low-carbon technologies increased by ited improvements in human capital and more than 50 percent from 2021 to 2022 (figure 6.2, mobility of workers; and not relinquish- panel a).12 Growth was rapid in three-quarters of ing state control of productive assets now the countries studied in Asia, Europe, and North being decommissioned ahead of time. To America, but stronger in Europe. Driving this be sure, workers displaced from the tran- growth were electric vehicles (EVs), solar energy, sition will need targeted and time-bound insulation, EV  charging, heat pumps, and wind support, but it is not a panacea for avoid- energy. In 2022, openings related to low-carbon ing difficult reforms. technologies grew rapidly in almost every industry, Destruc tion | 137 Box 6.1 The diffusion of low-carbon technologies as defined and measured in this chapter This chapter defines low-carbon technologies in line with the Y02 classification of patents related to “climate change mitigation technologies” adopted by the European Patent Office. Low-carbon technologies are defined as “technologies or applications which can be considered as countering the effects of climate change.” This classification has seven main categories: energy, greenhouse gas capture, buildings, industry (including agricul- ture), transport, waste management, and wastewater management. Some examples of the technologies included under these categories are: • Upstream energy supply technologies: renewable energy, combustion technologies with the potential to mitigate carbon emissions, energy storage technologies, decen- tralized energy, efficient electrical power generation technologies, and smart grids. • Downstream end-user technologies: energy-efficient lighting, energy-efficient heating, energy-efficient appliances, heat pumps, electric vehicles, electric vehicle charging, hybrid vehicles; technologies related to processing of metals and minerals and low- ering emissions in agriculture (such as solar water pumping and greenhouses); and technologies related to solid waste management, biopackaging, and bioplastics. Detailed microdata on the adoption of these granular technologies across a wide range of countries, regions, industries, and firms are not yet readily available. In the absence of such data, this chapter relies on a growing literature that takes advantage of real-time data sources and uses text analysis to infer the spread of new technologies through their footprint in the demand for new technology-related tasks or skills in online job post- ings.a Where possible, the chapter also complements these measures with country-level measures of specific technologies and data on energy intensity and carbon intensity at the country and industry levels. Finally, it also incorporates recent granular firm-level data on green technology adoption from the World Bank’s Firm-level Adoption of Technology (FAT) survey.b a. See, for example, Acemoglu et al. (2022); Goldfarb, Taska, and Teodoridis (2023). b. Cirera, Comin, and Cruz (2022). but  were particularly pronounced in manufac- measures, fuel switching, and accelerated deploy- turing, electricity, heat supply, and construction. ment of cleaner energy technologies. The energy Mentions of low-carbon technologies in the share- intensity of the gross domestic product (GDP) is holder meetings of the world’s largest firms also now 3.5 percent below levels before the pandemic doubled in 2022 (figure 6.2, panel b). in 2019—a rate of decline considerably higher Countries have responded to high energy than the 2 percent decline three years after the prices and energy security stemming from the onset of the 2008 financial crisis. Emissions invasion of Ukraine with energy conservation ­ atural gas also fell by 1.6 percent in 2022. from n 138 | WORLD DEVELOPMENT REPORT 2024 Figure 6.1 Learning by doing in the and 2023. The United States saw a historic shift manufacture of key low-carbon technologies in climate policy with passage of the Inflation has resulted in rapid cost declines Reduction Act (IRA) in late 2022, which intro- duced a wide variety of incentives for producing 100 and deploying low-carbon technologies.14 The Investment price index relative to price in 1985 (log scale) IRA could have large impacts on power sector investments and electricity prices, lowering retail electricity rates and resulting in negative prices in some wholesale markets.15 It could also signifi- 10 cantly hasten the adoption of renewable energy in the United States, increasing renewable pene- tration by about 13 percent by 2030, and it could spur adoption in other countries as the higher US investment drives capital prices lower.16 The 0 European Union also made substantial shifts in climate policy in 2022 and early 2023. These 85 90 95 00 05 10 15 20 19 19 19 20 20 20 20 20 included the REPowerEU strategy to end the Onshore wind Solar photovoltaics bloc’s reliance on Russian fossil fuels through Computers and peripheral equipment lower fossil fuel use; the Green Deal Industrial Software Plan, which aims to boost low-carbon manufac- Communication equipment turing and industry in Europe; and an increase in Sources: WDR 2024 team elaboration from Arkolakis the bloc’s binding renewable energy target from and Walsh (2023); FRED: Software (Federal Reserve 32 percent to 43 percent. China has also pushed Economic Data: Software) (database), Federal Reserve its scale-up of renewables and is now on track to Bank of St. Louis, St. Louis, https://fred.stlouisfed​ .org​ /tags/series?t=software; IRENA (2023); Prices and meet its 2030 renewable energy generation target Output for Information and Communication Technologies five years ahead of time, by 2025.17 (dashboard), Bureau of Economic Analysis, Suitland, MD, This acceleration of the clean energy transition https://www.bea.gov/prices-and-output-information-and​ -communication-technologies; Solar (photovoltaic) panel is driving the emergence of new spatial clusters prices (grapher), Our World in Data, Global Change Data and jobs (map 6.1). Evidence extending the work Lab and Oxford Martin Program on Global Development, of Bastos et al. (2023) to analyze 1 billion online University of Oxford, Oxford, UK, https://ourworldindata​ .org/grapher/solar-pv-prices. job postings across 86 countries over the last Note: The figure compares the price of investment for the decade reveals that in 2022 alone 3.3 million new three sectors with the fastest declines according to data from job openings were related to low-carbon technol- the Bureau of Economic Analysis with that for onshore wind ogies—just under 2 percent of jobs posted online and solar photovoltaics, with 1985 normalized to 1. in 2022.18 In absolute terms, new openings have In  light of policy changes, the International been highest in the United States, Europe, and Energy Agency (IEA) revised its forecast for China (figure 6.3, panel a). The new spatial clus- renewable capacity additions for 2023 and 2024, ters emerging in terms of low-carbon technology raising it by 38  percent from its expectations jobs are Catalonia and Madrid in Spain, home before the war, in December 2021. Countries to major solar industries; Guangzhou, Beijing, more dependent on imports of natural gas before Shanghai, Suzhou, and Shenzhen in China, home the war were more exposed to the price shock and to China’s largest EV and clean energy manu- increased hiring for jobs related to low-carbon facturing hubs; Rhône-Alpes and Île-de-France technologies.13 (the Paris region), home to France’s largest clean The global energy crisis also spurred a energy clusters; and California, the clean energy reshaping of the climate policy landscape in 2022 pioneer in the United States.19 Destruc tion | 139 Figure 6.2 The diffusion of low-carbon technologies is rapidly accelerating a. Share of jobs related to low-carbon technologies in 35 countries 2.5 2.0 Share of job postings related to LCTs (%) 1.5 1.0 0.5 0 2014 2015 2016 2017 2018 2019 2020 2021 2022 b. Share of publicly listed firms mentioning low-carbon technologies in shareholder meetings 35 mentioning LCTs in meetings (%) 30 Share of publicly listed firms 25 20 15 10 5 0 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 Source: Bastos et al. 2023. Note: Panel a displays the share of online job postings in 35 countries that mention low-carbon technologies (LCTs) as defined by the Y02 classification of patents adopted by the European Patent Office. See Classification of Patents: Climate Change Mitigation Technologies (dashboard), European Patent Office, Munich, https://www.epo.org/en/news-events​ in-focus​ /­­ /classification/climate-change. Panel b displays the share of shareholder meeting transcripts from publicly listed firms that mention these technologies. 140 | WORLD DEVELOPMENT REPORT 2024 Map 6.1 In 2022, one-third of online job postings related to low-carbon technologies were in middle-income countries Source: WDR 2024 team analysis extending Bastos et al. (2023) to 86 countries. Innovation is also driving the creation of start- near petroleum fields. The trade in fossil fuels has ups and capital flows (figure 6.3, panels b and c). been a driving component of global trade and geo- In 2021, about 1,500 new clean energy start-ups politics for decades, with fuel exports accounting were listed on Crunchbase, drawing on IEA data for 11.7 percent of total merchandise exports covering 40 countries, and about one-fourth of the globally in 2019. However, this picture is start- start-ups were in middle-income countries.20 Of ing to change—and more rapidly than had been these, more than half were in China and one-third imagined as the following shifts occur: in India, although growth has been most rapid in India, with a tripling of start-up creation over the • Shifts between countries and between regions last decade. Globally, total clean energy financing within countries. Low-carbon technolo- passed US$1 trillion for the first time in 2022.21 gies and fossil fuels tend to be produced These shifts are triggering a reallocation of in different countries, and low-carbon economic activity across countries, regions, technologies favor countries where manu- industries, occupations, and firms. Just as fossil facturing capabilities are already in place. fuels have shaped the geopolitical map over the Even within some countries, a spatial real- last two centuries, the clean energy transforma- location of production and jobs is under tion will alter the economic geography of manu- way. In China, for example, although fossil facturing and global trade and the landscape of fuel jobs have been highly concentrated international trade policy. Access to fossil fuels in the inland provinces highly reliant on has fundamentally shaped the world’s economic coal mining, low-carbon technology jobs geography for nearly a century, with heavy indus- are growing in the manufacturing hubs try close to coal beds and petrochemical plants on the east coast.22 Spatial disparities Destruc tion | 141 will be altered and potentially intensified most rapidly in centers of research and because the geographical determinants agglomeration.23 of low-carbon energy sources differ from • Shifts between industries. Because low-­carbon those of fossil fuel–based economies, and technologies are generally being mass technological innovation is likely to occur manufactured, the low-carbon transition is Figure 6.3 Low-carbon innovation is driving the emergence of new spatial clusters, start-ups, and financing a. Online job postings related to low-carbon technologies in 2022, by city, state, region, or country 3.3 million online job openings were related to low-carbon technologies in 2022, with one-third in middle-income countries Number of online job postings 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 Île Sh zh i, C tes Su gha Sta e de Na Ki rla y la jin ngd nd a, e C in st n, G rm in t, rm ny rm ny Sh nit , Fr ina Sl itt av ue dy ina Lo Z uttg nite ga lic Ca h B dr Ch in Ge e-F zh u, C ina w ra a te ce pe Pr mb u, C ina St , U Sin pub d ia d, , G ec y ni wi rm s y or -A jin Sp a Ch zh n, C nd lo wi m a bu lus Fra a Da B rg, ia, S nce Lo ngd , C ina va ce Ch a k erl y on ich t, G Sta re An Lo g, om an ha r e te tz an an ak lan ria Cz al va tz an , U , S e te c , S , F in lif ône ei id, in , S S er hi am a , in ne ran en, hin H Wu itze nc Re an ni lp g, a rm erli Ge pa M hou Spa nd ur ar d po Ge a ad e a an ed an la ov el a , , It U s h rn M r B ag ar h -d en o h h H d ire Ch u h d e , , gz ia o an lon n a Gu ata g e s C xa Te ys R Pa es a p W Esp U Middle-income High-income b. Creation of clean energy start-ups 1,506 clean energy start-ups were listed on Crunchbase in 2021, one-quarter in middle-income countries Number of clean energy start-ups 1,600 1,400 1,200 1,000 800 600 400 200 0 19 2 19 3 19 5 20 0 20 1 20 5 20 6 20 8 2011 20 3 20 5 20 6 20 7 20 7 18 20 9 19 0 91 19 6 19 8 20 2 20 4 09 20 0 20 2 20 4 94 97 99 20 3 21 20 0 1 1 9 9 9 9 9 9 0 0 0 0 0 0 0 1 1 1 1 1 1 0 20 19 19 19 19 20 20 20 Middle-income countries High-income countries (Figure continues next page) 142 | WORLD DEVELOPMENT REPORT 2024 Figure 6.3 Low-carbon innovation is driving the emergence of new spatial clusters, start-ups, and financing (continued) c. Growth of clean energy financing US$1 trillion in clean energy financing was extended globally in 2022 1,200 1,110 Investment (US$, billions) 1,000 853 800 627 600 524 471 483 396 425 400 313 271 245 217 214 200 123 159 156 55 82 36 0 04 06 07 08 05 09 10 11 12 13 14 15 16 17 18 19 20 21 22 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Sources: Panel a: WDR 2024 team analysis extending Bastos et al. (2023) to 86 countries. Panel b: Data and Statistics (dashboard), International Energy Agency, Paris, https://www.iea.org/data-and-statistics. Panel c: 2022 data from BloombergNEF (dashboard), Bloomberg, New York, https://about.bnef.com/. Note: Panel a displays the number of online job postings by city, state, region, or country that mention low-carbon technologies as defined by the Y02 classification of patents adopted by the European Patent Office. See Classification of Patents: Climate Change Mitigation Technologies (dashboard), European Patent Office, Munich, https://www.epo.org/en​ /news-events/in-focus/classification/climate-change. Panel b displays the number of new clean energy start-ups by year for middle- and high-income countries. Panel c displays the value of clean energy financing by year globally. driving a reallocation of economic activity • Shifts between firms. Low-carbon technolo- across industries from extractives to gies are more likely to be adopted by new manufacturing and ancillary services. Jobs entrants, exporters, private firms, and firms are most likely to be in the manufacturing, that use research and development (R&D) construction, and sales industries, but low- more intensively.26 Low-carbon technology carbon technology jobs are growing rapidly job postings are also highly concentrated in the high-skilled white-collar industries in multinational firms and their  supply able to develop, modify, adapt, apply, and chains.27 By contrast, high-­carbon jobs are maintain new technologies. more likely to be in older firms and state- • Shifts between occupations. The manufac- owned enterprises. ture of low-carbon technologies is more skill-intensive than that for the high-­carbon The energy transition is also shaping a new alternatives,24 resulting in a reallocation of global trade and industrial policy landscape, which economic activity between occupations is disrupting the last three decades of trade policy and skill types. Jobs are also more likely coordination. After decades of such coordination, to be filled by younger, college-educated which yielded significant growth dividends, the white-collar workers, whereas high-carbon world’s major economies now have divergent cli- jobs are more likely to be occupied by older mate and trade policy approaches. The IRA in the blue-collar workers.25 United States, the Net-Zero Industry Act in the Destruc tion | 143 European Union,28 and the growing number of a higher value on the Regulatory Indicators for similar policies globally increasingly include local Sustainable Energy, or RISE, Index); higher car- content requirements and other reshoring efforts bon pricing (as measured using the net effective to support local industries. Although the previous carbon rate of the Organisation for Economic waves of middle-income countries transitioned to Co-operation and Development, OECD); and lower high-income status against the backdrop of trade fossil fuel reserves.30 Countries with higher solar policy coordination, today’s middle-income coun- potential (measured using the World Bank PVOUT tries will need to navigate a more complex land- Index) also have higher solar deployment rates.31 scape in which key trade rules have not yet been agreed on. Diffusion of downstream low-carbon energy technologies in middle-income countries Adoption of low-carbon technologies by firms Productive incumbents and new and households in middle-income countries entrants are driving the diffusion of remains far more limited than in high-income cleaner technologies countries. In 2021, the emissions intensity of Diffusion of upstream clean energy energy consumption in middle-income countries ­ technologies in middle-income countries was 49 percent higher than that in high-­income As the costs of low-carbon technologies decline, countries, while energy consumption per unit growth using cleaner energy sources and energy of GDP was 2.5 times higher. Measures of efficiency technologies is for the first time a downstream technology adoption across a wide possibility in today’s middle-income countries. range of countries are not readily available, but Currently, these countries’ rate of adoption of emissions intensity within narrowly defined key low-carbon energy technologies is, in fact, industries can serve as a proxy for technology growing more rapidly than for that of high-­ adoption.32 On average, of 63 middle-income income countries, but levels of adoption remain countries, two-thirds (41) have higher carbon significantly lower. In levels, middle-income intensity across industries—in terms of direct countries are still lagging on the deployment of emissions that are owned or controlled by a three key low-carbon energy technologies—solar company (Scope 1 emissions)—than the high- energy, wind energy, and EVs—compared with income country average, while about one-third high-­income countries (figure 6.4). In 2021, the (22) have lower carbon intensity. average share of electricity generated from wind Firms in middle-income countries vary and solar power in middle-income countries widely in their adoption of low-carbon tech- was about half of that in high-income countries nologies, even within narrowly defined indus- (respectively, 4.1  percent versus 9.5 percent for tries. Firms’ overall management practices and wind and 2.7 ­ percent versus 5.3 percent for solar). technological sophistication, skill intensity, The uptake of EVs in middle-income countries and international orientation are correlated also remains about half as much as that in high-­ with their adoption of low-carbon technolo- income countries (measured as number of EVs gies and energy-saving practices. In Argentina, per  million inhabitants). But today in middle-­ for example, firms’ capacity to adopt more income countries, the average growth rates of key advanced low-carbon technologies has been clean energy technologies have overtaken those shown to be correlated with their share of of high-income countries.29 skilled workers.33 Exporters usually have lower Research for this Report found that after con- emissions intensity relative to nonexport- trolling for income, countries with higher rates of ers.34 Foreign-owned firms generally have bet- deployment of solar and wind energy have more ter environmental performance, as has been favorable renewable energy policies and a more shown for Côte d’Ivoire, Mexico, and República favorable regulatory environment (captured by Bolivariana de Venezuela.35 144 | WORLD DEVELOPMENT REPORT 2024 Figure 6.4 The rate of adoption of clean energy technologies is growing more rapidly in middle-income countries than in high-income countries, but the level of adoption is lower a. Wind energy generation 10 100 9 from wind energy 80 Share of power generation (%) Growth rate (%) 8 7 60 6 5 40 4 3 20 2 0 1 0 –20 94 96 98 00 06 04 14 90 02 08 10 12 16 18 20 92 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 19 Share from wind energy generation, HICs Growth rate, HICs (right axis) Share from wind energy generation, MICs Growth rate, MICs (right axis) b. Solar energy generation 6 180 160 from solar energy 5 140 Growth rate (%) Share of power generation (%) 120 4 100 80 3 60 2 40 20 1 0 –20 0 –40 94 96 98 00 06 14 18 90 02 04 08 10 12 16 20 92 20 19 19 19 19 20 20 20 20 20 20 20 20 20 20 19 Share from solar energy generation, HICs Growth rate, HICs (right axis) Share from solar energy generation, MICs Growth rate, MICs (right axis) c. Electric vehicles Number of electric vehicles 0.0020 250 (per million inhabitants) Growth rate (%) 0.0016 200 0.0012 150 0.0008 100 0.0004 50 0.0000 0 14 10 11 12 13 15 16 17 18 19 20 21 20 20 20 20 20 20 20 20 20 20 20 20 Number of electric vehicles, HICs Growth rate, HICs (right axis) Number of electric vehicles, MICs Growth rate, MICs (right axis) Source: WDR 2024 team analysis using data from Statistics Data (portal), International Renewable Energy Agency, Abu Dhabi, United Arab Emirates, https://www.irena.org/Data. Note: Comparing middle- and high-income countries, panels a and b display the shares and growth rates of wind energy and solar energy, respectively, in electricity generation, and panel c the number of battery electric vehicles per million inhabitants. HICs = high-income countries; MICs = middle-income countries. Destruc tion | 145 Production of low-carbon technologies is China. Only Bangalore and São Paolo were the concentrated in a few middle-income countries ­ other ­middle-income country cities in the top 30. Substantial manufacturing or innovation in These sectors and jobs are clustered both within low-carbon energy technologies is occurring in and across countries. For middle-income coun- only a handful of middle-income countries with tries to successfully seize the opportunities of the competitive manufacturing sectors. In terms of green economy, the right policy mix, financing, production, only one middle-income ­ country— regulatory environment, and infrastructure need China—is competitive in all three key low-­carbon to be in place. technologies in terms of breadth (across the full value chain from raw materials, processed materi- Incumbent state-owned enterprises, als, subcomponents, to the end product of the sup- legacy policies, and path ply chain) and depth (a high average market share) dependence—all block creative (figure 6.5). By contrast, other ­ middle-income destruction countries have either depth or breadth and typi- cally only in one product. For example, countries Incumbents in high-carbon industries erect with well-established manufacturing sectors such barriers to entry of low-carbon technologies as India and Türkiye show a high breadth of export The low-carbon transition will create winners competitiveness in the production of wind tur- and losers—and the losses are more concentrated bines. Middle-income countries with large min- than the gains, resulting in a political economy eral deposits (such as Brazil and Russia) have high prone to inertia. Although 80 percent of the depth in the export of critical minerals for EVs. global population lives in a country that imports Patenting for low-­carbon technologies is also fossil fuel, fossil fuel revenues are highly con- highly concentrated in just a handful of coun- centrated. In 18 countries, fuel exports account tries, with China in the lead, according to data for more than 50 percent of total merchandise from the International Renewable Energy Agency exports; seven countries generate 90 percent or (IRENA). Over the last two decades, China has more of their export earnings from fossil fuel filed the largest number of such patents, fol- exports.36 Fossil fuel resources are more concen- lowed by the United States, Japan, Korea, and trated than renewable resources.37 Within coun- Germany. Such patenting has been very limited in tries, the losses to fossil fuel companies are also other ­middle-income countries, with the second-­ highly concentrated. Studies have shown that the highest middle-income country, Brazil, filing only concentrated nature of the losses from the low-­ 2 percent as many patents as China over the last carbon transition, compared with the distrib- 20 years. uted nature of the gains, makes the clean energy Costa Rica and China are global front-­ runners transition particularly prone to obstructionism in terms of job creation related to low-carbon by entrenched interests.38 Moreover, the incen- technologies. In 2022, Costa Rica accounted tives for lobbying are asymmetric. Incumbents, for about 8 percent of all online job postings as a tightly defined group of actors, have a greater related to low-carbon technologies and China incentive to mobilize, and they face lower coor- for 5 ­percent—just below the global maximum dination costs. Smaller firms and individuals, by of 8 percent for Denmark (figure 6.6). Five other contrast, are more dispersed and face higher coor- countries have exceeded the high-income country dination costs. This results in political inertia, as average: Brazil, Georgia, Senegal, South Africa, has been evident from the persistent difficulty in and the Dominican Republic. In terms of emerg- phasing out fossil fuel subsidies globally.39 ing spatial clusters in 2022, 28 of the 30 cities Incentives, both explicit and implicit, that with the largest number of new online job post- favor higher-carbon industries and production ings related to low-carbon technologies were in processes over lower-carbon ones are higher in 146 | WORLD DEVELOPMENT REPORT 2024 Figure 6.5 Clean energy technology value chains are still dominated by high-income countries and China a. Solar photovoltaics 50 Japan Germany value chain products) Breadth (number of China 40 Austria Italy Korea, Rep. Spain United States 30 France Türkiye United Kingdom 20 Mexico Malaysia 10 Indonesia Congo, Dem. Rep. 0 0 5 10 15 20 Depth (average market share, %) b. Wind turbines 60 Czechia India value chain products) Italy Germany Breadth (number of Romania 50 Bulgaria China Türkiye Japan France United States 40 Tunisia Korea, Rep. United Kingdom 30 Mexico Malaysia Thailand 20 Russian Federation 10 Congo, Dem. Rep. 0 0 5 10 15 20 Depth (average market share, %) c. Electric vehicles 20 South Africa value chain products) China Breadth (number of Japan Spain Belgium Germany 15 Brazil Austria United States India Australia 10 United Kingdom Norway Russian Federation 5 Congo, Dem. Rep. 0 0 5 10 15 20 Depth (average market share, %) Low-income Lower-middle-income Upper-middle-income High-income Source: Rosenow and Mealy 2024. Note: Each panel measures countries’ competitive dominance across traded products in a given supply chain in terms of two key measures. Breadth represents a country’s export competitiveness across the raw materials, processed materials, subcomponents, and end products of the supply chain. Depth measures a country’s export competitiveness in terms of its average market share across supply chain products. The sample includes countries whose total number of value chain products with a revealed comparative advantage exceeds 1. Destruc tion | 147 Figure 6.6 Costa Rica and China are the global front-runners in jobs related to low-carbon technologies Job creation leaders Job creation laggards Higher demand Lower demand than than high-income high-income country average country average 800,000 Number of online LCT job postings 8 Share of online job postings 7 700,000 related to LCTs (%) 6 600,000 5 500,000 4 400,000 3 300,000 2 High-income country average (1.7%) 200,000 1 100,000 0 0 A c ai n pp n ge ico M er d ca M eru al na m ia a an am n ut pub a n a or l Br ia a a l la Jo dor au bia ia Co Lan s Fe m a In s P a Co C il ua ia Ph era bia Ge ega az h li e i iu st h i n Se fric T h rda ili tio So Re ibi di ic N ano n lo k Le aic n Bu ma Ja ays Ec gar g in Ri M nti la Sr rit Ar ex e b S at d Gu in ss m Ru Do LCT share of job postings Number of LCT job postings (right axis) Source: WDR 2024 team analysis extending Bastos et al. (2023). Note: The figure displays the share of online job postings of middle-income countries in 2022 that mention low-carbon technologies (LCTs) as defined by the classification of patents related to “climate change mitigation technologies” adopted by the European Patent Office. It also displays the total number of low-carbon technology postings in 2022. See Classification of Patents: Climate Change Mitigation Technologies (dashboard), European Patent Office, Munich, https://www.epo.org/en​ /news-events/in-focus/classification/climate-change. middle-income countries than in high-income of providing incentives and regulatory support for countries. Middle-income countries, p ­ articularly renewable energy and planning for the expansion energy exporters, have significantly lower carbon of renewable energy. In addition, middle-­ income prices.40 Consumer fossil fuel subsidies in m ­ iddle- countries generally provide highly polluting income countries totaled over US$800 ­ billion industries with higher corporate tax incentives. in 2022, of a global total of US$900 billion.41 Such incentives are particularly high in the Middle Estimates of explicit consumer and producer East and North Africa, as measured by the World ­fossil fuel subsidies in 2022 from the International Bank’s Global Corporate Income Tax Incentives Monetary Fund (IMF) also show that middle-­ Database. income countries account for 65  percent of the Incentives that support incumbent firms total.42 Fossil fuel subsidies have a sizable fiscal and energy sources, as well as major barriers to cost and exacerbate air pollution, contributing entry in power markets, also severely limit pri- directly to premature deaths.43 vate innovation in middle-income countries. Middle-income countries also score lower on Over the last 35 years, many countries have the RISE Index, particularly on the components introduced competitive markets in parts of the 148 | WORLD DEVELOPMENT REPORT 2024 electricity system to reduce costs and improve High-carbon inertia curtails reliability. In many middle-income countries, innovation, slows planning, and locks however, power markets remain a monopoly: a in behavior state-owned entity operating under a vertically integrated utility. This state-owned entity carries A wide body of literature has now demonstrated out all functions in the electricity sector, includ- that patents for low-carbon technologies “build ing generation, transmission, distribution, and on the shoulders” of earlier developments. Thus retail supply. Such an arrangement has generally patenting is path-dependent, meaning that inno- hindered competition, and in many countries it vations are more likely to follow existing inno- has resulted in the inefficient use of resources. In vations, which can impede innovations in new addition, in many middle-income countries the technologies at an early stage.46 This literature has first generators dispatched are often not those generally pointed out the need for initial subsidies with the lower marginal prices (that is, power to jump-start the innovation process and correct dispatch often does not follow merit order), serv- for the positive externalities that result in under- ing as a barrier to the penetration of renewables investment in R&D on low-carbon technologies, with rapidly declining costs. Generally, the shift as well as the need for carbon pricing to correct to more competitive market structures has low- for the negative externality of carbon emissions. ered costs and enabled more innovation and pen- Inertia in keeping up with, and planning for, etration of renewables.44 In addition, the fact that exponential progress in low-carbon technologies energy supply technologies tend to be big, com- has also slowed the changes needed in energy sys- plex, expensive, and slow to develop and that new tems. The rapid technical progress in a range of entrants must sell into entrenched markets dom- low-carbon technologies described earlier has gen- inated by incumbents severely limits incentives erally outpaced the expectations of leading agen- for private innovation.45 cies and energy-economy models. For example, Power purchase agreements with long time for many years the IEA forecasted linear growth horizons and inflexible terms that create “lock- in the supplies of solar under its business-as-usual ins” also impede change in energy systems. PPAs scenarios, even as supplies continued to rise expo- are widely used to procure power by establish- nentially. Not only the IEA but historically most ing a contract between a seller of power and a energy-economy models have underestimated the buyer, often a utility. If well-structured, PPAs deployment rates for renewable energy technolo- offer certainty for buyers as well as sellers, pro- gies and overestimated their costs, as outlined by tecting them from volatility in energy prices by Grubb et al. (2021) and Way et al. (2022). The rea- locking in the price buyers pay for electricity for son is that most national energy-economy models decades to come. This kind of long-term cer- and large-scale global integrated assessment mod- tainty offers sellers a steady source of revenue els label energy technology cost developments as and improves the chances of securing low-cost exogenous. However, energy systems investments financing. However, by their very nature, when are often large and indivisible (lumpy)—that is, used to generate power in emissions-intensive they are not easily divided or sold in parts—and ways, such as coal-fired power, these agreements they are made over long time horizons, often lock in polluting assets, often for decades at a spanning multiple decades. Thus such downside time, resulting in significant inertia in energy forecast inaccuracy has deterred investments in systems. Against the backdrop of the rapidly low-carbon technologies and resulted in inertia declining costs of clean energy technologies, in planning. inflexible PPAs with “take or pay” clauses are Studies have found that peer effects and also resulting in economically—and environ- social learning are important factors in decision-­ mentally—suboptimal energy systems choices. making related to climate change.47 Preferences Destruc tion | 149 on technology adoption or social change are not at certain times of the day or in certain seasons. cast in stone, but they change in response to deci- Such intermittency poses challenges for their inte- sions made by peers. The result is inertia in adop- gration into power systems, particularly before tion initially and then rapid adoption later once electricity storage is fully developed. Integration a critical mass of peers adopts the technology. also requires new approaches to demand-side Similarly, the production of low-carbon technol- management. The scale-up of the electrification ogies is subject to external economies of scale, of transport and buildings, which will increase which lead to lower production and operating the demand for electricity, also depends on the costs for all companies in the industry. These, in reliability of the power system. Scale-up requires turn, can increase profitability and competitive- accurate forecasting and forward planning. All ness. However, these cost reductions do not occur these aspects of the energy transition create bar- until a sufficient number of industry players or riers and necessitate strong systemwide coordina- scale is reached—meaning costs are higher for tion and institutional capacity. first movers. Legacy transmission networks built to serve large fossil fuel plants, along with outdated regulations, also create barriers to entry in Legacy transmission networks built ­ middle-income countries. Typically, legacy net- to serve large fossil fuel plants slow works have been designed for traditional energy diffusion of low-carbon energy sources for which generation can be located close Because of the market structure of electricity gen- to the source of demand (map 6.2). However, the eration and transmission, the deployment of clean renewable generation capacity for wind and solar energy technologies is more challenging and more must be built at the decentralized sites where complex than that for other technologies. Variable these natural resources are found. Thus networks renewable energy sources provide energy only to transmit power are critical for the scale-up of Map 6.2 Limited or outdated electricity transmission networks serve as barriers to the entry of renewable sources Source: Arkolakis and Walsh 2023. Note: The map displays electricity transmission networks in 2023. The two insets focus on areas of particular interest due to the geographic disparities in coverage of transmission networks in these regions. 150 | WORLD DEVELOPMENT REPORT 2024 variable renewable energy. Because electricity value becomes lower than their current expected transmission has elements of a public good that value.50 This disconnect could lead to overly high often result in its underprovision, transmission investment in and maintenance of infrastructure networks in many countries are not keeping pace that supports, and is supported by, the burn- with the ambitions of governments or the plans ing of fossil fuels, which later becomes stranded of firms. This inertia in transmission networks capital. In middle-income countries, coal-fired serves as a major barrier to the energy transi- power plants are the most exposed to the risk of tion in many middle-income countries. In addi- becoming stranded and may have to be retired tion, outdated regulations related to the siting of 10–30 years earlier  than they did in the past.51 renewables or permitting for rooftop solar also Several ­ expectations influence whether assets create inertia, hindering diffusion. become stranded. They include expectations about the implementation of climate policies, about technological progress, and about legal Destruction without creation: action  against high emitters. Stranding could The risks of becoming occur based on today’s projections of fossil fuel stranded nations production solely due to the current rate of tech- nological change, which is faster than expected.52 Preservation worsens obsolescence; The extent of stranding will depend on how pol- dynamic firms and mobile people are icy choices today shape expectations about the needed future, along with efforts today to facilitate the The low-carbon transition poses a major risk of reallocation of resources from sunsetting indus- accelerating the obsolescence of capital, skills, tries to growing ones. Efforts by countries, indus- and industries. It is a form of directed technical tries, or firms to preserve sunsetting industries change away from carbon-intensive production longer than is economically viable rather than processes.48 In perfectly competitive markets, reallocating resources to expanding ones risk the associated reallocation of labor and capital increasing transition risks and obsolescence. would have minimal transition costs because Overvalued assets also pose the risk of a car- workers and capital would smoothly and quickly bon bubble in financial markets. The scale of the adjust by switching jobs, moving to areas with fossil fuel industry is large enough to potentially growing demand, and supporting expansion of trigger broader financial crises. The exposure to greener firms. In practice, however, search com- correlated risks within and across portfolios, with plications, costs to acquire human capital, or ties many potentially stranded assets at risk of being to particular geographic areas may give rise to devalued simultaneously, alongside the under­ significant transition costs.49 Higher social, occu- exposure to assets in low-carbon technologies pational, and geographic mobility facilitates the with potentially higher returns, also pose sys- rapid movement of people out of declining indus- temic risks that could affect the financial system tries and into expanding ones, which in many as a whole.53 If undermanaged, this instability countries requires a move to another geographic could result in policy swinging away from tran- location. Likewise, dynamic enterprises and low sition to protecting the financial system more barriers to entry can lower transition costs within broadly. Ensuring an orderly transition and clar- countries. ity in direction is necessary to limit the potential Meanwhile, fossil fuel resources that cannot of this crisis.54 be burned and fossil fuel infrastructure no lon- These financial risks are heightened in coun- ger used risk becoming “stranded assets.” Stocks tries with more vulnerable financial systems. of unburnable carbon (such as coal reserves) Countries more dependent on resource revenues could become stranded resources if their future would be particularly exposed, limiting their Destruc tion | 151 capacity to respond to the shock. The degree to brown products, such as internal combustion which sectors are linked within the economy is engine vehicles, could find it relatively easy to also associated with the exposure of the financial transition, those with exports highly concentrated system, with greater links acting to potentially in a few low-complexity brown products (such multiply the effects of an initial shock, creating as commodities) have fewer opportunities for ripples through the wider economy, and exposing diversification. Of the top 10 countries facing the financial systems to further pressures.55 greatest risk of such “brown lock-ins,” as measured Countries with rich reserves of critical green by an index constructed by these authors, six are minerals will also need to ensure they avoid middle-income countries. Panel a of figure  6.7 repeating the mistakes of past natural resource displays the 10 middle-income countries most booms and contracting “Dutch disease” or a green exposed: Iraq, Libya, Angola, Equatorial Guinea, “resource curse.”56 The transition is also expected Azerbaijan, Nigeria, Algeria, Turkmenistan, to significantly increase the demand for minerals Timor-Leste, and Gabon. that are critical inputs in low-carbon technologies Middle-income countries with a high degree such as lithium, cobalt, copper, and rare essential of specialization of physical, institutional, and earths. This demand could present countries with human capital in declining sectors that cannot large reserves of such minerals, such as Chile, the be easily transitioned to new opportunities face Democratic Republic of Congo, and Namibia, the highest risks. The production of some carbon-­ with a potential economic opportunity. However, intensive products is similar to the production of critical minerals are often located in regions green ones—for example, there is a high degree of characterized by poor labor and environmen- overlap in the manufacture of internal combus- tal standards and considerable political fragility tion engine (ICE) vehicles and electric vehicles. In and corruption. Thus these increases in demand general, the risk of brown lock-ins is negatively could contribute to a proliferation of problematic and significantly associated with the ease of tran- mining practices and conflict. The failure of many sitioning to green or overall non-brown prod- hydrocarbon exporters to use export revenue to ucts. This risk is displayed in panel b of figure 6.7, diversify their economies, which now renders which shows the correlation between the Brown them vulnerable to the transition, should serve as Lock-in Index and the Transition Outlook, also a cautionary tale to mineral exporters. To avoid constructed by Andres et al. (2023) to measure another resource curse, revenue from a higher a country’s ease of transitioning from brown to demand for minerals could be used to support green products. investment in education, infrastructure, and the development of economic sectors that are higher Workers with obsolete occupations in value added. and skills will need support to avoid stranding Infusion of global technologies and The energy transition will lower the demand for diversification will be crucial for workers who extract and refine coal, natural gas, middle-income countries and oil. Employment in these activities tends Andres et al. (2023) have compiled a list of traded to be disproportionately occupied by workers “brown products,” whose use is likely to decline without a college education.58 Workers at legacy if the world is to mitigate climate change.57 They suppliers of electricity, such as the coal-fired then explore which countries are most at risk of power plants that are now being retired, will also seeing their productive capabilities “stranded.” be affected. At risk as well are those employed They find that, on average, brown products tend in energy-intensive manufacturing industries to be less complex than green products. Although such as basic chemicals, nonmetallic minerals, countries that export technologically sophisticated and primary metals, and in industries like ICE 152 | WORLD DEVELOPMENT REPORT 2024 Figure 6.7 Most of the countries currently “locked in” to declining brown industries are middle-income countries a. Brown Lock-in Index for top 10 highest-risk middle-income countries 3.5 3.0 Brown Lock-in Index 2.5 2.0 1.5 1.0 0.5 0 q a la a n n ia ria te n ne by ija ta bo Ira go er es ge is ba Li ui ig Ga -L An en Al lG N er or m Az ria m rk Ti to Tu ua Eq b. Brown Lock-in Index versus Transition Outlook for middle-income countries 2.5 2.0 KHM VUT 1.5 1.0 Transition Outlook LCA 0.5 VCT MDV 0 BLR TKM SWZ UZB EGY IRQ –0.5 NAM BRA IDN MMR TUV CIV RUS IRN –1.0 ARG MNG ECU AZE GHA COL KAZ DZA –1.5 CMR NGA PNG AGO MRT –2.0 GNQ TLS BOL COG LBY –2.5 NIC GAB –3.0 –1.0 –0.5 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Brown Lock-in Index Lower-middle-income Upper-middle-income Source: WDR 2024 team analysis based on Andres et al. (2023). Note: Panel a displays the Brown Lock-in Index for the 10 middle-income countries with the highest index values. See BLI (Brown Lock-in Index) (dashboard), Green Transition Navigator, London School of Economics and Political Science, London, https://green-transition-navigator.org/. Panel b displays the correlation between the Brown Lock-in Index and the Transition Outlook for all middle-income countries. The Brown Lock-in Index measures the risk that a country will be locked in to carbon-intensive industries. A higher index value indicates greater risk. The Transition Outlook measures a country’s ease of transitioning from brown to green products. A higher value indicates greater ease of transitioning. See Green Transition Navigator (dashboard), London School of Economics and Political Science, London, https://green-transition-navigator.org/. For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. Destruc tion | 153 vehicle manufacturing. Park et al. (2023) explore of high-carbon jobs.59 Using microdata represent- fossil fuel jobs in China and how their economic ing more than 130 million online work profiles, geography compares with that of low-carbon Curtis, O’Kane, and Park (2023) explore transitions technology jobs. In China, fossil fuel jobs are away from carbon-intensive production technol- concentrated around Shanxi and in the north ogies. They find that in 2021 only 0.7 percent of of China, whereas low-carbon technology jobs workers who transitioned out of a carbon-­intensive are concentrated in the cities on the east coast job found work in a green job. Conversely, the vast and particularly in the south, demonstrating the majority of workers obtaining green jobs do not challenge for workers to transition (map 6.3). come from carbon-intensive industries, but from Recent research on job transitions has demon- a wide range of other industries and occupations strated the challenge of transitioning workers out (such as sales manager, software developer, and Map 6.3 Low-carbon technology jobs in China are growing in manufacturing hubs on the southeast coast, whereas fossil fuel jobs are close to coal mines Source: Park et al. 2023. Note: In panel b, the share of fossil fuel jobs in 2019 is based on data from China Statistical Database, National Bureau of Statistics of China, Beijing, https://www.stats.gov.cn/english/. 154 | WORLD DEVELOPMENT REPORT 2024 marketing manager). On average, 20 percent of tran- low even though these states have a high density of sitions out of carbon-intensive jobs are into other existing carbon-intensive jobs. carbon-intensive jobs, whereas transitions into In conclusion, middle-income countries will manufacturing are the most common, accounting need to amplify the forces of creation, weaken for more than 25 percent of all transitions out of the forces of preservation, and manage the forces carbon-intensive jobs. Although in some US states, of destruction to advance technological prog- such as California, the rates of transition from dirty ress. The chapters that follow focus on policies to green jobs are relatively high, in others, such as that these countries can implement to help them West Virginia, the rates of green transitions are achieve these goals. Notes  1. Krueger and Yoo (2002). 21. See BloombergNEF (dashboard), Bloomberg, New York,  2. Davis and Haltiwanger (1990). https://about.bnef.com/.  3. Crises do not always benefit reallocation and have 22. Park et al. (2023). also been shown to have “sullying” effects, leading to 23. Bridge et al. (2013). the exit of otherwise productive firms or high-quality 24. See, for example, Saussay et al. (2023). jobs, with long-lasting scarring effects (Haltiwanger 25. Curtis, O’Kane, and Park (2023). et al. 2021). 26. Cirera, Comin, and Cruz (2024).  4. Haltiwanger et al. (2021). 27. Bastos et al. (2023).  5. Bresnahan and Raff (1991). 28. “Net zero” refers to the balance between the amount of  6. Hershbein and Kahn (2018). greenhouse gas produced and the amount removed  7. Bruhn, Demirgüç-Kunt, and Singer (2021). from the atmosphere. It can be achieved through a  8. Peters et al. (2012). combination of emissions reduction and removal.  9. UNEP (2009). 29. In the five years before the COVID-19 pandemic (2015– 10. Kwon, Lowry, and Verardo (2023). 19), the growth rates in the share of electricity gener- 11. Way et al. (2022). ated from solar and wind energy were higher in 12. Bastos et al. (2023). middle-income countries than in high-income coun- 13. Bastos et al. (2023). tries. The average growth rate in middle-income coun- 14. Climate and low-carbon competitiveness are tries was about 13 percent for wind and 44 percent for also  now covered in a wide number of new solar, compared with 11 percent and 16 percent, p olicies, including the CHIPS and Science Act US  ­ respectively, for high-income countries. The average of 2022 and Infrastructure Investment and Jobs Act growth rate of EVs per capita in middle-income coun - of 2021. tries over the five years preceding the COVID-19 pan - 15. Bistline et al. (2023). demic was in line with that of high-income countries, 16. Arkolakis and Walsh (2023). but in 2021 it was more than twice as high in middle-­ 17. Jones (2023); Xue (2024). income countries than in high-income countries. 18. This is similar to the IEA’s finding that clean energy jobs 30. See OECD (2023); RISE (Regulatory Indicators for now account for about 2 percent of all jobs globally Sustainable Energy), Data Catalogue, World Bank, (IEA and IFC 2023). Washington, DC, https://datacatalog.worldbank.org​ 19. Several smaller cities and those in middle-income /se a r c h /d a t ase t /0 0 4 0 4 47/ Wo r l d - - - Re g u la to r y​ countries are experiencing sizable job creation from -Indicators-for-Sustainable-Energy. low-carbon technologies. For example, 5 percent of 31. PVOUT (Photovoltaic Power Potential) (dashboard), online job postings in 2022 in Rio de Janeiro were by Data Catalogue, World Bank, Washington, DC, https:// low-carbon technologies, while in small provincial cit- datacatalog.worldbank.org/search/dataset/0038641. ies in China, such as Ningde, home to one of China’s 32. This emissions intensity could also reflect the compo - largest battery manufacturing clusters, 8 percent of all sition of products produced within these industries, so online job postings were related to low-carbon it is an imperfect proxy. technologies. 33. Albornoz et al. (2009). 20. Crunchbase is a US-based company that provides 34. See, for example, Holladay (2016); Richter and information about start-ups using data sourced from Schiersch (2017). investors and community contributors, such as start- 35. Eskeland and Harrison (2003). ups themselves. Although it includes companies in 36. Volz et al. (2021). more than 200 countries, it may underrepresent start- 37. Overland, Juraev, and Vakulchuk (2022). ups in emerging markets that are funded only domesti- 38. See, for example, Kwon, Lowry, and Verardo (2023); cally and do not have a global presence. Thus they do Srivastav and Rafaty 2022; Stokes (2020). not choose to list themselves and are not disclosed by 39. Skovgaard and van Asselt (2018). investors. For more information, see Dalle, den Besten, 40. Agnolucci et al. (2023); Agnolucci, Gencer, and Heine and Menon (2017). (2024). Destruc tion | 155 41. IEA (2023). 50. Caldecott et al. (2021). 42. Black et al. (2023). It is much more difficult to define the 51. Fofrich et al. (2020). tax benchmark. As a result, individual country data 52. Mercure et al. (2019). should not be compared or aggregated, and caution 53. Caldecott et al. (2021). should be applied when comparing producer subsidies 54. Daumas (2024). across countries. 55. Hiebert and Monin (2023). 43. Damania et al. (2023). 56. The “resource curse” is the phenomenon in which 44. See IFC (2023); Welch-Phillips and Goldenberg (2022). countries with an abundance of natural resources (such 45. Grubb et al. (2021). as  fossil fuels  and certain  minerals) have less  eco - 46. See, for example, Acemoglu et al. (2012, 2016); Aghion nomic growth, less  democracy, or worse  develop - et al. (2016); Grubb et al. (2021). ment  outcomes than countries with fewer natural 47. See, for example, Dechezlepretre et al. (2022); Talevi resources. et al. (2022). 57. Brown products are based on traditional convention 48. Curtis, O’Kane, and Park (2023). Directed technical production styles, whereas green products are sensi- change refers to technological progress steered by pol- tive to the environment, emphasizing zero waste and icy choices or other factors in a certain direction and net zero carbon. so affects factors of production differently, as first 58. Jacobsen and Parker (2016); Raimi (2021). defined in Acemoglu et al. (2012). 59. Curtis, O’Kane, and Park (2023). 49. See, for example, Manning (2021). References Acemoglu, Daron, Philippe Aghion, Leonardo Bursztyn, and Working paper, Development Research Group, World David Hémous. 2012. “The Environment and Directed Bank, Washington, DC. Technical Change.” American Economic Review 102 (1): Bistline, John, Geoffrey Blanford, Maxwell Brown, Dallas 131–66. Burtraw, Maya Domeshek, Jamil Farbes, Allen Fawcett, Acemoglu, Daron, Ufuk Akcigit, Douglas Hanley, and William et al. 2023. “Emissions and Energy Impacts of the Kerr. 2016. “Transition to Clean Technology.” Journal of Inflation Reduction Act.” Science 380 (6652): 1324–27. Political Economy 124 (1): 52–104. Black, Simon, Antung A. Liu, Ian Parry, and Nate Vernon. Acemoglu, Daron, David H. Autor, Jonathon Hazell, and 2023. “IMF Fossil Fuel Subsidies Data: 2023 Update.” Pascual Restrepo. 2022. “Artificial Intelligence and IMF Working Paper 169, International Monetary Fund, Jobs: Evidence from Online Vacancies.” Journal of Washington, DC. Labor Economics 40 (S1): S293–S340. Bresnahan, Timothy F., and Daniel M. G. Raff. 1991. “Intra- Aghion, Philippe, Antoine Dechezleprêtre, David Hémous, Industry Heterogeneity and the Great Depression: The Ralf Martin, and John Van Reenen. 2016. “Carbon Taxes, American Motor Vehicles Industry, 1929–1935.” Journal Path Dependency, and Directed Technical Change: of Economic History 51 (2): 317–31. Evidence from the Auto Industry.” Journal of Political Bridge, Gavin, Stefan Bouzarovski, Michael Bradshaw, and Economy 124 (1): 1–51. Nick Eyre. 2013. “Geographies of Energy Transition: Agnolucci, Paolo, Carolyn Fischer, Dirk Heine, Mariza Space, Place and the Low-Carbon Economy.”  Energy Montes de Oca Leon, Joseph Pryor, Kathleen Patroni, Policy 53 (February): 331–40. and Stéphane Hallegatte. 2023. “Measuring Total Bruhn, Miriam, Asli Demirgüç-Kunt, and Dorothe Singer. Carbon Pricing.” World Bank Research Observer. https:// 2021. “Competition and Firm Recovery Post-COVID-19.” doi.org/10.1093/wbro/lkad009. Policy Research Working Paper 9851, World Bank, Agnolucci, Paolo, Defne Gencer, and Dirk Heine. 2024. Washington, DC. “Total Carbon Pricing for Energy Consumption: The Caldecott, Ben, Alex Clark, Krister Koskelo, Ellie Mulholland, Importance of Energy Taxes and Subsidies. Energy and Conor Hickey. 2021. “Stranded Assets: Environmental Subsidy Reform in Action Series.” ESMAP Technical Drivers, Societal Challenges, and Supervisory Responses.” Report. Washington, DC: World Bank. Annual Review of Environment and Resources 46 Albornoz, Facundo, Matthew A. Cole, Robert J. R. Elliott, and (October): 417–47. Marco G. Ercolani. 2009. “In Search of Environmental Cirera, Xavier, Diego A. Comin, and Marcio Cruz. 2022. Bridging Spillovers.” World Economy 32 (1): 136–63. the Technological Divide: Technology Adoption by Firms in Andres, Pia, Penny Mealy, Nils Handler, and Samuel Developing Countries. Washington, DC: World Bank. Fankhauser. 2023. “Stranded Nations? Transition Cirera, Xavier, Diego A. Comin, and Marcio Cruz. 2024. Risks and Opportunities towards a Clean Economy.” “Anatomy of Technology and Tasks in the Establishment.” Environmental Research Letters 18 (4): 045004. NBER Working Paper 32281 (March), National Bureau of Arkolakis, Costas, and Conor Walsh. 2023. “Clean Growth.” Economic Research, Cambridge, MA. NBER Working Paper 31615 (August), National Bureau Curtis, E. Mark, Layla O’Kane, and R. Jisung Park. 2023. of Economic Research, Cambridge, MA. “Workers and the Green-Energy Transition: Evidence Bastos, Paulo, Jacob Greenspon, Katherine Stapleton, and from 300 Million Job Transitions.” NBER Working Paper Daria Taglioni. 2023. “Did the 2022 Global Energy Crisis 31539 (August), National Bureau of Economic Research, Accelerate the Diffusion of Low-Carbon Technologies?” Cambridge, MA. 156 | WORLD DEVELOPMENT REPORT 2024 Dalle, Jean-Michel, Matthijs den Besten, and Carlo Menon. Holladay, J. Scott. 2016. “Exporters and the Environment.” 2017. “Using Crunchbase for Economic and Managerial Canadian Journal of Economics 49 (1): 147–72. Research.” OECD Science, Technology, and Industry Hopenhayn, Hugo A. 1992. “Entry, Exit, and Firm Dynamics Working Paper 2017/08, Organisation for Economic in Long-Run Equilibrium.” Econometrica 60 (5): 1127–50. Co-operation and Development, Paris. IEA (International Energy Agency). 2023. Fossil Fuels Damania, Richard, Esteban Balseca, Charlotte de Consumption Subsidies 2022. Paris: IEA. Fontaubert, Joshua Gill, Kichan Kim, Jun Rentschler, IEA (International Energy Agency) and IFC (International Jason Russ, and Esha Zaveri. 2023. Detox Development: Finance Corporation). 2023. Scaling Up Private Finance Repurposing Environmentally Harmful Subsidies. for Clean Energy in Emerging and Developing Economies. Washington, DC: World Bank. Paris: IEA. Daumas, Louis. 2024. “Financial Stability, Stranded Assets IFC (International Finance Corporation). 2023. “Power and the Low-Carbon Transition—A Critical Review of the Markets for Development.” IFC, Washington, DC. Theoretical and Applied Literatures.” Journal of Economic IRENA (International Renewable Energy Agency). 2023. Surveys 38 (3): 601–716. Renewable Power Generation Costs in 2022. Abu Dhabi, Davis, Steven J., and John C. Haltiwanger. 1990. “Gross United Arab Emirates: IRENA. Job Creation and Destruction: Microeconomic Jacobsen, Grant D., and Dominic P. Parker. 2016. “The Evidence and Macroeconomic Implications.” In Economic Aftermath of Resource Booms: Evidence NBER Macroeconomics Annual 1990, vol. 5, edited by from Boomtowns in the American West.” Economic Olivier Jean Blanchard and Stanley Fischer, 123–86. Journal 126 (593): 1092–128. Cambridge, MA: National Bureau of Economic Research; Jones, Florence. 2023. “China to ‘Shatter’ Renewable Cambridge, MA: MIT Press. Installation Targets by 2030.” https://www.power​ Dechezleprêtre, Antoine, Adrien Fabre, Tobias Kruse, -technology.com/news/china - shat ter- renewable​ Bluebery Planterose, Ana Sanchez Chico, and -installation-target/. Stefanie Stantcheva. 2022. “Fighting Climate Change: Krueger, Anne O., and Junghoo Yoo. 2002. “Chaebol International Attitudes toward Climate Policies.” Capitalism and the Currency-Financial Crisis in Korea.” NBER Working Paper 30265 (July), National Bureau of In Preventing Currency Crises in Emerging Markets, Economic Research, Cambridge, MA. edited by Sebastian Edwards and Jeffrey A. Frankel, Eskeland, Gunnar S., and Ann E. Harrison. 2003. “Moving 601–62. National Bureau of Economic Research to Greener Pastures? Multinationals and the Pollution Conference Report. Cambridge, MA: National Bureau Haven Hypothesis.” Journal of Development Economics of Economic Research; Chicago: University of Chicago 70 (1): 1–23. Press. Fofrich, Robert, Dan Tong, Katherine Calvin, Harmen Sytze Kwon, Sungjoung, Michelle Lowry, and Michela Verardo. De Boer, Johannes Emmerling, Oliver Fricko, Shinichiro 2023. “Firms’ Transition to Green: Innovation versus Fujimori, Gunnar Luderer, Joeri Rogelj, and Steven Lobbying.” ECGI Finance Working Paper 921/2023, J. Davis. 2020. “Early Retirement of Power Plants in European Corporate Governance Institute, Brussels. Climate Mitigation Scenarios.” Environmental Research Manning, Alan. 2021. “Monopsony in Labor Markets: Letters 15 (9): 094064. A Review.” ILR Review 74 (1): 3–26. Goldfarb, Avi, Bledi Taska, and Florenta Teodoridis. 2023. Mercure, Jean-François, Florian Knobloch, Hector Pollitt, “Could Machine Learning Be a General Purpose Leonidas L. Paroussos, Silviu Serban Scrieciu, and Technology? A Comparison of Emerging Technologies Richard Lewney. 2019. “Modelling Innovation and the Using Data from Online Job Postings.”  Research Policy Macroeconomics of Low-Carbon Transitions: Theory, 52 (1): 104653. Perspectives and Practical Use.” Climate Policy 19 (8): Grubb, Michael, Paul Drummond, Alexandra Poncia, 1019–37. Will McDowall, David Popp, Sascha Samadi, Cristina OECD (Organisation for Economic Co-operation and Penasco, et al. 2021. “Induced Innovation in Energy Development). 2023. “Effective Carbon Rates 2023: Technologies and Systems: A Review of Evidence and Pricing Greenhouse Gas Emissions through Taxes and Potential Implications for CO2 Mitigation.” Environmental Emissions Trading.” OECD Series on Carbon Pricing and Research Letters 16 (4): 043007. Energy Taxation. OECD, Paris. Haltiwanger, John C., Henry R. Hyatt, Erika McEntarfer, Overland, Indra, Javlon Juraev, and Roman Vakulchuk. and Matthew Staiger. 2021. “Cyclical Worker Flows: 2022. “Are Renewable Energy Sources More Evenly Cleansing vs. Sullying.” NBER Working Paper 28802 Distributed than Fossil Fuels?” Renewable Energy 200 (May), National Bureau of Economic Research, (November): 379–86. Cambridge, MA. Park, Geunyong, Ande Shen, Katherine Stapleton, and Hershbein, Brad, and Lisa B. Kahn. 2018. “Do Recessions Zhenxuan Wang. 2023. “The Economic Geography of Accelerate Routine-Biased Technological Change? Low-Carbon Technology Jobs in China.” Policy Report Evidence from Vacancy Postings.” American Economic (August), World Bank, Washington, DC. Review 108 (7): 1737–72. Peters, Glen P., Gregg Marland, Corinne Le Quéré, Thomas Hiebert, Paul, and Pierre Monin. 2023. “Climate-Related Boden, Josep G. Canadell, and Michael R. Raupach. Systemic Risks and Macroprudential Policy.” INSPIRE 2012. “Rapid Growth in CO2 Emissions after the Policy Briefing Paper 14 (August), London School of 2008–2009 Global Financial Crisis.” Nature Climate Economics and Political Science, London. Change 2: 2–4. Destruc tion | 157 Raimi, Daniel. 2021. “Mapping County-Level Exposure and Talevi, Marta, Subhrendu K. Pattanayak, Ipsita Das, Jessica Vulnerability to the US Energy Transition.” RFF Working J. Lewis, and Ashok K. Singha. 2022. “Speaking from Paper 21-36 (December), Resources for the Future, Experience: Preferences for Cooking with Biogas in Washington, DC. Rural India.” Energy Economics 107 (March): 105796. Richter, Philipp M., and Alexander Schiersch. 2017. “CO2 UNEP (United Nations Environment Programme). 2009. Emission Intensity and Exporting: Evidence from Firm- “The Global Financial Crisis and Its Impact on Level Data.” European Economic Review 98 (September): Renewable Energy Finance.” Division of Technology, 373–91. Industry, and Economics, UNEP, Nairobi, Kenya. Rosenow, Samuel, and Penny Mealy. 2024. “Turning Risks Volz, Ulrich, Emanuele Campiglio, Etienne Espagne, into Reward: Diversifying the Global Value Chains Jean-François Mercure, William Oman, Hector Pollitt, of Decarbonization Technologies.” Policy Research Gregor Semieniuk, and Romain Svartzman. 2021. Working Paper 10696, World Bank, Washington, DC. “Transboundary Climate-Related Risks: Analysing Saussay, Aurélien, Misato Sato, Francesco Vona, and the Impacts of a Decarbonisation of the Global Layla O’Kane. 2023. “Who’s Fit for the Low-Carbon Economy on International Trade, Finance, and Money.” Transition? Emerging Skills and Wage Gaps in Job Ad Paper prepared for the 9th IMF Statistical Forum, Data.” Centre for Climate Change Economics and Policy “Measuring Climate Change: The Economic and Working Paper 406, Grantham Research Institute on Financial Dimensions,” Washington, DC, November Climate Change and the Environment Working Paper 17–18, 2021. 381, London School of Economics and Political Science, Way, Rupert, Matthew C. Ives, Penny Mealy, and J. Doyne London. Farmer. 2022. “Empirically Grounded Technology Schumpeter, Joseph Alois. 1942. Capitalism, Socialism and Forecasts and the Energy Transition.” Joule 6 (9): Democracy. New York: Harper and Brothers. 2057–82. Skovgaard, Jakob, and Harro van Asselt, eds. 2018. The Welch-Phillips, Ian, and Cara Goldenberg. 2022. “Market Politics of Fossil Fuel Subsidies and Their Reform. Structures: Global Energy Transformation Guide, New York: Cambridge University Press. Electricity.” RMI Innovation Center, Rocky Mountain Srivastav, Sugandha, and Ryan Rafaty. 2022. “Political Institute, Basalt, CO. Strategies to Overcome Climate Policy Obstructionism.” Xue, Yujie. 2024. “China to Meet Its 2030 Renewable Energy Perspectives on Politics 21 (2): 640–50. Target by End of This Year: State-Owned Researcher.” Stokes, Leah Cardamore. 2020. Short Circuiting Policy: South China Morning Post, July 1, 2024. https://www​ Interest Groups and the Battle over Clean Energy and .scmp.com/business/china-business/article/3268707​ Climate Policy in the American States. Studies in /china-meet-its-2030-renewable-energy-target-end​ Postwar American Political Development. New York: -year-state-owned-researcher?campaign=3268707&m Oxford University Press. odule=perpetual_scroll_0&pgtype=article. 158 | WORLD DEVELOPMENT REPORT 2024 Part 3 Making Miracles P art 2 of this Report describes how the energy markets and other sectors now forces of creation—the protagonist of eco- dominated by favored entities, especially nomic growth—are weak in middle-income state-owned enterprises. In addition, countries. Many large incumbents do not innovate middle-income countries will need to or infuse global technologies, and many entrants strengthen the capacity of competition do not disrupt. By contrast, the forces of preser- agencies to identify and rein in firms that vation—the antagonist to creation—are strong in abuse their incumbency advantage. Insti- middle-income countries. Incumbent firms and tuting progressive tax policies to discipline social elites often preserve the status quo. Mean- incumbent elites while still incentivizing while, the forces of destruction—which are often innovation will be needed as well. needed to remove the obstacles to creation—are • Rewarding merit (chapter 8). To strengthen held back by the forces of preservation. Together, the forces of creation, middle-income the forces of creation, preservation, and destruc- countries should reward merit activities— tion are unbalanced in middle-income countries those with positive externalities; that struggling with slower growth. is, those with positive effects on general Part 3 of this Report examines how well-being. This entails redesigning poli- middle-income countries can accelerate growth cies to focus on assessing the value added by balancing these three forces, which entails of firms in jobs, exports, technology infu- the following: sion, and innovation instead of designing policies that simply focus on the size of • Disciplining incumbency (chapter 7). firms. Middle-income countries will also Middle-income countries will need to need to design social policies that avoid weaken the forces of preservation that targeting the distribution of incomes protect incumbents from healthy com- and instead focus on allocating talent and petition. Contestable markets—and the human capital more efficiently by ensur- policies that enable them—give incum- ing that all those who have talent and bent firms the incentives to compete and display acquired ability have access to edu- upgrade their capabilities because their cation, employment, and business oppor- products and processes could be replaced tunities. Furthermore, energy policies by producers from other countries. This will need to shift from targeting the dis- effort also entails opening access to tribution of energy sources to supporting  | 159 activities that reduce emissions and local crisis, guaranteeing free access to pub- air pollution. lic health care, an old-age pension, and • Capitalizing on crises (chapter 9). The expanded unemployment insurance for all destructions of outdated arrangements— citizens—particularly women, who bear a essential for creation—are often weak disproportionate burden of caregiving. during boom times. Crises play an important role in reallocating resources The development strategies outlined in chap- and making room for new arrangements— ters 7, 8, and 9 must be tailored to the transi- in enterprises, jobs, technologies, policies, tions needed for a middle-income country to and public institutions. For example, the achieve high-income status. For each transition, energy crisis and the climate crisis are middle-income countries must, to escape being driving rapid technological progress and trapped in slower growth, adapt their institu- the deployment of low-carbon technol- tions to balance the forces of creation, preserva- ogies. Middle-income countries have an tion, and destruction—the key dynamic for any opportunity to infuse global technologies country aspiring to combine investment with and produce green intermediates for global infusion and innovation. Balancing these forces markets. Crises can also foreshadow social will require policy makers to make miracles. But change. In the Republic of Korea, the gov- only a handful of countries have succeeded. It will ernment adopted comprehensive welfare require implementing difficult and painstaking reforms following the 1997–98 financial reforms, but the payoff will be significant. 160 | WORLD DEVELOPMENT REPORT 2024 7 Disciplining Incumbency Key messages • By promoting contestable markets, middle-income countries can strike a balance between supporting incumbents and ensuring that they do not abuse their market power. • Middle-income countries will need to update their institutional arrangements that favor i­ncumbents, including by retracting the government’s direct involvement productive enterprises, challenging state-owned incumbents in the electricity in  ­ industry, and w ­ eakening the norms that discriminate against women and other marginalized groups. • Middle-income countries can encourage incumbents to strengthen creation through policies that enhance trade openness, support upgrading firm capabilities, and incentiv- ize the free movement of ideas and people, including highly skilled talent. • Middle-income countries can ensure the effectiveness of competition authorities to rein in errant incumbents and use fiscal policy to make elites contestable. Balancing incumbents’ strike a balance between supporting the growth of market leaders and ensuring that these firms innovation and abuse of do not abuse the market power that comes with dominance their larger scale in their market. For example, in Part 2 of this Report highlighted that incumbent Italy labor productivity has stagnated for more firms, especially market leaders, can be the van- than two decades. Recent research has high- guards of technical progress. But they can also lighted how incumbents use their dominance to use their incumbency advantage to block other hold back progress.1 Market leaders tend to bol- firms from entering the market and competing, ster political connections by hiring more local thereby resisting progress. In designing policies politicians as they gain larger market shares; at to facilitate investment alongside infusion and the same time, they reduce their efforts to inno- innovation, middle-income countries will need to vate (figure 7.1). 161 Figure 7.1 In Italy, market leaders increase their political connections while reducing innovation 60 Market leader 2.7 Number of patents per 100 workers Number of local politicians employed per 100 workers 50 2.6 40 2.5 30 2.4 Market leader 20 2.3 0 5 10 15 20 Firm’s market rank Patents Politicians (right axis) Source: Akcigit, Baslandze, and Lotti 2023. Note: The x-axis shows the top 20 firms in a market ranked by employment share, with the rank of 1 indicating the highest employment share (“market leader”). The assessment covers markets defined at the six-digit industry level for 20 regions in Italy from 1993 to 2014. The dark blue line is the line of best fit for patents per 100 workers. The orange line is the line of best fit for the number of local politicians employed per 100 workers. “Six-digit” refers to ATECO (Classification of Economic Activity) (database), Italian National Institute of Statistics, Rome, https://www.istat.it/en/classification​ /­ateco-classification-of-economic-activity-2007/. The challenges of incumbency extend beyond Murthy (Infosys). Many in the economic elite have firms. The economic elite has grown rapidly in amassed their fortunes through their innovative many middle-income countries. In 2004, only work—a reward for their ability, grit, risk-taking, 20  percent of the billionaires in the Forbes list and ingenuity. But others have inherited wealth were from emerging markets.2 By 2014, this share or thrived in environments where business had risen to 43 percent and by 2023 to 48 percent. ­ regulations, government patronage, and limited China ranks second in the number of billionaires international competition have protected them.3 (496) after the United States (735). The recent Dominant incumbents can block the policy surge of billionaire wealth in emerging markets reforms needed to promote social mobility and can be traced to the creation and management talent development. In early 2013, Mexican pres- of new companies. Some company founders and ident Enrique Peña Nieto tried to take on one executives are likened to superstar billionaires by of the world’s most powerful teacher unions, creating popular products known by millions the National Union of Workers in Education or introducing innovative production methods (Sindicato Nacional de Trabajadores de la that expand consumer choices and reduce prices. Educación, SNTE). The union was so large and Examples include Jack Ma (Alibaba) and Narayana powerful that it operated as a political machine. 162 | WORLD DEVELOPMENT REPORT 2024 Hundreds of members held administrative posi- Targeted interventions, often focusing on tions in the education system (including at the specific firms or individuals, are used as instru- highest levels), and the union exercised control ments of first resort by policy makers. But these locally over hiring decisions and, in some states, instruments end up being ineffective and can hin- over teacher payroll.4 Reforming teachers’ career der innovation if not designed with care. Policy paths and replacing underperforming teachers makers should recalibrate their strategies by first with those needed to promote children’s achieve- strengthening the institutions that weaken the ment were enormous challenges under those forces of preservation and then aligning the incen- circumstances. tives to strengthen forces of creation. Targeted State-owned enterprises (SOEs) are the interventions are effective when they follow insti- dominant incumbents in fossil fuel power tutions and incentives. generation. As a result, they block entry of new players in renewables using their outsize Updating institutions to market share, control of the grid, and influence on regulation. weaken the forces of Contestable markets—and the policies that preservation enable them—are vital for middle-income coun- Institutions—formal rules and informal norms— tries to discipline incumbents. Contestability are often well intentioned when they are designed means that incumbents feel pressure to compete in their specific contexts. But they may end up and upgrade because their existing products and persisting beyond their usefulness, protecting the processes can be displaced by technologically status quo. Such persistence hurts the economic sophisticated producers within their own coun- prospects of middle-income countries that need try or from other countries. Such contestability to rapidly change their growth model by adding is central to creative destruction. Contestability infusion and innovation to investment. Institutional is fostered by three sets of policies: (1) institutions arrangements that favor incumbents will then that weaken the forces of preservation; (2) incen- need updating. tives that strengthen the forces of creation; and (3) interventions that target errant incumbents to Retracting protection of incumbents, destroy harmful arrangements (figure 7.2). including state-owned enterprises Figure 7.2 Promoting contestability through In middle-income countries, institutional inertia institutions, incentives, and interventions protects incumbents, strengthening the forces of preservation. Turnover, especially among market Disciplining leaders, is low. In many middle-income countries, incumbency a small number of companies dominate markets, a survey suggest (figure 7.3). Policy makers will need to retract govern- ment’s direct involvement in productive enter- Contestability prise, while removing outdated regulations that favor and protect incumbents. Often, incumbents are protected by licenses that limit the number of market participants or directly restrict mar- ket entry. Standards and minimum firm sizes can further protect incumbent interests. These Institutions Incentives Interventions dangers should be weighed against benefits, such as the incentives to improve quality that Source: WDR 2024 team. Disciplining Incumbency | 163 Figure 7.3 In many middle-income countries, markets are dominated by a few business groups, as a survey suggests 7 Extent of market dominance (0–7) 6 (survey responses) 5 4 3 2 1 0 d la y ng ia Ar ma . ge nia rm s In d CD C ia Af a n kh lic Ko e bia a o a S e Po om ite er n er a ak et ia Tü tan e e ov ia In rtu a Fe Br a Co Ch e Co ne l So sta sia Ka ep am lo ile r il do ga p ite Ma an de az Ge tate iy d ag h ic in av hin Cz ag i ric Bu atio Un av apa n re xic Ki ys d ov Vi lgar Sl ch Po en Ro , Re za ub la rk ut R nt m R N s d p J M 5 n OE To ia Un Sl ss Ru Middle-income High-income Source: WDR 2024 team based on Schwab (2019). Note: The survey question: “In your country, how do you characterize corporate activity?” [1 = dominated by a few business groups; 7 = spread among many firms]. OECD = Organisation for Economic Co-operation and Development. standards  provide.5 For example, the standard business dynamism, and the effects were most connector (SC), a fiber-optic connector devel- prominent in states with labor market institu- oped by Tyco Electronics (now known as TE tions that favored employers.7 Connectivity Ltd) used in data networking and Often, dominant incumbents in local markets telecommunications, has become the dominant lobby local authorities to erect ad hoc entry bar- international standard. How? By joining national riers. In Italy, such local entry regulations in the standardization organizations in many countries, retail market increased price margins by 8 per- the company influenced the standard-making cent and reduced the productivity of incumbent process, with the result that European and inter- firms by 3 percent in the early 2000s.8 In Peru, national standards refer to the SC connector. when the national competition agency (Indecopi) Tyco thus gained a significant global market strengthened its powers to dismantle local and share, earning an additional US$50–$100 million sector-specific regulatory and administrative in profits between 1995 and 2004.6  entry barriers in 2013, productivity increased sig- India’s License Raj—a system of central con- nificantly, including for firms operating in down- trols introduced in 1951 regulating entry and pro- stream sectors or in the same municipality.9 duction activity in the registered manufacturing Product market regulations (PMRs), inten- sector—is another example. The system favored tionally or inadvertently, protect incumbents and incumbents and stifled Indian entrepreneurs constrain competition (figure 7.4). These regula- for more than four decades. Its dismantlement tions include regulatory barriers to firm entry and during the 1980s and 1990s amplified entry and competition in a broad range of key policy areas, 164 | WORLD DEVELOPMENT REPORT 2024 Figure 7.4 In middle-income countries, restrictive product market regulations are pervasive 3.5 Economywide PMR score (0–6) 3.0 2.5 2.0 1.5 1.0 0.5 0 rm e e a Fe t N n av m ov y rtu a C l er e Ja ge ak Po an pu d Ko e lic m es Ka olo ria kh ia Co Tür on So sta iye h ica In enti il do na Ch ia ite , R o S . C lga a de am al a Ar Br a a ga d ep Sl an g z av hil Ge rag Cz eni Po chi si in Bu ani M fric ia Vie sta Re lan Un ea ic s za mb 5 do a b Ro tat p a ti k ne ay ut R x ra A p ing e M To d K r CD ite n ov OE Un Sl ss Ru Middle-income High-income Source: WDR 2024 team based on 2018 data from PMR Database (OECD-WBG Product Market Regulation Database) (dashboard), Data Catalogue, World Bank, Washington, DC, https://prosperitydata360.worldbank.org/en/dataset​ OECDWBG+PMR; PMR Indicators (Indicators of Product Market Regulation) (dashboard), Organisation for Economic /­ Co-operation and Development, Paris, https://www.oecd.org/en/topics/sub-issues/product-market-regulation.html#:~:text​ =The%20PMR%20economy%2Dwide%20indicators,existing%20regulations%2C%20and%20foreign%20trade. Note: The economywide Product Market Regulation (PMR) indicators measure the regulatory barriers to firm entry and competition in a broad range of cross-sector policy areas, as well as in specific services and network sectors. The PMR indicators range from 0 to 6. A lower value indicates that a regulatory regime is friendlier to competition. OECD = Organisation for Economic Co-operation and Development. ranging from licensing and public procurement these inputs and those closer to the produc- to governance of SOEs, price controls, evalua- tivity frontier. Reforms in services sectors can tion of new and existing regulations, and foreign positively influence the productivity of manu- trade. They also include restrictive regulations facturing firms, emphasizing the importance of in key network and services sectors (figure 7.5). regulatory reform in enhancing overall economic Regulations tend to become less restrictive as performance. country incomes rise. Public ownership and its weak governance are Regulatory restrictions in critical input sectors significant entry barriers. Businesses of the state can lead to adverse effects that constrain firm (BOSs)—enterprises with majority or minority performance, job creation, and productivity in ­ state shareholdings—act as powerful incum- downstream industries. Anticompetitive regu- bents in many middle-income countries. Several lations in key upstream sectors such as energy, middle-income countries score much higher than ­ transport, and communications, as well as pro- the Organisation for Economic Co-operation fessional services such as legal and accounting, and Development (OECD) average on the pres- ­ hinder productivity growth and export perfor- ence and weak governance of SOEs in the econ- mance in manufacturing firms.10 This effect is omy (figure 7.6). Recent research presented in the more pronounced in sectors heavily reliant on World Bank’s The Business of the State report11 Disciplining Incumbency | 165 Figure 7.5 In middle-income countries, both economywide and sectoral input and product market regulations are more restrictive than in high-income countries Economywide Distortions induced by state involvement Barriers to domestic and foreign entry Network sectors Sector Professional services Retail trade 0 0.5 1.0 1.5 2.0 2.5 3.0 PMR score (0–6) Middle-income High-income Source: WDR 2024 team based on data from PMR Database (OECD-WBG Product Market Regulation Database) (dashboard), Data Catalogue, World Bank, Washington, DC, https://prosperitydata360.worldbank.org/en/dataset/OECDWBG+PMR; PMR Indicators (Indicators of Product Market Regulation) (dashboard), Organisation for Economic Co-operation and Development, Paris, https://www.oecd.org/en/topics/sub-issues/product-market-regulation.html#:~:text=The%20PMR​ %20economy%2Dwide%20indicators,existing%20regulations%2C%20and%20foreign%20trade. Note: The economywide Product Market Regulation (PMR) indicators measure the regulatory barriers to firm entry and competition in a broad range of cross-sector policy areas, as well as in specific services and network sectors. The sector PMR indicators measure the regulatory barriers to firm entry and competition at the level of individual sectors, with a focus on network industries, professional services, and retail distribution. The PMR indicators range from 0 to 6. A lower value indicates that a regulatory regime is friendlier to competition. The figure includes data for 38 high-income countries and 22 middle-income countries. reveals that a doubling of states’ market share in a to upper-middle-income countries (figure 7.7, given sector is associated with 5–35 percent lower panel b). Mechanisms to improve SOE governance entry (figure 7.7, panel a). BOSs’ operational and and efficiency are still limited in many countries, financial performance lags behind that of their including separation of ownership from regula- private peers, and they, on average, have lower tion, as well as transparent and reliable informa- labor productivity, profitability, and return on tion on performance-based measures. investments, while operating with higher levels of debt vis-à-vis private counterparts.12 Advantages such as subsidies, exclusion from competition Challenging state-owned incumbents laws, preferential access to finance, restrictions on in the electricity industry foreign direct investment (FDI), and import bans Disciplining the incumbency advantage of pub- in sectors dominated by SOEs are more preva- lic ownership is most pressing in the electric- lent in lower-middle-income countries compared ity industry. Although the cost of lower-carbon 166 | WORLD DEVELOPMENT REPORT 2024 Figure 7.6 The BRICS and large middle-income countries have a significant presence of publicly owned enterprises and governance frameworks that stifle competition 5 4 ownership score (0–6) Economywide public 3 2 1 0 Ka er sia Co ma nd e l a, ic Ar a R a er y Ko epu co So ge ica h na k n Br p. do n a m rt e Ja es Tü ria P iye a ite m ia St a o n ak e ge n a ca am e e et an CD rm ia lg il Cz u g a av an Bu az si av il st ni d bi Sl pa za atio Po rag re bl Un Co ch OE G ven Re do 5 Ch at ia M Afri In h i ut nti Vi hst R i Ro o l a a a rk ne d y x N Fe la C ng o M e l Ki d p ite ov To Un Sl ss Ru Middle-income High-income Source: WDR 2024 team based on 2018 data from PMR Database (OECD-WBG Product Market Regulation Database) (dashboard), Data Catalogue, World Bank, Washington, DC, https://prosperitydata360.worldbank.org/en/dataset/OECDWBG​ +PMR; PMR Indicators (Indicators of Product Market Regulation) (dashboard), Organisation for Economic Co-operation and Development, Paris, https://www.oecd.org/en/topics/sub-issues/product-market-regulation.html#:~:text=The%20PMR​ %20economy%2Dwide%20indicators,existing%20regulations%2C%20and%20foreign%20trade. Note: The BRICS nations are Brazil, the Russian Federation, India, China, and South Africa. The public ownership indicator shows the extent of the presence of state-owned enterprises (SOEs) in the economy and their governance. A higher score indicates higher presence of SOEs with weaker governance of SOEs. OECD = Organisation for Economic Co-operation and Development. energy is declining rapidly, private providers in competitive with larger, utility-owned power many countries face major barriers to entry. SOEs plants. Arguments calling for economies of scope dominate fossil fuel power generation, mainly in no longer hold because advances in communica- coal-fueled power plants, where they account for tion technology have reduced coordination costs, 84 percent of total installed capacity (figure 7.8). enabling competition in bulk (wholesale) power By contrast, the private sector owns about an markets and then in retail sales. The retail choice equal share (80 percent) of the installed capacity is similar to how consumers choose a cell phone of renewable energy. plan; consumers can buy electricity from any sup- Historically, the electricity industry was consid- plier just as they would purchase mobile phone ered to be a “natural monopoly” with one dominant services from a telecom provider.13 However, as firm due to economies of scale in transmission and of 2020 in 125 developing countries 60 percent of distribution (networks) and the need to coordinate consumers still relied on a public distribution util- generation with the grid. Arguments related to ity for electricity. economies of scale became obsolete when smaller In South Africa, the vertically integrated co-generators and gas-fired power plants became power company Eskom dominates the power Disciplining Incumbency | 167 Figure 7.7 A state presence has important effects on firm entry, market concentration, and preferential treatment a. Effects of doubling state presence b. Direct (explicit) and indirect (implicit) on firm entry and market concentration advantages for businesses of the state 30 35 20 30 Number of countries 10 25 Percent 0 20 –10 15 –20 10 –30 5 –40 0 Entry Market concentration Explicit Implicit Explicit Implicit (HHI) LMICs UMICs Brazil Ecuador Romania Advantage identified Türkiye Viet Nam No advantage identified Source: World Bank 2023a. Note: Panel a: Entry is based on the rate of entry of new firms in Romania and Türkiye and on the share of revenue accounted for by young firms (less than five years old) in Brazil, Ecuador, and Viet Nam. Market concentration is captured by the Herfindahl- Hirschman Index (HHI) (Herfindahl 1950; Hirschman 1964). For the World Bank Global Businesses of the State Database, see Dall’Olio et al. (2022). Panel b: Direct (explicit) advantages are legal provisions that explicitly favor a group of market players such as state-owned enterprises. These typically involve taxes, public debt, public procurement conditions, state support, and exemptions to legal frameworks. Indirect (implicit) advantages are regulations and enforcement conditions that exist at the product, sector, or economywide level but that, in practice, unlevel the playing field in favor of a group of players. These typically involve import restrictions, bans on licenses, price or quota regulation, and poor antitrust enforcement. Businesses of the state are enterprises with majority or minority state shareholdings. The figure is based on a sample of 58 World Bank Country Private Sector Diagnostics, including 33 for lower-middle-income countries (LMICs) and 16 for upper-middle- income countries (UMICs). market with a 90 percent market share. In Poland, SOEs are the largest providers of energy finance— four companies—Polska Grupa Energetyczna accounting for 60 percent of energy investment (PGE), Tauron Polska Energia, Energa, and (figure 7.9). A concern is that SOEs account for Enea—­ control nearly three-fourths of the mar- 50  percent of fossil fuel power investment glob- ket share in electricity production, with PGE ally, notably due to expanding coal plants in India ­ holding 40 percent. Poland’s transmission grid is and South Africa and gas plants in the Middle owned and operated by state-owned Polskie Sieci East and North Africa. Elektroenergetyczne.14 SOEs use their dominant SOEs can block entry of new market players position to thwart entrants and protect their using their outsize market share and control of markets by blocking technological change. the grid. To compete for customers, rival firms SOEs are also significant investors in energy, need access to the grid controlled by a large contributing 36 percent of global energy invest- incumbent. Thus access pricing becomes import- ment. In low- and middle-income countries, ant. Access prices are set by the sectoral regulator. 168 | WORLD DEVELOPMENT REPORT 2024 Figure 7.8 State-owned enterprises Figure 7.9 In low- and middle-income dominate coal power generation, while the countries, state-owned enterprises are private sector leads in modern renewable the largest investors in fossil fuel energy energy generation a. Coal b. Renewable energy 100 Investment by SOEs (%) 3% 7% 80 13% 13% 60 40 84% 80% 20 0 Low- and High-income China SOEs Private sector middle-income Source: Vagliasindi (2023) based on analysis of power plant Oil and gas Fossil fuel generation data for countries, including Pakistan, Poland, South Africa, RE&EE Electricity networks Average and Türkiye. Note: The figures for the private sector and state-owned Source: Vagliasindi (2023) based on IEA (2020). enterprise (SOE) shares of coal and renewable energy are Note: The figure shows the percentage share of investment the average shares of total installed capacity across the set by state-owned enterprises (SOEs) by fuel type. RE&EE = of countries included in the analysis, excluding ownership renewable energy and energy efficiency. by China, which is shown separately in dark blue. To set a fair access price, the regulator may need focusing only on simplistic concentration mea- the incumbent utility to reveal information about sures. These include examining its cost function. The incumbent has an incen- tive to overstate the cost of supplying the input to • The incentives of producers. In the near its competitors in order to raise the access price term, it is likely that electricity markets and try to eliminate the competition—that is, will feature a diverse set of firms, includ- “foreclose” the downstream market. Concerns ing publicly owned utilities, unregulated about anticompetitive foreclosure—through high generation companies, and traditional ver- prices, discriminatory conditions, and low qual- tically integrated regulated utilities. Each ity of service—are an important incentive for type of firm is likely to respond differently regulation of access in network industries such to a given competitive environment. as energy supply and telecommunications that • The price responsiveness (elasticity) of rely on networks to transmit and distribute their demand. In markets in which customers services (network industries). can easily choose not to consume a prod- Incumbents can collude to block new entrants. uct or to consume a substitute instead, Price collusion and other forms of collusion in producers cannot raise their prices far wholesale markets are frequent in small power above costs without significantly reducing systems and illiquid markets and are difficult to sales. Conversely, a producer that knows detect in advance by sectoral regulators and after that buyers find its product essential can the fact by competition agencies. Various market profitably raise prices to very high levels. analyses are employed to capture the strategic • The potential for expansion of output by aspects of competition in this industry and avoid competitors and potential competitors. Just Disciplining Incumbency | 169 as a producer with very price-responsive successful careers for women as employees and customers cannot exercise much mar- entrepreneurs.16 However, women still lack basic ket power, neither can a producer faced legal rights in many countries related to running with many price-responsive competitors. a business, such as to sign a contract, register a Transmission capacity in a region and business, and open a bank account. In Papua New available competitive generation capac- Guinea, social norms that hinder female labor ity are the main factors determining the force participation and productivity are esti- potential for short-term competitive entry mated to cost the economy about 0.5 percent of or output expansion. the gross domestic product (GDP) a year.17 Legal reforms that grant better property and inher- Incumbents can be disciplined using less restric- itance rights also improve women’s social and tive PMRs. Ex ante rules on access to essential economic outcomes. For example, in India the infrastructure and open wholesale and retail elec- 2005 Hindu Succession (Amendment) Act, which tricity markets can enable entry by private pro- increased a daughter’s share of land inheritance viders, avoiding cases of “curtailment,” respecting from 8 percent to 16 percent,18 led to an increase the merit order (chapter 8), and encouraging entry in women’s education rates, labor force partici- through global competition. pation, entrepreneurship, and autonomy within their marriages.19 The starting point to bringing Weakening patriarchal gender norms more women into the labor force is implementing Norms that discriminate against women per- institutional reforms that grant women rights to petuate the hold of men (the incumbents) in the property ownership, inheritance, and other basic workplace. In the labor market, women remain rights to access economic opportunities. an underused resource, particularly in the Middle Social norms shape personal attitudes toward East and North Africa and in South Asia, where women’s participation in the labor force. The like- female labor force participation is only half the lihood of a wife being employed increases if her level expected given these regions’ income levels. husband’s mother worked during his childhood.20 However, some factors have contributed to an Parents also have a major impact on their chil- ongoing rise in female labor force participation: dren’s attitudes toward gender (even more than an increase in the kinds of jobs requiring brains their peers), with mothers exerting more influ- rather than brawn (such as professional posi- ence than fathers.21 But personal attitudes tend to tions); the increase in part-time jobs; the adoption be more progressive than collective social expec- of labor-saving household technologies; the grow- tations, leading to misconceptions about social ing number of educated women; the availabil- norms themselves.22 ity of contraceptives; the elimination of policies Broad-based education and information inter- that punished married women; and the decline ventions can help address both personal attitudes in social stigma against women working outside and misperceptions about social attitudes. For the home.15 In the United States, for example, the example, a study in India found that two years of gender gap in labor force participation has nearly classroom discussions about gender equality led to been eliminated. Countries with strong prefer- improved attitudes toward gender equality among ences for male children, such as China, Japan, and teenage boys and girls.23 An intervention that the Republic of Korea, have experienced similar encouraged teenage girls to question restrictive increases in female labor force participation as social norms in India combined with connecting their economies have grown. girls to changemakers in the wider community Equal treatment of women in the law is asso- reduced school dropout rates and early marriage ciated with higher female labor force partic- and improved mental health.24 In China, the Spring ipation, smaller gender wage gaps, and more Bud Project, initiated in 1989, promotes equitable, 170 | WORLD DEVELOPMENT REPORT 2024 inclusive, and quality education (primary to the assemblies in India, which, in turn, has increased higher education level) for girls by means of sub- female labor force participation.29 The rise in the sidies for those from low-income ­ families, large- number of female entrepreneurs has also boosted scale advocacy and awareness-­ raising, and skills female labor force participation and the number building for adolescent girls, including digital of women in higher management positions.30 In competencies for income ­ generation and employ- addition, educating the public about the costs and ment. In Viet Nam, early exposure to female 25 benefits of working is an important way in which classmates led to more egalitarian gender atti- female labor force participation and other mod- tudes in adulthood—even more so for men who ern practices can spread through society.31 grew up in conservative households.26 Bringing women into labor markets will mean The misperception about support for gen- providing them with support services such as child der norms is widespread across the world: many and elderly care, safety in the workplace, and trans- people think social support for working women port. Both men and women in Indonesia also cite is much lower than it is.27 For example, in Saudi harassment in public transport and commuting, Arabia misperceptions about gender norms as well as in the workplace, as important factors in restrict women’s basic rights to work outside of the women not wanting to work.32 For women, difficulty home, and yet most Saudi men privately support in finding childcare is the primary reason given for women working but underestimate the extent not working. A review of 22 studies across low- and to which others share this view. Correcting this middle-income countries reveals that increasing misperception can lead to a significant increase access to and reducing the cost of childcare can in female labor force participation.28 Saudi Arabia improve maternal labor market outcomes, includ- has experienced an unprecedented surge in female ing employment, hours worked, income, produc- labor force participation since 2017 as a result of tivity, and job type.33 Governments can use various changing regulations and shifting social norms, options that range from free state-provided care to the implementation of sound structural reforms, offering providers and parents financial subsidies, and effective government communications. Saudi tax incentives, or other forms of support. The chal- Arabia’s success in increasing female labor force lenge of accessing affordable childcare dispropor- participation from 17.4 percent in 2017 to 36 per- tionately affects poor families. Childcare subsidies cent in 2023 may contain important lessons for in low-income countries can help make childcare other countries and regions. more affordable. However, such targeted policies In Bangladesh, the robust growth of the gar- are not very common in lower-middle-income and ment manufacturing industry has generated upper-­ middle-income countries, where only 9 per- more than 5 million jobs, amounting to 60 per- cent and 41 percent of countries, respectively, have a cent of female employment, one-quarter of law that establishes some form of financial s­ upport industrial employment, and three-quarters of for families for childcare services, compared with the country’s export earnings in recent years. 80 percent of high-income countries.34 Employment in the garment industry has enabled In countries where women lack equal eco- migrant women from rural areas to earn cash for nomic and social rights, support for them will their families back home and has had far-reaching increase their chances of contributing to the labor positive effects on the welfare and empowerment market, including through gender-based affirma- of women in rural and urban areas. tive action policies in education or employment Role models, including women who challenge (box  7.1). Upskilling and training programs are traditional behavior, can change the behavior and more effective when complemented with other aspirations of other women. Their actions have led interventions that address social, family, and to an increase in female participation in political logistical constraints that women can face, such positions in local village councils ( panchayats) and as household and childcare responsibilities. Disciplining Incumbency | 171 Box 7.1 A digital tool helps female entrepreneurs obtain capital and training in rural Mexico To improve opportunities for indigenous rural women in some of Mexico’s southeastern states, a local nongovernmental organization, Pro Mujer, partnered with Google to help women access capital and entrepreneurship training. In February 2023, they launched a joint initiative, “Women: Force of the Southeast,” which includes an online platform to provide female entrepreneurs with free, customized online training. By means of hybrid sessions, participating entrepreneurs learn how to use social media to sell their products, prepare a budget, handle their finances, and communicate effectively. The platform also offers small loans with minimum requirements and no collateral requirements, as well as other financial services. In this region of Mexico, where about 80 percent of women live in poverty, digital technologies offer the promise of boosting skills, entrepreneurship, and incomes. The program is expected to benefit 6,000 women through small loans and about 2,000 women through entrepreneurship and financial literacy training. Sources: Google 2023; Pro Mujer 2023. Incentives for incumbents to Knowledge acquired from the global econ- omy holds the key to the economic catch-up of strengthen creation middle-income countries. Countries that have Openness in goods and product markets is criti- experienced sustained high growth have rapidly cal for firms in middle-income countries seeking absorbed know-how (knowledge that is not easily to absorb and infuse global knowledge. And yet transmittable), technology, and, more generally, openness to trade and foreign investment works knowledge from the rest of the world.36 Middle- two ways. Although openness provides firms income countries far from the knowledge fron- with access to larger markets, international value tier should act quickly and furiously to infuse chains, technology, and know-how, it also exposes knowledge. The World Bank Productivity Project domestic firms to competition with international highlights that roughly half of overall productiv- firms closer to the technology frontier. Firms can ity growth is driven by incumbent firms adopt- either scale up or be eased out. ing new technologies, products, and processes.37 Yet middle-income countries face an “innovation paradox”—returns on infusion are believed to be Upgrading by trading high in middle-income countries, and yet firms in A key part of contestability is trade openness. these countries appear to invest little.38 Because the firms in middle-income countries The incentives for and ability of middle-­ most able to adopt global technologies are often income country firms to absorb knowledge can large incumbents (see chapter 4), these incum- be enhanced by government policies that sup- bents will be more likely to compete by enhancing port upgrading the capabilities of firms. For this, products and processes to the extent that markets industrial policy will need to connect global are globally connected.35 and local firms and support adoption of modern 172 | WORLD DEVELOPMENT REPORT 2024 Figure 7.10 Foreign technology licensing is limited among middle-income country firms Share of firms using foreign technology licensing (%) KSV 30 AZE SVK MNG MDA ARM MAR TUR 20 BIH ROM LTU HRV TJK SVN GEO POL MKD SRB CZE KGZ EST BGR LVA BGD PAK ALB MNE 10 BLR JOR LBN UKR RUS GHA KAZ ZMB YEM TUN HUN UGA KEN ISR UZB IND COD EGY SSD NGA NAM NPL TZA SDN 0 6 7 8 9 10 11 GDP per capita (log scale) Low-income Middle-income High-income Source: Cirera and Maloney (2017) based on data from WBES (World Bank Enterprise Surveys) (dashboard), World Bank, Washington, DC, https://www.enterprisesurveys.org/en/enterprisesurveys. Note: For country abbreviations (except KSV, here representing Kosovo), see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. GDP = gross domestic product. organizational models. Because knowledge is engage in licensing (figure 7.10). Of considerable often embodied in machinery and equipment, concern is that rising geopolitical tensions and for many countries purchasing equipment from protectionism can potentially worsen the diffu- a foreign-owned company is an important way sion of knowledge to middle-income countries.40 to acquire knowledge. In fact, 45 percent of firms In view of the positive effects associated with in Asia and 29 percent in Africa, on average, are deploying advanced technologies, a govern- doing so.39 Technology licensing is also import- ment can help firms by using industrial policy to ant for infusion, as seen in Korea (see chapter 2). make it easier for them to license technologies. However, licensing increases with proximity to Countries can use infusion incentives, such as the technology frontier, with most firms unable Korea’s temporary subsidies for the adoption of to license foreign technologies, possibly reflecting foreign technology,41 while pursuing more general weak human capabilities in using these technolo- policies such as investing in upgrading domestic gies (chapter 8), as well as weak institutions pro- skills (chapter 8) and protecting intellectual prop- tecting intellectual property. Less than 10 percent erty. Market contestability also promotes faster of firms in middle-income countries such as the foreign technology adoption at a lower cost to Arab Republic of Egypt, India, and Tunisia use consumers.42 licensed technology. And even in other coun- Government policies to open markets and facil- tries that are more economically advanced, such itate economic integration can also help domestic as Croatia and Türkiye, only 20 percent of firms firms connect with multinational corporations Disciplining Incumbency | 173 (MNCs), thereby introducing a country’s firms to Local conditions and institutions also matter, foreign production techniques, overseas markets, such as national innovation systems and train- and international supply chains. The gains from ing programs, access to financing, and contest- such connections are evident in Poland. After able domestic markets. For example, the success Poland joined the European Union, the benefits of Norwegian FDI for salmon farming in Chile to direct suppliers in sectors that use research and stemmed in large part from the favorable local development (R&D) more intensively spread to the business conditions.45 broader economy through positive vertical spill- Increased participation in global value chains overs, while positive horizontal spillovers were also promotes diversification and sophistica- boosted as domestic firms accumulated intangible tion and magnifies the gains of traditional trade. assets such as licenses and patents.43 In Costa Rica, Exporting transmits knowledge back to the domestic firms increased productivity by 4 per- domestic economy through value chain links and cent and expanded their workforce by 26 percent learning by exporting.46 And long-term supply within four years of joining an MNC supply chain. relationships between exporters and foreign buy- Supplying MNCs also led to improvements in ers can promote upgrading in product quality and domestic firms’ production processes, supported management practices.47 However, mixed results by capacity building from the MNCs.44 have emerged from promoting specific industries Domestic rules can encourage sharing tech- or sectors through tax breaks, direct subsidies, nology with MNCs, such as licensing agree- import tariff exemptions, cheap credit, dedicated ments and the ability of local suppliers to acquire infrastructure, or the bundling of all of these in competitive or licensed technologies (box 7.2). export zones.48 Box 7.2 Technology for market access Quid pro quo policies mandate that multinational firms entering the domestic market establish direct partnerships with local firms through joint ventures. The objective is to amplify the positive spillovers from multinational firms to domestic ones, solidify their connections, and facilitate knowledge transfers. Despite the potential, little is known about the impacts of such policies. Two recent studies have examined a policy implemented in 1978 in China that required international automakers wishing to enter the Chinese market to establish joint ventures with domestic firms for production facilities.a The policy facilitated the transmission of knowledge between foreign and domestic firms through workers. The research indicates that the likelihood of workers transitioning from a joint venture to a domestic firm was 18 percentage points higher than a benchmark in which worker movements were random. Domestic firms affiliated with multinationals through a joint venture shared, on average, 12 common suppliers, or nearly seven more than the suppliers shared between unaffiliated domestic firms and multinationals. The enhanced network explains about 65 percent of knowledge spillover via joint venturing. Finally, this policy contributed to an enhancement in product quality, measured as a 3.8 percent reduction in defects per car model for affil- iated domestic firms. a. Bai et al. (2022); Zuniga (2024). 174 | WORLD DEVELOPMENT REPORT 2024 Improving firms’ capabilities—including variety of products traded.51 Direct support pro- through exporter training, country promotion, grams to encourage the integration of small firms and market research—has been shown to increase with large ones have been shown to increase the exports by reducing fixed costs and enabling export capabilities of small firms (box 7.3). But to firms to initiate exports, explore new interna- ensure that the benefits are durable, complemen- tional markets, or introduce new products.49 tary measures to facilitate technology absorption Openness to trade also matters. For example, in are also needed.52 Pakistan increases in upstream markets’ tariff The integration of economic principles in duties reduced the productivity of firms in the industrial policy design and implementation is downstream markets.50 In Peru, firms that were essential for infusion for three reasons. First, helped to enhance their capabilities in the early assessing the need for industrial policy to address 2000s experienced a 17 percent higher export identified market failures, as well as the opportu- growth rate than firms that did not participate. nity costs of state support, is important. Second, The higher rate was driven primarily by expan- ensuring that design and implementation foster sion in the number of countries served and the contestability and merit will mitigate the risks Box 7.3 Supplier development programs to connect small firms with large firms Chile’s Supplier Development Program establishes two-way connections between poten- tial suppliers that are small and medium enterprises (SMEs) and their large firm custom- ers. In such connections, SMEs benefit from higher sales and employment, resulting in a lower likelihood of exit, while large buying firms enjoy higher sales and bolster their export capabilities. The program provides the government with a subsidy to execute projects sponsored by large firms, on the condition that SMEs act as suppliers for these larger companies. The program also subsidizes activities to enhance SMEs’ technical capabilities, such as providing professional advice and facilitating technology transfers. In the agribusi- ness sector, supplier firms saw a 16 percent increase in sales and an 8 percent increase in employment within one year after the program was approved. And the large firm cus- tomers enjoyed a 19 percent increase in sales and a 3 percent gain in the probability of becoming an exporter.a In the early 2000s, Costa Rica’s new PROPYME program began promoting innovation among Costa Rican SMEs by facilitating their connections with research units. These units were associated with local or foreign universities or private research centers with no university affiliation. Together, they collaborate on projects geared toward technol- ogy development, innovation, growth of human capital, or technology transfer. In addi- tion to fostering these partnerships, the government is financing up to 80 percent of the projects’ total cost. PROPYME resulted in a substantial 19 percentage point increase in labor demand and a 9.6 percentage point boost in the probability of SMEs engaging in exporting.b Source: Zuniga 2024. a. Arráiz, Henríquez, and Stucchi (2011). b. Monge-González, Hewitt, and Torres-Carballo (2015). Disciplining Incumbency | 175 of market distortions and favoritism. And, third, higher exposure to foreign competition translated monitoring and evaluating the direct intended into higher value, quantity, and product shares of impact, as well as market and spillover effects, high-quality exports.58 will maximize effectiveness. The European Evidence from Chile and Mexico highlight the Union uses these principles in its state aid con- following: trol framework, and they are now being used and • In Chile, as imports of Chinese products implemented in different ways in various coun- rose at an average pace of 27 percent each tries, including Chile, Colombia, Moldova, and year from 2001 to 2007, Chilean firms the Philippines.  that were market leaders increased their product innovation by 15 percent and Upgrading to ward off foreign product quality by 22 percent. The lag- competitors gards scaled back their process innovation Trade openness improves access to larger mar- by 11 percent and product innovation by kets, international value chains, technology, and 13 percent.59 know-how. But it also exposes domestic firms to • In Mexico, firms making a “peripheral” competition from international firms closer to product (one of small importance in the the technology frontier. Such competition forces firm’s total sales) were more likely to pull domestic leaders to upgrade to “escape” competi- it off the market if it faced strong competi- tion and puts pressure on laggards to quit, espe- tion. However, products that were core to cially when domestic markets are threatened by a firm’s business (those with large shares of imports. Such competition can induce a reallo- total output) were less vulnerable.60 Larger cation of resources toward firms that are more plants and “core” products benefited from productive.53 It can also incentivize innovation by expanded access to cheaper imported well-positioned incumbents54 and upgrading of intermediates, helping firms improve the product quality, including by introducing greater competitiveness of core products. Further- variety.55 more, Mexican firms that faced competi- China’s accession to the World Trade tion from Chinese imports used existing Organization (WTO) is a clear and compelling information technologies to increase pro- example of how market leaders in different coun- ductivity, while firms that did not face tries have responded to competition in domes- competition as intense did not, even if tic markets. China’s overall exports grew from they acquired the same technology.61 US$62 billion in 1990 to US$1.2 trillion in 2007, a staggering average increase of about 20 percent a year. China entered the WTO in 2001. By 2009, Learning by moving it had become the world’s largest exporter and by Connecting with the world by means of trade 2010 the second-largest economy in the world. is a basic requirement for incentivizing incum- How have firms responded to rising competition? bents to upgrade, thereby sustaining growth at In 12  European countries, imports from China the middle-income level. But so is the free move- increased the innovative activity of European ment of ideas and people. Together, the move- firms that survived the competition, while reduc- ment of ideas, along with goods, services, capital, ing employment and lowering overall chances of and people, are critical for advancing technology firm survival.56 In heavily exposed sectors, low-tech diffusion.62 firms suffered declines in jobs and survival rates, Middle-income countries should not restrict while high-tech ones remained relatively safe. In the movement of their highly skilled individuals. Argentina, there was a positive association between Their emigration can be an opportunity for the competition and technology upgrading.57 In Peru, origin country rather than a loss. The extent to 176 | WORLD DEVELOPMENT REPORT 2024 which this can happen depends on how strongly By  engaging the Albanian diaspora in those emigrants remain connected to the origin countries, NanoAlb is able to deliver a variety ­ countries—or even return—and on emigrants’ of products, engage in R&D activities, and offer ability to accumulate knowledge about modern classes in applications of nanotechnology. The production processes and technologies in their center also disseminates knowledge to society destination countries (box 7.4). This is particu- at large and provides educational activities for larly relevant in conflict-affected countries such young students. as Ukraine that have experienced a large out- Convincing highly skilled emigrants to return flow of highly skilled individuals. Origin coun- home has been a priority for countries such as tries can also create a conducive environment China, where the Thousand Talents program for knowledge transfer involving the diaspora brings back researchers for permanent employ- by creating conditions such as political stability, ment or short-term visits.65 However, policies to institutional quality, and a favorable investment bring such migrants back can have mixed results. climate.63 In addition, origin countries can lever- R&D workers are highly mobile and respond to age immigrants’ earnings for economic growth monetary incentives as well as research support by enabling a safe, efficient flow of remittances. and the proximity of other researchers.66 Reducing Some countries have adopted measures to the tax rates for returning R&D workers may lead facilitate engagement with their diaspora and to an increase in the number of inventors through foster collaboration on R&D. For example, in both the retention of domestic inventors and the 2019 the Academy of Sciences of Albania created immigration of foreign inventors, although it may NanoAlb, a virtual center to coordinate nano- also reduce knowledge spillovers and productivity science and nanotechnology research in institu- in the countries from which they are returning.67 tions located in Albania as well as Israel, Italy, Regardless of the specific policies used, turning Kosovo, Montenegro, North Macedonia, Spain, brain drain into brain gain remains an imperative the United Kingdom, and the United States.64 in countries’ talent agendas. Box 7.4 Turning brain drain into brain gain A highly skilled migrant who moves to a high-income country but cannot find a job—or must work as a cab driver—cannot gain new skills, whereas one who works using his or her skills (as a manager, professional, or technician) is more likely to do so. Figure B7.4.1 measures the potential of origin countries to realize gains from knowledge spillovers from their diaspora. To quantify this potential for each origin country, the figure measures the extent to which tertiary-educated workers migrate to high-income countries and the share of tertiary-educated migrants who succeed by working in “good” occupations in the destination country. Countries in the upper right of the figure have the greatest poten- tial because their diaspora is large and successful. To leverage this potential, they should promote knowledge exchange and connections between the diaspora and local industry leaders and investors. Other countries send fewer highly skilled emigrants to advanced countries (lower values on the x-axis), or their migrants are less successful (lower values on the y-axis). Although these features can complicate knowledge transfers, a small number of skilled and successful emigrants may nevertheless be sufficient to transmit ideas or knowledge back to their sending communities. (Box continues next page) Disciplining Incumbency | 177 Box 7.4 Turning brain drain into brain gain (continued) Figure B7.4.1 Some countries are strongly positioned to benefit from knowledge spillovers from their diaspora Share of highly skilled migrants in “good” occupations (%) 70 LSO ZMB NZL AUS SUR ZAF 60 ZWE ISR FRA MWI KEN PNG SGP BHS BELESP VUT MLT TZA MYS CAN ISL SLE MUS UGA SWE USA LUX LBN NLD DEU IND NAM BRN FIN 50 SLB SWZ BWA BEN LBR CYP DNK GRC CHE IRL JAM ITA VNM BHR KWT MDGKHM AUT NGA GBR QAT GUY TTO IRN CHN MOZ ARE TGO CMR CRINOR PAN GHA JPN TJK VCT BLZ FJI ARG NER BFA BRB LAO SYR TUN PRT VEN TUR DZA DJI CPV COG RWA LBY MMR ETH LKAHTI PHL CIV HRV 40 WSM TCD CAF LCA URY BLR MKD JOR SEN BIH CHL IDN BRA MAR MDV GAB SVK EGY COL BDI SVN CZE HUN MRT ERI NPL CUB PAK GMB AGO UZB NIC BGR DOM RUS OMN SDN PER UKR POL YEM ARM SLV THA KGZ AFG GTM ROM MEX AZE SAU 30 MLI PRY SOM ALB ECU BGD TKM FSM PSE COD COM BOL HND IRQ GIN EST TON LVA MDA GNB GEO KAZ STP GNQ LTU 20 TMP MNG BTN 10 0 5 10 15 Number of migrants to Western Europe and North America (log scale) Low-income Middle-income High-income China and India Sources: WDR 2024 team calculations; DIOC (Database on Immigrants in OECD and Non-OECD Countries), reference years 2010/11, Organisation for Economic Co-operation and Development, Paris, https://www.oecd.org/els/mig/dioc.htm. Note: DIOC 2010/11 provides data on migration flows by skill and current occupation and covers migration flows from 200 origins to 34 OECD country destinations. Each scatter point in the figure represents an origin (or birth) country. For each birth country, the x-axis shows the number of tertiary-educated migrants who now live in destination countries in Western Europe or North America (log scale). The y-axis shows the share of these tertiary-educated migrants who work as a manager, professional, or technician in their destination country. These occupations are labeled “good.” They represent the top three one-digit International Standard Classification of Occupations (ISCO) codes. The sample is restricted to persons at least 15 years old. Tertiary education is defined as a completed tertiary education. The analysis includes the following 15 high-income Western European and North American countries as destination countries: AUT (Austria), BEL (Belgium), CAN (Canada), DNK (Denmark), FRA (France), DEU (Germany), IRL (Ireland), ITA (Italy), NLD (Netherlands), NOR (Norway), ESP (Spain), SWE (Sweden), CHE (Switzerland), GBR (United Kingdom), and USA (United States). For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. OECD = Organisation for Economic Co-operation and Development. 178 | WORLD DEVELOPMENT REPORT 2024 Interventions to correct errant a diverse set of markets, including manufactur- ing, construction, wholesale and retail trade, and behavior by incumbents transportation and storage. In middle-income Governments can also discipline incumbents countries in Latin America, elimination of anti- through targeted interventions. In countries competitive practices has increased innovation where competition laws are in place (typically and productivity (box 7.5). rules against abuse of dominance and anticom- For competition rules to be effective, threats petitive agreements, as well as merger controls), of enforcement must be credible. A few middle-­ competition authorities monitor, discourage, and income countries (such as Brazil, Egypt, Mexico, punish anticompetitive behavior by firms and and South Africa) have successfully decided cases prevent mergers that could harm competition. about cartels and abuse of dominance involv- According to the World Bank Anti-Cartel ing digital platforms, as per the Global Digital Enforcement Database, competition authorities Antitrust Database.69 And yet the staffing and in 34 middle-income countries sanctioned 406 budget limits of competition authorities in many cartels from 2017 to 2022.68 These cartels cover middle-income countries reduce their capacity Box 7.5 Tackling anticompetitive practices increases incumbents’ innovation incentives Economic cartels are prevalent across various economies affecting many markets. Unlike other forms of anticompetitive practices, such as abuses of dominance in which efficien- cies could counterbalance some of the negative effects, cartels in middle-income countries constitute the most harmful anticompetitive practices. Failure to address cartel activity limits productivity growth.a Anticompetitive agreements weaken efficiency incentives, and the presence of cartels can cut the growth of labor productivity by as much as 20–30 percentage points when compared with that for industries without cartels.b Evidence from a 40-year-old cartel in the United States suggests that sectoral output declined by 22 percent over the counterfactual. Cartels also harm export competitiveness by raising the cost of inputs, with negative implications for the development of both domestic and international value chains, thereby diminishing the benefits of trade liberalization. Among countries in the Pacific Alliance (which have the lowest trade barriers in Latin America), at least 67 cartels operate in trad- able sectors, and one-third of them have been in place for more than five years.c Cartels and abuse of dominance have been associated with lower wages in Mexico and lower incumbent productivity or innovation efforts in Chile, Colombia, and Uruguay. In Mexico, after antitrust sanctions were put in place wages grew by 1.4 percentage points a year and productivity rose by 2.4 percentage points a year.d In Colombia, following sanctions on a sugar cartel and a separate intervention to sanction abuse of dominance, sales and value added in the affected market increased, while markups fell (figure B7.5.1). Furthermore, the leading firms increased efforts to raise productivity as they dealt with growing competition in the market after price-fixing behavior was curtailed (figure B7.5.2).e (Box continues next page) Disciplining Incumbency | 179 Box 7.5 Tackling anticompetitive practices increases incumbents’ innovation incentives (continued) Figure B7.5.1 In Colombia, after a Figure B7.5.2 In Colombia, after an cartel is sanctioned, market outcomes abuse of dominance case, positive improve through the entry and growth market outcomes are driven by of previously lagging firms improvements in leading firms 0.4 1.6 0.2 1.4 1.2 0 Percent 1.0 –0.2 0.8 –0.4 Percent 0.6 0.4 –0.6 0.2 –0.8 0 Laggards Leaders –0.2 Sales Value added –0.4 TFPR Markups Laggards Leaders Source: Sampi, Urrutia Arrieta, and Vostroknutova Sales Value added 2022. TFPR Markups Note: The figure shows the changes in various market outcome variables after a cartel is Source: Sampi, Urrutia Arrieta, and Vostroknutova sanctioned for previously lagging firms and leading 2022. firms in the unaffected market. TFPR = total factor Note: The figure shows the changes in various productivity ratio. market outcome variables after an abuse of dominance case for previously lagging firms and leading firms in affected markets. TFPR = total factor productivity ratio. Sources: Bridgman, Qi, and Schmitz 2009; OECD 2014; Petit, Kemp, and van Sinderen 2015; Reed et al. 2022; Sampi, Urrutia Arreita, and Vostroknutova 2022; Vostroknutova et al. 2024; World Bank 2021. a. Petit, Kemp, and van Sinderen (2015). b. OECD (2014). c. World Bank (2021). d. Reed et al. (2022). e. Sampi, Urrutia Arrieta, and Vostroknutova (2022). 180 | WORLD DEVELOPMENT REPORT 2024 to act (figure 7.11). According to another sample, Competition authorities should also have the in those countries more than two-thirds of com- power to advocate the elimination of regulatory petition authorities have annual budgets of less restrictions of competition that can ultimately than US$5 million, while the average number facilitate anticompetitive practices. Competition of staff per million inhabitants in high-income laws that do not exclude certain firms (such as countries is more than 70 percent higher than businesses of the state) or specific sectors71 are in middle-income countries.70 Upper-middle- also essential. income countries, in particular, should invest in The challenges with competition are even building independent and accountable competi- more pronounced in partially contestable mar- tion authorities that are adequately funded and kets and natural monopolies. For example, to staffed. ensure efficient pricing, controls are needed for Figure 7.11 Competition authorities in middle-income countries need more capacity to deal with sophisticated policy problems 250 25 Number of competition staff per competition Average budget of competition authority 200 20 (real euros, millions) authority, 2022 150 15 100 10 50 5 0 0 Americas Asia-Pacific Europe Middle East and Africa Number of competition staff per competition authority Average budget of competition authority (right axis) Source: OECD Competition Statistics Survey data (OECD 2024). Note: OECD Competition Trends cover 77 OECD and non-OECD jurisdictions. For a list of countries included in each region shown in the figure, see OECD Competition Statistics (OECD CompStats) Survey 2024 (web page), Organisation for Economic Co-operation and Development, Paris, https://survey.oecd.org/index.php?r=survey/index&sid=371986&lang=en; OECD Competition Trends (portal), Organisation for Economic Co-operation and Development, Paris, https://www.oecd.org/fr​ /­corruption/oecd-competition-trends.htm. Disciplining Incumbency | 181 final services and products in the case of natural policy and support coalition building to make monopolies and only for essential inputs in the elites contestable. A society’s wealthiest members case of partially contestable markets. By contrast, often use their social and political power to slow price controls can completely undermine market creation and preserve the systems that benefit signals in competitive markets. them. Meanwhile, wealth is frequently inherited As economies (or sectors) move closer to the or acquired through rent-seeking, which does not technology frontier, competition agencies will create and add value to the economy. Instead, it need to consider a possible trade-off between manipulates the social and political conditions of innovation incentives and market power.72 When economic activity to distribute wealth upward. market power reduces the incentive to innovate, By adopting a progressive income taxation firms may resort to anticompetitive behavior, system, countries can compress the after-tax necessitating the intervention of antitrust poli- income distribution, reduce inequality, and pro- cies. Here, competition and innovation policies mote social mobility.76 Tax rates that are too high, will have to be coordinated to achieve the optimal however, can dampen incentives to undertake outcomes for innovation.73 Again, for the sake of high-­ return, high-risk innovation activities. For coordination, competition authorities must have example, in response to higher income taxes, adequate independence, budget, capacity, and innovators or entrepreneurs can reduce their technical sophistication. efforts, evade taxes, or migrate to lower-tax local- For upper-middle-income countries shifting ities. Inventors prefer to locate in the same places to innovation, a special concern should be the as other inventors in their specific domain.77 containment of killer acquisitions—that is, when Countries can use inheritance or estate taxes incumbents acquire innovative firms specifically to reduce wealth inequality while financing to kill future competing products and technolo- social protection programs. Progressive inher- gies.74 Not all acquisitions are deadly: many young itance taxes can motivate charitable giving by entrepreneurs try deliberately to be acquired by allowing tax deductions for donations by wealthy an incumbent, producing complementary inno- ­ individuals—and others—just as progressive vations that an incumbent can scale up. Antitrust income taxes often do. Charitable giving has agencies must use a risk-based approach to care- gained momentum in some middle-income fully examine the effects of risky acquisitions on countries, including the BRICs (Brazil, Russia, corporate innovation and future competition. India, China). And yet in terms of gifts per donor, Even if competition authorities use counter- Asia and Latin America rank the lowest (about factuals to anticipate the potential effects of a US$200,000). deal, building such scenarios in rapidly evolv- Overall, policy makers in middle-income coun- ing markets such as digital ones may result in tries, like those in advanced economies, must errors. Thus distinguishing good concentration strike a fine balance in disciplining economic elites from bad is hard.75 In these instances, a dynamic, without getting rid of the geese that could lay the forward-looking perspective is essential, building ­ golden eggs. What is critical is finding the optimal on credible data. Authorities should consider the tax rate that will balance disincentive effects with future potential of the acquired firms, recogniz- steps to lower inequality. Governments can also ing the possibility that these start-ups will grow offset some of the disincentive effects of progres- rapidly and become tomorrow’s superstar firms. sive taxation by supporting an enabling innova- While strengthening competition authorities tion environment, with universities, high-quality and regulatory institutions to discipline incum- infrastructure, urban amenities, and direct incen- bent firms, a government should also use fiscal tives for innovation (R&D subsidies). 182 | WORLD DEVELOPMENT REPORT 2024 Notes  1. Akcigit, Baslandze, and Lotti (2023). 37. For more information on the World Bank Productivity  2. Emerging markets are defined here as countries or Project, see https://www​.world​bank​.org​/en/topic​ territories that have a billionaire but are not one of the /­c ompetitiveness/brief​/ the ​- world ​- bank​- productivity​ high-income member countries of the Organisation for -project. Economic Co-operation and Development (OECD). They 38. Cirera and Maloney (2017). are mostly middle-income countries. The data for 2004 39. Cirera and Maloney (2017). and 2014 are from Freund (2016). 40. Melitz and Redding (2021); World Bank (2023b); and dis-  3. A study indicates that as of 2014 a substantial share cussion in chapter 3. of the extremely wealthy individuals in developing 41. Choi and Shim (2023). nations had amassed their wealth through inheritance 42. Phumpiu Chang and Castillo (2024). (29 p ­ ­ercent), resource exploitation, or preferential 43. Kolasa (2008). government ties and monopolies (16 percent), rather 44. Jordaan, Douw, and Qiang (2020). than through genuine productive investments (Freund 45. Fløysand and Barton (2014). 2016). More recent data are not available. 46. Akcigit and Melitz (2022); World Bank (2020).  4. Schneider (2022). 47. Verhoogen (2023).  5. Arayavechkit, Jooste, and Urrutia Arrieta (2022). 48. Commission on Growth and Development (2008);  6. de Vries 2006; Dixit and Gill (2024). Defever, Riaño, and Varela (2020); Defever et al. (2020);  7. Aghion et al. (2008). Lovo and Varela (2023).  8. Schivardi and Viviano (2011). 49. Lederman, Olarreaga, and Payton (2006).  9. Schiffbauer, Sampi, and Coronado (2022). 50. Varela et al. (2022). 10. Dauda and Drozd (2020). 51. Volpe Martincus and Carballo (2008). 11. World Bank (2023a). 52. Cadot et al. (2015). 12. Sanchez Navarro (2024). On average, BOSs are less 53. Melitz and Redding (2021). productive in terms of revenue per worker, have lower 54. Aghion et al. (2022). profit margins, and for every dollar in assets in a BOS 55. Verhoogen (2023). firm, the return is lower than the median private peer in 56. Bloom, Draca, and Van Reenen (2016). the same industry. 57. Bustos (2011). 13. Ellig (2020). 58. Medina (2022). 14. For details, see Vagliasindi (2023) and World Bank 59. Cusolito, Garcia-Marin, and Maloney (2023). (2023a). 60. Iacovone, Rauch, and Winters (2013). 15. Goldin (2006, 2014, 2021); Goldin and Katz (2002). 61. Iacovone, Pereira López, and Schiffbauer (2023). 16. Hyland et al. (2021). 62. Buera and Oberfield (2020); Coe and Helpman (1995); 17. Piontkivsky and Nikijuluw (2023). Coe, Helpman, and Hoffmaister (1997); Eaton and 18. Deininger, Goyal, and Nagarajan (2013). Kortum (2001); Hsieh et al. (2019); Rachapalli (2021). 19. Deininger et al. (2019); Heath and Tan (2020); 63. World Bank (2023c). Naaraayanan (2022); Roy (2008). 64. See García-Sanchez and Crawley (2024). 20. Fernández, Fogli, and Olivetti (2004). See also World 65. See Shi, Liu, and Wang (2023) and references therein. Bank (2011, 2018, 2024). 66. Akcigit, Baslandze, and Stantcheva (2016). 21. Dhar, Jain, and Jayachandran (2019). 67. Prato (2023). 22. World Bank (2022a, 2022b). 68. Although the Anti-Cartel Enforcement Database covers 23. Dhar, Jain, and Jayachandran (2022). a much broader time period and 75 countries, complete 24. Andrew et al. (2022). data for middle-income countries are available only for 25. For more information, see https://www.unesco.org/en​ 34 countries between 2017 and 2022. For more infor- /­articles/unesco-laureate-china-children-and ​-­teenagers​ mation on the database, see World Bank (2021). -fund - helps - disadvantaged - girls -thrive ​ - through​ 69. World Bank (2021). -education. 70. These data are based on a sample of 43 high-income 26. Garcia-Brazales (2021). countries and 52 middle-income countries and on 27. Bursztyn et al. (2023). public information for 2021 or the latest year available 28. Bursztyn, González, and Yanagizawa-Drott (2020). (Begazo and Licetti 2024). Statistics on budget exclude 29. Beaman et al. (2012); Priyanka (2020). Cambodia, Kuwait, and Nigeria, which only publish staff 30. Chiplunkar and Goldberg (2021). information. 31. Fernández (2013). 71. Begazo and Licetti (2024). 32. Cameron, Contreras Suarez, and Tseng (2023). 72. Cheng (2021); Gal et al. (2019). 33. Halim, Perova, and Reynolds (2021). 73. Aghion et al. (2001, 2009). 34. World Bank (2024). 74. Cunningham, Ederer, and Ma (2021). 35. Aghion et al. (2001, 2005); Akcigit, Alp, and Peters 75. Syverson (2019). (2021). 76. Diamond and Saez (2011). 36. Commission on Growth and Development (2008). 77. Akcigit et al. (2022). Disciplining Incumbency | 183 References Aghion, Philippe, Antonin Bergeaud, Matthieu Lequien, Firm Customers.” OVE Working Paper OVE/WP–04/11 Marc Melitz, and Thomas Zuber. 2022. “Opposing (May), Office of Evaluation and Oversight, Inter- Firm-Level Responses to the China Shock: Horizontal American Development Bank, Washington, DC. Competition versus Vertical Relationships?” NBER Bai, Jie, Panle Jia Barwick, Shengmao Cao, and Shanjun Li. Working Paper 29196, National Bureau of Economic 2022. “Quid Pro Quo, Knowledge Spillover, and Industrial Research, Cambridge, MA. Quality Upgrading: Evidence from the Chinese Auto Aghion, Philippe, Nick Bloom, Richard Blundell, Rachel Industry.” NBER Working Paper 27644 rev. (November), Griffith, and Peter Howitt. 2005. “Competition and National Bureau of Economic Research, Cambridge, Innovation: An Inverted-U Relationship.” Quarterly MA. Journal of Economics 120 (2): 701–28. Beaman, Lori A., Esther Duflo, Rohini Pande, and Petia Aghion, Philippe, Richard Blundell, Rachel Griffith, Peter Topalova. 2012. “Female Leadership Raises Aspirations Howitt, and Susanne Prantl. 2009. “The Effects of Entry and Educational Attainment for Girls: A Policy on Incumbent Innovation and Productivity.” Review of Experiment in India.” Science 335 (6068): 582–86. Economics and Statistics 91 (1): 20–32. Begazo, Tania Priscilla, and Martha Martínez Licetti. 2024. Aghion, Philippe, Robin Burgess, Stephen J. Redding, Market Institutions: Beyond Competition Authorities and Fabrizio Zilibotti. 2008. “The Unequal Effects of and the Need for a Broader Institutional Ecosystem. Liberalization: Evidence from Dismantling the License Washington, DC: World Bank. Raj in India.” American Economic Review 98 (4): Bloom, Nicholas, Mirko Draca, and John Van Reenen. 1397–412. 2016. “Trade Induced Technical Change? The Impact Aghion, Philippe, Christopher Harris, Peter Howitt, and John of Chinese Imports on Innovation, IT and Productivity.” Vickers. 2001. “Competition, Imitation and Growth with Review of Economic Studies 83 (1): 87–117. Step-by-Step Innovation.” Review of Economic Studies Bridgman, Benjamin, Shi Qi, and James A. Schmitz, Jr. 68 (3): 467–92. 2009. “The Economic Performance of Cartels: Evidence Akcigit, Ufuk, Harun Alp, and Michael Peters. 2021. “Lack from the New Deal U.S. Sugar Manufacturing Cartel of Selection and Limits to Delegation: Firm Dynamics 1934–74.” Research Department Staff Report 437 in Developing Countries.” American Economic Review (November), Federal Reserve Bank of Minneapolis, 111 (1): 231–75. Minneapolis. Akcigit, Ufuk, Salomé Baslandze, and Francesca Lotti. Buera, Francisco Javier, and Ezra Oberfield. 2020. “The 2023. “Connecting to Power: Political Connections, Global Diffusion of Ideas.” Econometrica 88 (1): 83–114. Innovation, and Firm Dynamics.” Econometrica 91 (2): Bursztyn, Leonardo, Alexander W. Cappelen, Bertil 529–64. Tungodden, Alessandra Voena, and David H. Akcigit, Ufuk, Salomé Baslandze, and Stefanie Stantcheva. Yanagizawa-Drott. 2023. “How Are Gender Norms 2016. “Taxation and the International Mobility of Perceived?” NBER Working Paper 31049 (March), Inventors.” American Economic Review 106 (10): National Bureau of Economic Research, Cambridge, 2930–81. MA. Akcigit, Ufuk, John Grigsby, Tom Nicholas, and Stefanie Bursztyn, Leonardo, Alessandra L. González, and David Stantcheva. 2022. “Taxation and Innovation in the Yanagizawa-Drott. 2020. “Misperceived Social Norms: 20th Century.” Quarterly Journal of Economics 137 (1): Women Working Outside the Home in Saudi Arabia.” 329–85. American Economic Review 110 (10): 2997–3029. Akcigit, Ufuk, and Marc J. Melitz. 2022. “International Bustos, Paula. 2011. “Trade Liberalization, Exports, and Trade and Innovation.” In Handbook of International Technology Upgrading: Evidence on the Impact of Economics: International Trade, vol. 5, edited by Gina MERCOSUR on Argentinian Firms.” American Economic Gopinath, Elhanan Helpman, and Kenneth Rogoff, Review 101 (1): 304–40. 377–404. Handbooks in Economics Series. Amsterdam: Cadot, Olivier, Alan Asprilla, Julien Gourdon, Christian North-Holland, Elsevier. Knebel, and Ralf Peters. 2015. “Deep Regional Andrew, Alison, Sonya Krutikova, Gabriela Smarrelli, and Integration and Non-Tariff Measures: A Methodology Hemlata Verma. 2022. “Gender Norms, Violence and for Data Analysis.” Policy Issues in International Trade Adolescent Girls’ Trajectories: Evidence from a Field and Commodities Research Studies Series 69, United Experiment in India.” IFS Working Paper W22/41, Nations Conference on Trade and Development, Institute for Fiscal Studies, London. Geneva. Arayavechkit, Tanida, Charl Jooste, and Ana Urrutia Cameron, Lisa, Diana Contreras Suarez, and Yi-Ping Tseng. Arrieta. 2022. “How Regulation and Enforcement of 2023. “Women’s Transitions in the Labour Market as Competition Affects ICT Productivity: Evidence from a Result of Childbearing: The Challenges of Formal Matched Regulatory-Production Surveys in Peru’s Sector Employment in Indonesia.” IZA Discussion ICT  Sector.” Policy Research Working Paper 10151, Paper Series 16136, Institute of Labor Economics (IZA), World Bank, Washington, DC. Bonn, Germany. Arráiz, Irani, Francisca Henríquez, and Rodolfo Stucchi. Cheng, Thomas K. 2021. The Patent-Competition Interface 2011. “Impact of the Chilean Supplier Development in Developing Countries. Oxford, UK: Oxford University Program on the Performance of SMEs and Their Large Press. 184 | WORLD DEVELOPMENT REPORT 2024 Chiplunkar, Gaurav, and Pinelopi Koujianou Goldberg. Effects from India.” Journal of Development Studies 2021. “Aggregate Implications of Barriers to Female 55 (12): 2549–71. Entrepreneurship.” NBER Working Paper 28486 de Vries, Henk J. 2006. “Standards for Business: How (February), National Bureau of Economic Research, Companies Benefit from Participation in International Cambridge, MA. Standards Setting.” In International Standardization Choi, Jaedo, and Younghun Shim. 2023. “Technology as a Strategic Tool: Commended Papers from the Adoption and Late Industrialization.” STEG IEC Centenary Challenge 2006, 130–41. Geneva: Working Paper WP033 rev. (September), Structural International Electrotechnical Commission. Transformation and Economic Growth, Centre for Dhar, Diva, Tarun Jain, and Seema Jayachandran. 2019. Economic Policy Research, London. “Intergenerational Transmission of Gender Attitudes: Cirera, Xavier, and William F. Maloney. 2017. The Innovation Evidence from India.” Journal of Development Studies Paradox: Developing-Country Capabilities and the 55 (12): 2572–92. Unrealized Promise of Technological Catch-Up. Dhar, Diva, Tarun Jain, and Seema Jayachandran. 2022. Washington, DC: World Bank. “Reshaping Adolescents’ Gender Attitudes: Evidence Coe, David T., and Elhanan Helpman. 1995. “International from a School-Based Experiment in India.” American R&D Spillovers.” European Economic Review 39 (5): Economic Review 112 (3): 899–927. 859–87. Diamond, Peter, and Emmanuel Saez. 2011. “The Case Coe, David T., Elhanan Helpman, and Alexander W. for a Progressive Tax: From Basic Research to Policy Hoffmaister. 1997. “North-South R&D Spillovers.” Recommendations.” Journal of Economic Perspectives Economic Journal 107 (440): 134–49. 25 (4): 165–90. Commission on Growth and Development. 2008. The Growth Dixit, Siddharth, and Indermit S. Gill. 2024. “AI, the Report: Strategies for Sustained Growth and Inclusive New  Wingman of Development.” Background paper Development. Washington, DC: World Bank. https:// ­prepared for World Development Report 2024, World openknowledge.worldbank.org/handle/10986/6507. Bank, Washington, DC. Cunningham, Colleen, Florian Ederer, and Song Ma. 2021. Eaton, Jonathan, and Samuel Kortum. 2001. “Technology, “Killer Acquisitions.” Journal of Political Economy Trade, and Growth: A Unified Framework.” European 129 (3): 649–702. Economic Review 45 (4–6): 742–55. Cusolito, Ana Paula, Alvaro Garcia-Marin, and William F. Ellig, Jerry. 2020. “Retail Electric Competition and Natural Maloney. 2023. “Proximity to the Frontier, Markups, and Monopoly: The Shocking Truth.” In Regulation and the Response of Innovation to Foreign Competition: Economic Opportunity: Blueprints for Reform, edited Evidence from Matched Production-Innovation Surveys in by Adam Hoffer and Todd Nesbit, 277–302. Logan, Chile.” American Economic Review: Insights 5 (1): 35–54. UT: Center for Growth and Opportunity, Utah State Dall’Olio, Andrea, Tanja Goodwin, Martha Martinez Licetti, University. Ana Cristina Alonso Soria, Maciej Drozd, Jan Orlowski, Fernández, Raquel. 2013. “Cultural Change as Learning: The Fausto Patiño Peña, and Dennis Sanchez-Navarro. Evolution of Female Labor Force Participation over a 2022. “Are All State-Owned Enterprises Equal? A Century.” American Economic Review 103 (1): 472–500. Taxonomy of Economic Activities to Assess SOE Fernández, Raquel, Alessandra Fogli, and Claudia Olivetti. Presence in the Economy.” Policy Research Working 2004. “Mothers and Sons: Preference Formation and Paper 10262, World Bank, Washington, DC. Female Labor Force Dynamics.” Quarterly Journal of Dauda, Seidu, and Maciej  Drozd. 2020. Barriers to Economics 119 (4): 1249–99. Competition in Product Market Regulation: New Insights Fløysand, Arnt, and Jonathan R. Barton. 2014. “Foreign on Emerging Market and Developing Economies. Direct Investment, Local Development, and Poverty Washington, DC: World Bank. Reduction: The Sustainability of the Salmon Industry in Defever, Fabrice, José-Daniel Reyes, Alejandro Riaño, and Southern Chile.” In Alternative Development: Unravelling Gonzalo J. Varela. 2020. “All These Worlds Are Yours, Marginalization, Voicing Change, edited by Cathrine Except India: The Effectiveness of Cash Subsidies Brun, Piers Blaikie, and Michael Jones, 55–71. Farnham, to Export in Nepal.” European Economic Review 128 UK: Ashgate Publishing. (September): 103494. Freund, Caroline L. 2016. Rich People Poor Countries: The Defever, Fabrice, Alejandro Riaño, and Gonzalo J. Varela. Rise of Emerging-Market Tycoons and Their Mega Firms. 2020. “Evaluating the Impact of Export Finance Support Assisted by Sarah Oliver. Washington, DC: Peterson on Firm-Level Export Performance: Evidence from Institute for International Economics. Pakistan.” CESifo Working Paper 8519 (August), Munich Gal, Peter, Giuseppe Nicoletti, Theodore Renault, Stéphane Society for the Promotion of Economic Research, Center Sorbe, and Christina Timilioti. 2019. “Digitalisation for Economic Studies, Ludwig Maximilian University and Productivity: In Search of the Holy Grail; Firm- and Ifo Institute for Economic Research, Munich. Level Empirical Evidence from EU Countries.” Deininger, Klaus W., Aparajita Goyal, and Hari K. OECD Economics Department Working Paper 1533 Nagarajan. 2013. “Women’s Inheritance Rights and (February  6), Organisation for Economic Co-operation Intergenerational Transmission of Resources in India.” and Development, Paris. Journal of Human Resources 48 (1): 114–41. Garcia-Brazales, Javier. 2021. “Changing Gender Attitudes: Deininger, Klaus W., Songqing Jin, Hari K. Nagarajan, and The Long-Run Effects of Early Exposure to Female Fang Xia. 2019. “Inheritance Law Reform, Empowerment, Classmates.” Paper presented at the virtual 33rd and Human Capital Accumulation: Second-Generation Annual Conference of the European Association of Disciplining Incumbency | 185 Labour Economists, Department of Economics and Developing and Emerging Economies: Evidence and Management Marco Fanno, University of Padua, Padua, Policy Making.” Policy Research Working Paper 9364, Italy, September 16–18, 2021. World Bank, Washington, DC. García-Sanchez, Cristobal, and Edward Crawley. 2024. Kolasa, Marcin. 2008. “How Does FDI Inflow Affect “Accelerating Innovation for Development: How Productivity of Domestic Firms? The Role of Horizontal Universities Leverage Knowledge Exchange to Boost and Vertical Spillovers, Absorptive Capacity and Economic Growth in Middle-Income Countries.” Competition.” Journal of International Trade and Background paper prepared for World Development Economic Development 17 (1): 155–73. Report 2024, World Bank, Washington, DC. Lederman, Daniel, Marcelo Olarreaga, and Lucy Payton. Goldin, Claudia. 2006. “The Quiet Revolution that 2006. “Export Promotion Agencies: What Works and Transformed Women’s Employment, Education, and What Doesn’t.” Policy Research Working Paper 4044, Family.” American Economic Review 96 (2): 1–21. World Bank, Washington, DC. Goldin, Claudia. 2014. “A Grand Gender Convergence: Its Last Lovo, Stefania, and Gonzalo J. Varela. 2023. “Internationally Chapter.” American Economic Review 104 (4): 1091–119. Linked Firms, Integration Reforms and Productivity: Goldin, Claudia. 2021. Career and Family: Women’s Century- Evidence from Pakistan.” Policy Research Working Long Journey toward Equity. Princeton, NJ: Princeton Paper 9349, World Bank, Washington, DC. University Press. Medina, Pamela. 2022. “Import Competition, Quality Goldin, Claudia, and Lawrence F. Katz. 2002. “The Power Upgrading, and Exporting: Evidence from the Peruvian of the Pill: Oral Contraceptives and Women’s Career Apparel Industry.” Review of Economics and Statistics. and Marriage Decisions.”  Journal of Political Economy Published ahead of print, July 26, 2022. https://​ 110 (4): 730–70. doi.­org/10.1162/rest_a_01221. Google. 2023. “Unidos por el Bienestar Económico de las Melitz, Marc J., and Stephen J. Redding. 2021. “Trade and Mujeres del Sureste Mexicano.” February 21, 2023, Innovation.” NBER Working Paper 28945 (June), National blog. https://blog.google/intl/es-419/unidos-por-el​ Bureau of Economic Research, Cambridge, MA. -bienestar-economico-de-las-mujeres-del-sureste​ Monge-González, Ricardo, John Hewitt, and Federico -mexicano/. Torres-Carballo. 2015. “Do Multinationals Help or Halim, Daniel, Elizaveta Perova, and Sarah Reynolds. 2021. Hinder Local Firms? Evidence from the Costa Rican ICT “Childcare and Mothers’ Labor Market Outcomes in Sector.” CAATEC Working Paper, Comisión Asesora en Lower- and Middle-Income Countries.” Policy Research Alta Tecnología, San José, Costa Rica. Working Paper 9828, World Bank, Washington, DC. Naaraayanan, Sankararama Lakshmi. 2022. “Women’s Heath, Rachel, and Xu Tan. 2020. “Intrahousehold Inheritance Rights and Entrepreneurship Gender Gap.” Bargaining, Female Autonomy, and Labor Supply: Paper presented at the Indian School of Business– Theory and Evidence from India.” Journal of the National Bureau of Economic Research conference, European Economic Association 18 (4): 1928–68. “Economic Policy and the Indian Economy,” Indian School Herfindahl, Orris C. 1950. “Concentration in the U.S. of Business, Hyderabad, India, December 17–18, 2022. Steel  Industry.” PhD dissertation, Columbia University, OECD (Organisation for Economic Co-operation and New York. Development). 2014. “Factsheet on How Competition Hirschman, Albert O. 1964. “The Paternity of an Index.” Policy Affects Macro-Economic Outcomes.” October, American Economic Review 54 (5): 761–62. OECD, Paris. Hsieh, Chang-Tai, Erik Hurst, Charles I. Jones, and Peter OECD (Organisation for Economic Co-operation and J. Klenow. 2019. “The Allocation of Talent and U.S. Development). 2024.  OECD Competition Trends 2024. Economic Growth.” Econometrica 87 (5): 1439–74. Paris: OECD Publishing. https://doi.org/10.1787​ Hyland, Marie Caitriona, Nona Karalashvili, Silvia Muzi, /e69018f9-en. and Domenico Viganola. 2021. “Female-Owned Firms Petit, Lilian T. D., Ron G. M. Kemp, and Jarig van Sinderen. during the COVID-19 Crisis.” Global Indicators Brief 2015. “Cartels and Productivity Growth: An Empirical 2 (July 29), World Bank, Washington, DC. Investigation of the Impact of Cartels on Productivity Iacovone, Leonardo, Mariana De La Paz Pereira López, in the Netherlands.” Journal of Competition Law and and Marc Tobias Schiffbauer. 2023. “Competition Economics 11 (2): 501–25. Makes IT Better: Evidence on When Firms Use IT More Phumpiu Chang, Paul, and Jose Castillo. 2024. “Shaping Effectively.” Research Policy 52 (8): 104786. Market Dynamics: The Role of Competition Policy in Iacovone, Leonardo, Ferdinand Rauch, and L. Alan Winters. Fostering Contestability and Technological Adoption 2013. “Trade as an Engine of Creative Destruction: in the Mobile Telecommunications Sector in Middle- Mexican Experience with Chinese Competition.” Journal Income Countries.” Background paper prepared of International Economics 89 (2): 379–92. for World Development Report 2024, World Bank, IEA (International Energy Agency). 2020. World Energy Washington, DC. Investment 2020. July. Paris: IEA. Piontkivsky, Ruslan, and Ruth Nikijuluw. 2023. “Papua New Jordaan, Jacob Arie, Willem Douw, and Christine Zhenwei Guinea Country Economic Memorandum: Pathways Qiang. 2020. “Multinational Corporation Affiliates, to Faster and More Inclusive Growth.” World Bank, Backward Linkages, and Productivity Spillovers in Washington, DC. 186 | WORLD DEVELOPMENT REPORT 2024 Prato, Marta. 2023. “The Global Race for Talent: Brain Drain, Syverson, Chad. 2019. “Macroeconomics and Market Knowledge Transfer, and Economic Growth.” Working Power: Context, Implications, and Open Questions.” Paper, Ettore Bocconi Department of Economics, Journal of Economic Perspectives 33 (3): 23–43. Bocconi University, Milan. Vagliasindi, Maria. 2023. “The Role of SOEs in Climate Priyanka, Sadia. 2020. “Do Female Politicians Matter for Change.” Background paper prepared for The Business Female Labor Market Outcomes? Evidence from State of the State, World Bank, Washington, DC. Legislative Elections in India.” Labour Economics 64 Varela, Gonzalo, Zehra Aslam, Andreas Eberhard, (June): 101822. Aroub Farooq, Veronica Michel Gutierrez, Alen Pro Mujer. 2023. “Pro Mujer and Google Join Forces to Mulabdic, Rafay Khan, et al. 2022. “From Swimming Impact the Lives of Thousands of Indigenous Women in Sand to High and Sustainable Growth: Pakistan’s in Southeast Mexico.” Pro Mujer (blog), February 23, Country Economic Memorandum.” World Bank, 2023. https://promujer.org/en/blog/with-support-from​ Washington, DC. https://documents1.worldbank​ -the - ministr y- of- economy- pro - mujer-and - google​ . o r g / c u r a t e d / e n / 0 9 9 8 2 0 4 1 0 112 2 6 7 3 5 4 / p d f​ - join -forces -to - impact-the - lives - of-thousands - of​ /P1749040fc80a70ca0b6f70f7860c4a1034.pdf. -indigenous-women-in-southeast-mexico/. Verhoogen, Eric A. 2023. “Firm-Level Upgrading in Rachapalli, Swapnika. 2021. “Learning between Buyers and Developing Countries.” Journal of Economic Literature Sellers along the Global Value Chain.” Paper presented 61 (4): 1410–64. at the Cowles Foundation for Research in Economics Volpe Martincus, Christian, and Jerónimo Carballo. 2008. virtual 2021 Conference on International Trade, Yale “Is Export Promotion Effective in Developing Countries? University, New Haven, CT, June 7, 2021. Firm-Level Evidence on the Intensive and the Extensive Reed, Tristan, Mariana De La Paz Pereira López, Ana Margins of Exports.” Journal of International Economics Francisca Urrutia Arrieta, and Leonardo Iacovone. 76 (1): 89–106. 2022. “Cartels, Antitrust Enforcement, and Industry Vostroknutova, Ekaterina, James Sampi, Charl Jooste, Performance: Evidence from Mexico.” Policy Research and Jorge Thompson Araujo. 2024. Competition Working Paper 10269, World Bank, Washington, DC. and Productivity in Latin America and the Caribbean. Roy, Sanchari. 2008. “Female Empowerment through Washington, DC: World Bank. Inheritance Rights: Evidence from India.” Paper World Bank. 2011. World Development Report 2012: presented at the Second Riccardo Faini Doctoral Gender  Equality and Development. Washington, DC: Conference on Development Economics, University World Bank. of Milan, Milan, September 7–9, 2008. https://www​ World Bank. 2018. Women Economic Empowerment Study. .dagliano.unimi.it/media/roy.pdf. May. Washington, DC: World Bank. Sampi, James, Ana Francisca Urrutia Arrieta, and Ekaterina World Bank. 2020. World Development Report 2020: Trading Vostroknutova. 2022. “Antitrust Enforcement, Markups, for Development in the Age of Global Value Chains. and Productivity in South America.” World Bank, Washington, DC: World Bank. Washington, DC. World Bank. 2021. “Fixing Markets, Not Prices: Policy Sanchez Navarro, Dennis. 2024. “What Happens When Options to Tackle Economic Cartels in Latin America the State Is Bossing around Markets? An Analysis of the and the Caribbean.” World Bank, Washington, DC. Performance Differentials between Businesses of the World Bank. 2022a. Poverty and Shared Prosperity 2022: State (BOS) and Private-Owned Enterprises (POEs).” Correcting Course. Washington, DC: World Bank. Policy Research Working Paper 10820, World Bank, World Bank. 2022b. Reshaping Norms: A New Way Forward. Washington, DC. South Asia Economic Focus (April). Washington, DC: Schiffbauer, Marc Tobias, James Sampi, and Javier World Bank. Coronado. 2022. “Competition and Productivity: World Bank. 2023a. The Business of the State. Washington, Evidence from Peruvian Municipalities.” Review of DC: World Bank. Economics and Statistics. Published ahead of print, World Bank. 2023b. “Protectionism Is Failing to Achieve Its November 15, 2022. https://doi.org/10.1162/rest​ Goals and Threatens the Future of Critical Industries.” _a_01257. Feature Story, August 29, 2023. https://www.worldbank​ Schivardi, Fabiano, and Eliana Viviano. 2011. “Entry Barriers .org/en/news/feature/2023/08/29/protectionism-is​ in Retail Trade.” Economic Journal 121 (551): 145–70. -failing-to-achieve-its-goals-and-threatens-the-future​ Schneider, Ben Ross. 2022. “Teacher Unions, Political -of-critical-industries. Machines, and the Thorny Politics of Education Reform World Bank. 2023c. World Development Report 2023: in Latin America.” Politics and Society 50 (1): 84–116. Migrants, Refugees, and Societies. Washington, DC: Schwab, Klaus, ed. 2019. Insight Report: The Global World Bank. Competitiveness Report 2019. Geneva: World World Bank. 2024. Women, Business and the Law 2024. Economic Forum. http://www3.weforum.org/docs​ Washington, DC: World Bank. /­W EF_TheGlobalCompetitivenessReport2019.pdf. Zuniga, Maria Pluvia. 2024. “The Impact and Shi, Dongbo, Weichen Liu, and Yanbo Wang. 2023. “Has Effectiveness of Innovation Policy: Evidence from China’s Young Thousand Talents Program Been Middle-Income Countries.” Background paper Successful in Recruiting and Nurturing Top-Caliber prepared for World  Development Report 2024, Scientists?” Science 379 (6627): 62–65. World Bank, Washington, DC. Disciplining Incumbency | 187 8 Rewarding Merit Key messages • Middle-income countries can allocate talent and human capital more efficiently by opening up education, employment, and business opportunities to all who have talent and ­ ­ display acquired ability. Economic development policies should not target an ideal income distribution but pay attention to enhancing mobility. ­ • Middle-income countries should avoid solely targeting firms for support by size. Instead, they could also assess the value that firms add to the economy through jobs, exports, technology infusion, and innovation. Countries’ policies should target support to those firms that display the most potential for growth, while letting go of unproductive firms, modernizing organizational models to manage firms, and connecting entrepreneurs with money (financing opportunities), mentors, and markets. • Middle-income countries will need to adopt policies that support the diffusion of lower-carbon energy technologies, incentivize the efficient use of energy, and consider the social and ecological costs of greenhouse gas emissions. They should consider all options for decoupling their growing economies from the growth of these emissions. Moving forward by • The economic and social mobility of peo- ple. Governments should adopt policies promoting merit activities that will enable an economy to allocate To strengthen the forces of creation, talent and human capital more effi- middle-income countries will need to shift their ciently—in particular, by opening up policies and institutions toward promoting merit education, employment, and business activities—that is, those with positive effects on opportunities to all those who have tal- general well-being and that aid in the efficient use ent and display acquired ability. Because of talent, capital, and energy. advancing the sophistication of an econ- Specifically, middle-income countries should omy can increase income inequality, put three considerations at the center of economic economic development policies should policy making: the economic and social mobility of not target an ideal income distribution. people; the value added by firms; and the greenhouse Instead, by rewarding talent and effort, gases (GHGs) emitted by the economy. policies can generate both higher social | 189 and economic mobility, which leads to two related challenges. First, talent is scarce faster growth.1 However, perceptions of in those countries because they do not accu- mobility matter. When perceptions of mulate human capital. Second, middle-income the opportunities for future mobility are countries are not as effective as high-income high, social tolerance of inequality is also countries in allocating talent to tasks. Middle- high, and vice versa.2 income countries must then focus on improving • The value added by firms. A similar princi- social mobility—going beyond narrowly focus- ple applies to firms. Dynamic, productive ing on inequality. In doing so, these countries young firms should be able to expand, will need to revise how they reward talent by and less productive firms should be able upgrading their talent pool, selecting efficient to contract and exit, thereby enhancing learners, expanding occupation choices and value added across the economy through rewards, and nurturing scientific inquiry and efficient resource utilization. Policies enhancing research capabilities, and installing should not target firms for support by a socially responsible safety net to protect those size, but instead should assess the value who may lose their livelihoods in the creative that firms add to the economy through destruction process. jobs, exports, technology infusion, and innovation. Upgrading the talent pool • The GHG emissions of an economy. A similar principle also applies to energy. Support high-quality secondary and higher Government policies should aid in the dif- education fusion of lower-carbon energy technolo- Countries that transitioned from middle- to gies, encourage the efficient use of energy, high-income status in recent decades sought and consider the social and ecological to accumulate human capital, which begins costs of GHG emissions. Because today’s with developing foundational skills. As a result, middle-income countries account for their upper-secondary enrollment and gradu- two-thirds of global GHG emissions and ation rates, as well as their tertiary education all new emissions growth, they will need enrollment rates, steadily increased (figure 8.1). to consider all options to decouple their Countries that did not transition to high-income growing economies from higher GHG status have attained levels of tertiary education emissions. that are close to the levels achieved by countries at the time of their transition, but enrollment in and completion of upper-secondary education The economic and social have lagged for a wider swath of the population.3 mobility of people For significant shares of the population in coun- The introduction of global technologies must tries that have not transitioned to high-income go hand in hand with a search for the technical status, education stops in the early teen years, workers and specialized professionals needed or earlier, especially for girls and members of for firms to adopt technologies, along with com- minorities. Unless these countries secure foun- petent managers to run firms. Meanwhile, more dational skills, their talent pipeline will remain engineers are needed to build and maintain weak. infrastructure, teachers to educate students, How should countries allocate education doctors to treat patients, and highly qualified spending between foundational and advanced men and women to run governments. Chapter 5 skills? In setting priorities, countries may highlighted that middle-income countries face want to consider the principle of progressive 190 | WORLD DEVELOPMENT REPORT 2024 Figure 8.1 Middle-income countries that transitioned to high-income status first focused on foundational skills a. Enrollment rate b. Completion rate 100 100 Share of population (%) Share of population (%) 80 80 60 60 40 40 20 20 0 0 –15 –10 –5 0 5 –15 –10 –5 0 5 Years since transition Years since transition Higher education, MICs to HICs Upper-secondary, MICs to HICs Higher education, MICs’ current average Upper-secondary, MICs’ current average Source: WDR 2024 team calculations based on data of UIS.Stat (dashboard), Institute for Statistics, United Nations Educational, Scientific, and Cultural Organization, Montreal, http://data.uis.unesco.org/. Note: Solid lines show averages for countries that were middle-income countries in the early 1990s and transitioned to high-income status any time before 2019 (23 countries for enrollment rates and 15 countries for completion rates). Dashed lines indicate the current average for middle-income countries that have not yet transitioned. Gross enrollment rate is the ratio of total enrollment, regardless of age, to the population of the official age group for the education level. Completion rate is the percentage of individuals three to five years older than the theoretical age of completion of the level’s last grade who have completed the level’s last grade. For countries that have transitioned, data are for 1981–2022 for enrollment rates and 1990–2021 for completion rates. The number of years since graduation is computed as the calendar year minus transition year. For countries not yet transitioned, data are for the most recent year available. HICs = high-income countries; MICs = middle-income countries. universalism, which advocates investing pro- spending to secondary education and attained gressively in higher education as quality for equally rapid success. Only later did it invest all is reached at lower educational levels.4 The substantially in tertiary education. Throughout “universal” component advocates high-quality this period, Korea focused not only on enroll- education for all. The “progressive” component ment but also on learning—for all.5 Because recognizes that efforts and resources are limited its efforts to improve foundational skills were and prioritizes the early years of a child’s life, as focused and deliberate, Korea was able to pivot well as outreach to disadvantaged students who to advanced skills in a relatively short period of face the greatest learning hurdles. The Republic time. An emphasis on foundational skills for all of Korea adopted this approach. In the 1950s, was key to the success of school reforms in other it enforced compulsory education and devoted countries as well (box 8.1). nearly 80 percent of its education budget to Strengthening foundational skills does not primary education, thereby increasing enroll- always require more money, but it does require effi- ment rates from about 40 percent to 90 percent ­ ountries that spend more cient spending. In fact, c in 10  years. Korea then shifted its efforts and do not necessarily have better learning outcomes.6 Rewarding Merit | 191 Box 8.1 Developing foundational skills: Learning from Finland and Chile The successful reforms implemented by countries that graduated from middle- to high-income status have shared three elements.a First, they broadened access to foun- dational skills to create a large, deep talent pool of high school graduates. Second, they relied on student assessments to collect information on learning and evaluate progress toward achieving policy goals. Third, education reforms were part of a deliberate long- term growth strategy. Finland. In the early 1970s, Finland replaced its two-track school system (students were divided into two tracks, general secondary and vocational, after the first four years of education) with a nine-year comprehensive school system (all students followed the same national curriculum until age 16 and then chose a track). In addition, teacher training became more selective and rigorous. Finally, curricula setting went from highly central- ized to more decentralized.b These reforms broadened the talent pool, which may have helped weaken the relationship between parental income and, for example, the likelihood of becoming an inventor.c Reforms also targeted higher education, making access more equitable and holding institutions accountable for their contributions to the economy. Chile. Since the 1980s, students in Chile have had access to universal vouchers to attend private schools. In 2008, a reform was enacted to raise the voucher amount for disadvantaged students, provide additional funding to schools with large shares of such students, and create an accountability system for schools receiving vouchers. Critically, a test-based assessment system was created in the 1980s and is still being used to gauge results. The opening of private universities was encouraged, new sources of student fund- ing became available through scholarships and loans, and the supply of short-cycle higher education programs grew rapidly.d Against the backdrop of these reforms, the share of high school students scoring above minimum proficiency levels in the worldwide assess- ment, the Programme for International Student Assessment (PISA), rose from 53 percent to 68 percent between 2000 and 2018, and net enrollment rates in higher education have grown from less than 5 percent to about 45 percent since 1970. Furthermore, college access has become more equitable for men and women and for students from differ- ent socioeconomic backgrounds.e However, the share of disadvantaged students among top performers in the high school math exit exam has barely changed over the past five decades, suggesting that, although many disadvantaged students have gained access to college, their elementary and high schools are not yet preparing them for admission to Chile’s top universities. See, for example, Aghion et al. (2023) and Pekkarinen, Uusitalo, and Pekkala Kerr (2009) for Finland; a.  World Bank (2021) for Ireland; Lee, Jeong, and Hong (2014) for the Republic of Korea; Guyon, Maurin, and McNally (2012) for Northern Ireland; and Jakubowski et al. (2016) for Poland. Pekkarinen, Uusitalo, and Pekkala Kerr (2009); Saavedra-Chanduví, Alasuutari, and Gutiérrez Bernal b.  (2018).  ghion et al. (2023). c. A  guirre (2021); Fontaine and Urzúa (2018); Murnane et al. (2017); Solis (2017); Vegas (2018). d. A Gallegos, Barrios-Fernández, and Neilson (2024). e.  192 | WORLD DEVELOPMENT REPORT 2024 A comparison across 46 c ­ ountries of 150 technical workers in manufacturing but also ­interventions in early childhood and ­primary and middle managers (who, according to firms, are secondary education reveals that some of the most undertrained), as well as sales and banking spe- cost-effective programs deliver the equivalent of cialists.11 The co-op model alternates between three additional years of high-quality schooling terms of classroom-based instruction and (comparable with that of the highest-performing terms of paid work, and it relies on partnerships education systems) for just US$100 per child.7 between institutions and employers. Institutions The three most cost-­ effective interventions are in middle-income countries offering co-ops targeted information campaigns about the ben- include the Universidade de São Paulo in Brazil, efits, costs, and quality of schooling; interven- Nelson Mandela University in South Africa, and tions to target teaching instruction by learning TOBB University of Economics and Technology level rather than grade; and improved pedagogy in Türkiye.12 The integrated curriculum model through structured lesson plans providing stu- is exemplified by Indonesia’s Merdeka Belajar dent materials, teacher professional develop- (Emancipated Learning) initiative. Educational ment, and monitoring.8 Other policies—such as authorities oversee large-scale curriculum coor- early childhood development and merit-based dination among universities, firms, and other scholarships—are costlier and yet are still highly institutions, which allows students to spend up effective. to a third of their time in work settings inside or outside the country. Universities also train prac- Develop advanced skills titioners who are interested in learning how to To build a pool of talent with advanced skills, teach and pair them with mentors.13 middle-income countries must substantially Countries can use public funding for higher upgrade their higher education systems. education to pursue country-level strategic goals These systems need to be aware of the skills and promote equitable access to higher education. in demand or they run the risk of producing Expanding the base of science, technology, engi- unemployable graduates whose skills are not neering, and mathematics (STEM) graduates may relevant to the labor market. In Latin America be one such strategic goal. China, for example, and the Caribbean, shorter-term higher edu- has dramatically expanded its tertiary-educated cation programs that interact more with population over the last few decades by means of ­ employers and help students in their job search ­policies at home and training overseas, as well as its contribute more to students’ labor market out- base of STEM graduates.14 Targeted scholarships comes than other such “short-cycle” programs.9 can increase student interest in STEM careers, Although short-cycle programs pursue such including the life sciences, physical sciences, engi- connections more often than bachelor’s pro- neering, mathematics, computer science, and the grams,10 the ­connections are equally important health sciences. When coupled with information for both. and mentoring interventions, they are particularly Lack of experience often holds back univer- effective at attracting females to STEM fields.15 sity  graduates from finding jobs. Education To promote equitable access to higher education, ­ models that incorporate work experience some countries provide need-based financial attempt to break this cycle and facilitate aid—such as Brazil’s ProUni and South Africa’s the school-to-work transition. Examples National Student Financial Aid Scheme (NSFAS); include dual training programs, apprentice- unsubsidized student loans—such as those pro- ships, co-operative education, and integrated vided by Tanzania’s Higher Education Students’ curricula. In Ecuador, Corporación Formados Loan Board (HESLB); and subsidized student provides students with dual training based loans—such as Malaysia’s Student Loan Fund on the German model to produce not only Corporation (PTPTN). Rewarding Merit | 193 Countries can provide incentives to align  the institutions and majors, and they are the back- attainment and use of actual skills in the work- bone of the increasingly popular centralized place and entrepreneurship. For example, short- admission systems used in many countries term technical and vocational education and (box  8.2). On the other hand, entrance exams training (TVET) programs depend for their suc- can perpetuate the very inequities they seek to cess on connecting to employers,16 and offering eliminate. Because students with more educated incentives to industry can help.17 In Finland, parents and from higher-income households institutions sign student performance contracts enjoy access to higher-quality basic education,20 with the government and are held accountable they score higher on university entrance exams for their outcomes. The government of Denmark and therefore appear to have more “merit” than also signs performance contracts with universi- others. Students are more likely to confuse an ties. The country relies on local advisory coun- actual “aristocracy of privilege” with an imag- cils formed by representatives of governments, ined “aristocracy of talent.”21 institutions, and enterprises for short-term pro- Unless all students develop foundational skills grams and training. Councils periodically discuss and receive excellent basic education, merit-based local skill needs and decide how many programs schemes can detract from—rather than pro- are needed to address them, closing programs as mote—social mobility. Furthermore, entrance needed. Denmark also has a well-developed sys- exams are not perfect measures of student prepa- tem to forecast skills, as well as an information ration or the potential for advanced training. As system that provides data on job opportunities for a result, institutions may reject applicants who hundreds of occupations by region.18 would perform well if they were admitted. A more holistic approach to selection may be needed. In principle, “merit” could be defined in Selecting efficient learners broader terms than just exam scores. For exam- Some institutions and programs—particularly ple, institutions could reserve a share of their those teaching highly advanced material—may classroom seats for students from disadvantaged require advanced academic readiness and estab- minorities, who would also be chosen based on lish selective, merit-based admission crite- merit—as in Brazil. They may also choose to ria. Countries need an effective mechanism to admit the top share of students from every high identify, select, and promote talent and ability. school—as in the Top Ten Percent Program in Efficient learners with “merit” can be identified the US state of Texas and similar programs in using the test scores of higher education entrance California and Florida. All these variants, how- exams (or high school exit exams), sometimes in ever, have their own trade-offs. More important, combination with other criteria. capacity at highly selective institutions is limited. Entrance exams are a standardized, compa- The challenge is to build tertiary education sys- rable way to evaluate students from different tems that provide high-quality opportunities at backgrounds, especially when used nationally. all institutions, not only the most selective ones. On the one hand, exams are a transparent and In the United States, many institutions—not simple way to sort students across institutions only top-ranked universities where merit-based and programs. In 2005, for example, the coun- admission is most prevalent—promote social try of Georgia established a university entrance mobility.22 Although countries may want to adopt exam because its previous admission system— merit-based admission to some higher education inherited from the Soviet era—was obscure, institutions, the emphasis should be on ensur- prone to corruption, and widely perceived as ing high-quality education at all institutions and inequitable.19 Entrance exams also establish educational levels in order to build a broad, deep clear rules for the selection of students across talent pool. 194 | WORLD DEVELOPMENT REPORT 2024 Box 8.2 Promoting better student choices with digital tools Countries have increasingly adopted centralized choice and admission systems (CCAS) for primary, secondary, and higher education. CCAS can enhance students’ decision-making by providing information about the available choices and helping students use it. Currently, 57 developing countries rely on such mechanisms. Chatbots powered by artificial intelli- gence have helped students gain admission to higher-quality colleges than they would have attended otherwise in cities such as Bogotá and Palmira (Colombia), Recife (Brazil), Tacna (Peru), and various cities across Chile.a Digital tools are particularly powerful when combined with administrative data. For example, trackers can use administrative data on graduates to follow them through higher education and beyond. In Chile, an online portal reports the average labor market out- comes for every higher education program in the country,b allowing prospective students to identify programs with high returns. Although necessary for good decision-making, information alone is not sufficient because disadvantaged students often need help to interpret or utilize information. Advice and counseling—including through chatbots—can further enhance student decision-making.c Source: Neilson 2024.  rteaga et al. (2022). a. A See Subsecretaría de Educación Superior (Subsecretariat of Higher Education) (portal), Ministry of b.  Education, Santiago, Chile, https://www.mifuturo.cl/.  or more on information-related interventions for higher education, see Ferreyra et al. (2021). c. F Leverage digital technologies children, ensuring a critical advantage for them Digital technologies—such as the internet, mobile to improve their test scores. To level the play- phones, social media, and web-based informa- ing field and give students from more disadvan- tion systems—have a large capacity to promote taged backgrounds opportunities, countries can both social mobility and talent development. leverage digital technologies to deliver instruc- The internet, for example, allows individuals to tional materials to many more students online find information, which makes them more pro- (box 8.3). ductive; email allows for the exchange of ideas, which leads to knowledge diffusion; smartphones can be used for mobile banking, which promotes Expanding occupation choice and entrepreneurship; and data systems provide gov- rewards ernment with a wealth of information that can be Climbing the economic and social ladder may not used to efficiently target interventions. be possible when individuals cannot gain access In selecting students for higher education to jobs or realize their entrepreneurial potential programs, merit-based selection requires stan- on the basis of merit. In many countries, attend- dardized measures of merit such as test scores. ing an elite school is not sufficient to secure a top Higher-income parents typically have the job because connections—not merit—determine means to invest in tutoring services for their recruitment (see chapter 5). Although combating Rewarding Merit | 195 Box 8.3 Improving students’ test scores by using online studying assistance from the Khan Academy In 2004, Sal Khan, a Bangladeshi American, began tutoring his cousin Nadia in mathemat- ics using a phone and Yahoo Doodle. As Nadia improved her performance in mathematics, word of Khan’s service spread, and he began tutoring a handful of his cousins and family members. In 2006, lacking the time for one-on-one tutoring, he began recording videos and posting them on YouTube, offering them to everyone to watch at their own pace. In 2008, this initiative turned into the Khan Academy. It has since produced more than 8,000 video lessons on a wide range of academic subjects, including mathematics, sciences, lit- erature, and computer science, as well as supplementary practice exercises and materials for educators. The learning materials, all provided free of charge, are available in many dif- ferent languages. Today, they are a supplement to in-class learning, giving teachers more time to focus on individual students’ needs. In July 2017, the Khan Academy became the official practice partner for the College Board’s Advanced Placement courses. Students who study for the college entrance exam, the SAT, for at least 20 hours via the Khan Academy increase their scores, on average, by 115 points (of a possible 1,600). Source: Khan Academy (website), Mountain View, CA, https://www.khanacademy.org/. such discriminatory practices is a necessary first better align their beliefs about their job perfor- step, it is not sufficient because it does not address mance and their search strategies.23 the inherent information asymmetry between job Although lack of information creates unreal- candidates and employers. Job candidates rely on istic expectations for many job-seekers about the social networks to learn about job openings and type of work and wages they can find, some inter- employers’ characteristics. Meanwhile, employers ventions can help.24 In Ethiopia, a job fair experi- rely on these networks to ascertain workers’ skills ment that brought firms and job-seekers together and trustworthiness. This is particularly true in allowed firms to advertise openings more widely countries with a high degree of informality and and job-seekers to apply for more positions. It also limited information about labor markets. improved employment outcomes for less-educated To address these issues, countries need insti- job-seekers.25 In South Africa, young job-seekers tutions that can serve as reliable, timely conduits who live far from the city centers where jobs are of information. Online job portals—proliferating located overestimated their employment pros- quickly across middle-income countries—are a pects and underestimated actual commuting cost-effective mechanism to communicate job costs. By increasing their access and exposure to openings to job-seekers. But information on job the broader labor market job-seekers were able opportunities needs to be combined with credi- to adjust their expectations and accept jobs closer ble certifications of skills. In South Africa, a job to home.26 matching program provided candidates with Giving job-seekers access to “better” networks— information about job openings and a credible those of more influential individuals—can also assessment of their own skills. Giving job-­ seekers boost their opportunities.27 Mentoring improves this assessment to share with firms increased outcomes mostly by teaching mentees about employment and earnings and enabled them to entry-level jobs and labor market dynamics.28 196 | WORLD DEVELOPMENT REPORT 2024 A mentoring program in Uganda that assisted the National Mentoring Program matches vocational students during their school-to-work advanced learners in 10th and 11th grades with transitions increased their employment prospects top ­professionals in students’ areas of interest to three months after graduation, and their earnings collaborate on a project of mutual interest.34 In a were higher one year later. similar vein, exposing young children—particu- larly those from disadvantaged backgrounds—to inventors and scientists can widen this pipeline. Nurturing scientific inquiry and In the United States, several private initiatives enhancing research capabilities bring children together with inventors; in Spain, The process of nurturing innovation can start workshop initiatives bring children and scientists early. Identifying high-potential, high-perform- together.35 ing students—advanced learners—in the early Building and expanding high-quality universi- school grades and inculcating in them a mind- ties that can train top talent and contribute to inno- set for scientific inquiry is crucial. In India, Atal vation requires an efficient system of public funding Tinkering Labs, with the sponsorship of the for research, as well as fluid ­university-industry con- government, sets up in schools physical labora- nections to promote the exchange of knowledge. tories that are equipped with scientific kits and Public spending on research and development (R&D) apparatuses for use by students between the sixth ­ is lower in middle-income countries—0.3 percent and twelfth grades. The opportunity to “tinker” of the gross domestic product (GDP) of the median and learn by doing is intended to sow the seeds of middle-income country—than in high-­ income a scientific mindset and an entrepreneurial spirit countries—1.4 percent of GDP in the median from an early age.29 Between 2016 and mid-2022, high-income country.36 Thus it is more efficient the program funded 9,600 spaces in 34 states and for middle-income countries to focus their public Union Territories. funding on a few strategic areas of research such as Mechanisms to identify advanced learners are STEM, health, and the energy transition, with funds also important. Testing plays an important role. allocated in a competitive fashion. For example, the Instead of using a single absolute measure such Pakistan Science Foundation gives competitive as the student’s place in the national distribu- research grants to scientists, engineers, technol- tion of test scores, teachers can identify students ogists, innovators, academics, and entrepreneurs with the greatest potential or the best perform- who need support to build on their initial research ers in every classroom or school. In fact, they can findings and develop new products, prototypes, and do so repeatedly and not just in one high-stakes pilot-scale production in nanotechnology, material test, which leads to a larger, more equitable talent science, and artificial intelligence.37 pool.30 Advanced learners can be offered opportu- There are many examples of partnerships nities—such as participating in advanced classes aimed at expanding countries’ research and edu- or attending selective schools—that match their cational capacities. In Argentina, the Instituto interests and abilities without necessarily hurting Balseiro is a highly selective public institution other students’ outcomes.31 that trains undergraduate and graduate students Developed countries such as Finland, France, in physics, nuclear engineering, and other STEM Japan, the Netherlands, Spain, the United fields by means of a partnership between Cuyo Kingdom, and the United States provide special National University and the National Atomic publicly backed programs for advanced learn- Energy Commission.38 Admitted students receive ers. Developing countries such as Colombia full scholarships and have access to state-of-the- and Mexico have followed suit.32 Programs vary art labs and highly personalized training, which in format. In the United States, many schools allows them to pursue highly successful careers, and school districts run a broad array of pro- both domestically and abroad. In Israel, sev- grams for advanced learners,33 whereas in Israel, eral higher education institutions—such as the Rewarding Merit | 197 Technion and the Weizmann Institute—have had universities can enhance oversight and monitor- a strong STEM orientation since their inception. ing processes, adding robustness to the gover- In 2013, Israel’s eight universities produced more nance framework. patents than the country’s firms, military labs, and One outcome of university-industry collabora- private labs combined.39 Partnerships with world- tion is venture creation by university faculty, staff, class universities can serve as a strategy to develop students, and postdoctoral students, with pri- a research base. The Egypt-Japan University of vate investors serving as venture capitalists. For Science and Technology (E-JUST) offers gradu- example, the Tshimologong Digital Innovation ate courses on electronics and communications, Project in Johannesburg, South Africa, is a prod- mechatronics and robotics, energy and envi- uct of collaboration between the University of ronment, computer science, industrial science the Witwatersrand, the province of Gauteng, and manufacturing, chemicals and petrochem- and firms such as IBM, Cisco, Microsoft, and icals, and materials science.40 The Indonesian Telkom.45 Tshimologong (“place of new begin- Biodiversity Research Center (IBRC) was cre- nings”), located in an inner-city neighborhood of ated in 2010 by Udayana University, Diponegoro Johannesburg, is close to a major research uni- University, the State University of Papua, the versity and urban infrastructure and is accessible University of California at Los Angeles, and the not only to students from the university but also Smithsonian Institution to promote biodiver- to disadvantaged youth who enjoy free training sity research and build educational and ­ scientific in computer programming, cybersecurity, and capacity in Indonesia. 41 digital animation through the program’s Digital Governments can also provide tax incentives Skills Academy. The project has created more to companies that collaborate with universi- than 105 start-ups and seeks to inspire additional ties, such as a generous tax deduction, as in Sri university-based incubators in Africa. Lanka.42 Establishing a regulatory framework Universities can also partner with local for knowledge exchange is key, particularly in companies to provide services. In Manizales, relation to the intellectual property produced by Colombia, for example, the Universidad Nacional universities with public resources.43 Universities, de Colombia and Universidad Autónoma de in turn, typically establish technology trans- Manizales are participating in a partnership, fer offices (TTOs) to promote university-based Innvestiga, to encourage innovation and produc- innovation and entrepreneurship. In 2003, the tivity enhancement among local firms. It gathers State University of Campinas (Unicamp) estab- scientists and engineers from these institutions lished Brazil’s first TTO, Inova; four years later, and supports the needs of small and medium Unicamp was the second most frequent patent enterprises (SMEs) by generating solutions in applicant in Brazil.44 Governments can also pro- materials science and processes and providing vide land and infrastructure close to universities services such as research, prototyping, and lab to attract firms, usually in science and technology tests.46 parks and incubators. The significant additional advantages of government support through uni- versities include funding projects that are near The value added by firms the knowledge frontier, promoting the design of An economy that is functioning well allocates more industry-relevant education, fostering the factors of production to the most productive development of an entrepreneurship ecosystem, firms. In developing countries, the reallocation facilitating workforce development, aiding in of factors of production has been an important technology commercialization, and mitigating driver of productivity growth, accounting for the risks of corruption through mechanisms such about 25 percent of the growth in efficiency.47 The as peer review, whistleblower protection, and ability of middle-income countries to catch up public disclosure of information. Furthermore, will depend on how well they attract and adopt 198 | WORLD DEVELOPMENT REPORT 2024 leading technologies and facilitate reallocation in middle-income countries does not solely mir- of resources toward growing, productive firms ror the challenges they face. Instead, it indicates a and industries. Countries at lower levels of devel- deficiency in competition, originating from larger opment have more opportunities for potentially firms that would have displaced them in the mar- productivity-enhancing reallocation as workers ket if they had expanded.51 Blanket support for move from less valuable and less productive activ- small firms can curtail the exit of unproductive ities toward more productive ones (figure 8.2). small businesses, perpetuate smallness, crowd Evidence from 21 European countries reveals that out other firms, and misallocate resources.52 In countries that experience the highest growth in Sri Lanka and Viet Nam, the provision of targeted terms of increases in GDP per capita are those subsidized loans and financial assistance reduced with the highest job reallocation rates. total factor productivity (TFP)—a measure of the Business dynamism is characterized by an efficient allocation of resources.53 In Mexico, pol- up-or-out dynamic; entrants exit at disproportion- icies to support SMEs have been associated with ally high rates, but those that survive grow quickly, talent misallocation and reduced labor productiv- on average. The most successful firms mature ity.54 A long-running program in Japan documents and grow larger, displacing less productive firms. that SME support programs reduce incentives for Potentially high-productivity firms must decide SME growth, thereby reducing job creation.55     whether to invest in the managerial and technical Ideally, government support would help SMEs capabilities or the R&D required to raise efficiency grow into larger, more productive companies that and product quality. Distortions in the operating pay higher wages and adapt knowledge instead of environment have a substantial impact on firms’ perpetually supporting small firms. To help SMEs decisions.48 The benefits of reallocating resources grow, support programs must identify firms with through potentially high-productivity firms not genuine constraints to expansion and produc- only helps the firms themselves, but also boosts job tivity growth and alleviate specific constraints. and output growth and creates positive spillovers Improved data and information, along with for other businesses along the value chain.49 analytic capability, are necessary to identify firm-­ ­ Policies in many middle-income countries, specific constraints.  however, are not compatible with rewarding merit And yet enforcement of tax codes often tends activities. Such policies thus need to be revisited to lump firms together by size. Even where tax and upgraded. codes do not create explicit provisions based on firm size, middle-income countries may be cre- ating a practical subsidy to SMEs through size-­ Moving away from coddling small dependent tax enforcement: governments with firms or vilifying large firms weak tax collection capacity may concentrate Subsidies to SMEs are widespread in middle-income enforcement on larger firms.56 In Türkiye, enter- countries. Governments may enact such subsi- prises with 50 or more workers must comply with dies hoping small firms will grow, creating more labor and safety laws that include establishing a jobs  and growth. Many subsidies seek to reduce health and safety board and hiring physicians and small producers’ operating costs through special other health staff and setting up a health unit.57 In credit terms and tax exemptions. But the same fact, between 1994 and 2014 more than 70 coun- distortions also reduce the incentives for a produc- tries created special enforcement units for large tive firm to expand, deterring it from scaling up taxpayers, in addition to the 18 countries that production. already had such units in 1994 (­ figure 8.3).58 This Many firms in middle-income countries remain development is part of a growing trend of taxpayer small even when long-established; they simply do segmentation, recommended by i ­nternational not aspire to grow.50 The abundance of small firms institutions. Rewarding Merit | 199 Figure 8.2 Countries at lower levels of development have more opportunities for potentially productivity-enhancing job reallocation a. Job reallocation rate and GDP per capita 90,000 CHE 80,000 GDP per capita (2015 US$), 2022 70,000 DNK 60,000 NLD SWE 50,000 GBR FIN DEU 40,000 FRA ITA SVN MLT 30,000 ESP CZE 20,000 PRT SVK LTU LVA HUN HRV ROM 10,000 POL 0 14 15 16 17 18 19 20 21 22 23 24 Job reallocation rate (%) b. Job reallocation rate and GDP per capita growth rate 45 LTU 40 LVA Cumulative GDP per capita 35 growth rate (%), 2010–17 MLT ROM 30 POL 25 HUN 20 SVK 15 CZE GBR DEU SVN 10 HRV ESP SWE DNK NLD 5 FRA PRT FIN CHE 0 ITA –5 14 15 16 17 18 19 20 21 22 23 24 Job reallocation rate (%) Source: WDR 2024 team using the 9th Vintage CompNet Dataset (dashboard), Competitiveness Research Network, Halle Institute for Economic Research, Halle, Germany, https://www.comp-net.org/data/9th-vintage/. Note: Small firms with fewer than 20 employees are excluded to allow consistency across countries and sectors. The unit of analysis is the legal unit (firm). Entrants and exiters are excluded from the analysis. Data on gross domestic product (GDP) per capita are from National Accounts (dashboard), Organisation for Economic Co-operation and Development, Paris, https://www.oecd.org/sdd/na/; WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https:// datacatalog.worldbank.org/search/dataset/0037712. GDP per capita growth is computed as the growth rate between 2010 and 2017 because the job reallocation rate is computed over 2010–17. Twenty-one European countries are included. The diagonal line indicates linear fitted values. For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. 200 | WORLD DEVELOPMENT REPORT 2024 Figure 8.3 The number of countries creating Letting go of unproductive firms special enforcement units for large taxpayers Letting inefficient firms and business models has increased fail is a core principle of creative destruction. 120 Literature on firm exit—stemming from seminal work by Hopenhayn (1992)—shows that the exit 100 of less productive firms contributes substantially Number of countries to raising aggregate productivity. In many coun- 80 tries, during periods of trade liberalization the exit of the least productive firms has boosted growth.61  60 In middle-income countries, however, bureau- cratic frictions prolong the survival of zombie 40 firms—inefficient, debt-ridden companies that 20 crowd out investment by productive firms.62 Reforms to bankruptcy laws can prevent unpro- 0 ductive incumbents from turning into zombie firms and should focus on enabling failed busi- 80 20 0 60 70 90 00 14 1 19 19 19 19 20 20 nesses to exit swiftly and predict­ ably and on Source: Bachas, Fattal Jaef, and Jensen 2019. allowing viable businesses to restructure.63 Note: The figure includes data collected by Bachas, Fattal More generally, the shedding of outdated Jaef, and Jensen (2019) for all 113 countries with more than arrangements—enterprises, jobs, technologies, 1 million inhabitants. private contracts, policies, and public institu- tions—is essential for reallocation, innovation, Research has revealed that if firms were and growth. Economic downturns sometimes to comply with size-dependent tax policies, create a much-needed opportunity for older, less in 140  countries employment growth would productive firms to make way for newer, more drop by 25 percent. Conversely, removing productive ones. A key Schumpeterian insight is size-­ dependent taxation would lead to TFP that the creation of new industries does not go gains of about 1 percent, on average, and up to ­ ­ forward without sweeping away past realities.64 2.3 ­ percent for more distorted economies.59 For example, in Mexico eliminating distortions cre- ated by size-dependent ­ t axation ­policies Modernizing organizational models to favoring  small firms could boost output by manage firms   9  ­percent.60 In Chile, China, and India, reduc- As barriers to growth are removed and resources tions in distortions helped these ­ economies are better allocated toward firms that add value, close the gap between actual and  poten- firms will need to make some strategic decisions. tial productivity  by  10 percent. Figure  8.4 Firm owners aspiring to expand must delegate ­ i llustrates  the  value of the actual level  of managerial tasks to professionals. Delegation manufacturing TFP in each country relative necessitates strong legal institutions to estab- to the potential level if resources were allo- lish and enforce contracts between owners and cated efficiently—that is, if the economy was managers. These contracts should give managers on the production ­ possibilities frontier.  More incentives to work hard and should hold them important,  reforms of size-­ dependent policies accountable for any wrongdoing. Such arrange- increase the return to skills  and encourages ments encourage firms to increase their invest- technology adoption and productivity in the ments in expansion. As productive firms expand, longer term.  creative destruction eliminates unproductive Rewarding Merit | 201 Figure 8.4 Improvements in allocative efficiency in Chile, China, and India have been driven by reducing productivity-dependent distortions 1.3 Relative to first year of growth acceleration 1.2 1.1 1.0 0.9 0 5 10 15 20 Number of years since first year of growth acceleration Chile China India Source: WDR 2024 team based on the methodology in Hsieh and Klenow (2009). Note: The first year in each country’s growth acceleration covered by the firm-level data is 1980 in Chile, 1998 in China, and 2000 in India. Each line in the figure illustrates the evolution of the gap between observed total factor productivity (TFP) in a country relative to the potential TFP if resources had been allocated efficiently across firms, normalized by the gap observed at the beginning of the sample in 1998. firms, driving up market compensation for good employees, setting production targets, managers within those expanding firms. The deciding on product varieties, and monitoring opportunity to earn more, in turn, encourages performance. The impacts of better manage- capable young individuals to invest in accumulat- ment are critical not only to individual firms ing human capital to ensure a brighter future and but also to countries.66 Managerial quality enhances opportunities for everyone. accounts, on average, for about one-third of the Capable professional managers can make gap in TFP between the country at the lead of sound decisions in their daily operations, think the technology frontier, the United States, and strategically, and manage human resources other countries.67 In fact, Spain’s rapid growth efficiently.65 In small firms, good management from an economic backwater in Europe to a practices include keeping separate household developed nation was backed by technology and business accounts and monitoring inven- adoption and improving managerial capabili- tories. In larger firms, they include rewarding ties (box 8.4).  202 | WORLD DEVELOPMENT REPORT 2024 Box 8.4 Catching up by opening up and modernizing firms: The Spanish growth miracle  Spain grew rapidly between 1950 and 1980, becoming a high-income economy in the late 1970s. Over those three decades, its output per capita increased from 27 percent to 57 percent of that of the United States. How did Spain achieve rapid catch-up? Political stability, a necessary condition, helped. But economic policies remained unreformed, state-owned enterprises remained a significant part of the economy, and rent-seeking by incumbents remained rampant. Spain’s auto industry—which is one of the world’s largest car exporters, the country’s largest export sector, and one of the country’s largest employ- ers—offers three notable lessons. First, although the domestic market for cars remained protected, Spain opened up to foreign direct investment to attract large investment in a new plant by Ford. The author- ities reduced the local content requirement but set a minimum size of production, estab- lished an export ratio, and capped the number of cars that Ford could sell domestically. Overall, incentives were aligned to learn from global technology and organizational struc- tures. The venture was disciplined by the need to export in global markets.a Second, addressing complementary bottlenecks also helped.b Ford was able to export its large production of cars because the port of Valencia operated smoothly. The port was capitalized enough to respond to the demands placed on it. Third, technology adoption and a greater focus on management techniques proceeded rapidly, and efficiency gains quickly materialized, supported by a well-trained cadre of local professionals and entrepreneurs who were ready to adopt and adapt the new technology to the realities of Spain.   Eventually, the opening up of Spain’s economy helped speed up the “creative destruc- tion” that is central to structural transformation.   Source: Based on Calvo-González (2021). a. There are significant parallels with the operation of Korean chaebols as conduits for technology diffusion. b. See Cusolito and Maloney (2018). Improving general education is the first step yields high benefits to larger firms, such as in toward promoting managerial skills, as discussed Colombia, India, and several other countries.71 earlier in this chapter. In emerging economies, Consultants diagnose management practices and firms with more educated owners tend to have provide recommendations accordingly. Positive better business practices,68 and more educated effects can emerge within one year and tend to entrepreneurs start firms that are larger and grow last for several years—although they sometimes faster.69 disappear when the manager who adopted them A second step—which can provide benefits leaves the firm. These interventions are highly sooner—involves training and information inter- effective, but they are also relatively costly. The ventions. Training can be provided in a formal costs can be reduced by providing group-based classroom or by means of in-firm consulting. consulting.72  Classroom-based training is the most popular Entrepreneurs in middle-income countries method among small firms.70 In-firm consulting often rely on family and social networks to Rewarding Merit | 203 maintain business relationships and run their hobble their potential. For this and other reasons, enterprises, but this practice hampers their ability there is a paradox of low entrepreneurship amid to grow. Even for large publicly traded firms, how great opportunity in emerging economies: a firm is controlled can be traced back to the coun- try’s legal origins. 73 The share of family-controlled Like any investor, an entrepreneur is fun- firms is highest in countries that have adhered to damentally placing a bet, comparing an the French civil law system, followed by countries entrepreneurial project with an expected that have adopted the German civil law system and range of returns and risks against other civil law countries in Scandinavia.74 Shareholder alternatives, such as “safe” salaried work, protection rights—provisions of corporate law which is the opportunity cost of entre- that allow shareholders to take legal action against preneurship. This implies both a process managers who abuse their position—are system- of managing risk and a process of learn- atically linked to dispersed patterns of owner- ing—about the investment, about run- ship, which can lead to better management. In the ning a firm, and about evaluating and absence of solid shareholder protection, family managing risk.79 control and management of a firm offer protection against abuse by management.75 However, in fam- Approaches aimed at supporting entrepre- ily firms with lower-quality governance, manage- neurship in developing countries are hampered ment quality is substantially lower. 76   by a “missing middle.” On the one hand, there has Even among publicly listed corporations, been a drive in development policy to focus on family control is more likely in countries where informal businesses or microentrepreneurs ori- organized labor is more powerful and collective ented toward survival.80 On the other hand, there dispute-resolution mechanisms are stronger. has been a fascination with high-tech entrepre- After all, it is easier to collude when corporate neurs in a handful of hot spots.81 What is missing control is concentrated in the hands of families is a clear-eyed assessment of the barriers facing and organized labor.77 Management quality also growth-oriented entrepreneurs in middle-income varies with labor regulations and organization. countries—the protagonists of Schumpeterian In the United States, stronger union power weak- growth. Three barriers are paramount: money, ens the ability of managers to use desirable per- mentors, and markets. formance incentives.78 Across countries, people Money. Lack of targeted finance is a fun- management practices are weaker in those with damental reason why many opportunities in more restrictive labor regulations. Revising laws middle-income countries do not lead to more and regulations related to corporate governance growth-oriented entrepreneurship. Improving may be necessary to support entrepreneurship access to finance is not only about extending and good management.  credit; it is also about backing a money-making idea implemented by capable founders and man- agers. But it is difficult to gauge these dimensions. Connecting entrepreneurs with money, For example, a study of a large business plan com- mentors, and markets petition in Nigeria found that even after an ini- Entrepreneurs in middle-income countries are tial screening, expert judges, machine learning the main protagonists of Schumpeterian growth models, and economic models had very low abil- through creative destruction, but these entre- ities to predict which firms would grow the fast- preneurs are often disconnected from finance, as est over the next three years.82 Female founders well as from networks of other entrepreneurs who of firms face particular biases and obstacles. In can mentor them and help them access markets. general, project evaluators place a lower value on Information asymmetries and lack of collateral the competence or leadership potential of women 204 | WORLD DEVELOPMENT REPORT 2024 than of men, and investors inquire more about surprised by what they’re capable of.”90 A highly risks  when dealing with female founders than competitive process allows founders to join a with men.83  three-month program of capability assessment Information asymmetries and lack of collat- and upgrading in which their ideas are pitched eral are especially constraining for entrepreneurs to investors. After the program’s inception, 5,000 with new ideas, regardless of whether those ideas US-based start-ups accelerated between 2005 are imitative or innovative. Equity markets can be and 2015 raised nearly US$20 billion in venture instrumental in supporting innovative activities, capital.91  especially in private firms, which typically face Accelerators have made their way into emerg- larger financing gaps than publicly listed firms.84 ing markets, helping entrepreneurs access However, private markets for equity financing early-stage investment. But beyond invest- lack depth and access in emerging economies ments,  accelerators are “schools for entre- figure  8.5). Private equity markets also make it (­ preneurs” that give competitively selected easier for entrepreneurs to cash in on their invest- participants opportunities to build their capa- ments and move on to a new project, should they bilities—including through business training, choose to do so. Often, entrepreneurial ventures networking, and mentoring—and sometimes do not perform as planned, but the costs of exit- supply funding.92 ing are very high, and the entrepreneur is held Researchers associated with the Global responsible for the entire downside risk of a failed Accelerator Learning Initiative (GALI) com- endeavor, leading to reputational and finan- piled information on 2,455 ventures that cial downfalls.85 This is a great disincentive to applied to 43 accelerator programs between risk-taking.   2013 and 2015, about half in high-income Mentors. Although money is important, entre- countries and half in emerging markets.93 preneurial success is not all about the money. After one year, participating ventures report Most entrepreneurs need to be connected with higher revenue and employee growth, as well networks of entrepreneurs—those at their as higher equity and debt investment growth, stage as well as successful ones—to fully assess compared with ventures rejected from the whether they and their ideas are fit for entre- application pool. Surprisingly, the major gains preneurship. Data reveal that start-ups collab- for accelerated ventures in emerging markets orating with seasoned venture capitalists tend are in leveraging debt, not equity. Furthermore, to exhibit superior performance.86 Accelerator emerging market entrepreneurs rarely indicate programs are a relatively recent addition to that connections made during a program help the variety of programs that direct knowledge, grow their networks. They also tend to place social, and financial capital to promising people more emphasis on business skill development. and ideas.87 They provide training and techni- However, program managers in emerging mar- cal assistance, along with mentorship and net- kets report difficulties in recruiting mentors working support, and offer certification, thereby and advisers.94   reducing information asymmetries between Markets. In middle-income countries, creative entrepreneurs and investors.88 destruction is amplified by better connections Y Combinator, launched in 2005, is widely between opportunity and entrepreneurship. regarded as the first accelerator program.89 Its Governments can help create and sustain con- website stated: “You can’t make people something testable markets by weakening the forces of pres- they’re not, but the right conditions can bring out ervation, and they can work with investors and the best in them. And since most people have way growth-oriented entrepreneurs to nurture infu- more potential than they realize, they’re often sion and innovation.    Rewarding Merit | 205 Figure 8.5 In emerging market and developing economies, few companies are funded through venture capital or private equity a. Venture capital 90 SGP ISL Number of companies funded ISR 80 EST IRL per 1 million persons 70 SWE 60 GBR CHE LUX FIN DNK 50 MLT USA 40 NOR BEL CAN 30 NZL NLD BMU HKG AUS FRA 20 ASM LTU ESP AUT PRT CYP HUN LVA KOR DEU 10 LBN SVN BHS ARE QAT CZE ITA JPN SAU KWT 0 20,000 40,000 60,000 80,000 100,000 120,000 GDP per capita (US$), 2020 b. Private equity 35 CAN 30 Number of companies funded per 1 million persons 25 FRO LUX ISL 20 SWE AUS NOR 15 CYM AND FIN DNK GBR NLD 10 ESP FRA SGP IRL BHR SAU EST NZL USA MLT BEL CHE LCA 5 CYP ISR DEU MUS BRB PRT HKG CZE ITA ARE AUT SVN KOR JPN QAT 0 20,000 40,000 60,000 80,000 100,000 120,000 GDP per capita (US$), 2020 EMDEs High-income economies Source: Didier and Chelva 2023.  Note: The figure displays the number of companies funded through venture capital (panel a) and private equity (panel b) investments from deals concluded during 2018–19. Economies are classified according to the World Bank’s income classification as of June 2020 (Serajuddin and Hamadeh 2020). For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. EMDEs = emerging market and developing economies; GDP = gross domestic product. 206 | WORLD DEVELOPMENT REPORT 2024 Reducing an economy’s follows first examines economywide ­fiscal ­policies, then turns to sectoral policies, and finally explores greenhouse gas emissions the options for scaling up renewable technologies. In middle-income countries, emissions drive the impact of development on global climate change. Policy debates in middle- and high-income coun- Economywide fiscal policies tries should be concerned primarily with decou- Regarding energy, the ultimate impact of growth pling GDP growth from emissions growth in all on global greenhouse gas emissions will depend ways feasible. One important consideration is on the carbon intensity of GDP, reflecting both arrangements that lock middle-income countries GDP’s energy mix and energy intensity. Fiscal into particular types of production, such as long- policies can influence both dimensions. term investments in coal and other fossil fuels. To Carbon pricing is an essential policy for mit- the extent that so-called lock-in effects are pre- igating emissions,95 while helping to raise public venting progress toward actually meeting global revenue in an efficient and less distortive way environmental goals, all efforts should be directed than the alternatives. It signals to markets the toward reducing the path dependence from the social cost of emitting GHGs, creating financial specific energy source, while better protecting incentives to abate emissions, reduce fossil fuel natural resources, including forests (box 8.5). consumption, and innovate low-carbon products Policies and technologies to mitigate carbon and processes. Some economists believe that car- emissions will vary among middle- and high-­ bon taxation is the most efficient instrument for income countries, depending on their economic reducing emissions in a growth-friendly way. On structures, resource endowments, and institu- January 16, 2019, 43 of the world’s most promi- tional and technical capabilities. A good starting nent economists, including 27 Nobel Laureates, point is fiscal policies. These can be surgically issued a statement published in the  Wall Street sharp if designed with care. The discussion that Journal (2019) arguing that a carbon tax in the Box 8.5 Productivity growth can slow deforestation in Brazil Brazil’s Amazon region provides the world with immense ecosystem services. Estimates peg the services to be worth, at a minimum, US$317 billion a year, US$210 billion of which is accounted for by carbon dioxide (storage) alone.a Deforestation is among Brazil’s lead- ing sources of greenhouse gas emissions and a major threat to biodiversity.b In fact, eco- system collapse in the Amazon stemming from deforestation and climate change ranks among the most catastrophic tipping points for the planet.c In response, Brazil has made much progress in protecting the Amazon, improving forest and land governance by means of, for example, the Action Plan for the Prevention and Control of Deforestation in the Legal Amazon, first launched in 2004 and most recently reinvigorated in 2023. And yet Brazil’s growth model also matters for the Amazon. The fact that it remains anchored in factor accumulation and land accumulation is synonymous with agricultural frontier expansion and deforestation. To overcome the middle-income trap Brazil will have to raise productivity—and growing through productivity rather than factor expansion would also slow deforestation.d Indeed, there is a strong relationship between Brazilian productivity and the change in forest cover in the country’s Amazon (figure B8.5.1). (Box continues next page) Rewarding Merit | 207 Box 8.5 Productivity growth can slow deforestation in Brazil (continued) Figure B8.5.1 Amazon deforestation falls when Brazilian productivity rises 130 0 125 Net change in forest cover –5 TFP index (1996 = 100) 120 (thousands of km2) 115 –10 110 105 –15 100 –20 95 90 –25 85 80 –30 2002 2004 2007 2011 2014 2017 2018 20 0 1996 2003 20 9 22 1997 1998 2099 2001 2005 2006 2008 2009 2010 20 2 20 3 2015 2016 20 0 20 1 1 0 1 1 2 2 19 Total factor productivity Net change in forest cover, Legal Amazon (right axis) Source: WDR 2024 team extending Ferreira Filho, De Souza, and Hanusch (2022). Note: Legal Amazon is the largest sociogeographic division in Brazil, containing all nine states in the Amazon basin. km2 = square kilometer; TFP = total factor productivity. a. Hanusch (2023). b. World Bank (2023a). c. Lenton et al. (2023). d. Ferreira Filho, De Souza, and Hanusch (2022); Hanusch (2023). United States “offers the most cost-effective but the greater a program’s scope, the more effective lever to reduce carbon emissions at the scale and it can be.97 The report points out that the number speed that [are] necessary.” Others have proposed of countries that have adopted direct carbon pric- a strategic combination of temporary research ing schemes through ETSs or carbon taxes is lim- subsidies and carbon taxes that could steer tech- ited.98 Perhaps more important, so is the coverage nological advancements toward more environ- of such programs. With the introduction of the ETS mentally sustainable solutions.96 in China in 2021, the share of global carbon dioxide Direct carbon pricing instruments include car- (CO2) emissions from fossil fuels covered by direct bon pricing signals sent through carbon taxes and carbon pricing schemes rose to about 31 percent emissions trading systems (ETSs). According to (amounting to about one-quarter of global GHG the World Bank’s State and Trends of Carbon Pricing emissions). Along with their coverage, average car- 2023 report, these schemes currently cover a rela- bon prices have been rising over the last few years. tively limited portion of global carbon emissions, The carbon price in the European Union  (EU) 208 | WORLD DEVELOPMENT REPORT 2024 through the EU ETS rose sharply from 2019 to higher carbon price signals than their direct coun- 2021. Nevertheless, the carbon prices prevailing in terparts. By contrast, energy subsidies send strong most jurisdictions and their estimated global aver- signals in the opposite direction, undermining age remain quite modest. the positive signals sent from direct and indirect Because the overall carbon price signal is not instruments, as illustrated in figure 8.6.100 confined to direct carbon pricing, the concept Removing inefficient fossil fuel subsidies is an of the total carbon price (TCP) has been intro- integral part of the policy mix to reduce carbon duced—a metric intended to assess the price signal emissions. This market distortion discourages resulting from a combination of direct and indi- the adoption of clean energy because regulated rect carbon pricing instruments, including energy prices or taxes favor fossil fuels. After a notice- excise taxes and fuel subsidies.99 Illustrative able dip in 2020 stemming from the COVID-19 TCP calculations carried out using the best avail- pandemic, global fossil fuel subsidies for 2022 able global data sets relying on annual data for doubled from the previous year to an all-time 142 countries covering the last 30 years find that high of US$1 trillion, as indicated by preliminary indirect carbon pricing instruments play a much estimates.101 According to a global tracking effort, more prominent role in sending price signals on at least 60 countries increased (or even reintro- carbon emissions. Among indirect carbon pricing duced) general fuel price subsidies (as opposed to instruments, an analysis of illustrative TCP calcu- targeted compensation), and at least 98 countries lations finds that energy taxes, in particular, send announced energy-related measures, including the strongest price signal. These taxes cover a sig- subsidies for fuel, electricity, transport, and elec- nificant share of global emissions and send much tric vehicles, as well as price controls for fuel.102 Figure 8.6 Indirect carbon pricing such as energy taxes is the strongest price signal 50 40 30 US$/tCO2 20 10 0 –10 –20 91 19 93 97 99 01 13 95 03 17 05 07 09 11 15 21 20 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 ETSs Carbon taxes Energy taxes Energy subsidies VAT deviations Source: Agnolucci, Gencer, and Heine 2024. Note: The figure presents illustrative calculations for the global aggregate total carbon price using the best available global data. The figure covers 142 countries. ETSs = emissions trading systems; tCO2 = metric tons of carbon dioxide; VAT = value added tax. Rewarding Merit | 209 Well-designed taxes can be a starting point market for renewable energy and allow policy to incentivize citizens and businesses to make makers to manage the quantity and the quality of cleaner choices, thereby reducing climate damage renewable energy projects by setting the volume, and air pollution. Taxes also raise much-needed technology, and the criteria for the bids. Auctions revenue, which can be used to fund vital govern- can also foster innovation and diversification of ment services and support vulnerable groups in renewable energy sources by creating different adjusting to higher energy prices, including by categories or segments for the bids. Auctions have introducing or strengthening social safety nets. led to low utility-scale prices for solar photovol- Policy makers, however, face strong resistance taics (PV) in middle-income countries (figure 8.7). from industrial and residential consumers to Technology adoption and deployment are aided removing subsidies and raising the social cost of by scale. In Europe, the companies most diver- carbon. Often, resistance comes from interest sified in renewables are the largest firms—they groups representing specific sectors concerned have a lower risk because of their size advantage. about bearing the brunt of the economic and Those companies able to obtain the cheapest cap- job losses associated with the loss of those rents. ital to replace energy from the Russian Federation Indeed, the literature simulating the impact of by means of additional renewable generation are removing subsidies finds that reductions in GDP those already most invested in renewables. and welfare will occur in countries in the short Raw materials are a significant element of the term. In China, models indicate the GDP could cost structure of many technologies, and any dis- decline by up to 3.8 percent.103 In the Islamic ruption in supply can increase the cost of capital. Republic of Iran, GDP could decrease by 2.2 per- For lithium-ion batteries, technology advance- cent and welfare by 5.2 percent, and the non‑​ ment and economies of scale have reduced overall energy price index could increase by 26 percent.104 costs by 90 percent over the last decade. However, However, those short-term effects are reversed in if both lithium and nickel prices were to double at the long term, leading to a substantial increase the same time, it would offset all the anticipated in GDP and welfare, particularly when subsidy cost reductions associated with a doubling of bat- reforms are accompanied by complementary pol- tery production capacity. For electricity networks, icies such as cash transfers and other productive copper and aluminum currently represent about activities (including electrification, public trans- 20 percent of total grid investment costs. Higher port, and investment in education and health). prices arising from a tight supply could constrain the level of grid investment. The most efficient way to scale up the deploy- Complementary sectoral policies ment of low-carbon energy is to respect merit Countries such as Germany and Spain have used order: the sequence followed by grid operators feed-in tariffs (FiTs) to support the deployment selling power to the market. The starting point is of new clean energy technologies, but as tech- set by the cheapest offer (made by the power sta- nologies mature, competitive bidding through tion with the lowest running costs), which deter- auctions in middle-income countries (including mines the wholesale market prices. Any provider Brazil, India, and South Africa) have been more that can offer renewable energy at zero marginal cost-efficient in procuring renewable capacity cost—that is, with insignificant operating costs— (for example, by achieving lower prices per unit should have priority in meeting demand. When of electricity, as assigned in the power purchase the merit order functions as designed, it shifts agreements).105 FiTs guarantee a fixed price for the prices along the supply curve, which energy econ- electricity produced by renewable energy sources. omists accordingly call the “merit order curve.”106 However, they may result in overpayment to pro- The design and enforcement of effective reg- ducers if the cost of renewable energy technolo- ulations are vital in ensuring that the merit gies falls over time. Auctions create a competitive order is respected. Clear, transparent rules on 210 | WORLD DEVELOPMENT REPORT 2024 Figure 8.7 In some middle-income countries, the prices of renewable energy through competitive auctions have reached record lows 9 PPA tariff estimates based on 8 auctions (US$0.01 per kWh) Serbia, 8.2 7 6 Poland, 5.2 Philippines, 4.5 5 Kazakhstan, 4.0 Zambia, 3.9 4 Chile, 3.7 Uzbekistan, 2.7 India, 2.7 India, 3.0 3 Ethiopia, 2.6 India, 2.8 2 Qatar, 1.6 Colombia, 1.8 Tunisia, 2.4 Saudi Arabia, 1.0 1 Dubai, 1.7 Abu Dhabi, 1.4 0 2018 2019 2020 2021 2022 2023 Asia (South and East) Europe and Central Asia Latin America Middle East and North Africa Sub-Saharan Africa Source: WDR 2024 team based on results of auctions reported in PV magazine, http://www.pv-magazine.com. Note: The emirates of Abu Dhabi and Dubai are shown separately. kWh = kilowatt-hour; PPA = power purchase agreement. interconnection will be needed to ensure entry in detecting abuse of dominance in the Bulgarian by low-carbon energy providers and respect of and German wholesale markets. In Mexico, the the merit order. Allegations of abuse of domi- Federal Economic Competition Commission nance are common in wholesale electricity mar- (COFECE) issued a public report in 2021 about kets, which are more susceptible to the exercise improving the regulation of Clean Energy of market power. This abuse includes both phys- Certificates, an instrument that provides incen- ical withholding (not offering available capacity tives for generating electricity using clean tech- to the market that could be profitably produced nologies under the framework of the Electricity at the market price) and economic withholding Industry Law and the Energy Transition Law. (offering available capacity at a price that does not In Colombia, the Superintendence of Industry reflect its marginal cost—including the opportu- and Commerce (SIC) developed an advocacy ini- nity cost). Both cases of withholding make the tiative in the energy sector in 2019 encouraging merit order curve steeper and shifts its intersec- competitive bidding through renewable energy tion with the demand curve, resulting in a higher auctions. For example, SIC recommended price. Customers are then worse off while produc- arranging subsequent rounds of auctions (ten- ers benefit. ders), instead of a single tender, to enable the Worldwide, promising regulatory initiatives participation of companies that were still pre- are under way. In the European Union, regula- paring to enter the market and would only be tors began investigating 109 cases of abuse of able to provide renewable energy sources after wholesale energy market integrity and transpar- the initial date indicated in the first auction ency in 2021. Box 8.6 describes successful cases model. Rewarding Merit | 211 Box 8.6 Correcting abuses of dominance in electricity markets Liberalizing and opening up a sector that historically has been governed by the state and state-controlled actors are challenges that go far beyond updating the regulatory frame- work, as witnessed in Germany and Bulgaria. Bulgaria. In 2019, the Bulgarian Commission for Protection of Competition (BCPC) fined the National Electricity Company EAD (NEK) for abuse of dominance. NEK is part of Bulgarian Energy Holding, which owns 80 percent of all hydropower plants in Bulgaria. It is historically the last-resort supplier and the coordinator of special balancing energy groups.a BCPC found that NEK abused its dominant position in the balancing market of electricity producers from renewable energy sources at preferential prices in two ways.b First, it unilaterally altered the forecasted hourly electricity production schedules submit- ted by renewable energy producers. NEK submitted a different schedule, usually unilat- erally providing for less production (that is, without informing the respective producer). These changes made it practically impossible for the renewable producers to meet the amounts set by NEK. Second, the BCPC found that unilaterally the renewable energy supply producers were financially burdened by the artificially increased imbalances.c Germany. In May 2021, the German Federal Network Agency (Bundesnetzagentur) imposed fines of €200,000 on Energi Danmark A/S and €175,000 on Optimax Energy GmbH for manipulation of the wholesale electricity market. The penalties were the out- comes of investigations opened in September 2020 after significant imbalances were observed in the system in June 2019. The Bundesnetzagentur’s analysis of trading activ- ities indicated market manipulation involved sales of electricity that was not available. The companies placed offers to sell electricity on the intraday market shortly before the electricity was due to be supplied without intending to supply it. Their incentive to do so stemmed from the difference between the unusually high intraday price and the lower expected imbalance price in the balancing market.d The practice distorted market signals at a time when transmission system operators had to make full use of balancing energy and  take other measures to ensure the stability of the German system. The practice not only allowed the companies to realize unjustified profits but also threatened system stability.e Source: Mateina and Grunova 2020. a. A balancing group is a group of participants on the free market, both consumers and producers which optimize their electricity costs by netting their counter hourly deviations (imbalances) and reducing the overall deviation between the projected and reported electricity consumption. A special balancing group includes only participants of the regulated market. b. See Decision No. 833/18.07.2019. c. Mateina and Grunova (2020). d. Balancing services are reactive short-term means of leveling out frequency deviations in the power grid. These services (sometimes also called control reserve) are one of many ancillary services that system operators must provide for a secure power supply. e. See Bundesnetzagentur (2021). 212 | WORLD DEVELOPMENT REPORT 2024 In many middle-income countries, policies to enhance value added across the ­ economy need to be revisited and upgraded to reward through the efficient resource utilization of merit activities. But this will require a change in talent, capital, and energy. All these efforts mindset. Policy makers should think in terms of will help middle-income countries escape the adding value (merit): economic, social, and envi- ­m iddle-income trap. ronmental. That requires changes in policies Notes  1. World Bank (2018). STEM professors, significantly increased the likelihood  2. Graham (2013). of college students expressing STEM career aspira-  3. China is an exception. By 1990, China and India were tions. Mentoring and information interventions are similar in terms of their gross domestic product (GDP) among the most effective ways to boost the interest of per capita, upper-secondary completion rate, and ter- female students in STEM (Muñoz-Boudet et al. 2017). tiary enrollment rate. Today, China surpasses India in all 16. World Bank, UNESCO, and ILO (2023). three indicators. A comprehensive study of educational 17. In their review of job training programs, Carranza and upgrading in Brazil, China, India, and Indonesia starting McKenzie (2023) note the importance of design with the 1950 cohorts finds that, although China fol- issues, given that the returns to most TVET programs lowed a strategy of progressive universalism, India did are modest. They highlight the success of Colombia’s not (Schady, Isaacs, and Parra 2024). Jóvenes en Acción program and of others led by non-  4. Education Commission (2016). governmental organizations. The scalability and gen-  5. As the focus on basic education declined, so did its eral equilibrium effects of these programs remain to budget share. But because the country was growing, be seen. the absolute amount of resources spent on basic edu- 18. Material on Finland is drawn from Nieminen and cation did not decline. Kaukonen (2001). Information on Denmark is from  6. Angrist et al. (2023); World Bank (2018). Cedefop (2023), Peters et al. (2010), and the sources  7. Angrist et al. (2020). This comparison includes only cited therein. those reforms that have been rigorously evaluated. 19. See, for example, Gorgodze and Chakhaia (2021).  8. These findings are also consistent with the “Smart 20. World Bank (2018). Buys” recommendations of the GEEAP (2023). In addi- 21. Khanna and Szonyi (2022). tion, Angrist et al. (2023) conduct a cost-benefit analy- 22. Chetty et al. (2020). sis for two interventions—structured pedagogy and 23. Carranza et al. (2022). teaching at the right level—and find that, if applied to 90 24. Abebe et al. (2021); Alfonsi, Namubiru, and Spaziani percent of the nearly 470 million students in low-in- (2022); Bandiera et al. (2022); Groh et al. (2015); Kelley, come and lower-middle-income countries, they would Ksoll, and Magruder (2023). cost on average US$18 per student and yield US$65 in 25. Abebe et al. (2021). benefits for every dollar spent. Overall, they would 26. Banerjee and Sequeira (2023). increase spending by a mere 6 percent and yet raise 27. Chetty et al. (2022a, 2022b). learning by 120 percent. See also Dixit and Gill (2024). 28. Alfonsi, Namubiru, and Spaziani (2022).  9. Dinarte-Díaz et al. (2023); Ferreyra et al. (2021). 29. Khanna (2023). 10. Paredes (2024). 30. Thomas B. Fordham Institute (2023). 11. La Hora (2022). 31. Card and Giuliano (2016) documented large achieve- 12. In the United States, the co-op model was first imple - ment gains for students who are tracked in separate mented at the University of Cincinnati and Northeastern “gifted/high achiever” classrooms. Benefits were over- University. Today, additional institutions include whelmingly concentrated among minority participants Georgia Tech, Purdue, and Drexel. without negative spillovers to the students who were 13. Garcia and Crawley (2024). left behind. Van Reenen (2021) reviews other evidence 14. China’s number of science and engineering (S&E) grad - for the United States. uates grew from 225 per million population in 2000 to 32. Rutigliano and Quarshie (2021); Tirri and Kuusisto (2013). 1,057 in 2014, and the number of PhD S&E graduates 33. For a full discussion on advanced learners in the United increased from 5.6 per million population in 2000 to 23 States, see Thomas B. Fordham Institute (2023). in 2014. The source of these figures is WDR 2024 team 34. Zorman, Rachmel, and Bashan (2016). calculations based on World Bank (2020). 35. See Baeza (2020) for Spain and Invention Programs 15. For example, Kitchen, Sonnert, and Sadler (2018) find (portal), Kid Museum, Bethesda, MD, https://kid​ that in the United States the National Science museum.org/invention-programs/ and NIHF (2022) -­ Foundation’s STEM Talent Expansion Program for the United States. increased high school students’ interest in a STEM 36. WDR 2024 team calculations based on OECD (2021). career. In addition, Kitchen, Sonnert, and Sadler (2020) Data are available for eight middle-income countries show that campus visits, including meetings with and 30 high-income countries. Rewarding Merit | 213 37. For a description of Pakistan’s Competitive Research (2019); Higuchi, Nam, and Sonobe (2015); Karlan, Program, see https://psf.gov.pk/crp.aspx. Knight, and Udry (2015). 38. See About Us, Instituto Balseiro, San Carlos de 72. Iacovone, Maloney, and McKenzie (2022). Bariloche, Argentina, https://www.ib.edu.ar/english​ 73. Glaeser and Shleifer (2002); La Porta et al. (1998). _­version/about_us.php. 74. Aminadav and Papaioannou (2020); La Porta et al. 39. Drori and Netivi (2013). (1999). 40. Bond et al. (2012). 75. La Porta et al. (1997, 1999). 41. Shetty et al. (2014). 76. Scur et al. (2021). 42. The deduction is 300 percent. The company reduces 77. Pagano and Volpin (2005). taxable income by three times the amount of R&D 78. Bloom et al. (2019). expenditure (Mendes 2015). 79. Cusolito and Maloney (2018, 79). 43. Brazil, China, Malaysia, the Philippines, and 80. Naudé (2011). South Africa, as well as other member countries of the 81. Aldrich and Ruef (2018). Organisation for Economic Co-operation and 82. McKenzie and Sansone (2019). Development (OECD), have followed the United States 83. Miller et al. (2023). in allowing a university, research institute, small firm, or 84. Didier and Cusolito (2024). nonprofit institution to claim ownership of an invention 85. Cusolito and Maloney (2018). funded with public resources. For evidence on the 86. Akcigit et al. (2022). effectiveness of providing scientists with greater own- 87. Lall, Chen, and Roberts (2020). ership of innovations, see Hvide and Jones (2018) for 88. Kim and Wagman (2014); Plummer, Allison, and Norway and Lach and Schankerman (2008) for the Connelly (2016). For more on the highly successful United States following the Bayh-Dole Act. accelerator program of Y Combinator, widely regarded 44. For example, see Bueno (2009). as the first, see Hathaway (2016). 45. Garcia and Crawley (2024). 89. Hathaway (2016). 46. Garcia and Crawley (2024). 90. See https://www.ycombinator.com/about. 47. Cusolito and Maloney (2018). 91. Hathaway (2016). 48. Cusolito and Maloney (2018). 92. Gonzalez-Uribe and Hmaddi (2022). 49. Grover, Medvedev, and Olafsen (2019). 93. Roberts et al. (2017). 50. Eslava and Haltiwanger (2020); Hsieh and Olken (2014). 94. Roberts et al. (2017). 51. Akcigit, Alp, and Peters (2021). 95. Stern (2022). 52. Bertoni, Colombo, and Quas (2023); Kersten et al. 96. See Acemoglu et al. (2012, 2016). (2017). 97. World Bank (2023b). 53. Aivazian and Santor (2008); Vu and Tran (2021). 98. World Bank (2023b). 54. López and Torres (2020). 99. Agnolucci et al. (2023); Agnolucci, Gencer, and Heine 55. Tsuruta (2020). (2024). TCP components labeled as “energy taxes” and 56. Bachas, Fattal Jaef, and Jensen (2019); see discussion “energy subsidies” are based on “net” computed val - in chapter 4. ues, as proxies for actual values of energy taxes and 57. Akcigit et al. (2023). subsidies, due to data limitations. Energy taxes and 58. Bachas, Fattal Jaef, and Jensen (2019). subsidies are estimated based on the “price gap” 59. Bachas, Fattal Jaef, and Jensen (2019). between retail prices and supply costs for a particular 60. López and Torres (2020). energy carrier, used in a specific sector in a jurisdiction 61. Melitz (2003). in a given year. The net energy taxes and subsidies are 62. Didier and Cusolito (2024). then aggregated across sectors, fuels, and countries to 63. McGowan and Andrews (2016). yield a global value, as illustrated in figure 8.6. More 64. Schumpeter (1942). details on this methodology are discussed in Agnolucci, 65. McKenzie et al. (2023); Scur et al. (2021). Gencer, and Heine (2024). 66. Bloom and Van Reenen (2010); McKenzie and 100. Agnolucci, Gencer, and Heine (2024). Woodruff (2017). 101. IEA (2023). 67. Bloom, Sadun, and Van Reenen (2016). 102. Gentilini et al. (2022). 68. McKenzie and Woodruff (2017). 103. Lin and Li (2012). 69. Queiró (2022). 104. Shahmoradi, Haqiqi, and Zahedi (2011). 70. McKenzie et al. (2023). 105. Becker and Fischer (2013); Cozzi (2012); Eberhard and 71. Back, Parboteeah, and Nam (2014); Bloom et al. (2013, Kåberger (2016). 2020); Bruhn, Karlan, and Schoar (2018); Giorcelli 106. Acemoglu, Kakhbod, and Ozdaglar (2017). References Abebe, Girum, Stefano Caria, Marcel Fafchamps, Paolo G²LM|LIC Working Paper 49 (March), Gender, Growth, Falco, Simon Franklin, Simon Quinn, and Forhad and  Labour Markets in Low-Income Countries Shilpi. 2021. “Matching Frictions and Distorted Programme, Institute of Labor Economics, Bonn, Beliefs: Evidence from a Job Fair Experiment.” Germany. 214 | WORLD DEVELOPMENT REPORT 2024 Acemoglu, Daron, Philippe Aghion, Leonardo Bursztyn, and Angrist, Noam, Elisabetta Aurino, Harry A. Patrinos, George David Hémous. 2012. “The Environment and Directed Psacharopoulos, Emiliana Vegas, Ralph Nordjo, and Technical Change.” American Economic Review 102 (1): Brad Wong. 2023. “Improving Learning in Low- and 131–66. Lower-Middle-Income Countries.” Journal of Benefit- Acemoglu, Daron, Ufuk Akcigit, Douglas Hanley, and William Cost Analysis 14 (S1): 55–80. Kerr. 2016. “Transition to Clean Technology.” Journal of Angrist, Noam, David K. Evans, Deon Filmer, Rachel Political Economy 124 (1): 52–104. Glennerster, F. Halsey Rogers, and Shwetlena Sabarwal. Acemoglu, Daron, Ali Kakhbod, and Asuman Ozdaglar. 2017. 2020. “How to Improve Education Outcomes Most “Competition in Electricity Markets with Renewable Efficiently? A Comparison of 150 Interventions Using Energy Sources.” Energy Journal 38 (1): 137–55. the New Learning-Adjusted Years of Schooling Metric.” Aghion, Philippe, Ufuk Akcigit, Ari Hyytinen, and Otto Policy Research Working Paper 9450, World Bank, Toivanen. 2023. “2022 Klein Lecture: Parental Education Washington, DC. and Invention: The Finnish Enigma.” International Arteaga, Felipe, Adam J. Kapor, Christopher A. Neilson, Economic Review 64 (2): 453–90. and Seth D. Zimmerman. 2022. “Smart Matching Agnolucci, Paolo, Carolyn Fischer, Dirk Heine, Mariza Platforms and Heterogeneous Beliefs in Centralized Montes de Oca Leon, Joseph Dixon Callisto Pryor, School Choice.” Quarterly Journal of Economics Kathleen Patroni, and Stéphane Hallegatte. 2023. 137 (3): 1791–848. “Measuring Total Carbon Pricing.” Policy Research Bachas, Pierre, Roberto N. Fattal Jaef, and Anders Working Paper 10486, World Bank, Washington, DC. Jensen. 2019. “Size-Dependent Tax Enforcement Agnolucci, Paolo, Defne Gencer, and Dirk Heine. 2024. and Compliance: Global Evidence and Aggregate “Total Carbon Pricing for Energy Consumption: The Implications.”  Journal of Development Economics  140 Importance of Energy Taxes and Subsidies.” ESMAP (September): 203–22. Technical Report, Energy Subsidy Reform in Action Back, Yujin, K. Praveen Parboteeah, and Dae-il Nam. Series, World Bank, Washington, DC. 2014. “Innovation in Emerging Markets: The Role of Aguirre, Josefa. 2021. “Long-Term Effects of Grants and Management Consulting Firms.” Journal of International Loans for Vocational Education.” Journal of Public Management 20 (4): 390–405. Economics 204 (December): 104539. Baeza, Cristóbal. 2020. “Female Role Models for the Aivazian, Varouj Aram, and Eric Santor. 2008. “Financial Scientists of the Future.” Gender Equality (blog), January Constraints and Investment: Assessing the Impact of 27, 2020. https://www.bbva.com/en/sustainability​ a World Bank Credit Program on Small and Medium /­female-role-models-for-the-scientists-of-the-future/. Enterprises in Sri Lanka.”  Canadian Journal of Bandiera, Oriana, Ahmed Elsayed, Anton Heil, and Economics 41 (2): 475–500. Andrea Smurra. 2022. “Presidential Address 2022: Akcigit, Ufuk, Yusuf Emre Akgündüz, Harun Alp, Seyit Mumin Economic Development and the Organisation of Cilasun, and Jose M. Quintero. 2023. “Good Firms, Bad Labour, Evidence from the Jobs of the World Project.” Policies: The Dynamics of Informality and Industry Journal of the European Economic Association 20 (6): Policy in Shaping Economic Growth.” University of 2226–70. Chicago Working Paper, University of Chicago, Chicago. Banerjee, Abhijit Vinayak, and Sandra Sequeira. 2023. Akcigit, Ufuk, Harun Alp, and Michael Peters. 2021. “Lack “Learning by Searching: Spatial Mismatches and of Selection and Limits to Delegation: Firm Dynamics Imperfect Information in Southern Labor Markets.” in Developing Countries.” American Economic Review Journal of Development Economics 164 (September): 111 (1): 231–75. 103111. Akcigit, Ufuk, Emin Dinlersoz, Jeremy Greenwood, Becker, Bastian, and Doris Fischer. 2013. “Promoting and Veronika Penciakova. 2022. “Synergizing Renewable Electricity Generation in Emerging Ventures.” Journal of Economic Dynamics and Control Economies.” Energy Policy 56 (May): 446–55. 143 (October): 104427. Bertoni, Fabio, Massimo G. Colombo, and Anita Quas. 2023. Aldrich, Howard E., and Martin Ruef. 2018. “Unicorns, “The Long-Term Effects of Loan Guarantees on SME Gazelles, and Other Distractions on the Way to Performance.”  Journal of Corporate Finance 80 (June): Understanding Real Entrepreneurship in America.” 102408. Academy of Management Perspectives 32 (4): 458–72. Bloom, Nicholas, Erik Brynjolfsson, Lucia Foster, Ron Alfonsi, Livia, Mary Namubiru, and Sara Spaziani. 2022. Jarmin, Megha Patnaik, Itay Saporta-Eksten, and John “Meet Your Future: Experimental Evidence on the Michael Van Reenen. 2019. “What Drives Differences in Labor Market Effects of Mentors.” Paper presented Management Practices?” American Economic Review at the virtual 17th Institute of Labor Economics and 109 (5): 1648–83. 4th Institute of Labor Economics/Center for Research Bloom, Nicholas, Benn Eifert, Aprajit Mahajan, David J. in Economics and Statistics Conference, Labor McKenzie, and John Roberts. 2013. “Does Management Market Policy Evaluation, October 6–7, 2022. https:// Matter? Evidence from India.” Quarterly Journal of conference.iza​ . org/conference _files/E VAL _ 2022​ Economics 128 (1): 1–51. /alfonsi ​_ l24410.pdf. Bloom, Nicholas, Aprajit Mahajan, David McKenzie, and Aminadav, Gur, and Elias Papaioannou. 2020. “Corporate John Roberts. 2020. “Do Management Interventions Control around the World.”  Journal of Finance  75 (3): Last? Evidence from India.” American Economic 1191–246. Journal: Applied Economics 12 (2): 198–219. Rewarding Merit | 215 Bloom, Nicholas, Raffaella Sadun, and John Michael Van Center for International Environment and Resource Reenen. 2016. “Management as a Technology?” NBER Policy, Tufts University, Medford, MA. Working Paper 22327 (June), National Bureau of Cusolito, Ana Paula, and William F. Maloney. 2018. Economic Research, Cambridge, MA. Productivity Revisited: Shifting Paradigms in Analysis Bloom, Nicholas, and John Michael Van Reenen. 2010. and Policy. Washington, DC: World Bank. “Why Do Management Practices Differ across Firms and Didier, Tatiana, and Beulah Chelva. 2023. “Private Equity Countries?”  Journal of Economic Perspectives  24  (1): Markets in EMDEs.” World Bank, Washington, DC. 203–24. Didier, Tatiana, and Ana Paula Cusolito. 2024. Unleashing Bond, Michael, Heba Maram, Asmaa Soliman, and Riham Productivity through Firm Financing. Washington, DC: Khattab. 2012. Science and Innovation in Egypt. London: World Bank. Royal Society. https://royalsociety.org/~/media/policy​ Dinarte-Díaz, Lelys, María Marta Ferreyra, Sergio Urzúa, and /­projects/atlas-islamic-world/atlas-egypt-report.pdf. Marina Bassi. 2023. “What Makes a Program Good? Bruhn, Miriam, Dean S. Karlan, and Antoinette Schoar. Evidence from Short-Cycle Higher Education Programs 2018. “The Impact of Consulting Services on Small and in Five Developing Countries.” World Development Medium Enterprises: Evidence from a Randomized Trial 169 (September): 106294. in Mexico.” Journal of Political Economy 126 (2): 635–87. Dixit, Siddharth, and Indermit S. Gill. 2024. “AI, the New Bueno, Rachel. 2009. “The Inova Success Story: Technology Wingman of Development.” Background paper pre - Transfer in Brazil.” WIPO Magazine 6 (December): 24–25. pared for World Development Report 2024, World Bank, Bundesnetzagentur. 2021. “Fines for Manipulation of Washington, DC. Wholesale Energy Market.” Press Release, October Drori, Gili S., and Avida Netivi. 2013. “STEM in Israel: The 5 ,   2 0 2 1.   h t t p s : // w w w. b u n d e s n e t z a g e n t u r. d e​ Educational Foundation of ‘Start-Up Nation.’” Securing /­S haredDocs​/Pressemitteilungen/EN/2021/20211005​ Australia’s Future Series, Australian Council of Learned _ ­BussgeldMarktmanipulation.html. Academies, Melbourne. Calvo-González, Oscar. 2021. Unexpected Prosperity: How Eberhard, Anton, and Tomas Kåberger. 2016. “Renewable Spain Escaped the Middle Income Trap. Oxford, UK: Energy Auctions in South Africa Outshine Feed-In Oxford University Press. Tariffs.” Energy Science and Engineering 4 (3): Card, David E., and Laura Giuliano. 2016. “Can Tracking Raise 190–93. the Test Scores of High-Ability Minority Students?” Education Commission. 2016. The Learning Generation: American Economic Review 106 (10): 2783–816. Investing in Education for a Changing World. New Carranza, Eliana, Robert Garlick, Kate Orkin, and Neil Rankin. York: International Commission on Financing Global 2022. “Job Search and Hiring with Limited Information Education Opportunity. about Workseekers’ Skills.” American Economic Review Eslava, Marcela, and John C. Haltiwanger. 2020. 112 (11): 3547–83. “The  ­ Life-Cycle Growth of Plants: The Role of Carranza, Eliana, and David J. McKenzie. 2023. “Job Training Productivity, Demand and Wedges.” NBER Working and Job Search Assistance Policies in Developing Paper 27184 (May), National Bureau of Economic Countries.” Policy Research Working Paper 10576, Research, Cambridge, MA. World Bank, Washington, DC. Ferreira Filho, Joaquim Bento De Souza, and Marek Cedefop (European Centre for the Development of Hanusch. 2022. “A Macroeconomic Perspective of Vocational Training). 2023. “Skills Anticipation in Structural Deforestation in Brazil’s Legal Amazon.” Denmark (2022 Update).” Data Insights (March 21), Policy Research Working Paper 10162,  World Bank, Skills Anticipation in Countries Series, Cedefop, Washington, DC. Thessaloniki, Greece. https://www.cedefop.europa.eu​ Ferreyra, María Marta, Lelys Dinarte Díaz, Sergio Urzúa, /en/data-insights/skills-anticipation-denmark. and Marina Bassi. 2021. The Fast Track to New Chetty, Raj, John N. Friedman, Emmanuel Saez, Nicholas Skills: Short-Cycle Higher Education Programs in Turner, and Danny Yagan. 2020. “Income Segregation Latin  America and the Caribbean. Washington, DC: and Intergenerational Mobility across Colleges in the World Bank. United States.” Quarterly Journal of Economics 135 (3): Fontaine, Arturo, and Sergio Urzúa. 2018. Educación con 1567–633. patines. Santiago, Chile: Ediciones El Mercurio. Chetty, Raj, Matthew O. Jackson, Theresa Kuchler, Johannes Gallegos, Sebastian, Andres Barrios-Fernández, and Stroebel, Nathaniel Hendren, Robert B. Fluegge, Sara Christopher A. Neilson. 2024. “Closing Gaps in Higher Gong, et al. 2022a. “Social Capital I: Measurement Education Opportunities for Underrepresented Students and Associations with Economic Mobility.” Nature 608 over the Past Five Decades in Chile.” Background paper (7921): 108–21. prepared for World Development Report 2024, World Chetty, Raj, Matthew O. Jackson, Theresa Kuchler, Johannes Bank, Washington, DC. Stroebel, Nathaniel Hendren, Robert B. Fluegge, Sara Garcia, Cristobal, and Edward Crawley. 2024. “Accelerating Gong, et al. 2022b. “Social Capital II: Determinants of Innovation for Development: How Universities Leverage Economic Connectedness.” Nature 608 (7921): 122–34. Knowledge Exchange to Boost Economic Growth in Cozzi, Paolo. 2012. “Assessing Reverse Auctions as a Middle-Income Countries.” Background paper ­ prepared Policy Tool for Renewable Energy Deployment.” Report for World Development Report 2024, World Bank, 007 (May), Energy, Climate, and Innovation Program, Washington, DC. 216 | WORLD DEVELOPMENT REPORT 2024 GEEAP (Global Education Evidence Advisory Panel). Hsieh, Chang-Tai, and Benjamin A. Olken. 2014. “The 2023. “Cost-Effective Approaches to Improve Global Missing ‘Missing Middle’.” Journal of Economic Learning:  What Does Recent Evidence Tell Us Are Perspectives 28 (3): 89–108. ‘Smart Buys’ for Improving Learning in Low- and Hvide, Hans K., and Benjamin F. Jones. 2018. “University Middle-Income Countries?” May, GEEAP, World Bank, Innovation and the Professor’s Privilege.” American Washington, DC. https://documents1.­worldbank​.org​ Economic Review 108 (7): 1860–98. /­curated/en/099420106132331608/pdf​/­IDU0977f73d7 Iacovone, Leonardo, William F. Maloney, and David J. 022b1047770980c0c5a14598eef8.pdf. McKenzie. 2022. “Improving Management with Gentilini, Ugo, Mohamed Almenfi, Hrishikesh T. M. M. Individual and Group-Based Consulting: Results from Iyengar, Yuko Okamura, John Austin Downes, Pamela a Randomized Experiment in Colombia.”  Review of Dale, Michael Weber, et al. 2022. Social Protection and Economic Studies 89 (1): 346–71. Jobs Responses to COVID-19: A Real-Time Review of IEA (International Energy Agency). 2023. “Fossil Fuels Country Measures. Living Paper Version 16 (February 2, Consumption Subsidies 2022.” February, IEA, Paris. 2022). Washington, DC: World Bank. http://hdl.handle​ https://www.iea.org/reports/fossil-fuels​- consumption​ .net/10986/37186. -subsidies-2022. Giorcelli, Michela. 2019. “The Long-Term Effects of Jakubowski, Maciej, Harry Anthony Patrinos, Emilio Ernesto Management and Technology Transfers.” American Porta, and Jerzy Wiśniewski. 2016. “The Effects of Economic Review 109 (1): 121–52. Delaying Tracking in Secondary School: Evidence Glaeser, Edward L., and Andrei Shleifer. 2002. “Legal Origins.” from the 1999 Education Reform in Poland.” Education Quarterly Journal of Economics 117 (4): 1193–229. Economics 24 (6): 557–72. Gonzalez-Uribe, Juanita, and Ouafaa Hmaddi. 2022. “The Karlan, Dean S., Ryan Knight, and Christopher R. Udry. Multi-Dimensional Impacts of Business Accelerators: 2015. “Consulting and Capital Experiments with What Does the Research Tell Us?” June, London School Microenterprise Tailors in Ghana.” Journal of Economic of Economics and Political Science, London. Behavior and Organization 118 (October): 281–302. Gorgodze, Sophia, and Lela Chakhaia. 2021. “The Uses Kelley, Erin Munro, Christopher Ksoll, and Jeremy Magruder. and Misuses of Centralised High Stakes Examinations: 2023. “How Do Online Job Portals Affect Employment Assessment Policy and Practice in Georgia.” and Job Search? Evidence from India.” Working paper, Assessment in Education: Principles, Policy, and Department of Economics, University of California– Practice 28 (3): 322–42. Berkeley, Berkeley, CA. Graham, Carol L. 2013. “Pursuing Happiness: Social Kersten, Renate, Job Harms, Kellie Liket, and Karen Maas. Mobility and Well-Being.” Commentary (blog), October 2017. “Small Firms, Large Impact? A Systematic Review 24, 2013. https://www.brookings.edu/articles/­ pursuing​ of the SME Finance Literature.”  World Development 97 -happiness-social-mobility-and-well-being/. (September): 330–48. Groh, Matthew, David J. McKenzie, Nour Shammout, and Khanna, Tarun. 2023. “Science-Based Entrepreneurship in Tara Vishwanath. 2015. “Testing the Importance of India: A Policy Glass (as Yet) Quarter-Full.” India Policy Search Frictions and Matching through a Randomized Forum 2022, vol. 19, edited by Poonam Gupta, Pravin Experiment in Jordan.” IZA Journal of Labor Economics Krishna, and Karthik Muralidharan, 1–53. New Delhi: 4 (1): 7. Academic Foundation. Grover, Arti, Denis Medvedev, and Ellen Olafsen. 2019. Khanna, Tarun, and Michael Szonyi, eds. 2022. Making High-Growth Firms: Facts, Fiction, and Policy Options for Meritocracy: Lessons from China and India, from Antiquity Emerging Economies. Washington, DC: World Bank. to the Present. New York: Oxford University Press. Guyon, Nina, Eric Maurin, and Sandra McNally. 2012. Kim, Jin-Hyuk, and Liad Wagman. 2014. “Portfolio Size “The Effect of Tracking Students by Ability into Different and Information Disclosure: An Analysis of Startup Schools.” Journal of Human Resources 47 (3): 684–721. Accelerators.”  Journal of Corporate Finance 29 Hanusch, Marek. 2023. A Balancing Act for Brazil’s (December): 520–34. Amazonian States: An Economic Memorandum. Kitchen, Joseph A., Gerhard Sonnert, and Philip M. Sadler. Washington, DC: World Bank. 2018. “The Impact of College- and University-Run Hathaway, Ian. 2016. “What Startup Accelerators Really High  School Summer Programs on Students’ End Do.” Entrepreneurship (blog), March 1, 2016. https://hbr​ of High School STEM Career Aspirations.” Science .org/2016/03/what-startup-accelerators-really-do. Education 102 (3): 529–47. Higuchi, Yuki, Vu Hoang Nam, and Tetsushi Sonobe. 2015. Kitchen, Joseph Allen, Gerhard Sonnert, and Philip “Sustained Impacts of Kaizen Training.”  Journal of Sadler. 2020. “Campus Visits: Impact of a College Economic Behavior and Organization  120 (December): Outreach Strategy on Student STEM Aspirations.” 189–206. Journal of Student Affairs Research and Practice Hopenhayn, Hugo A. 1992. “Entry, Exit, and Firm 57 (3): 266–81. Dynamics in Long-Run Equilibrium.” Econometrica 60 Lach, Saul, and Mark Schankerman. 2008. “Incentives and (5): 1127–50. Invention in Universities.” RAND Journal of Economics Hsieh, Chang-Tai, and Peter J. Klenow. 2009. “Misallocation 39 (2): 403–33. and Manufacturing TFP in China and India.” Quarterly La Hora. 2022. “Hay una desconexión entre lo que las Journal of Economics 124 (November): 1403–48. empresas necesitan y lo que se está formando en el Rewarding Merit | 217 sistema educativo” [There’s a disconnect between what “Training Entrepreneurs.” VoxDevLit  1 (3). https:// enterprises need and what the education system pro- voxdev.org/sites/default /files/2023 -11​/ ­Training​ vides]. País, November 29, 2022. https://www.lahora​ _ ­Entrepreneurs_Issue_3.pdf. .com.ec/pais/hay-una-desconexion-entre-lo-que-las​ Melitz, Marc J. 2003. “The Impact of Trade on Intra-Industry -empresas-necesitan-y-lo-que-se-esta-formando-en​ Reallocations and Aggregate Industry Productivity.” -el-sistema-educativo/. Econometrica 71 (6): 1695–725. Lall, Saurabh A., Li-Wei Chen, and Peter W. Roberts. 2020. Mendes, Philip. 2015. “Integrating Intellectual Property “Are We Accelerating Equity Investment into Impact- into  Innovation Policy Formulation in Sri Lanka.” Oriented Ventures?” World Development 131 (July): January, World Intellectual Property Organization, 104952. Geneva. La Porta, Rafael, Florencio López‐de‐Silanes, Andrei Shleifer, Miller, Amisha, Saurabh A.  Lall, Markus P. Goldstein, and and Robert W. Vishny. 1997. “Legal Determinants of Joao H. C. Montalvao. 2023. “Asking Better Questions: External Finance.” Journal of Finance 52 (3): 1131–50. The Effect of Changing Investment Organizations’ La Porta, Rafael, Florencio López‐de‐Silanes, Andrei Evaluation Practices on Gender Disparities in Shleifer, and Robert W. Vishny. 1998. “Law and Finance.” Funding Innovation.”  Policy Research Working Paper Journal of Political Economy 106 (6): 1113–55. 10625, World Bank, Washington, DC. La Porta, Rafael, Florencio López-de-Silanes, Andrei Muñoz-Boudet, Ana María, Lourdes Rodríguez-Chamussy, Shleifer, and Robert W. Vishny. 1999. “Investor Cristina Chiarella, and Isil Oral Savonitto. 2017. “Women Protection: Origins, Consequences, and Reform.” NBER and STEM in Europe and Central Asia.” World Bank, Working Paper 7428 (December), National Bureau of Washington, DC. Economic Research, Cambridge, MA. Murnane, Richard J., Marcus R. Waldman, John Lee, Ju-Ho, Hyeok Jeong, and Song-Chang Hong. 2014. “Is B. Willett,  Maria Soledad Bos, and Emiliana Korea Number One in Human Capital Accumulation? Vegas. 2017.  “The Consequences of Educational Education Bubble Formation and Its Labor Market Voucher  Reform in Chile.” NBER Working Paper Evidence.” KDI School Working Paper 14-03 (August), 23550 (June), National Bureau of Economic Research, KDI School of Public Policy and Management, Seoul, Cambridge, MA. Republic of Korea. Naudé, Wim. 2011. “Entrepreneurship Is Not a Binding Lenton, Timothy M., David I. Armstrong McKay, Sina Constraint on Growth and Development in the Poorest Loriani,  Jesse F. Abrams, Steven J. Lade, Jonathan Countries.” World Development 39 (1): 33–44. F. Donges, Manjana Milkoreit, et al. 2023.  The Global Neilson, Christopher A. 2024. “The Rise of Coordinated Tipping Points Report 2023. Exeter, UK: University of Choice and Assignment Systems in Education Exeter. Markets around the World.” Background paper pre - Lin, Boqiang, and Aijun Li. 2012. “Impacts of Removing pared for World Development Report 2024, World Bank, Fossil Fuel Subsidies on China: How Large and How to Washington, DC. Mitigate?” Energy 44 (1): 741–49. Nieminen, Mika, and Erkki Kaukonen. 2001. Universities López, José Joaquín, and Jesica Torres. 2020. and R&D Networking in a Knowledge-Based Economy: “Size-Dependent Policies, Talent Misallocation, and A Glance at Finnish Developments. Sitra Report 11. the Return to Skill.”  Review of Economic Dynamics  38 Helsinki: Sitra. (October): 59–93. NIHF (National Inventors Hall of Fame). 2022. “Behind the Mateina, Eleonora, and Anastasiya Grunova. 2020. “The NIHF Scenes: Camp Invention Is for Every Kind of Kid!” Bulgarian National Electricity Company Sanctioned for NIHF, North Canton, OH. https://www.invent.org/blog​ Abuse of Dominance on the Market of Balancing Energy /­behind-nihf-scenes/camp-invention-2022-all-kids. of Renewables.” Kluwer Competition Law Blog (blog), OECD (Organisation for Economic Co-operation and January 7, 2020. https://competitionlawblog.kluwer​ Development). 2021. “Man Enough? Measuring competitionlaw.com/2020/01/07/. Masculine Norms to Promote Women’s Empowerment.” McGowan, Müge Adalet, and Dan Andrews. 2016. “Insolvency Social Institutions and Gender Index Series, OECD Regimes and Productivity Growth: A Framework for Development Centre, Paris. Analysis.” OECD Economics Department Working Paper Pagano, Marco, and Paolo F. Volpin. 2005. “Managers, 1309, Organisation for Economic Co-operation and Workers, and Corporate Control.” Journal of Finance Development, Paris. 60 (2): 841–68. McKenzie, David J., and Dario Sansone. 2019. Paredes, Ricardo. 2024. “Duoc and the Higher Vocational “Predicting  Entrepreneurial Success Is Hard: Technical Education in Chile.” Background paper pre - Evidence from a Business Plan Competition in Nigeria.” pared for World Development Report 2024, World Bank, Journal of Development Economics 141 (November): Washington, DC. 102369. Pekkarinen, Tuomas, Roope Uusitalo, and Sari Pekkala Kerr. McKenzie, David J., and Christopher M. Woodruff. 2017. 2009. “School Tracking and Intergenerational Income “Business Practices in Small Firms in Developing Mobility: Evidence from the Finnish Comprehensive Countries.” Management Science 63 (9): 2967–81. School Reform.” Journal of Public Economics 93 (7–8): McKenzie, David J., Christopher M. Woodruff, Kjetil Bjorvatn, 965–73. Miriam Bruhn, Jing Cai, Juanita Gonzalez Uribe, Simon Peters, Marjolein, Kees Meijer, Etienne van Nuland, Quinn, Tetsushi Sonobe, and Martin Valdivia, eds. 2023. Thijs  Viertelhauzen, and Ekim Sincer. 2010. 218 | WORLD DEVELOPMENT REPORT 2024 Sector  Councils on Employment and Skills at EU Stern, Nicholas. 2022. “A Time for Action on Climate Change Level:  Country Reports. March. Rotterdam: ECORYS and a Time for Change in Economics.” Economic Journal Nederland BV. 132 (644): 1259–89. Plummer, Lawrence A., Thomas H. Allison, and Brian L. Thomas B. Fordham Institute. 2023. “Building a Wider, More Connelly. 2016. “Better Together? Signaling Interactions Diverse Pipeline of Advanced Learners: Final Report of in New Venture Pursuit of Initial External Capital.” the National Working Group on Advanced Education.” Academy of Management Journal 59 (5): 1585–604. June, Thomas B. Fordham Institute, Washington, DC. Queiró, Francisco. 2022. “Entrepreneurial Human Capital Tirri, Kirsi, and Elina Kuusisto. 2013. “How Finland Serves and Firm Dynamics.” Review of Economic Studies 89 (4): Gifted and Talented Pupils.” Journal for the Education of 2061–100. the Gifted 36 (1): 84–96. Roberts, Peter W., Genevieve Edens, Abigayle Davidson, Tsuruta, Daisuke. 2020. “SME Policies as a Barrier to Growth Edward Thomas, Cindy Chao, Kerri Heidkamp, and of SMEs.” Small Business Economics 54 (4): 1067–106. Jo-Hannah Yeo. 2017. “Accelerating Startups in Emerging Van Reenen, John Michael. 2021. “Innovation and Human Markets: Insights from 43 Programs.” May, Aspen Capital Policy.” NBER Working Paper 28713 (April), Network of Development Entrepreneurs, Washington, DC; National Bureau of Economic Research, Cambridge, MA. Social Enterprise at Goizueta, Goizueta Business School, Vegas, Emiliana. 2018. “5 Lessons from Recent Educational Emory University, Atlanta; Deloitte, London. Reforms in Chile.” Brookings Research (blog), March 20, Rutigliano, Alexandre, and Nikita Quarshie. 2021. “Policy 2018. https://www.brookings.edu/articles/5-lessons​ Approaches and Initiatives for the Inclusion of Gifted -from-recent-educational-reforms-in-chile/. Students in OECD Countries.” OECD Education Working Vu, Quang, and Tuyen Quang Tran. 2021. “Government Paper 262 (December 17), Organisation for Economic Financial Support and Firm Productivity in Co-operation and Development, Paris. Vietnam.” Finance Research Letters 40 (May): 101667. Saavedra-Chanduví, Jaime, Hanna Alasuutari, and Wall Street Journal. 2019. “Economists’ Statement on Marcela Gutiérrez Bernal. 2018. “Finland’s Education Carbon Dividends: Bipartisan Agreement on How to System: The Journey to Success.” Education Combat Climate Change.” Opinion (blog), January for Global Development (blog), August 14, 2018. 16, 2019. https://www.wsj.com/articles/economists​ https://blogs.worldbank.org/en/education/finland​ -­statement​- on-carbon-dividends-11547682910. -s-education-system-journey-success. World Bank. 2018. World Development Report 2018: Schady, Norbert R., Margarita Isaacs, and Diego Learning to Realize Education’s Promise. Washington, Parra. 2024. “50 Years of Educational Upgrading DC: World Bank. in the Largest Developing Countries.” World Bank, World Bank. 2020. “Promoting Innovation in China : Washington, DC. Lessons from International Good Practice.” Finance, Schumpeter, Joseph Alois. 1942. Capitalism, Socialism and Competitiveness, and Innovation Insight: Firms, Democracy. New York: Harper and Brothers. Entrepreneurship, and Innovation, World Bank, Scur, Daniela, Raffaella Sadun, John Michael Van Reenen, Washington, DC. Renata Lemos, and Nicholas Bloom. 2021. “The World World Bank. 2021. “Ireland’s Human Capital: The Management Survey at 18: Lessons and the Way Forward.” Contribution of Education and Skills Development to Oxford Review of Economic Policy 37 (2): 231–58. Economic Transformation.” August, Building Human Serajuddin, Umar, and Nada Hamadeh. 2020. “New World Capital: Lessons from Country Experiences Series, Bank Country Classifications by Income Level: 2020– World Bank, Washington, DC. 2021.” Data Blog (blog), July 1, 2020. https://blogs​ World Bank. 2023a. Country Climate and Development .worldbank.org/opendata/new-world-bank- ­c ountry​ Report: Brazil. Washington, DC: World Bank. -classifications-income-level-2020-2021. World Bank. 2023b. State and Trends of Carbon Pricing Shahmoradi, Asghar, Iman Haqiqi, and Raziyeh Zahedi. 2023. Washington, DC: World Bank. 2011. “Impact Analysis of Energy Price Reform and World Bank, UNESCO (United Nations Educational, Cash Subsidy Payment in Iran: CGE Approach.” Scientific, and Cultural Organization), and ILO Quarterly Journal of Economic Research and Policies (International Labour Organization). 2023. Building 19 (57): 5–30. Better Formal TVET Systems: Principles and Practice Shetty, Priya, Husein Akil, Trina Fizzanty, and Grace in Low- and Middle-Income Countries. Washington, DC: Simamora. 2014. Indonesia: The Atlas of Islamic World World Bank; Paris: UNESCO; Geneva: ILO. Science and Innovation, Country Case Study. London: Zorman, Rachel, Shlomit Rachmel, and Zipi Bashan. 2016. Royal Society. “The National Mentoring Program in Israel: Challenges Solis, Alex. 2017. “Credit Access and College Enrollment.” and Achievements.” Gifted Education International Journal of Political Economy 125 (2): 562–622. 32 (2): 173–84. Rewarding Merit | 219 9 Capitalizing on Crises Key messages • The climate and energy crises are providing an opportunity for middle-income countries to infuse global technologies domestically and become producers of green intermediates for global markets. By doing so, they will help reduce the cost of decarbonization worldwide. • To accelerate the diffusion of technologies, middle-income countries will need to improve their investment climate so that contestable markets will give domestic incumbents an incentive to upgrade their technology. Alongside these efforts, high-income countries will need to reduce the scope of industrial policies that protect domestic incumbents and prevent middle-income countries from accessing technologies and markets. • Middle-income countries should assess viable investment opportunities for renewable energy technologies and the cost of capital. • To reduce emissions intensity, middle-income countries will need to create a market for renewable technologies, take into account their own natural resource endowment, and ­ optimize the choice of technologies. Using crises to destroy Because middle-income countries will need to recalibrate their mix of investment, infusion, and outdated arrangements innovation, crises are, in a sense, a necessary evil To ensure it has the appropriate balance of because they weaken the forces of preservation investment, infusion, and innovation, a coun- that maintain the status quo. Today, the climate try must pursue the destruction of outdated crisis is one of the most pressing challenges facing arrangements—enterprises, jobs, technologies, not only the global economy, but also humanity private contracts, policies, and public institu- at large. tions. In many countries, however, the forces of To effectively tackle the climate crisis, destruction are weak during boom times. middle-income countries will need to direct 221 investment, infusion, and innovation toward to retrofit green.4 But there is a challenge—despite reducing greenhouse gas emissions, bringing the falling cost of green infrastructure and poten- them to net zero.1 In support of climate action, tial for high economic returns in middle-income more than 140 countries have set a net zero tar- countries, capital does not flow from high-income get, covering about 88 percent of global emis- to low- and middle-income countries to under- sions.2 This chapter explores what it will take take these infrastructure investments. for ­middle-income countries to capitalize on the Finally, where capital costs are high, low-­ climate crisis to overcome structural stasis and carbon energy may be unaffordable. Moreover, advance decarbonization both locally and globally. the intermittency of variable renewables and Low-carbon sources of energy such as wind high storage costs pose risks to energy security if and solar are technologically sophisticated and not managed well. Middle-income countries will benefit from increasing returns to scale in pro- need flexibility in choosing emissions-reducing duction and deployment. Thus middle-income strategies. countries have an opportunity to infuse global technologies domestically and become producers Globalizing decarbonization of green intermediates for global markets, thereby reducing the cost of decarbonization worldwide. Upgrade and compete through global To accelerate diffusion of such technologies, these low-carbon value chains countries will need to improve their investment Izmir, Türkiye, a vibrant city of nearly 4.5 million climate so that contestable markets give domestic people on the Aegean Sea, has been historically incumbents the incentive to upgrade. Alongside significant for more than 5,000 years. Its settle- these efforts, major Group of Twenty (G20) and ment goes back to the third millennium BCE. Organisation for Economic Co-operation and Now it is poised to become a significant producer Development (OECD) countries will need to and exporter of intermediate products in Europe’s reduce the scope of industrial policies that protect wind energy value chain. Thirteen factories pro- domestic incumbents and prevent middle-income duce towers, blades, gearboxes, and generators, countries from accessing technologies and mar- and 80 percent of the production from this value kets. Indeed, the globalization of protectionist chain is exported—primarily to Europe. Türkiye industrial policy poses the risk of slowing down is increasing its competence and competitive- the globalization of decarbonization. Global ness in the wind turbine value chains, as well as coordination of the use of green industrial policy developing capabilities in a broader range of tech- will ensure that industrial policy does not ­hinder nically sophisticated green products (so-called the climate transition of or penalize low- and green complex products). ­middle-income countries. As this example suggests, middle-income Middle-income countries could also diffuse countries could join global low-carbon energy low-carbon technologies at home by build- value chains by supplying intermediate prod- ing green to meet their significant demand for ucts, thereby reducing the cost of green energy infrastructure systems. For example, one-third faster than would be possible through national of projected urban growth will occur in large efforts alone. Moreover, through their participa- middle-income countries such as China, India, ­ tion, middle-income countries can infuse global and Nigeria by 2050, and three-quarters of the knowledge into their own industries and subse- world’s urban infrastructure that will exist in quently increase their “economic complexity.”5 30  years has not yet been built.3 By building An example is the solar photovoltaic (PV) green, investments in middle-income coun- industry (figure 9.1). From 2008 to 2020, the glo- tries can help reduce emissions at a lower cost balized market for PV modules saved PV installers than in high-­income countries that would have in China US$36 billion, those in the United States 222 | WORLD DEVELOPMENT REPORT 2024 Figure 9.1 Use of globalized value chains for solar panels results in faster learning and lower global prices a. China b. Germany c. United States 6,000 6,000 6,000 Learning rate: 32.3% Learning rate: 20.2% Learning rate: 26.4% 5,000 5,000 5,000 Price per kW (2020 US$) Price per kW (2020 US$) Price per kW (2020 US$) 4,000 4,000 4,000 3,000 3,000 3,000 2,000 2,000 2,000 1,000 1,000 1,000 0 0 0 06 08 10 12 14 16 18 20 06 08 10 12 14 16 18 20 06 08 10 12 14 16 18 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Modeled price, global market scenario Modeled price, national market scenario Historical module price Source: Helveston, He, and Davidson 2022. Note: The figure displays the estimated module prices for global versus national market scenarios for 2006–20. The dots represent historical module prices, and the two solid lines reflect the modeled prices using global (blue) versus national (orange) market scenarios. In each modeled curve, the learning rates are held constant by country, and the prices of silicon follow historical global trends. The global market scenario uses global capacities, and the national market scenario uses a weighted sum of national and global capacities that reflects a gradual transition to fully domestically supplied markets over a 10-year period. The shaded uncertainty bands represent 95 percent confidence intervals from the estimated learning models, which were computed using a simulation. kW = kilowatt. US$24 billion, and those in Germany US$7 billion Wright’s Law (learning curves): costs fall as a when compared with the cost of having domes- power function of cumulative deployment due tic manufacturers supply an increasing share of to the positive effects of learning by doing or installed capacity over a 10-year period.6 China increasing returns to scale in the production of produced about 80 percent of solar PV cells and technologies.8 By contrast, more complex tech- modules globally in 2023. From 2020 to 2023, nologies and those that require a greater level of China’s PV module prices plummeted to US$0.15 customization to local environments (type 3 tech- per watt, or to more than 60 percent lower than nologies in figure 9.2) tend to “learn” more slowly. the US price of US$0.40 cents per watt. It suc- Small “granular” technologies (type 1 and type 2 ceeded by infusing global technologies through technologies in figure 9.2) can diffuse much more start-ups backed by foreign finance, international rapidly. Middle-income countries can create value collaboration among researchers, and licensing by becoming suppliers in global value chains of and the mass production of technologies devel- type 1 and type 2 technologies. oped in foreign labs.7 To date, however, participation by most As highlighted in chapter 6, four technologies— ­ middle-income countries in value chains produc- solar panels, wind turbines, lithium-ion batteries, ing low-carbon energy technologies has been lim- and electrolyzers used for green hydrogen—follow ited. These value chains are highly concentrated, Capitalizing on Crises | 223 Figure 9.2 Middle-income countries can support global decarbonization by becoming global suppliers of “granular” (type 1 and type 2) energy technologies Standardized Mass-customized Customized Platform-based complex Standardized complex product systems Complex product product systems (for example, small systems Complex (for example, CCGT modular reactor nuclear (for example, nuclear Degree of design complexity power plants) power plants, carbon power plants, BECCS) capture and storage) Platform-based Complex-customized Mass-produced complex complex products products products (for example, wind (for example, biomass Design-intensive (for example, electric turbines concentrating power plants, vehicles) solar power) geothermal power) Mass-produced products Mass-customized products Small-batch products (for example, solar (for example, rooftop (for example, building Simple PV modules, LEDs) solar PV) envelope retrofits) Need for customization Technology type: Type 1 Type 2 Type 3 Source: Malhotra and Schmidt 2020. Note: Type 1 technologies, such as solar photovoltaic (PV) modules and efficient light emitting diode (LED) lighting, are simple to assemble and distribute at scale (although individual components may exhibit complexity) and have rapid learning rates and scale economies. Type 2 technologies, such as wind turbines and electric vehicles, involve relatively more complex designs and move toward scalability more slowly, although they include standardized components, have the potential for scale economies and increasing replication, and have learning cycles of a few years. Type 3 technologies, such as nuclear power, require extensive customization and involve a high degree of complexity. They are susceptible to cost overruns and have limited scope for rapid learning by doing. BECCS = bioenergy with carbon capture and storage; CCGT = combined-cycle gas turbine. as discussed in chapter 6. Many inputs into clean overall technological sophistication (as discussed energy value chains are also highly concentrated. in c­hapter  6). Middle-income countries tapping For example, the Democratic Republic of Congo into manufacturing opportunities for clean energy supplies 74 percent of cobalt, China 68 percent of technologies are typically those already competitive rare earth elements, and Indonesia 49 percent of in manufacturing or exporters of high-­ technology nickel. Australia accounts for 47 percent of lith- products. Research conducted for this Report 9 ium mining, and Chile for 24 percent. Processing shows that the share of online job postings for all of these minerals is also highly concentrated, with disruptive technologies is highly correlated with the China refining 90 percent of rare earth elements share of online job postings related to low-­ carbon and 65 percent and 74 percent, respectively, of technologies.10 This finding suggests that coun- lithium and cobalt (figure 9.3). tries experiencing a rapid diffusion of all emerging One pathway for middle-income coun- ­ disruptive technologies are also witnessing a rapid tries is to improve their firm capabilities and diffusion of low-carbon technologies. 224 | WORLD DEVELOPMENT REPORT 2024 Figure 9.3 Extraction and processing of critical minerals for the clean energy transition remain highly concentrated in certain countries a. Extraction Congo, Copper Chile Peru Dem. Rep. Russian Federation Nickel Indonesia Philippines Indonesia Minerals Cobalt Congo, Dem. Rep. Australia Lithium Australia Chile China United Rare earths China Australia States 0 10 20 30 40 50 60 70 80 90 100 Share of global extraction (%) Top three share, 2019 b. Processing Copper Chile Japan Russian Nickel Indonesia China Federation Minerals Cobalt China Finland Canada Lithium China Chile Argentina Rare earths China Malaysia Estonia 0 10 20 30 40 50 60 70 80 90 100 Share of global processing (%) Top three share, 2019 Source: IEA 2023a. In fact, low-carbon production has been higher export a wide range of green complex products in countries with greater overall competitiveness.11 include China, Bulgaria, India, Mexico, Türkiye, The green complexity index—which measures Serbia, Belarus, Thailand, Bosnia and Herzegovina, a country’s current ability to export green com- and Tunisia12 (figure 9.4, panel a). The green com- plex products competitively—reveals that middle-­ plexity potential index measures countries’ poten- income countries currently able to competitively tial to export green, technologically sophisticated Capitalizing on Crises | 225 Figure 9.4 Many middle-income countries have untapped potential to manufacture green products a. Middle-income countries with b. Middle-income countries with highest green complexity index highest green complexity potential index 3.5 4.5 Green complexity potential index 3.0 4.0 Green complexity index 3.5 2.5 3.0 2.0 2.5 1.5 2.0 1.5 1.0 1.0 0.5 0.5 0 0 Tu na Bu ina ia Tü co nd M a Se e Be ia d Th rus ze nd a Tü na e Bu ia Th aria ab a do . Be ia et s am In Rep Vi ru iy iy di si Ar bi ar rb d s i vi i la er ila rk rk ni ne ex Ch Ch er In In la la N lg lg go ai a t, S H yp Eg an ia sn Bo Source: WDR 2024 team based on Mealy and Teytelboym (2022). Note: Panel a shows the green complexity index (GCI), which is aimed at capturing the extent to which countries are able to competitively export green, technologically sophisticated products. A country is considered to be competitive in a product if its revealed comparative advantage for this product is greater than 1. Panel b shows the green complexity potential (GCP) index. This index aggregates the information contained in each country’s green adjacent possible (GAP)—which is aimed at identifying the green diversification opportunities for each country—into a single, comparable metric. The GCP index measures each country’s average relatedness to green complex products in which the country is not yet competitive. products in the future, and it reveals that many Resist uncoordinated and middle-income countries have untapped poten- protectionist industrial policy tial to export green complex products. China, In the climate context, there is a strong case Türkiye, India, Bulgaria, and Thailand have the highest untapped potential based on other prod- economically for countries to use industrial poli- ucts they are currently manufacturing (figure 9.4, cies to support far-from-market low-carbon tech- panel b). Türkiye, for example, has high potential nologies. Furthermore, countries are faced with in the value chains for wind turbines and electric legitimate concerns about energy security and vehicles. Although production is growing rapidly, the need for politically feasible climate policies, these industries could become even greater driv- which warrant certain unique considerations in ers of growth. making energy-related trade policy. Nevertheless, Only if middle-income countries are able to to ensure a successful low-carbon transition, join low-carbon value chains can they effectively it  is  essential to implement well-coordinated, contribute to global decarbonization and realis- ­balanced industrial policies that avoid protection- tically aspire to tilting their own markets green. ism and minimize market distortions. 226 | WORLD DEVELOPMENT REPORT 2024 Research using the New Industrial Policy spreading across low- and middle-income coun- Observer (NIPO) reveals the significant imple- tries. Many middle-income economies are design- mentation of new industrial policy measures ing and deploying state-led projects to foster green in 2023, particularly by high-income and large industrialization. Notably, the Arab Republic of middle-income countries (figure 9.5, panel a).13 Egypt, Kenya, Morocco, Namibia, South Africa, For industrial policies targeting low-carbon and Tunisia have launched initiatives to support technologies specifically, a similar pattern holds the development of green hydrogen. This shift (figure 9.5, panel b). Policies in the NIPO classi- introduces significant challenges, often impos- fied as potentially trade distorting and that target ing a “development tax” whether or not countries low-carbon technologies have been spearheaded engage in industrial policies. For example, policy by a few large G20 countries. These include the response options for International Development European Union’s Green Deal Industrial Plan, the Association and other low-income countries are US Inflation Reduction Act (IRA), Japan’s Green typically limited to nontariff measures (NTMs) Growth Strategy, and the Korean New Deal. due to financial constraints and World Trade However, green industrial policies are rapidly Organization (WTO) commitments. Subsidizing Figure 9.5 All industrial policy implementation and green industrial policy implementation are correlated with GDP per capita a. All industrial policies b. Green industrial policies 6 5 USA USA Number of measures, 2023 (log) Number of measures, 2023 (log) 5 CAN BRA ITA 4 CAN IND DEU ITA JPN AUS AUS BRA KOR DEU GBR IND CHN GBR CHN RUS KOR 4 JPN ESP FRA ARG TUR 3 ESP CHE ARG FRA POL NLD CHE 3 MEX RUS NLD AUT TUR IDN SAU DNK PRT AUT THA PAK ZAF CZE SWE BEL 2 MEX 2 GRC IDN POL FIN ZAF GRC HUN BEL VNM MYS HUN NZL MYS CZE DNK UKR THA NPL HRV SWE SGP 1 HRV SVN IRL SGP 1 EGY PRT SAU BOL BGD EGY HKG VNM UKR SVN FIN IRL 8 9 10 11 12 8 9 10 11 12 GDP per capita, 2021 (log) GDP per capita, 2021 (log) High-income EMDEs Fitted values Sources: Panel a: Barattieri, Mattoo, and Taglioni 2024. Panel b: WDR 2024 team analysis replicating Barattieri, Mattoo, and Taglioni (2024) for green technologies only. Note: The vertical axis in both panels is the log of numbers of all potentially trade distortive measures from the New Industrial Policy Observer (NIPO). The horizontal axis in both panels is gross domestic product (GDP) per capita in 2021 based on WDI (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank​ .org/search​ dataset/. Panel b includes all NIPO policies classified as targeting low-carbon technologies, hydrogen, or critical /­ minerals, according to NIPO definitions. EMDEs = emerging market and developing economies. For country abbreviations, see International Organization for Standardization (ISO), https://www.iso.org/obp/ui/#search. Capitalizing on Crises | 227 their own industries can divert funds from essen- to 2 degrees Celsius, low- and middle-income tial public services, and NTMs can strain econo- countries will have to undertake investments of mies, potentially leading to impoverishing growth 4.5 percent of their gross domestic product (GDP) and a global race to the bottom. each year.15 Additional investments will be needed Such policies may also risk slowing down to provide safe water and sanitation, as well as knowledge transfer and the diffusion of low-­ reliable electricity and transport to meet the ris- carbon technologies. Policy makers in the world’s ing demand based on growing incomes. major economies must therefore coordinate on Many of these investments can provide dou- the appropriate use of green industrial policies. ble dividends by enhancing living standards and Indeed, they should consider that the energy mitigating greenhouse gas (GHG) emissions. transition has outcomes that extend beyond its For example, city developers that integrate land effects on climate, but also on technology diffu- use and transport plans can enhance economic sion and the economic development of middle-­ productivity while reducing GHG emissions.16 income countries. Designing buildings with emissions and energy To support middle-income countries as they savings in mind is likely to be more cost-­ efficient move toward global decarbonization, policy mak- than retrofitting, the dominant practice in ers will need to update global trade policy rules high-income countries.17 Similarly, scaling up to clearly define the appropriate use of green sub- investment in energy production and distribu- sidies, export controls, and import controls. Such tion, as well as transportation systems in low- and reforms are already being discussed and designed middle-income countries, can provide important at the WTO level. The Villars Framework 2.0 benefits in structural transformation and eco- has also detailed various potential reforms of nomic productivity.18 the WTO to bring the international trading sys- Considering the development and decarbon- tem into harmony with a global “commitment ization potential of these investments, invest- to a sustainable future.” These reforms include, ment opportunities in middle-income countries among others, carefully distinguishing between appear to be untapped. By leveraging multilateral subsidies that benefit sustainability and subsi- resources, private capital in high-income coun- dies that impair it. Other possibilities include tries could alleviate the shortage of infrastructure modifying existing agreements with supplemen- in low- and middle-income countries, help achieve tary clauses in much the same way that Articles the SDGs, and contribute to economic growth, as 20  and  21 of the General Agreement on Tariffs suggested by the World Bank and International and Trade (GATT) were used to carve out excep- Monetary Fund, along with regional development tions.14 Such clauses could acknowledge that banks.19 countries may need to nurture emerging domestic Research conducted for this Report provides industries if they are to achieve a transition with the first set of estimates on “investment energy security, but the use of subsidies should potential” in renewable energy in middle- also be restricted to specific circumstances, such income countries by examining the expected as the commercialization of far-from-market rates of return and cost of capital.20 Capital low-carbon technologies. costs constitute the largest part of life cycle costs in renewable energy projects. In middle- income countries such as  Brazil  and India, the Expanding low-carbon cost of capital can even account for 50 percent infrastructure of the levelized cost of energy for solar PV.21 For To achieve the United Nations’ Sustainable fossil fuel–based power generation, fuel costs Development Goals (SDGs) related to infrastruc- and other operational costs make up the largest ture and stay on track to limit climate change proportion of costs. 228 | WORLD DEVELOPMENT REPORT 2024 To assess whether investing in a renewable ratio is greater than 1. The vertical axis in energy technology in a middle-income country figure  9.6 clears such a domestic hurdle, would be an efficient use of domestic and foreign so that to the right of the axis it is eco- savings, this study compares a middle-income nomically efficient to invest in low-carbon country’s social rate of return on infrastructure energy infrastructure. with its social rate of return22 on private capital • If a middle-income country’s infrastruc- and the social rate of return on foreign capital: ture is scarce relative to that of a wealth- ier ­ country, it becomes economically • If a middle-income country’s infrastruc- efficient for capital to move toward ture is scarce, its social rate of return on the ­ middle-income country. Above the infrastructure will exceed its social rate horizontal axis, the hurdle for foreign of return on private capital—that is, its ­investment is cleared. Figure 9.6 Countries must clear hurdles for both efficient domestic investment and foreign investment in renewable energy • Governments should • It is economically stop expropriating efficient for foreign savings and let capital to move to foreign savings Quadrant IV Quadrant I middle-income finance Fails domestic Clears both countries. infrastructure. hurdles. domestic and Clears foreign foreign hurdles. hurdles. Quadrant III Quadrant II Fails both Clears domestic • Either domestic and hurdles. • There may be a infrastructure foreign hurdles. need for is already over- Fails foreign concessional provided or hurdles. finance. • The investment climate is too poor. Source: Gardner and Henry 2023. Note: For a middle-income country, the dual-hurdle framework sorts each country-infrastructure observation into one of four quadrants according to whether it clears the hurdle for both efficient domestic investment and foreign investment. The framework allows evaluating whether investing in the infrastructure of a given developing country would be an efficient use of developing country (domestic) and developed country (foreign) savings. Capitalizing on Crises | 229 The unit cost of installing renewable energy countries in the study pass both hurdles, and country-specific rates of return to capital including Ecuador, Jordan, and Malaysia. are calculated using evidence from a new meta-­ • Quadrant II comprises countries in which analysis to derive the output elasticity of renew- the social rate of return on infrastructure able energy together with data from the World clears the domestic hurdle but falls Bank Private Participation in Infrastructure below the foreign hurdle. Although these (PPI) database23 as well as country estimates countries would benefit from additional from the International Energy Agency (IEA), the investment in renewable energy, it is not International Renewable Energy Agency (IRENA), efficient for foreign savings to finance it. the World Bank Electricity Planning Model These countries can tap into domestic (EPM), and renewable energy auctions.24 The fol- savings, and there may be a role for lowing insights then emerge from the framework concessional foreign financing due to the (figure 9.7): global co-benefits in GHG reductions. This category includes 12 of 31 middle-income • Quadrant I in figure 9.6 comprises coun- countries, including Brazil, Cambodia, tries in which the return on infrastruc- China, Colombia, and India. ture clears both the domestic and foreign • Quadrant III comprises countries in hurdles—that is, the minimum rate of which the social rate of return on renew- return that an investor needs to proceed able energy clears neither the domestic with a project. Five of 31 middle-income nor the foreign hurdle. Additional inves- tigation is warranted for these coun- Figure 9.7 In many middle-income tries. On the one hand, a country with a countries, it is economically efficient to vibrant private investment climate (and expand renewable energy therefore a high social rate of return on 50 renewable energy) may be well endowed with capital in renewable energy. Thus middle-income countries in data set) Share of middle-income countries 45 in each quadrant (as % of total the marginal benefit of installing another 40 unit is not an efficient use of either local 35 or foreign savings. On the other hand, 30 a country may have a poor investment 25 climate that leads to low social rates of 20 return on private investment, even as it 15 remains relatively undercapitalized in renewable energy. 10 5 Assess financial returns and cost of 0 Quadrant I Quadrant II Quadrant III capital for renewable energy (5 of 31 (12 of 31 (14 of 31 countries) countries) countries) For investments that add value in economic terms, investors—domestic or foreign—will Source: Lall and Vagliasindi 2024. want to know whether they can cover their cost Note: The sample includes 31 middle-income countries. of capital and secure sufficient revenue over an In quadrant I countries, the return on infrastructure clears both the domestic and the foreign hurdles. In quadrant II extended period to access finance on reason- countries, the economic rate of return on infrastructure able terms. Investors use the cost of capital to clears the domestic hurdle but falls below the foreign assess project risk. Costs vary among countries, hurdle. In quadrant III countries, the economic rate of return on renewable energy clears neither the domestic nor the with the spread often determining the compet- foreign hurdle. itiveness of renewable energy.25 Investment risk 230 | WORLD DEVELOPMENT REPORT 2024 also varies according to type of renewable tech- Figure 9.8 In low- and middle-income nology.26 Rotating equipment such as a wind countries, the cost of capital for renewables turbine is prone to more wear and tear than a is high photoelectric system such as solar PV, thereby raising operating costs and increasing uncer- 9 tainty about the costs of repairing the compo- 8 nents of wind turbines.27 Country-specific and 7 Cost of capital (%) technology-specific investment risks can vary 6 over time28 due to specific policy or cost changes 5 in a technology or economywide variations in 4 interest rates.29 Before entering long-term commitments, 3 investors consider the level of risk in a country, 2 such as macroeconomic stability and political 1 uncertainties.30 Where cash flow depends on pay- 0 ments from a state-owned enterprise, the credi- e e e e m m m m bility of the specific policy framework matters.31 co co co co -in in in in e- e- h- The cost of capital for low-carbon technologies w dl dl ig Lo id id H -m r-m affects the investment decisions of both finan- er pe cial institutions and private corporations. Data w Up Lo on the cost of capital in 45 countries using solar Source: IRENA 2023. PV or wind technologies reveal that the cost of Note: Data are for 2021 and 2022. capital in middle-income countries is twice that in high-income countries, averaging 3.8 p ­ ercent The trends in middle-income countries such as in high-income countries, but 7.2 percent in China are the opposite: low-carbon electric utilities upper-middle-income countries and more than have a higher cost of debt than high-carbon ones. 8.5 percent in lower-middle-income countries32 In other emerging markets, such as Latin America (figure 9.8).33 The high cost of capital has material and Asia, utilities focused on renewables are sub- implications for affordable energy. For example, ject to a higher cost of capital. And there is a clear for a representative solar PV project or onshore divide in the cost of capital between OECD and wind project, the total cost of electricity increases non-OECD countries, with average cost of debt in by 80 percent if the cost of capital is 10 percent 2021 of 3.9 percent for OECD countries compared rather than 2 percent.34 with 4.7 percent for non-OECD countries. The high cost of capital in many low- and Accelerating reductions in carbon emissions middle-income countries could increase the cost ­ will require reducing the cost of capital for of renewables by 50 percent or more. The Energy low-carbon technologies. Addressing technology Transition Risk and Cost of Capital Program of the risk, development risk, and pricing risk can help Oxford Sustainable Finance Group tracks the cost incentivize investors—utilities, banks, or other of capital across equities, syndicated loans, corpo- institutions—to invest in renewable projects. rate bonds, and accounting data, and it has revealed Derisking requires a whole-of-economy approach. a significant variation in trends across regions.35 It depends on licensing, policy stability, and social In European countries, the cost of capital for low-­ acceptance, along with technical, market, and carbon electric utilities is significantly lower than regulatory risks.36 Derisking not only makes the cost for high-carbon ones. In North America, renewable energy projects less expensive, but also the cost of debt and equity for low-carbon electric reduces the amount of public finance needed to utilities is comparable to that for high-carbon ones. support these projects. Capitalizing on Crises | 231 Decoupling economic growth Figure 9.9 Today’s upper-middle-income and emissions countries are more energy efficient than upper-middle-income countries in the past Middle-income countries aspire to grow their economies and achieve living standards closer to 1.2 those of high-income countries. But economic expansion in middle-income countries could lead Elasticity of energy intensity 1.0 to a significant surge in the demand for energy if efficiency is not improved.37 A country does not, 0.8 however, need to grow at the expense of steps to reduce emissions if it can reduce both the 0.6 amount of energy required to fuel its economy and the carbon emissions per unit of energy used 0.4 by the economy. Growth has become less energy- intensive (figure 9.9). Most notably, the amount 0.2 of energy needed to fuel the economy (per unit of GDP) is much lower for today’s upper-middle- 0 income countries than for upper-middle-income countries in the past. In other words, there is e e e 16 e m m m 20 m ) co co co 0– co evidence that leapfrogging is advancing over time -in in in 96 -in e- e- w (1 dle dl dl as middle-income countries move closer to the Lo id id id -m r-m r-m technology frontier. er pe pe w Up Up Economic growth is also accompanied by Lo structural and spatial transformations that help reduce the carbon emissions of an economy. The Source: WDR 2024 team calculations based on data from Fetter (2022). carbon emissions of the global economy—a com- Note: Elasticity of energy intensity refers to the energy bination of energy intensity (energy consumed used per unit of gross domestic product (GDP). Estimates per dollar of GDP) and carbon intensity (carbon of the elasticity of energy intensity in the long run are emissions per unit of energy)—have declined based on a fixed effect dynamic response lag model covering 136 countries over the period 1960–2017 (34,800 globally from about 0.69 million tons of car- observations), including log price index, two lags of end- bon dioxide (CO2) per billion dollars of GDP in user energy consumption, dummies for structural breaks 1980 to 0.46 million tons in 2018 (figure 9.10).38 in data series, and country-sector fixed effects. Values of the elasticity of energy intensity can be interpreted as In particular, China’s drop in carbon intensity follows: if GDP increases by 1 percent, energy consumption is globally material because China is the world’s increases only by 0.75 percent in upper-middle-income largest emitter.39 More broadly, the carbon emis- countries that achieved upper-middle-income status as of 2017, compared with 1.1 percent in upper-middle-income sions per unit of GDP over time of upper-​middle-​ countries that achieved upper-middle-income status income and lower-middle-income countries have within the period 1960–2016. declined, even when supply chain emissions are taken into account. A recent study cover- innovation and research into renewables. Those ing 137 countries also finds that countries with measures substantially compensated for the higher GDP per capita have lower energy inten- increase in these countries’ emissions arising sity. Furthermore, countries with a high level of from a growing economy (figure 9.11). Today, a energy intensity experience a stronger reduction combination of reductions in carbon emissions in energy intensity.40 and improvements in carbon efficiency are off- Past oil crises motivated high-income coun- setting the increase in emissions stemming from tries to lower energy intensity and to accelerate economic and population growth. By contrast, 232 | WORLD DEVELOPMENT REPORT 2024 Figure 9.10 Carbon emissions per unit of Although low-carbon and renewable energy GDP have been declining worldwide can improve energy security by reducing price volatility through lower exposure to fuel price shocks during supply disruptions, energy from Carbon emissions (million tons of CO2 per billion 2015 US$ of GDP) 3.5 these sources cannot be produced consistently 3.0 throughout the day. Because electricity systems 2.5 must always be balanced—that is, ensure that the 2.0 supply of electricity is meeting the demand at all 1.5 times—countries will have to consider a balanced mix of energy sources in which low-carbon and 1.0 renewable sources cannot provide 100 percent 0.5 of the supply and to incentivize energy-intensive 0 users to adopt energy-conserving technologies. What can these countries do to reduce energy 80 86 92 98 04 10 16 22 19 19 19 19 20 20 20 20 intensity and accelerate energy efficiency? China Europe India United States World How middle-income countries can Source: WDR 2024 team analysis based on adapting from and updating Pindyck (2021). Carbon emissions data reduce energy intensity are from Energy Institute (2023). GDP data are from WDI In 2022, middle-income countries were expe- (World Development Indicators) (Data Catalog), World Bank, Washington, DC, https://datacatalog.worldbank.org​ riencing energy intensity (energy consump- /­search/dataset/0037712. tion per unit of GDP) 2.5 times higher than Note: Carbon emissions, the product of energy intensity and that in high-­ income countries. The first set of energy efficiency, are measured in million metric tons of carbon dioxide (CO2) emissions per billion 2015 US dollars insights on middle-income countries’ adoption of GDP. GDP = gross domestic product. of energy-saving technologies is now available through the World Bank’s Firm-level Adoption of Technology (FAT) survey of seven countries— economic growth is the key driver of emissions Bangladesh, Brazil, Cambodia, Chile, Ethiopia, in middle-income countries, and those emis- Georgia, and India. sions  more than outweigh the reductions in Adoption of energy-saving technologies var- emissions from lowering energy intensity and ies significantly across the countries sampled for improving energy efficiency. The effect of eco- the FAT survey. For example, the primary green nomic growth on emissions is especially strong building certification, Leadership in Energy and in upper-­middle-income countries. According to Environmental Design (LEED), reflects the level projections from the Network for Greening the of technologies used in buildings.42 Research Financial System, if middle-income countries con- suggests that firms operating in green-certified tinue their current policies, there is an 83 ­ percent buildings use 8 percent less energy than those chance that even if today’s high-income countries in noncertified buildings.43 The adoption of the achieve their 2050 net zero goals, global emissions LEED certification varies significantly. The high- will exceed the remaining carbon budget required est adoption rates are in Brazil and Chile, where for limiting the change in global warming to less more than 20 percent of firms are LEED-certified. than 2 degrees Celsius by 2050.41 In Bangladesh, 7 percent of firms are certified, In deciding how to “decouple” emissions from whereas less than 3 percent of firms in Georgia a growing economy, middle-income countries and 1 percent in India are LEED-certified.44 will need to consider their country’s concerns Scale economies matter in the adoption of with energy security and access to reliable energy. energy-saving technologies. Technologies such Capitalizing on Crises | 233 Figure 9.11 High-income countries have succeeded in reducing overall emissions by curbing energy intensity a. Middle-income countries b. High-income countries Change in emissions (MtCO2) Change in emissions (MtCO2) 15 15 10 10 5 5 0 0 –5 –5 –10 –10 –15 –15 1 5 5 5 5 5 5 5 5 1 5 5 –2 –7 –8 –9 –0 –1 –2 –7 –8 –9 –0 –1 16 65 76 86 96 06 16 65 76 86 96 06 20 19 19 19 19 20 20 19 19 19 19 20 Population effect Economic growth effect Energy intensity effect Carbon intensity effect Overall change in emissions Source: WDR 2024 team analysis based on data from GCB (Global Carbon Budget) (data hub), Future Earth, Fort Collins, CO; University of Exeter, Exeter, UK, https://globalcarbonbudgetdata.org/#. Note: Decomposition according to the Kaya identity, which states that the overall change in emissions (blue line) is the sum of four factors: population; gross domestic product (GDP) per capita (economic growth); energy intensity (energy used per unit of GDP); and carbon intensity (emissions per unit of energy consumed). The analysis is based on a global sample of 182 middle- and high-income countries and a time horizon since 1965, which allows the analysis to incorporate major crises, including the oil price shocks in the 1970s. MtCO2 = million tonnes of carbon dioxide. One tonne is equal to 1,000 kilograms. as advanced end-of-pipe treatment entail large Such evidence indicates that market installation costs, making adoption more likely contestability, as well as opportunities for among larger firms. Competition also matters. value-adding firms to grow, is compatible with In Georgia, markets with a higher concentration adopting energy-­ saving technologies. As for have lower energy efficiency (concentration mea- renewable energy, firms’ adoption of energy sured as the average price markups of the top efficiency technologies has been hampered by the 25 percent of firms in the markup distribution high up-front costs of acquiring the technology, within their sector-size group and municipality). lack of access to accurate information about In addition, of firms with similar capital intensity, the technology and its costs and benefits, and entrants are more energy-efficient than incum- low returns from early adoption of technologies bents.  In Argentina, for example, firms with a that require a large network of users.48 Policies higher share of skilled workers are better able to that improve the efficient allocation of resources adopt advanced green technologies.45 Exporters by subsidizing technology adoption, to the also tend to have lower emissions intensity than extent that environmental benefits are not fully nonexporters.46 And foreign-owned firms gener- captured by adopters, or ease the financing ally perform better on environmental standards, constraints that prevent technology adoption such as in Côte d’Ivoire, Mexico, and República are important in decoupling emissions from Bolivariana de Venezuela.47  economic growth. 234 | WORLD DEVELOPMENT REPORT 2024 There is reason to be optimistic that ongoing How middle-income countries can energy price increases will encourage firms and reduce emissions intensity other users to reduce energy intensity. In the long term, countries spend 5–9 percent of GDP Access to reliable, affordable energy is an import- on energy, and increases in energy prices tend to ant consideration for growing middle-income be fully compensated by higher efficiency.49 In the countries, which must choose the appropriate short term, however, price shocks may pose a cost mix of energy sources that reduce emissions to an economy, so complementary policies may be while ensuring stable energy supplies. Moving needed to compensate for economic and welfare across energy sources is a slow process. Despite impacts. Governments can help provide firms commitments by high-income and low- and that use energy intensively with access to energy- middle-income countries, fossil fuels continue conserving technologies and finance for adoption to account for more than 75 percent of global as part of a reform program. Proactive programs energy consumption (figure 9.12). Although that offer a mix of information, finance, and there has been significant growth in renewable support can encourage changes in production energy in recent years—particularly wind, solar, technology and investment in ­ energy-saving and hydroelectric power—the relative contribu- equipment. tions are still small. Figure 9.12 The world is slowly transitioning away from fossil fuels 180,000 Other renewables Global primary energy consumption, by source (TWh) Modern biofuels Solar 160,000 Wind Hydropower Nuclear 140,000 120,000 Natural gas 100,000 80,000 Oil 60,000 40,000 Coal 20,000 Traditional biomass 0 1800 1850 1900 1950 2000 2022 Source: WDR 2024 team analysis of data from Our World in Data (dashboard), Global Change Data Lab and Oxford Martin Program on Global Development, University of Oxford, Oxford, UK, https://ourworldindata.org/; Statistical Review of World Energy (dashboard), BP, London, https://www.bp.com/en/global/corporate/energy-economics/statistical-review-of-world- energy.html. Note: TWh = terawatt-hour. One terawatt-hour is equal to 1 trillion watt-hours, or the amount of power generated by a 1-terawatt generator running for one hour. Capitalizing on Crises | 235 Create a market for technologies solar potential of a region—it is also influenced To hasten the expansion of today’s low- by temperature (the higher the temperature, the carbon technologies and support research and less efficient is the solar potential), the type of ter- development (R&D) to create new technologies rain, the extent and type of mountains and hills, in the future, should government policies in and so on. Taking these factors into account, the ­m iddle-income countries subsidize development solar potential of each location within a coun- of specific technologies? This Report has try can be computed—the so-called PVOUT, the highlighted that most middle-income countries ratio between energy obtained (in kilowatt-hours, would benefit from infusing global technologies, kWh) and installed power (in terms of kilowatt not innovating prematurely, before the peak, kWp). This ratio serves as an assessment of necessary complements are in place. To promote the efficiency and productivity of solar plants in energy efficiency, middle-income countries each region. need a cluster of technologies that affect energy To switch to a renewable energy source, a coun- production, storage, and consumption. try must have the potential to create an abundant Policy should be directed at creating markets supply of that energy. Meanwhile, countries with for low-carbon technologies. This would include a rich endowment of fossil fuels may find it more removing subsidies for fossil fuel–related technol- difficult to embrace an energy transition due to ogy (see chapter 8) so that low-carbon technologies the forces of inertia and the uneven playing field can compete on a level playing field. Furthermore, from fossil fuel subsidies. middle-income countries need complementary investments in transmission infrastructure, as Optimize the choice of technologies well as interoperability standards. As in China, Multiple sources of low-carbon technologies demand-side “pull” may be needed to create a offer a myriad of possibilities for alleviating the market for these technologies. Sectoral policies heavy reliance on fossil fuels. Among the options such as government feed-in tariff programs50 were are emerging and next-generation technologies, particularly significant in creating a market for including, on the supply side, carbon capture, renewable energy, first in Germany in the 1990s, utilization, and storage (CCUS), carbon capture followed by Italy, Spain, the United States, China, and storage (CCS), green hydrogen, blue hydro- and India by the 2010s. Notably, as technologies gen, and carbon removal technologies (DACCS have matured, feed-in tariffs have been replaced and BECCS),51 and, on the demand side, hydro- by more cost-efficient procurement methods, gen technologies in transport (hydrogen fuel such as auctions (for example, in Brazil, India, cell trucks, aircraft) and industry (green steel). South Africa, and, more recently, the Middle According to a comprehensive modeling exer- East and North Africa and Sub-Saharan Africa). cise focusing on Europe and Central Asia, these Auctions have achieved record lowest prices per new technologies have a critical role to play in unit of electricity. the energy transition.52 Their growth rates will need to replicate the rapid rates of solar and wind Consider resource endowment energy to stay in line with climate targets,53 as A country’s likelihood of switching from energy most net zero modeling of these new technologies sources related to fossil fuels depends heavily on indicates. But they are not on such a trajectory its resource endowment. For example, if solar now. radiation is not sufficiently powerful, the mar- Although the multiple sources of low-carbon ket value of producing such renewable energy technologies offer countries options for weaning would be low and would not justify the switch off carbon and other fossil fuels, infrastructure for an incumbent firm or energy company. But constraints related to generation, transmission, irradiance is not the full story in measuring the and storage, as well as limits on the availability 236 | WORLD DEVELOPMENT REPORT 2024 of renewable resources, present challenges. power through renewable sources requires not Renewable energy sources depend on geographic only investments in new transmission lines, but location, climate, and weather. Thus regions have also in renovating legacy lines to integrate them different needs for and availability of specific smoothly in new frameworks. As long as energy renewable sources. Moreover, connecting renew- storage options remain both economically and able sources to the electricity grid is a major issue technically limited, the appetite may increase for in terms of cost and efficiency. Currently, there other technologies to maintain baseload energy is a bottleneck in transmission lines because the supplies. Examples are natural gas and hydro existing transmission grids are supporting con- technologies, as well as other technologies at dif- ventional energy resources. Legacy power grids ferent levels of maturity (ranging from geother- that have been in place for several decades occupy mal to green hydrogen), along with fuels such as the routes and land needed to set up additional nuclear power. Box 9.1 discusses the role of alter- power lines supporting renewable sources. nate technologies as system stabilizers. Middle-income countries may be able to par- In view of the high cost of capital, the risk of tially leapfrog legacy grids located near large fos- intermittency, and the high storage cost for low sil fuel power plants in urban areas. But it may be carbon energy, middle-income countries will need difficult to leapfrog to a grid that is entirely pow- flexibility in managing how they go about reduc- ered by wind and solar plus energy storage—at ing their own emissions, while expanding energy least not while also expanding electricity access access and maintaining security. Furthermore, to and driving economic growth. Even in advanced support global decarbonization, they will need to countries such as the United States, transmis- ensure coordination on the use of green industrial sion lines need to increase by at least 25 percent policy so that it does not limit their participation over the next decade. Expanding distribution for in low-carbon value chains. Box 9.1 Technologies that can act as “stabilizers” of energy supply Different technologies can play different roles as stabilizers of energy supply. Hydropower. In addition to being a low-carbon source of energy, hydropower is a dis- patchable and flexible technology. Hydropower with storage is currently one of the most cost-effective low-carbon solutions for integrating large-scale variable renewable energy capacity. For example, 1 megawatt of hydropower in Bhutan and Nepal can help inte- grate 5–6 megawatts of variable renewable energy in India and Bangladesh. There is an urgent need to scale up investment in rehabilitating and upgrading the existing hydro- power installed capacity as well as restoring the current reservoir storage capacity. By 2030, more than 20 percent of the global hydropower infrastructure is expected to be more than 55 years old, the age at which major electromechanical equipment needs to be rehabilitated and upgraded. The risk-return profile and long gestation period of large hydropower projects has resulted in the limited participation of private financing in hydro- power projects. Only 27 percent of the added hydropower installed capacity from 2011 to 2020 was privately owned—a rate that is even lower for large hydropower projects. Geothermal. Geothermal power is one of the cleanest energy resources. It is also associated with some of the lowest land use of any energy technology, including other (Box continues next page) Capitalizing on Crises | 237 Box 9.1 Technologies that can act as “stabilizers” of energy supply (continued) renewable sources. Although traditionally a baseload source of electricity, geothermal power can offer flexibility because plants can run continuously or adjust quickly to match demand and supply. Adopting appropriate pricing structures that recognize geothermal power’s up-front costs can increase the flexibility in generating and dispatching geother- mal power. Geothermal energy and heat pumps can also play a key role in enhancing the stability and flexibility of the grid, particularly with the rise of renewable energy sources. Hydrogen. Hydrogen has a role to play in storing energy and providing grid flexibility, as well as serving as a fuel in sectors in which carbon emissions are hard to abate. The cost of renewable hydrogen production depends on the cost of renewable power and the capital cost of equipment—notably, electrolyzersa—as well as on the financing cost repre- sented by the cost of capital. In today’s best locations and under optimistic assumptions, the production cost can be as low as US$3 per kilogram, although this level cannot serve as a benchmark for low- and middle-income countries due to lower-quality renewable resources and high capital costs of equipment. For example, the cost of an electrolyzer system varies significantly, from less than US$500 per kilowatt in China to as much as US$2,000 per kilowatt elsewhere.b Electrolyzer costs are expected to fall rapidly in the coming years, and so future projects will benefit from large-scale electrolyzers that will be cheaper. Although the current investment in hydrogen is significantly less than that in mature renewable energy technologies such as wind and solar, hydrogen technology has seen strong inflows of early stage capital as well as high levels of national funding in recent years in Europe and the United States. Regions with abundant solar and wind endow- ments, especially in Africa, can provide cheap green hydrogen for both domestic use and export. Demand for green hydrogen can also reciprocally boost investments in renewables and provide a salient business case for investment in renewables while facilitating electri- fication in some middle-income countries. Natural gas. The use of natural gas reduces emissions if it displaces coal and if fugitive emissions are sufficiently low. Natural gas can empower industrial development as a chemi- cal feedstock, fertilizer component, direct energy source, and electricity provider. However, switching from coal to natural gas does not help to shift or avoid path-dependency. It can lead to a carbon lock-in—that is, a long-term reliance on the built fossil fuel infrastructure.c Wealth losses from stranded gas reserves could be significant. Estimates range from US$1.7 trillion to US$3.8 trillion based on climate targets that are aligned with the Paris Agreement on climate change.d This force for preservation is further exacerbated by the large share of government ownership of natural gas reserves—approximately 80 percent. As a result, governments may be reluctant to abandon these assets due to windfall profits and rents. Sources: ESMAP et al. 2023; Hansen 2022. a. An electrolyzer is  a device that  uses electricity to split water or other components into their con - stituent  elements. It is a critical technology for producing low-emission hydrogen from renewable electricity. b. ESMAP et al. (2023). c. Melekh, Grubb, and Dixon (2024). d. Hansen (2022). 238 | WORLD DEVELOPMENT REPORT 2024 Notes  1. Stern (2023). Pursuit of net zero is important for stabi - 30. Waissbein et al. (2013); WEF (2014). lizing global temperatures and keeping global surface 31. Egli, Steffen, and Schmidt 2018; Estache and Steichen temperature increases below 2 degrees Celsius. 2015; Lüthi and Wüstenhagen (2012).  2. See For a Livable Climate: Net-Zero Commitments 32. IRENA (2023). Must Be Backed by Credible Action (dashboard), 33. Estimates of the cost of capital are based on the cost United Nations, New York, https://www.un.org/en​ of debt and the cost of equity. The cost of debt is the /climatechange​/net-zero-coalition. cost to finance a loan for a renewable energy asset.  3. Dasgupta (2018). The cost of equity is the return on equity required by  4. Glennerster and Jayachandran (2023). the project developer (IRENA 2023).  5. Introduced by Hausmann et al. (2014), “economic com - 34. Iyer et al. (2015); Schmidt, Cancella, and Pereira (2016). plexity” refers to a measure of a society’s productive 35. Zhou et al. (2023). knowledge. Prosperous societies have the knowledge 36. Noothout et al. (2016). to make a larger variety of more complex products. 37. Kahn and Lall (2022). This surge in demand reflects a Hausmann et al. (2014) attempt to measure the amount combination of relatively high-income elasticities and of productive knowledge countries hold and how they modest price elasticities—see the meta-analysis in can move to accumulate more of it by making more Labandeira, Labeaga, and López-Otero (2017). This complex products.  6. Helveston, He, and Davidson (2022). combination implies that energy demand is likely to  7. Green (2019). continue growing, particularly for middle-income coun-  8. Way et al. (2022). tries, notwithstanding efforts to enhance energy  9. Mealy and Teytelboym (2022). efficiency. 10. Bloom et al. (2023). 38. Pindyck (2021). 11. Bettarelli et al. (2023). 39. IEA (2021). 12. Mealy and Teytelboym (2022). 40. Deichmann et al. (2019). 13. Barattieri, Mattoo, and Taglioni (2024). Any attempt to 41. See Scenarios Portal, Network for Greening the measure industrial policies is fraught with challenges, Financial System, Paris, https://www.ngfs.net/ngfs​ and this database reflects only one measure of indus - -scenarios-portal/. trial policies, which may overrepresent countries that 42. The LEED certification is based on a scored rating issue a relatively large quantity of legislative docu- mechanism that evaluates the environmental perfor- ments or those with greater regulatory transparency. mance of buildings, including location, sustainable 14. Jain et al. (2024). sites, water efficiency, energy efficiency and atmo - 15. As research conducted by Rozenberg and Fay (2019) sphere, material selection and resources used, indoor suggests. air quality, and integrative process. 16. Lall et al. (2023). 43. Qiu and Kahn (2019). 17. Glennerster and Jayachandran (2023). 44. Cirera, Lee, and Ding (2024). 18. Foster et al. (2023). 45. Albornoz et al. (2009). 19. Gardner and Henry 2023; World Bank (2015). 46. Holladay (2016); Richter and Schiersch (2017). 20. Lall and Vagliasindi (2024). 47. Eskeland and Harrison (2003). 21. Schmidt, Cancella, and Pereira (2016). 48. Bryan and Williams (2021). 22. The social rate of return refers to the extrafinancial 49. Bashmakov (2007); Bashmakov et al. (2023). value of an investment (such as the value of environ- 50. A feed-in tariff is a policy tool that encourages the use mental or social outcomes). of renewable energy technologies by guaranteeing 23. See PPI (Private Participation in Infrastructure Database), World Bank, Washington, DC, http://ppi​ customers a set price for the electricity they generate. .worldbank.org/. 51. Carbon capture, utilization, and storage (CCUS) is an 24. Using parameters from Lowe, Papageorgiou, and advanced iteration of the traditional carbon capture Pérez-Sebastián (2019). and storage (CCS) technology. CCS focuses mainly on 25. Egli (2020); Mazzucato and Semieniuk (2018); Steffen the capture and sequestration of carbon dioxide to mit- and Waidelich (2022). igate emissions, whereas CCUS takes a step further by 26. Polzin et al. (2019); Salm (2018). finding practical applications for the captured carbon. 27. Steffen et al. (2020). DACCS is direct air capture with carbon storage. 28. Egli 2020; Mazzucato and Semieniuk (2018). BECCS is bioenergy with carbon capture and storage. 29. Egli, Steffen, and Schmidt (2018);  Kirkpatrick and 52. World Bank and ESMAP (2024). Bennear (2014). 53. IEA (2023b). Capitalizing on Crises | 239 References Albornoz, Facundo, Matthew A. Cole, Robert J. R. Elliott, and Development), GIF (Global Infrastructure Facility), Marco G. Ercolani. 2009. “In Search of Environmental and Hydrogen Council. 2023. “Scaling Hydrogen Spillovers.” World Economy 32 (1): 136–63. Financing for Development.” ESMAP Paper, World Bank, Barattieri, Alessandro, Aaditya Mattoo, and Daria Washington, DC. Taglioni. 2024. “Trade Effects of Industrial Policies: Estache, Antonio, and Anne-Sophie Steichen. 2015. “Is Are Preferential Agreements a Shield?” World Bank, Belgium Overshooting in Its Policy Support to Cut the Washington, DC. Cost of Capital of Renewable Sources of Energy?” Bashmakov, Igor. 2007. “Three Laws of Energy Transitions.” Reflets et Perspectives de la Vie Économique 54 (1–45): Energy Policy 35 (7): 3583–94. 33–45. Bashmakov, Igor, Michael Grubb, Paul Drummond, Robert Fetter, T. Robert. 2022. “Energy Transitions and Technology Lowe, Anna Myshak, and Ben Hinder. 2023. “‘Minus 1’ and Change: ‘Leapfrogging’ Reconsidered.” Resource and Energy Costs Constants: Empirical Evidence, Theory and Energy Economics 70: 101327. Policy Implications.” SSRN Preprint (March 30), Social Foster, Vivien, Nisan Gorgulu, Dhruv Jain, Stéphane Science Research Network, Rochester, NY. https://papers​ Straub, and Maria Vagliasindi. 2023. “The Impact of .ssrn.com/sol3/papers.cfm?abstract_id=4401851. Infrastructure on Development Outcomes: A Meta- Bettarelli, Luca, Davide Furceri, Pietro Pizzuto, and Analysis.” Policy Research Working Paper 10350, World Nadia Shakoor. 2023. “Environmental Policies and Bank, Washington, DC. Innovation in Renewable Energy.” IMF Working Paper Gardner, Camille, and Peter Blair Henry. 2023. “The Global WP/23/180 (September), International Monetary Fund, Infrastructure Gap: Potential, Perils, and a Framework Washington, DC. for Distinction.” Journal of Economic Literature 61 (4): Bloom, Nicholas, Tarek Alexander Hassan, Aakash Kalyani, 1318–58. Josh Lerner, and Ahmed Tahoun. 2023. “The Diffusion Glennerster, Rachel, and Seema Jayachandran. 2023. of New Technologies.” NBER Working Paper 28999 rev. “Think Globally, Act Globally: Opportunities to Mitigate (November), National Bureau of Economic Research, Greenhouse Gas Emissions in Low- and Middle-Income Cambridge, MA. Countries.” Journal of Economic Perspectives  37 (3): Bryan, Kevin A., and Heidi L. Williams. 2021. “Innovation: 111–36. Market Failures and Public Policies.” Chapter 13 in Green, Martin A. 2019. “How Did Solar Cells Get So Cheap?” Handbook of Industrial Organization, vol. 5, edited by Joule 3 (3): 631–33. Kate Ho, Ali Hortaçsu, and Alessandro Lizzeri, 281–388. Hansen, Tyler A. 2022. “Stranded Assets and Reduced Amsterdam: North-Holland. Profits: Analyzing the Economic Underpinnings of Cirera, Xavier, Kyung Min Lee, and Yuheng Ding. 2024. the Fossil Fuel Industry’s Resistance to Climate “Green Technology Adoption, Energy Efficiency, and Stabilization.” Renewable and Sustainable Energy Carbon Emissions in India: A Firm Perspective.” World Reviews 158 (April): 112144. Bank, Washington, DC. Hausmann, Ricardo, César A. Hidalgo, Sebastián Bustos, Dasgupta, Ani. 2018. “IPCC 1.5° Report: We Need to Build Michele Coscia, Alexander James Gaspar Simoes, and and Live Differently in Cities.” Commentary, Topic: Cities (blog), October 31, 2018. https://www.wri.org/insights​ Muhammed A. Yildirim. 2014. The Atlas of Economic /ipcc-15deg-report-we-need-build-and-live-differently​ Complexity: Mapping Paths to Prosperity. Cambridge, -cities. MA: MIT Press. Deichmann, Uwe, Anna Reuter, Sebastian Vollmer, and Helveston, John Paul, Gang He, and Michael R. Davidson. Fan Zhang. 2019. “The Relationship between Energy 2022. “Quantifying the Cost Savings of Global Solar Intensity and Economic Growth: New Evidence from Photovoltaic Supply Chains.” Nature 612 (7938): a Multi-Country Multi-Sectorial Dataset.” World 83–87. Development 124 (December): 104664. Holladay, J. Scott. 2016. “Exporters and the Environment.” Egli, Florian. 2020. “Renewable Energy Investment Risk: An Canadian Journal of Economics 49 (1): 147–72. Investigation of Changes over Time and the Underlying IEA (International Energy Agency). 2021. “An Energy Sector Drivers.” Energy Policy 140 (May): 111428. Roadmap to Carbon Neutrality in China.” IEA, Paris. Egli, Florian, Bjarne Steffen, and Tobias S. Schmidt. 2018. IEA (International Energy Agency). 2023a. “Critical Minerals “A Dynamic Analysis of Financing Conditions for Market Review 2023.” IEA, Paris. Renewable Energy Technologies.” Nature Energy 3 IEA (International Energy Agency). 2023b. World Energy (November): 1084–92. Outlook 2023. Paris: IEA. Energy Institute. 2023. “Statistical Review of World Energy IRENA (International Renewable Energy Agency). 2023. 2023.” June, Energy Institute, London. “The Cost of Financing for Renewable Power.” IRENA, Eskeland, Gunnar S., and Ann E. Harrison. 2003. “Moving Abu Dhabi, United Arab Emirates. to Greener Pastures? Multinationals and the Pollution Iyer, Gokul C., Leon E. Clarke, James A. Edmonds, Brian Haven Hypothesis.” Journal of Development Economics P. Flannery, Nathan E. Hultman, Haewon C. McJeon, 70 (1): 1–23. and David G. Victor. 2015. “Improved Representation ESMAP (Energy Sector Management Assistance Program), of Investment Decisions in Assessments of CO2 OECD (Organisation for Economic Co-operation and Mitigation.” Nature Climate Change 5 (March): 436–40. 240 | WORLD DEVELOPMENT REPORT 2024 Jain, Gautam, Chris Bataille, Noah Kaufman, and Sagatom with an Investor Perspective.” Applied Energy 236 Saha. 2024. “Resuscitating WTO for the Energy (February 15): 1249–68. Transition: Why, How, and Who.” Background paper pre - Qiu, Yueming, and Matthew E. Kahn. 2019. “Impact of pared for World Development Report 2024, World Bank, Voluntary Green Certification on Building Energy Washington, DC. Performance.” Energy Economics 80 (May): 461–75. Kahn, Matthew E., and Somik Vinay Lall. 2022. “Will the Richter, Philipp M., and Alexander Schiersch. 2017. “CO2 Developing World’s Growing Middle Class Support Low- Emission Intensity and Exporting: Evidence from Carbon Policies?” NBER Working Paper 30238 (July), Firm-Level Data.” European Economic Review 98 National Bureau of Economic Research, Cambridge, MA. (September): 373–91. Kirkpatrick, A. Justin, and Lori S. Bennear. 2014. Rozenberg, Julie, and Marianne Fay, eds. 2019. Beyond “Promoting Clean Energy Investment: An Empirical the Gap: How Countries Can Afford the Infrastructure Analysis of Property Assessed Clean Energy.” Journal They Need while Protecting the Planet. Sustainable of Environmental Economics and Management 68 (2): Infrastructure Series. Washington, DC: World Bank. 357–75. Salm, Sarah. 2018. “The Investor-Specific Price of Labandeira, Xavier, José M. Labeaga, and Xiral López- Renewable Energy Project Risk: A Choice Experiment Otero. 2017. “A Meta-Analysis on the Price Elasticity of with Incumbent Utilities and Institutional Investors.” Energy Demand.” Energy Policy 102 (March): 549–68. Renewable and Sustainable Energy Reviews 82 (Part 1): Lall, Somik Vinay, Jon Kher Kaw, Forhad J. Shilpi, and Sally 1364–75. Beth Murray. 2023. Vibrant Cities: On the Bedrock of Schmidt, Johannes, Rafael Cancella, and Amaro O. Pereira, Stability, Prosperity, and Sustainability. Washington, DC: Jr. 2016. “An Optimal Mix of Solar PV, Wind and World Bank. Hydro Power for a Low-Carbon Electricity Supply in Lall, Somik Vinay, and Maria Vagliasindi. 2024. “The Green Brazil.” Renewable Energy 85 (January): 137–47. Lucas Paradox: Are There Renewable Energy Bargains Steffen, Bjarne, Martin Beuse, Paul Tautorat, and Tobias S. in Developing Countries?” Background paper prepared Schmidt. 2020. “Experience Curves for Operations and for World Development Report 2024, World Bank, Maintenance Costs of Renewable Energy Technologies.” Washington, DC. Joule 4 (2): 359–75. Lowe, Matt, Chris Papageorgiou, and Fidel Pérez-Sebastián. Steffen, Bjarne, and Paul Waidelich. 2022. “Determinants 2019. “The Public and Private Marginal Product of of Cost of Capital in the Electricity Sector.”  Progress in Capital.” Economica 86 (342): 336–61. Energy 4 (3): 033001. Lüthi, Sonja, and Rolf Wüstenhagen. 2012. “The Price of Stern, Nicholas. 2023. “Harnessing Creative Destruction Policy Risk: Empirical Insights from Choice Experiments to the Tackling of Climate Change: Purpose, Pace, with European Photovoltaic Project Developers.” Energy Policy.” In The Economics of Creative Destruction: New Economics 34 (4): 1001–11. Research on Themes from Aghion and Howitt, edited by Malhotra, Abhishek, and Tobias S. Schmidt. 2020. Ufuk Akcigit and John Michael Van Reenan, 365–97. “Accelerating Low-Carbon Innovation.” Joule 4 (11): Cambridge, MA: Harvard University Press. 2259–67. Waissbein, Oliver, Yannick Glemarec, Hande Bayraktar, and Mazzucato, Mariana, and Gregor Semieniuk. 2018. Tobias S. Schmidt. 2013. Derisking Renewable Energy “Financing Renewable Energy: Who Is Financing What Investment: A Framework to Support Policymakers in and Why It Matters.” Technological Forecasting and Selecting Public Instruments to Promote Renewable Social Change 127 (February): 8–22. Energy Investment in Developing Countries. March. New Mealy, Penny, and Alexander Teytelboym. 2022. “Economic York: United Nations Development Programme. Complexity and the Green Economy.” Research Policy Way, Rupert, Matthew C. Ives, Penny Mealy, and J. Doyne 51 (8): 103948. Farmer. 2022. “Empirically Grounded Technology Melekh, Yaroslav, Michael Grubb, and James Dixon. 2024. Forecasts and the Energy Transition.” Joule 6 (9): 2057–82. “Energy Technologies for Low-Carbon Development WEF (World Economic Forum). 2014. “Mitigation of in Middle-Income Countries: Assessment and Political and Regulatory Risk in Infrastructure Projects: Implications.” Background paper prepared for World Introduction and Landscape of Risk.” WEF, Geneva. Development Report 2024, World Bank, Washington, DC. World Bank. 2015. “From Billions to Trillions: MDB Noothout, Paul, David de Jager, Lucie Tesnière, Sascha van Contributions to Financing for Development.” July, Rooijen, Nikolaos Karypidis, Robert Brückmann, Filip World Bank, Washington, DC. Jirouš, et al. 2016. The Impact of Risks in Renewable World Bank and ESMAP (Energy Sector Management Energy Investments and the Role of Smart Policies: Final Assistance Program). 2024. “Net Zero Energy by Report. Karlsruhe, Germany: Fraunhofer Institute for 2060: Charting Europe and Central Asia’s Journey Systems and Innovation Research. toward Sustainable Energy Futures.” World Bank, Pindyck, Robert S. 2021. “What We Know and Don’t Know Washington, DC. about Climate Change, and Implications for Policy.” Zhou, Xiaoyan, Christian Wilson, Anthony Limburg, Gireesh Environmental and Energy Policy and the Economy 2 (1): Shrimali, and Ben Caldecott. 2023. “Energy Transition 4–43. and the Changing Cost of Capital: 2023 Review.” March, Polzin, Friedemann, Florian Egli, Bjarne Steffen, and Tobias Oxford Sustainable Finance Group, Smith School of S. Schmidt. 2019. “How Do Policies Mobilize Private Enterprise and the Environment, University of Oxford, Finance for Renewable Energy? A Systematic Review Oxford, UK. Capitalizing on Crises | 241 The Middle-Income Trap Middle-income countries are in a race against time. Many of them have done well since the 1990s to escape low-income levels and eradicate extreme poverty, leading to the perception that the last three decades have been great for development. But the ambition of the more than 100 economies with incomes per capita between US$1,100 and US$14,000 is to reach high-income status within the next generation. When assessed against this goal, their record is discouraging. Since the 1970s, income per capita in the median middle- income country has stagnated at less than a tenth of the US level. With aging populations, growing protectionism, and escalating pressures to speed up the energy transition, today’s middle-income economies face ever more daunting odds. To become advanced economies despite the growing headwinds, they will have to make miracles. Drawing on the development experience and advances in economic analysis since the 1950s, World Development Report 2024  identifies pathways for developing economies to avoid the “middle-income trap.” It points to the need for not one but two transitions for those at the middle-income level: the first from investment to infusion and the second from infusion to innovation. Governments in lower-middle-income countries must drop the habit of repeating the same investment-driven strategies and work instead to infuse modern technologies and successful business processes from around the world into their economies. This requires reshaping large swaths of those economies into globally competitive suppliers of goods and services. Upper- middle-income countries that have mastered infusion can accelerate the shift to innovation—not just borrowing ideas from the global frontiers of technology but also beginning to push the frontiers outward. This requires restructuring enterprise, work, and energy use once again, with an even greater emphasis on economic freedom, social mobility, and political contestability. Neither transition is automatic. The handful of economies that made speedy transitions from middle- to high-income status have encouraged enterprise by disciplining powerful incumbents, developed talent by rewarding merit, and capitalized on crises to alter policies and institutions that no longer suit the purposes they were once designed to serve. Today’s middle-income countries will have to do the same. ISBN 978-1-4648-2078-6 SKU 212078