38140 Global Economic Prospects Managing the Next Wave of Globalization 2007 ©2007 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 10 09 08 07 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-6727-7 ISBN-13: 978-0-8213-6727-8 eISBN-10: 0-8213-6728-5 eISBN-13: 978-0-8213-6728-5 DOI: 10.1596/978-0-8213-6727-8 ISSN: 1014-8906 Cover photo: Pallava Bagla/Corbis The cutoff date for data used in this report was November 22, 2006. Dollars are current U.S. dollars unless otherwise indicated. Contents Foreword vii Acknowledgments ix Overview xi Abbreviations xxvii Chapter 1 Prospects for the Global Economy 1 Summary of the medium-term outlook 1 Global growth surged to 3.9 percent in 2006 2 Regional outlooks 6 Financial markets 12 World trade 18 Commodity markets 20 Downside risks predominate 23 Notes 26 References 27 Chapter 2 The Coming Globalization 29 The evidence of globalization 30 The world in 2030--the big picture 36 The four channels of globalization 46 What will happen if growth is slower--or faster--in the next 25 years? 52 Challenges of the coming globalization 58 Notes 61 References 63 Chapter 3 Income Distribution, Inequality, and Those Left Behind 67 The global distribution of income 70 Within-country inequality and poverty reduction 80 Policy implications 89 Notes 94 References 97 iii C O N T E N T S Chapter 4 New Pressures in Labor Markets: Integrating Large Emerging Economies and the Global Sourcing of Services 101 The impact of globalization: the story so far 102 New challenge I--absorbing large emerging economies into the global market 109 New challenge II--global sourcing of services 120 Policies to confront the labor market challenges of globalization 125 Notes 133 References 136 Chapter 5 Managing the Environmental Risks to Growth 141 The immediate risk of epidemics 143 The medium-term risks to marine fisheries 146 The long-term risk of climate change 149 Conclusions and policy recommendations 160 Notes 163 References 164 Appendix: Regional Economic Prospects 167 Figures 1.1 Industrial production may be slowing 2 1.2 Regional growth trends 4 1.3 Inflation has increased moderately 13 1.4 Inflation is rising in high-income countries 13 1.5 Signs of overheating in some developing countries 13 1.6 Despite turbulence, financing conditions remain favorable 14 1.7 Private capital flows to developing countries remain strong 15 1.8 Ample liquidity keeps long-term interest rates low 15 1.9 Global demand shifts from the United States to Europe and developing countries 16 1.10 A start to orderly adjustment? 16 1.11 Interest rate spreads support the dollar 17 1.12 Turbulence resulted in sharp depreciations for some developing countries 17 1.13 Rotation in global trade 18 1.14 China's exports exceed those of the United States 19 1.15 Diverging trends in commodity prices 20 1.16 Oil prices continue to rise 21 1.17 Higher prices slow oil demand 22 1.18 A disappointing supply response 22 1.19 Spare production capacity remains low 22 1.20 After rising rapidly, housing price growth slows sharply 24 2.1 World trade has expanded dramatically . . . 31 2.2 . . . and become more diversified . . . 31 2.3 . . . than increase in migrants--in particular toward high-income countries . . . 32 2.4 . . . and a sharp rise in capital flows 32 2.5 Diffusion of traditional technologies has been slow, except in high-growth regions . . . 33 2.6 . . . but the uptake of new technologies has been faster 33 iv C O N T E N T S 2.7 World population growth will be concentrated in developing countries in coming decades 38 2.8 Developing countries will account for a larger portion of world output in coming decades 39 2.9 In some developing regions, per capita incomes will begin to converge with those in high-income countries 41 2.10 Labor force growth is slowing 43 2.11 Due to the demographic dividend, fewer resources will be needed for a declining youth population 44 2.12 More resources will be needed to take care of a growing elderly population 45 2.13 Future-flow securitizations in developing countries, 1990­2004 50 2.14a Past global growth . . . 52 2.14b . . . has been around 2 percent per capita for high-income regions . . . 53 2.14c . . . and much more volatile in developing countries 53 2.15 More acceleration in growth is possible 57 2.16 Wages outpace profit income 59 3.1 Middle-class expansion is sensitive to growth assumptions 75 3.2 World tourism is expected to double between 2004 and 2020 77 3.3 The world's poor may be concentrated in Africa 78 3.4 By 2030, East and South Asia are likely to move up the global income distribution ladder, while other regions will lag 79 3.5 Migration out of agriculture reduces poverty more when education is more equally distributed 82 3.6 Changes in inequality are mainly due to economic shifts 84 3.7 Inequality hampers the potential of growth to reduce poverty 84 3.8 Restricting intersectoral mobility can lead to large increases in inequality 86 3.9 Ending aid would hurt the poor 90 3.10 Global trade reform can be pro-poor 91 3.11 The inequality effects of trade liberalization are not large and depend on the structure of initial protection 93 4.1 Developed countries' imports of manufactures increasingly come from developing countries 103 4.2 In many developed countries the gap between high- and low-income earners has widened 106 4.3 Average wages in China have increased more than in other countries 113 4.4 China's imports from developing countries have surged over the last two decades 114 4.5 Developing-country exports of business services are growing rapidly 121 4.6 Low-income countries depend heavily on import duties for tax revenues 132 5.1 The SARS epidemic was contained in a matter of months 144 5.2 Total marine fish catch has leveled off 147 5.3 Temperatures have increased rapidly since the Industrial Revolution 149 5.4 Temperatures and greenhouse gas emissions have risen 150 5.5 Greenhouse gas emissions have long-term effects 152 5.6 Carbon emissions from developing countries are set to rise 157 5.7 Global trading in carbon emissions has mushroomed 159 v C O N T E N T S Tables 1.1 The global outlook in summary 3 2.1 Services exports rise in line with goods exports 34 2.2 Country rankings--1980­2005 40 2.3 Regional breakdown of poverty in developing countries 60 3.1 The global middle class is growing, its composition changing 73 3.2 Where the return to education is high, its poverty-reducing impact is also high 88 3.3 Some factors affect the probability of being in the lowest income decile more than others--and the differences are changing over time 88 4.1 Employment in developing countries has shifted out of agriculture into manufactures and services 104 4.2 In 2030 most workers will be in developing countries and unskilled 110 5.1 Progress in providing many global public goods is limited 143 5.2 Uncertainty and incentives affect international institutions 161 A.1 East Asia and the Pacific forecast summary 167 A.2 East Asia and the Pacific country forecasts 168 A.3 Europe and Central Asia forecast summary 169 A.4 Europe and Central Asia country forecasts 169 A.5 Latin America and the Caribbean forecast summary 171 A.6 Latin America and the Caribbean country forecasts 172 A.7 Middle East and North Africa forecast summary 174 A.8 Middle East and North Africa country forecasts 175 A.9 South Asia forecast summary 176 A.10 South Asia country forecasts 176 A.11 Sub-Saharan Africa forecast summary 177 A.12 Sub-Saharan Africa country forecasts 178 Boxes 2.1 Inside the box--the components of scenario building 37 2.2 Challenge of geopolitical shifts for long-term economic forecasts: lessons of history 55 3.1 Changes in demographic structure, occupational choices, and factor rewards determine the authors' hypothetical 2030 world income distribution 68 3.2 Aggregate economic performance: distribution matters 71 4.1 What causes the gap between skilled and unskilled labor--technology or trade? 105 4.2 Workers in the nontraded sector--the role of migration 107 4.3 Is the world flat . . . or just smaller? 111 4.4 Global production and the iPod 118 4.5 Does globalization lead to a race to the bottom on labor standards? 119 4.6 The number of services jobs liable to be moved abroad: large or small? 122 4.7 Trading goods and services or trading tasks? 126 4.8 Key challenges for education systems in the new global economy 129 4.9 Overview of the impact of active labor-market programs 131 5.1 The vanishing polar ice 151 5.2 Can efficiency and renewables be the answer? 152 5.3 Stern Review: The Economics of Climate Change 155 vi Foreword G LOBAL ECONOMIC PROSPECTS reports access to telecommunications and the Internet, have customarily aimed to stand back as well as through innovative forms of business from the Bank's day-to-day work and organization, often linked to foreign investment. explore existing or emerging debates in the The next globalization--deeper integration international arena that are of critical impor- with the world economy through trade, flows tance to developing countries. We have en- of information technology, finance, and deavored to focus on areas in which the Bank's migration--will offer renewed and enhanced researchers and technical experts may provide opportunities to increase productivity and insights based on their cross-country and raise incomes. Producers participating in global knowledge. Thus, past reports have bigger international markets will be able to helped to deepen the Bank contribution to pol- produce on a larger scale, access the most ap- icy debates in areas such as international and propriate technology and knowledge, and par- regional trade, investment, and, last year, mi- ticipate in increasingly integrated global gration and remittances. production chains. Consumers everywhere will The strong performance of the global have access to the latest international products. economy--and of developing countries in However, along with rising average incomes particular--in recent years led us to ask may come dislocations and environmental whether these higher rates of growth could be pressures. This Global Economic Prospects sustained for the long term. And if so, what analyzes three possible consequences-- would the implications be for the global econ- growing inequality, pressures in labor markets, omy and for the world's poor? Answering and threats to the global commons. All are ev- those questions leads us to explore the nature ident in the current globalization, but in com- of the "next globalization." ing years they are likely to become more acute. Three features are likely to be particularly If these forces are left unchecked, they could prominent in the next wave of globalization. slow or even derail globalization and thus First is the growing economic weight of develop- adversely affect growth and development in ing countries in the international economy, no- many developing countries. The report is tably the emergence of new trading power- premised on the idea that the threats to contin- houses such as China, India, and Brazil. Second ued global growth and poverty reduction from is the potential for increased productivity that is environmental damage, social unrest, or new offered by global production chains, particu- increases in protectionist sentiment are poten- larly in services, arguably the most dynamic sec- tially serious, and it is worth exploring ways tor of trade today. Third is the accelerated diffu- that these disruptive forces might be addressed sion of technology, made possible through now if we wish to see sustainable global falling communications costs and improved growth in the future. vii F O R E W O R D To analyze these problems, the report em- National policy makers must decide how best ploys a series of projections and simulations to respond to globalization--because the built around a central scenario of the evolu- growth and long-term competitiveness of tion of the global economy. The objective of their countries are at stake. And interna- the scenario-based approach is to analyze the tional policy makers must devise ways for benefits and stresses of integration. The pur- nations to work together to ensure that pose is not to predict the future--the actual growth is sustained and widely shared, and numbers for global or country performance does not cause irreparable damage to the may turn out to be higher or lower--but to environment. think about dynamics in the global economy in a coherent analytical framework. Focusing on the future helps bring into François Bourguignon sharper relief the choices facing policy Senior Vice President and Chief Economist makers in managing global integration today. The World Bank viii Acknowledgments T HIS REPORT WAS produced by staff from the World Bank's Development Prospects Group. Richard Newfarmer was the lead author and manager of the report. The principal author of chapter 1 was Andrew Burns. Chapter 2 was written by Dominique van der Mensbrugghe, with written contributions from Dilek Aykut and Sanket Mohapatra and with support from Sebnem Sahin. Chapter 3 was written by Maurizio Bussolo and Denis Medvedev, with the benefit of guidance from Francisco Ferreira and with support from Victor Sulla and Rafael De Hoyos. Chapter 4 was written by Julia Nielson, Paul Brenton, and Mombert Hoppe, with guidance from Gordon Betcherman. Chapter 5 was written by consultants Peter Sturm and William Shaw, with written contributions from Maureen Cropper and Philippe Ambrosi and with analytic work by Mombert Hoppe. The accompanying online publication, Prospects for the Global Economy (PGE), was produced by a team led by Andrew Burns and comprising Sarah Crow, Cristina Savescu, and Shuo Tan, with technical support from Gauresh Rajadhyaksha. The report was produced under the guidance of Uri Dadush and François Bourguignon. Several reviewers offered extensive advice and comment throughout the conceptualization and writing stages. These included Nancy Birdsall, Cornelis de Hann, Shahrokh Fardoust, Alan Gelb, Lidvard Gronnevet, Bernard Hoekman, Robert Holzmann, Eriko Hoshino, Kieran Kelleher, Jeffrey Lewis, William Maloney, Branko Milanovic, Moisés Naím, Ian Noble, Stefano Scarpetta, David Wheeler, and Roberto Zagha. Several people contributed substantively to the various chapters. In chapter 1, the Global Trends Team, under the leadership of Hans Timmer, was responsible for the projections, with contributions from John Baffes, Maurizio Bussolo, Betty Dow, Annette De Kleine, Donald Mitchell, Denis Medvedev, Mick Riordan, Cristina Savescu, and Shane Streifel. In chapter 2, the poverty numbers originated with Shaohua Chen and Martin Ravallion from the Development Research Group. In addition, Steven Kennedy and Bruce Ross Larsen edited the report. Dorota A. Nowak and Nigar Farhad Aliyeva managed the publication process for the Development Prospects Group, and Merrell Tuck managed the dissemination activities. Book production was coordinated by the World Bank Office of the Publisher. ix Overview T HE INTENSE PACE of globalization has high-income countries? How will global inte- improved living standards worldwide gration, interacting with demography, techni- on an unprecedented scale--but not cal change, and other forces, affect the distri- for everyone. Some countries and some social bution of income and labor markets in rich groups have been left behind. Even in coun- and poor countries? How will it affect the tries that have benefited greatly from global- global environmental and health threats that ization, tensions in labor markets have cloud long-term growth prospects? simmered, at times boiling over into civil dis- Global Economic Prospects 2007 explores turbances. Meanwhile economic growth, the next wave of globalization. The organiz- while essential to improving living standards, ing vehicle for discussion is a set of growth sce- is damaging what many call the "global com- narios covering the years 2006 to 2030. The mons," giving rise to concerns about the sus- objective of the scenario-based approach is to tainability of long-term growth. analyze the opportunities and stresses of inte- These pressures are likely to intensify in gration. The purpose is not to predict the fu- coming years. Why? Because as markets inte- ture but to bring into sharper relief the choices grate, competition among countries--and facing the world today. National policy their firms and workers--increases. Develop- makers must decide how best to respond to ing countries, once at the periphery of the globalization--because the growth and long- global economy, are now moving to center term competitiveness of their countries are at stage and are becoming serious competitors stake. And international policy makers must in the markets of high-income countries and devise ways for nations to work together to in each other's markets. Concerns about ensure that growth can continue without competition from China and other low-wage becoming destabilizing. suppliers now pepper the headlines in rich and poor countries alike. The loss of white- collar jobs from global sourcing of services, Prospects for 2007 and 2008-- often to India and other developing coun- tries, provides fodder for heated debate on bright, with a few dim spots T talk shows and as the theme of several best- he medium-term outlook for the world selling books.1 economy remains fairly bright (chap- Will global integration--of trade, finance, ter 1). While the pace of economic expansion technology, ideas, and people--continue into is slowing, developing economies are pro- the foreseeable future? If so, what will it jected to grow by 7.0 percent in 2006, more mean for developing countries and for today's than twice as fast as high-income countries xi O V E R V I E W Figure 1 Growth, though tapering off, will likely remain solid over the medium term Percent change in GDP 10 9 2005 8 2006 2007 7 2008 6 5 4 3 2 East Asia and Europe and Latin America Middle East and South Asia Sub-Saharan the Pacific Central Asia and the Caribbean North Africa Africa Source: World Bank. (3.1 percent), with all developing regions Disruptions in oil markets are always possible. growing by about 5 percent or more (figure 1). And the unwinding of the U.S. current account Looking forward, limited inflationary pres- deficit and its mirror surpluses in oil-exporting sures and high savings among oil exporters countries and East Asia could also be disruptive and in Europe (as Europeans prepare to meet if sudden movements in capital markets, per- the challenges of their aging societies) are ex- haps abetted by collective policy inaction, drive pected to keep long-term interest rates low. As the rebalancing. Even so, these risks appear a result, while growth in developing countries manageable, and the promising environment may slow somewhat over the next two years, for growth makes this an opportune time to it is expected to remain very robust--at more focus on long-term issues. than 6 percent in 2007 and 2008. Increases in supplies of key commodities, in combination Globalization's next 25 years-- with demand-side substitution and conserva- incomes up, poverty down, three tion measures, should allow for some easing of prices, including those for oil, but continuing big threats to growth D strong global growth is expected to keep com- emographic trends will be a major driver modity prices high by historical standards. of future events. The Earth's current pop- Even though a tapering down of growth to a ulation of some 6.5 billion is expected to rise to sustainable but robust rate remains most likely, 8.0 billion by 2030, an average increase of this positive outlook is subject to significant 60 million annually. More than 97 percent risks. Efforts to temper the expansion in some of this growth will take place in developing of the fastest-growing developing countries countries. Both the European Union and Japan may not succeed, leading to stronger growth in are likely to experience a decline in population, the short term but a sharper slowdown later. and most of the increase in other rich countries A faster-than-expected weakening of housing will be due to migration. The largest country in markets in high-income countries could brake the world, China, will see its population con- the economy much more abruptly than tinue to grow, but at a slower pace than the rest expected, thus weakening global demand. of the developing world. With more rapid xii O V E R V I E W population growth, India will likely surpass China as the world's most populous country Figure 2 Average incomes are likely to converge, if modestly sometime during this period. The global labor force will increase from just over 3 billion Per capita incomes as percent of high-income countries today to 4.1 billion in 2030, a rate of increase 24 2030 greater than that of population. Meanwhile 22 2020 the dependency ratio is likely to fall, providing a sustained boost to world growth. 20 2010 If this report's central scenario materializes, 2005 18 global economic growth will be somewhat 16 faster in 2006­30 than in 1980­2005. But growth in the global economy will be powered 14 increasingly by developing countries, where 12 per capita incomes will rise 3.1 percent a year 10 on average, up from 2.1 percent over the ear- Low- and middle-income countries lier period. That rate of increase will produce Source: World Bank simulations using the Linkage model. average per capita incomes in the developing Note: Ratio of Purchasing Power Parity (PPP)­adjusted world of $11,000 in 2030, compared with per capita incomes relative to high-income average. PPP is fixed at base year (2001) level. $4,800 today,2 roughly the level of the Czech Republic and the Slovak Republic today. Average income in rich countries will also rise dramatically: the average income of the chil- Figure 3 Developing-country share of the dren of today's baby boomer is likely to be global economy will rise in coming years nearly twice that of their parents. The output of the global economy will rise GDP, 2001 $ trillions 80 from $35 trillion in 2005 to $72 trillion (at con- stant market exchange rates and prices) in 2030, 70 an average annual increase of about 3 percent-- 60 2.5 percent for high-income countries and 50 4.2 percent for developing countries. Though 40 the incomes of developing countries will still be 30 less than one-quarter those in rich countries in 20 2030, they will continue to converge with those 10 31% 28% of wealthy countries (figure 2). This would 23% 25% 0 imply that countries as diverse as China, 2005 2010 2020 2030 Mexico, and Turkey would have average living High-income countries standards roughly comparable to Spain today. Low- and middle-income countries This is good news for the world's poor. The implications of sustained growth for reducing Source: World Bank simulations using the Linkage model. poverty around the world are nothing short of Note: GDP in 2001 $ trillions, using PPP exchange rates. astounding. Despite population growth, the number of people living in dire poverty--below the $1-a-day poverty line--is likely to fall to Developing countries, once considered the 550 million from 1.1 billion today. Similarly, periphery of the global economy, will become the number of people living on less than $2 a main drivers. Overall, developing countries' day should fall below 1.9 billion, 800 million share in global output will increase from fewer than today. The bottom line? Poverty will about one-fifth of the global economy to decline, despite continuing population growth. nearly one-third (figure 3). Their share of xiii O V E R V I E W global purchasing power would surpass half. Today, six developing countries have Figure 5 Globalization will intensify in coming years populations greater than 100 million and an annual gross domestic product (GDP) of Exports of developing and developed countries, 2005­30 more than $100 billion. By 2030, under rea- sonable assumptions of economic growth, at Share of world exports (%) trillion 80 30 least 10 countries will reach the twin 100s High-income High-income countries (left axis) countries (right axis) thresholds.3 70 25 Global integration is likely to enter a new 60 Developing 20 phase. In virtually every growing economy the 50 countries Developing (right axis) countries importance of trade--captured by the ratio of 40 (left axis) 15 trade to GDP--will rise, continuing the trend of 30 10 the past two decades. The growth in the trade 20 ratio over the next 25 years will be powered by 5 10 a new dynamism in services trade. Global trade 0 0 in goods and services, growing faster than out- 2005 2010 2020 2030 put, is likely to rise more than threefold to Source: World Bank simulations using the Linkage model. $27 trillion in 2030 (figures 4 and 5). Note: Exports in US$2001 trillion. Roughly half that increase will come from developing countries. This means that a growing share of global production of goods and services will be performed in those de- new opportunities. For example, agriculture veloping countries able to take advantage of now accounts for about 2 percent of the economic value added of most rich coun- tries; that share will shrink to boutique niches. A few resource-rich regions and Figure 4 Countries everywhere increase countries, including Latin America and exports as world trade outpaces other Australia, will be the source of 90 percent of sources of growth the world's sugar, 50 percent of its grain, Export-to-GDP share (%) and 40 percent of its dairy products. 40 Projected Whether countries exceed projections--or 35 fall short--depends heavily on the policies they adopt over this long period. 30 Several factors could alter this relatively 25 sanguine outcome for better or worse. The central long-term scenario in this report is 20 robust enough to resist periodic recessions, 15 isolated regional conflicts, and even many of 10 the destabilizing crises the world has experi- enced in the past 30 years. These threats are 5 likely to affect particular regional or national 0 economies more than the world economy, 1970 1975 1980 1985 1990 1995 20002005*2010*2015*2020*2025*2030* and if history is a guide, are likely to be of relatively short duration. Between 1980 and Sources: Development Data Platform (DDP) and staff calculations. *Indicates World Bank using the Linkage model. 2005 the world economy grew at a steady Note: Export-to-GDP ratio for the world. The export share pace despite several major disruptions-- is calculated in nominal dollar terms. Observations are including the Latin American debt crisis, smoothed using five-year moving averages. the demise of the Soviet Union, the East Asia xiv O V E R V I E W Figure 6 More acceleration is possible Income growth per capita, % per year, 2010­30 8 7 High 6 Central 5 4 3 2 1 0 High-income East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America countries the Pacific Central Asia North Africa Africa and the Caribbean Source: World Bank staff simulations using the Linkage model. crisis, two global downturns, and the tragedy scenario in this report (figure 6) is based on of September 11, 2001. These events had the assumption that countries perform only short-term effects on global growth and closer to their full potential over a longer a marginal impact on the steady advance of period of time. Predicated on maintaining globalization, even though regional ripples the solid growth rates of the last half- continued for years afterward. This suggests decade, the high-growth scenario sketched that the basic long-term trends discussed here would lead to incomes in 2030 that here, if not the assumed growth rates, are are some 45 percent higher than those pro- fairly impervious to all but the most severe jected under the central scenario and to and sustained shocks. declines in absolute poverty ($1 per day) At the same time, the possibility exists from about 20 percent of the world's popu- that the world will be even better than en- lation today to less than 4 percent in 2030. visioned in the central scenario--thanks Two points emerge from the discussion of possibly to unanticipated technological im- scenarios (chapter 2). First, policy matters. provements, more innovation in business The right domestic and international policies, processes that allow for an acceleration of sustained over long periods, have the power to globalization, and widespread adoptions of raise incomes around the world, especially in good policies within countries. Indeed, certain countries. Second, whether the under- greater integration promotes knowledge lying growth rates are low or high, the dynam- about policies that work. It also shortens ics underpinning any likely scenario will gen- the duration of bad policies, as investment erate stresses that require policy attention capital and human resources can more read- today. The report analyzes in detail three main ily flee poorly performing nations. That dis- stresses in the global economy that could cipline is likely to become more effective as threaten growth: widening inequality, growing financial, merchandise, and technological tensions in labor markets, and new environ- markets continue to integrate. The upside mental pressures. xv O V E R V I E W Income inequality could widen-- Figure 7 The "global middle class" will among and within countries expand significantly T he benefits of globalization are likely to Population in low- and middle-income countries be uneven across regions and countries earning $4,000­$17,000 per capita (PPP) (chapter 3). Africa, because of underlying Percent growth trends and the presence of many frag- 16 Sub-Saharan Africa 1.2 billion ile states, is the region most likely to fall farther South Asia 14 behind. But, it also has the most to gain from Middle East and North Africa 12 Latin America and the Caribbean integration, because it can take advantage of 10 technology and wage gaps to propel higher Europe and Central Asia 8 sustained growth. East Asia and the Pacific Of equal concern, strong forces in the 6 400 million global economy may tend to increase inequal- 4 ity in many national economies. Even though 2 a large segment of the developing world is 0 likely to enter what can be called the "global 2005 2030 middle class," some social groups may be left Source: World Bank staff calculations. behind or even marginalized in the growth process. Unskilled workers, in particular, may fall farther behind. Technological progress, by generating demand for greater skills, tends to widen the gap between the wages of skilled countries are classified as rich in a global con- and unskilled workers. Demographic patterns text, while many people viewed as wealthy that affect social dependency ratios (the ratio in developing countries are members of the of workers to youth and retirees) and educa- global middle class.) This large group will tional attainment are also important. participate actively in the global marketplace, Trade per se has been found generally to demand world-class products, and aspire to have no systematic effect across countries as a international standards of higher education. direct channel for wage-gap widening. How- That is, they would have the purchasing ever, in combination with technological change power to buy automobiles (perhaps second- and, to a lesser extent, foreign investment, hand), purchase many consumer durables, these globalization-related forces may combine and travel abroad. to increase inequality in many countries--at While still a minority in their own coun- the same time as they are raising average tries, the new members of the global middle incomes. class will place new and quite different de- mands on domestic political structures. Their A global middle class will emerge livelihoods and standards of consumption are By 2030, fully 1.2 billion people in develop- likely to be connected to the global market, ing countries--15 percent of the world so, as the studies reviewed in chapter 3 show, population--will belong to the global middle their political predilection may be more likely class, up from 400 million in 2005. Families of to favor access to the international market, if four in that class earn between $16,000 and not greater openness itself. They also are more $68,000 in PPP dollars (figure 7). (Because likely to demand transparency in political and the definition used here is absolute and based corporate governance, certainty of contracts, on a global scale, most of those who consider and property rights--all hallmarks of an themselves middle class in high-income improving investment climate. xvi O V E R V I E W Most who enter the middle class will do so potential. Most fundamental is a cessation of because they are able to move from agricul- crippling civil conflicts that have stunted de- ture into manufacturing and services or ac- velopment in several regions of Sub-Saharan quire valuable skills faster than their compa- Africa. In the high-case scenario explored in triots. For a given rate of growth, policies that chapter 2, Africa's income could be twice as allow mobility across sectors and that provide high as projected in the central scenario (see broader access to education can accelerate figure 6). economic growth by creating the opportunity While developing countries are closing the and incentives for all citizens to develop their income gap with rich countries, as many as productive potential. two-thirds--more than 80 percent of the devel- oping world outside China--may experience a Africa and some groups within countries worsening of within-country inequality, thus may lag behind muting the poverty-reducing effects of growth Sub-Saharan Africa will have to make a strong and fanning social tensions that could derail effort, with the support of the international growth. Demography plays a role, as aging so- community, if it is to avoid being left behind cieties tend to become more unequal. But the (figure 8). Today, half of the poorest tenth of main driver is the widening difference in earn- the world population lives in Asia; by 2030 ing potential between skilled and unskilled Asia's share in the lowest tenth will be reduced workers (figure 9). This is because investments to one-fifth under the central scenario. By con- in capital and technology create a more rapidly trast, Africa, now home to one-third of the growing demand for skilled workers. The sim- poorest people, is likely to see its share of ulations in this report suggest that the com- the lowest tenth double by 2030. To be sure, bined effects of all these forces--technology, the region has the potential for more rapid globalization, demography, and demand for growth, and sustained improvements in policy skilled labor--may widen income distribution and investment climate could bring out that in as many as two-thirds of all countries, Figure 8 Africa risks falling behind, as average incomes are unlikely to converge Per capita incomes as percent of high-income countries 40 35 30 25 2030 20 2020 2010 15 2005 10 5 0 East Asia and Europe and Latin America Middle East and South Asia Sub-Saharan the Pacific Central Asia and the Caribbean North Africa Africa Source: World Bank simulations using the Linkage model. Note: Ratio of PPP-adjusted per capita incomes relative to high-income average. PPP is fixed at base year (2001) level. xvii O V E R V I E W Figure 9 Returns to skilled workers are likely to rise faster than for unskilled workers in the global economy Ratio of skilled wages relative to unskilled wages 7 6 5 4 2030 2001 3 2 1 0 East Asia and Europe and Latin America Middle East and South Asia Sub-Saharan the Pacific Central Asia and the Caribbean North Africa Africa Source: World Bank staff calculations. including many of the most populous develop- health. Finally, increasing the access of poor ing countries. countries to global markets (and thus raising Since in some countries, girls are deprived living standards) by completing the now- of schooling, women in these countries are suspended Doha Round of world trade nego- more likely to enter the labor market without tiations and lowering barriers unilaterally skills. This discrimination in effect foredooms could boost incomes in poor countries. Mea- them to be denied access to the opportunities sures to expand trade should be coupled with afforded by global integration, and means aid to overcome supply-side constraints that the widening skill premium is likely to affect now weigh down the trade of poor develop- them disproportionately. ing countries. Of these constraints, counter- Several policies can lead to more egalitar- productive domestic policies are often the ian countries and a more egalitarian world. most binding. Governments can create new opportunities for the poor through additional investments in education. Investments in girls' education China, India, and global sourcing can be an important complement to reducing will put pressure on labor markets, gender discrimination in the workplace. Ad- ditional resources for education and other especially for the unskilled R pro-poor investments are likely to become apid technological progress, burgeoning available from a tax base centered on a grow- trade in goods, and growing international ing middle-class. Moreover, increasing devel- sourcing of services have come together to put opment assistance for lagging regions and the new pressures in labor markets, pressures that poorest countries--and making that assis- will only become more acute in the next tance more effective--is critical. Particularly 25 years (chapter 4). Globalization offers op- important are investments to overcome bot- portunities for export growth and access to a tlenecks in infrastructure, education, and wider range of cheaper imported products that xviii O V E R V I E W can fuel productivity growth and raise average living standards. But by creating a progres- Figure 10 Imports from developing countries could rise, spawning sively more integrated global market for labor, productivity growth...and tensions it imposes adjustment costs on certain groups within countries, exerting downward pressure Share of high-income countries' imports of manufactures originating in developing countries, on some wages, decreasing job security, and 1973­2003 making retraining and relocation necessary. Percent Even though wages of unskilled workers in vir- 80 tually all countries have risen as productivity 70 has increased with globalization, the unskilled have received wage increases that are lower 60 Other Asia than those for skilled workers--and they have 50 experienced greater difficulty in sustaining 40 China their employment. The projections in this report 30 offer little reason to believe that this will Europe and change in the coming decades. 20 Central Asia Particularly challenging is the rise of China, Latin America 10 Middle East India, and other developing countries as man- Africa 0 ufacturing powerhouses, and, with growing 1973 1983 2003 2020 2030 tradability of services, as suppliers of services Source: Bank staff projections using the Linkage model. to the global market. While the qualitative im- plications of increasing exports of manufac- tured products from China and India are the same as for the emergence of the Asian tigers countries of formerly nontradable service more than a decade ago, their sheer size raises activities imperils white-collar employment in the specter of intense export competition. Im- these activities in both high-income countries ports of high-income countries from all devel- and advanced developing countries. Services oping countries have rise from below 15 per- exports have grown by leaps and bounds in cent in the 1970s to nearly 40 today--but many developing countries (figure 11), creat- more important, their share is expected to rise ing opportunities for productivity gains in to more than 65 percent in 2030 (figure 10). both high-income and developing countries-- This has already exposed workers in rich but have led to more rapid job turnover in countries to competition from low-wage formerly nontraded white-collar jobs. The countries, a pressure that will only intensify global sourcing of such relatively high-paying over the next 25 years. skilled jobs, in contrast to the displacement Many developing countries fear that ex- of low-skilled manufactures, has the potential ports from these large new players may to destroy the investments of white-collar swamp their domestic markets, squeeze them workers in firm-specific knowledge. out of the global export market, foreclose The analysis in this report suggests that avenues of diversification in manufactures as a three factors are likely to mitigate these effects road to higher growth, and soak up the pool in the medium and even the long term. of foreign direct investment (FDI). High- income countries worry that if the large · First, the growth of the Chinese, Indian, emerging economies can readily acquire and and other emerging markets offers enor- master the newest technologies, their exports mous offsetting opportunities for other may soon take over high-tech markets. developing and developed countries to Global sourcing of services exerts similar increase exports. As China and India pressures. The transfer to firms in developing increase their exports, they will have xix O V E R V I E W Figure 11 Services markets are integrating Figure 12 China's non-oil imports from all even more rapidly developing countries have surged over the last two decades Cumulative services export growth, 1994­2003 (%) Imports in $ billions India 140 Estonia Sub-Saharan Africa 120 Romania South Asia Middle East and North Africa China 100 Israel Latin America and the Caribbean 80 Brazil Europe and Central Asia Argentina 60 East Asia and the Pacific Mauritius 40 EU15 20 United States Canada 0 Barbados 1985 1990 1995 2000 2005 Dominicaa Source: World Bank staff calculations. Australia Ghana Japan 0 100 200 300 400 500 600 700 800 Percent more flexible countries to progress faster in institutional development and Source: Data are from IMF Balance of Payment Statistics. Note: Business services are defined as total services minus for rich countries to continue to lead in transportation, travel, and government services. productivity-enhancing innovation. The a. Indicates 1994­2002. flow of service activities from rich to poor countries, which entails some trans- fer of know-how, will be slowed to the to increase imports of intermediate in- extent that institutional frameworks fail puts, energy, technology, and investment to protect the ownership of such assets goods. Driven by Chinese demand, Asia and thus discourage FDI. was the principal destination for acceler- ated export growth for Africa and Latin Despite this sanguine conclusion, the pol- America during the late 1990s and the icy response of countries will determine early years of this decade (figure 12). whether they will be among those able to take · Second, as exports and domestic living advantage of the new opportunities and im- standards rise in these emerging prove their living standards--or fall behind. economies, wages (and exchange rates) Policies to embrace, rather than resist, global rise as well, creating space for low- integration will lay the foundations for future income countries to move into low-skill growth and job creation. Openness to trade activities abandoned by producers in the and FDI will become ever more critical if the large emerging countries. Wages already poorest countries are to absorb technologies have risen in China faster than in many and know-how from abroad and seize the op- other developing countries, a trend portunities created by rising demand from, expected to continue (figure 13). and production shifts in, China and India. But · Third, developing the social institutions openness alone will not be sufficient to foster that support a dynamic market economy integration in the absence of an attractive in- in China and India will take time, vestment climate, with sound institutions and providing an opportunity for smaller, policies that allow resources (labor, capital, xx O V E R V I E W Figure 13 Average wages in China are rising rapidly, creating new opportunities for other low-income countries, 1995­2005 Internationally comparable average wage rates Index 1998 100 230 China 180 Mexico India 130 80 South Africa Philippines 30 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Sources: China Statistical Yearbook 2005, People's Bank of China, International Labour Organization (the Philippines, South Africa), IBGE (Brazil), Banco de Mexico, Ministry of Statistics and Programme Implementation (India); exchange rates from IMF International Financial Statistics. Note: Wages are average wages for China, the Philippines, and South Africa, average private sector wages in Brazil, and manufacturing wages for India and Mexico. Wages for 1998­2000 for the Philippines have been estimated using observed wages from 2001 and projecting them backward using GDP per capita growth rates. and knowledge) to flow from low-return Environmental threats will sectors to high-return sectors. Developing demand much more multilateral knowledge-intensive activities as future dri- collaboration vers of growth will require investing in the T institutions and policy frameworks that foster he gains from growth and globalization innovation and providing effective education could be undermined by their environ- and lifelong learning for all workers. mental side effects. Because increases in pro- Even in the most propitious economic en- duction magnify cross-border pollution, while vironments, policies are needed to cushion improvements in technology make it possible the adjustment costs associated with rapidly to expand or intensify exploitation of scarce changing work force demands and involun- global resources, decisions at the national level tary dislocation. Projections indicate that re- are having a growing impact on other coun- turns to skilled labor will continue to in- tries. International institutions will thus be re- crease more quickly than those to unskilled quired to play a larger role in a wide spectrum labor, extending today's natural wage-widen- of issues--all involving global public goods4-- ing tendencies evident in many, if not most, where exclusive reliance on the decisions of in- countries, and underscoring the need for dividual governments or the private market public policies to support workers at the low can lead to adverse outcomes. As developing end of the scale. Together, both volatility and countries enlarge their role on the global stage, rising wage inequality argue for labor-market their integration as full partners in multilateral policies focused on protecting workers rather solutions to global problems will be essential. than protecting jobs. These trends also argue Mitigating climate change, containing in- for creating opportunities for low-income fectious diseases, and preserving marine fish- people through educational and infrastruc- eries are three prominent global public goods ture investments while eschewing subsidies that demonstrate the need for--and benefits to inefficient activities. of--international policy cooperation. xxi O V E R V I E W Figure 14 Carbon emissions in developing countries are set to rise sharply Annual carbon emissions in the GREEN baseline scenario Million tons of carbon Million tons of carbon 18,000 18,000 16,000 16,000 Major 14,000 14,000 emitters 12,000 World 12,000 10,000 Annex 1 10,000 8,000 8,000 6,000 6,000 4,000 4,000 Industrial countries 2,000 2,000 0 0 1990 2000 2010 2020 2030 2040 2050 Source: OECD Green model simulations. Note: Annex 1 includes the industrial countries plus some countries in Eastern Europe and the former Soviet Union. Major emitters include Annex 1 countries plus large developing countries: China, India, Indonesia, Mexico, and South Africa. · Rising industrial output means that, implications for the welfare of future based on current trends with existing generations. technologies, annual emissions of green- · Technological progress and rising de- house gasses (GHGs) will increase mand have increased efforts to harvest roughly 50 percent by 2030 and will in- fish from the open seas, degrading ocean crease twofold by 2050 (figure 14). This environments and driving some valuable necessarily would sharply increase the species to near-extinction. Fish catches concentrations of GHGs in the atmos- already have leveled off (see figure 15). phere, with substantial risks of detri- mental effects on future productivity and--more generally--on human wel- Figure 15 Worldwide maritime catches fare around the globe. The problem is have leveled off how best to provide energy necessary Millions of tons for growth, while at the same time re- 110 ducing emissions toward levels that will Trendline, 1995­2004 eventually stabilize atmospheric concen- 90 Trendline, 1950­95 trations. Even in the next decade or two, scientists underscore the possibility-- 70 though still low probability--that 50 global warming could cause natural disruptions severe enough to push 30 growth rates perilously below historic trends. While decades will pass before 10 the most severe effects of climate 1950195319561959196219651968197119741977198019831986198919921995199820012004 change will begin to be felt, the collec- tive response of today's global leaders is Source: Food and Agriculture Organization's FISHSTAT database 2004. almost certain to have far-reaching xxii O V E R V I E W Recent scientific calculations project adoption, action to reduce deforestation, and near-complete depletion by 2048 in the assistance to poor developing countries to absence of collective international ef- promote adaptation to permanent climatic forts to limit fishing to sustainable levels changes. (see Worm and others 2006). Long- Similarly, failure to contain an epidemic standing efforts to limit marine catches could suddenly brake global commerce, iso- to sustainable levels have met with only late some populations, and impose huge losses a few successes, as institutional weak- on affected developing countries. Unrestrained nesses, technical difficulties, and inap- marine fishing, while less potentially calami- propriate incentives, such as fishing sub- tous than climate change or a flu pandemic, sidies, impede sustainable management. could permanently degrade a critical global · The growing interaction of national food source and destroy irreplaceable deep- economies through trade and movements sea habitats and biodiversity. of people, while broadly beneficial, has Effective multilateral collaboration is increased the risk of spreading conta- needed to ensure that economic growth and gious diseases. HIV/AIDS (human im- poverty reduction proceed without causing ir- munodeficiency virus/acquired immune reparable harm to future generations. Devel- deficiency syndrome) is one example. oping countries are central to the manage- The severe acute respiratory syndrome ment of these risks. Though these countries (SARS) is another. The most prominent are relatively small contributors to global current threat is posed by the avian in- warming today, the projections in this report fluenza virus. imply that soon enough they will become large contributors; moreover, if no action is These examples of the side effects of taken, the standard of living that they could globalization--one long-term, one medium- otherwise expect may well be put at risk. Sim- term, and one immediate--pose risks to the ilarly, given the limited supply of medical progressive expansion of the global economy, facilities and nursing care in the developing and to developing countries in particular. world, a flu pandemic could have horrific Some of the more catastrophic climate-change consequences. In many developing countries scenarios, if they materialize, could undermine people depend on fish for an important share the development prospects of whole countries of their diet, and the poor would suffer if the and even regions through their effects on agri- price of fish, as well as substitutes, were to culture, water, and ecosystems. According skyrocket as supplies dwindled. to the United Kingdom government's recent The three cases differ in the degree of comprehensive analysis, the Stern Review, agreement among policy makers--and, to less failure to address climate change could-- degree, scientists--regarding the risks potentially--lead to huge reductions in wel- involved. There is a large international con- fare (cuts in per capita consumption of 5­20 sensus on the need to protect against the percent), while the cost of stemming the rise spread of (selected) contagious diseases, and in GHG concentrations at a reasonable level on the right methods of doing so. The poten- could be about 1 percent of annual GDP by tial for exhausting marine fisheries is well un- 2050 (U.K. 2006). These estimated costs of in- derstood, although disagreement remains on action are substantially higher than previous the amount of resources to commit, the limits estimates. The report concludes that an inter- on fishing to impose, and how to allocate national framework should include emissions access to fisheries. There is an international trading to encourage energy efficiency, techno- consensus that human activity is contribut- logical cooperation to ensure more rapid ing to climate change, and that industrial xxiii O V E R V I E W emissions are directly related to atmospheric workers--all the while promoting change, concentrations of GHGs, although the precise not fighting it. implications of different levels of GHG con- Deepening economic interdependence also centrations for climate change remain places a new burden on the collective actions of uncertain. While disagreements over the facts the international community. Several positive in each case have hampered efforts at interna- responses are clear. First, increasing the tional cooperation, they have not been the amount and effectiveness of development major impediment to progress. assistance through both multilateral and The greatest policy challenge in safeguard- bilateral institutions can ease the tendency ing the global commons involves strengthen- of globalization to produce uneven growth. ing international agreements and institutions. Second, liberalizing trade in the framework of The World Health Organization has ad- the World Trade Organization can create new dressed the threat of global pandemics effec- opportunities for poor countries and poor peo- tively. The basic legal framework is in place to ple. The most immediate task is to reactivate safeguard the sustainability of marine fishing, the Doha Round and conclude an agreement but is often inadequately enforced by weak that lowers trade barriers to the products the institutions. Even more work is required to world's poor produce, especially agriculture establish the global institutions capable of re- and labor-intensive manufactures. And third, ducing the risks from climate change. Discus- deepening institutional mechanisms to deal sions aimed at replacing the Kyoto Protocol, with threats to the global commons can ensure which expires in 2012, with a more compre- that globalization is not undone by its own hensive and ambitious agreement are under- success--by providing forums in which dis- way within the framework of the United agreements about how to provide global public Nations Framework Convention on Climate goods in which all nations ultimately have an Change. Meanwhile, it may be useful to start interest can be resolved. Multilateral coopera- putting in place other vehicles, such as a tion will be even more important in the inte- global system for trading emission permits grating world of tomorrow than it is today. The and better means of monitoring emissions in way the international community, acting to- both high-income and developing countries, gether, manages the process of integration will so as to allow a rapid implementation of determine whether the world of 2030 will real- effective policies, once these are agreed. ize its potential. Achieving policy consensus is difficult, but it is now urgent. Notes 1. Several recent and provocative books deal with these themes or variations, and from quite different The world in 2030 perspectives. See, for example, Dervis (2005), Friedman (2005), Goldin and Reinert (2006), Mishkin A ll these developments are pieces of the (2006), Stiglitz (2006), and Wolf (2004) as well as new burden lying on the shoulders of na- various issues of Foreign Policy. tional policy makers: to manage globalization 2. This is measured in constant dollars adjusted or risk being run over by it. This requires for purchasing power parity. government policies to ensure that the poor 3. Today the six countries are Brazil, China, India, Indonesia, the Russian Federation, and, most recently, are incorporated into the growth process Mexico. By 2030, Bangladesh, Nigeria, Pakistan, the through pro-poor investments in education, Philippines, and Vietnam will reach both thresholds. infrastructure, and transfers. Similarly, it Already today the populations of Bangladesh, Nigeria, requires policies to support and invest in and Pakistan exceed 100 million. xxiv O V E R V I E W 4. Examples of global public goods, in addition Goldin, Ian, and Kenneth Reinert. 2006. Globalization to protecting the environment, include ensuring for Development: Trade, Finance, Aid, Migration, global security, keeping the trading system open and and Policy. Washington, DC: World Bank. nondiscriminatory, and maintaining global financial Mishkin, Frederic S. 2006. The Next Great Globaliza- stability. A useful overview of many of these can be tion. Princeton: Princeton University Press. found in Bhargava 2006. Stiglitz, Joseph. 2006. Making Globalization Work. New York: Norton. U.K. Government. 2006. Stern Review: Economics of References Climate Change. London: Government of the Bhargava, Vinay. 2006. Global Issues for Global United Kingdom. Citizens: An Introduction to Key Development Wolf, Martin. 2004. Why Globalization Works. New Challenges. Washington, DC: World Bank. Haven: Yale University Press. Dervis, Kemal. 2005. A Better Globalization: Legiti- Worm, Boris, E. Barbier, N. Beaumont, J. Duffy, C. macy, Governance and Reform. Washington, DC: Folke, B. Halpern, J. Jackson, H. Lotze, F. Center for Global Development. Micheli, S. Palumbi, E. Sala, K. Selkoe, J. Friedman, Thomas. 2005. The World Is Flat: A Brief Stachowicz, and R. Watson. 2006. "Impacts of History of the 21st Century. New York: Farrar, Biodiversity Loss on Ocean Ecosystem Services." Straus and Giroux. Science 314 (5800): 787­90. xxv Abbreviations ALMP Active labor market program APEC Asia-Pacific Economic Cooperation BOP Balance of payments CGE Computable general equilibrium (economic models) CPIA Country Policy and Institutional Assessment DDA Doha Development Agenda DDP Development data platform DPRs Diversified Payment Rights EU European Union FCC Federal Communications Commission FDI Foreign direct investment GDP Gross domestic product GATT General Agreement on Tariffs and Trade GATS General Agreement on Trade in Services GHG Greenhouse gas GTAP Global Trade Analysis Project IMF International Monetary Fund IPPC Intergovernmental Panel on Climate Change ITA Information Technology Agreement MDG Millennium Development Goal MNEs Multinational enterprises NAFTA North American Free Trade Agreement NIEs Newly industrialized economies ODA Official development assistance OECD Organisation for Economic Co-operation and Development OPEC Organization of the Petroleum Exporting Countries PPP Purchasing Power Parity RTAs Regional trade agreements TAA Trade adjustment assistance TFP Total factor productivity UNCTAD United Nations Conference on Trade and Development WHO World Health Organization WTO World Trade Organization xxvii 1 Prospects for the Global Economy Summary of the medium-term growth is projected to slow over the next two outlook years, it should remain robust at 6.1 percent in D espite high commodity prices, rising 2008. Mainly because of the continued expan- short-term interest rates, and a bout of fi- sion of developing economies, global growth nancial market volatility, global growth accel- will remain robust and this should keep com- erated in the first half of 2006. While there are modity prices high. Nevertheless, increases in indications that the pace of the expansion is supply, combined with demand-side substitu- already slowing, developing economies are tion and conservation measures, should allow projected to expand by 7 percent for the year for some easing of commodity prices, includ- as a whole, more than twice as fast as high- ing that of oil. income countries (3.1 percent), with all devel- This positive outlook is subject to signifi- oping regions growing by close to or more cant risks. Past episodes of fast growth and than 5 percent. favorable financial conditions have been fol- The very fast growth of developing coun- lowed by sharp and largely unanticipated re- tries over the past five years has been fueled versals. While stronger fundamentals in most by low interest rates and abundant global liq- developing countries reduce the likelihood uidity. This has led to rising commodity prices that a hard landing would be as severe as in and overheating in some high-income and the past, countries need to take particular developing countries. This, in turn, has pro- care to ensure that their fiscal, monetary, and voked a tightening of monetary policy that is structural policies are in order so as to mini- in part responsible for the slowdown that has mize the domestic consequences of external already begun. However, in most countries shocks--a point driven home by the financial strong productivity growth, due in part to the market turbulence observed in the spring of absorption of China and the former Eastern 2006, which affected most sharply those Bloc countries into the global economy, has countries whose fundamentals were most out checked inflationary pressures. of balance. Limited inflationary pressures and high A soft landing remains likely, but the savings among oil exporters and in Europe (as global economy has reached a turning point Europeans prepare to meet the challenges of and many factors could result in a more pro- their aging society) are expected to keep long- nounced slowdown. A faster-than-expected term interest rates low. Moreover, improved weakening of housing markets in high- fundamentals have boosted trend growth rates income countries could generate a much in many developing countries. Together these sharper downturn and even recession, with factors suggest that, while developing-country potentially significant effects for developing 1 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 countries. Much slower growth would likely cause commodity prices to weaken more than Figure 1.1 Industrial production may be slowing already projected, potentially placing many developing countries that have so far avoided 3-month annualized growth rate current-account problems in difficulty. In ad- Percent Percent dition, demand is expanding unsustainably 12 30 Developing countries, China (right axis) rapidly in many developing countries. Should excluding China 25 10 efforts to contain growth in these countries 20 8 fail, their economies could overheat, yielding 15 initially stronger growth outcomes and addi- 6 10 tional inflation, but a much sharper slow- 4 5 down later on. An oil-sector supply shock 0 2 could be similarly disruptive, driving up oil 5 0 prices even further, while simultaneously 10 High-income countries slowing growth and weakening the prices of 2 15 Jan. July Jan. July Jan. July non-oil commodities. Finally, although global 2004 2004 2005 2005 2006 2006 imbalances appear to be stabilizing, they re- Source: World Bank. main large. There is a continuing risk that they will be resolved in a more disruptive manner than is assumed in the baseline sce- nario outlined here. grew 5.5 percent (5 percent for small oil exporters), and all regions are expected to have grown by close to, or more than, 5 percent. Global growth surged to Most of the acceleration in global growth 3.9 percent in 2006 was concentrated in the first half of the year. D espite oil prices that topped $75 a barrel World industrial production grew 6.7 percent during the course of the year, world gross in the first six months of 2006, compared with domestic product (GDP) growth is estimated 4.3 percent in 2005 (figure 1.1). Among de- to have strengthened in 2006, coming in at veloping countries, rates of growth of indus- 3.9 percent, compared with 3.5 percent in trial production eased in the second and third 2005 (table 1.1). To a significant degree, this quarters, although this was partially offset by strong global performance reflects the very stronger growth among high-income coun- rapid expansion in developing economies, tries. Order books and business sector confi- which grew by 7 percent--more than twice as dence are strong in both Europe and Japan, fast as high-income countries (3.1 percent). suggesting that industrial activity should re- Overall, 38 percent of the increase in global main robust for the remainder of the year, output originated in developing countries, far while in the United States there are clear signs exceeding their 22 percent share in world GDP. that industrial production is slowing. As discussed in chapter 2, continued faster In the United States, the acceleration in in- growth among developing countries over the dustrial output was mirrored by GDP, which next two decades is expected to lift their share began 2006 expanding by a torrid 5.6 percent. of world output to about 31 percent in 2030. However, responding to higher short-term Very strong growth (10.4 percent) in China interest rates, residential investment spending played a significant role in the recent strength of has fallen sharply and a cooling housing mar- developing countries, contributing an expected ket has moderated consumer demand.1 As a 0.5 percentage points to global growth. Never- result, the economy slowed in the third quar- theless, the pickup was broadly based. Even ex- ter to a 1.6 percent annualized growth rate, cluding China and India, developing countries with most of the slowdown restricted to the 2 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y Table 1.1 The global outlook in summary Percentage change from previous year, except interest rates and oil price Estimate Forecast 1960­80 1980­2000 2004 2005 2006 2007 2008 2008­30 Global conditions World trade volume -- 5.8 10.4 7.7 9.7 7.5 7.8 Consumer prices G-7 countriesa,b 1.8 2.2 2.5 2.1 1.7 United States 2.7 3.4 3.4 2.5 2.1 Commodity prices (US$) Non-oil commodities 17.5 13.4 20.6 4.5 8.4 Oil price (US$ per barrel)c 37.7 53.4 64.0 55.9 52.7 Oil price (percent change) 30.6 41.5 19.9 12.7 5.7 Manufactures unit export valued 6.9 0.8 2.4 3.8 0.4 Interest rates $, 6-month (percent) 1.6 3.6 5.4 5.7 5.0 , 6-month (percent) 2.1 2.2 3.0 3.6 4.2 Real GDP growthe World 4.7 3.0 4.1 3.5 3.9 3.2 3.5 2.9 Memo item: World (PPP weights)f 5.2 4.7 5.1 4.5 4.6 High-income 4.5 2.9 3.3 2.7 3.1 2.4 2.8 2.4 OECD countries 3.2 2.6 3.0 2.3 2.7 Euro Area 1.7 1.4 2.4 1.9 1.9 Japan 2.7 2.6 2.9 2.4 2.5 United States 4.2 3.2 3.2 2.1 3.0 Non-OECD countries 6.4 5.8 5.3 4.7 4.8 Developing countries 6.2 3.4 7.2 6.6 7.0 6.4 6.1 4.0 East Asia and the Pacific 5.5 8.5 9.0 9.0 9.2 8.7 8.1 5.1 China 10.1 10.2 10.4 9.6 8.7 Indonesia 5.1 5.6 5.5 6.2 6.5 Thailand 6.2 4.5 4.5 4.6 5.0 Europe and Central Asia 10.7 0.6 7.2 6.0 6.4 5.7 5.5 2.7 Poland 5.3 3.4 5.4 5.1 5.2 Russian Federation 7.2 6.4 6.8 6.0 5.5 Turkey 8.9 7.4 6.0 5.0 5.0 Latin America and the Caribbean 5.5 2.2 6.0 4.5 5.0 4.2 4.0 3.0 Argentina 9.0 9.2 7.7 5.6 4.0 Brazil 4.9 2.3 3.5 3.4 3.8 Mexico 4.4 3.0 4.5 3.5 3.5 Middle East and North Africa 5.9 4.0 4.8 4.4 4.9 4.9 4.8 3.6 Algeria 5.2 5.3 3.0 4.5 4.3 Egypt, Arab Rep. of 4.2 4.9 5.8 5.6 5.8 Iran, Islamic Rep. of 5.1 4.4 5.8 5.0 4.7 South Asia 3.7 5.4 8.0 8.1 8.2 7.5 7.0 4.7 Bangladesh 6.3 6.2 6.7 6.2 6.5 India 8.5 8.5 8.7 7.7 7.2 Pakistan 6.4 7.8 6.6 7.0 6.5 Sub-Saharan Africa 4.4 2.2 5.2 5.5 5.3 5.3 5.4 3.3 Kenya 4.9 5.8 4.9 5.1 4.9 Nigeria 6.5 6.2 4.8 5.1 5.4 South Africa 4.5 4.9 4.6 3.9 4.3 Memorandum items Developing countries excluding transition countries 5.1 4.2 7.3 6.8 7.0 6.4 6.1 excluding China and India 6.6 2.3 6.1 5.1 5.5 4.9 4.9 Source: World Bank. Note: PPP purchasing power parity. a. Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States. b. In local currency, aggregated using 2000 GDP weights. c. Simple average of Dubai, Brent, and West Texas Intermediate. d. Unit value index of manufactured exports from major economies, expressed in U.S. dollars. e. GDP in 2000 constant dollars; 2000 prices and market exchange rates. f. GDP measured at 2000 PPP weights. 3 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 housing sector. Importantly profits, non- estimated to have expanded by 2.9 percent in residential investment, and consumption re- 2006. A slowdown in exports contributed to main robust and inflation and unemployment weaker growth in the second quarter of the low. As a consequence, although growth is ex- year, but growth rebounded in the third quarter pected to remain subdued, it should not de- led by a surge in investment spending. As of cline in the fourth quarter and the strong first August, exports were up 11 percent from a year quarter means that output for the year as a earlier, partly reflecting a 25.6 percent increase whole is expected to increase 3.2 percent. in sales to China, where import volumes have In high-income Europe, following several strengthened markedly. years of weakness, growth also accelerated in Developing economies grew 7 percent in the first half of 2006. GDP expanded by 2006. Much stronger European and continued about 3.3 percent in the first two quarters of robust Japanese growth, combined with low the year, as private consumption and invest- real interest rates and interest rate spreads, ment spending took over from exports as the made for robust activity among the world's main drivers of the recovery. Growth slowed developing economies, which are expected to to a 2 percent pace in the third quarter, with have expanded by 7 percent in 2006. This rep- growth in France having stalled as invest- resents the fourth consecutive year that their ment expenditures turned negative and ex- growth rates have exceeded 5 percent. ports weakened. However, both in France The expansion was particularly robust in and in the rest of Europe, consumer demand China and India, where output is estimated to remained robust and consumer and business have increased by 10.4 and 8.7 percent, res- surveys suggest that economic activity should pectively. But the strong performance was be robust in the fourth quarter, leading to an broadly based, with all developing regions estimated 2.5 percent increase in GDP for the growing by close to or by more than 5 percent year as a whole (2.4 percent for the Euro (figure 1.2). Despite further substantial in- Area). creases in the price of oil during the first half In Japan, the acceleration in output of the year, growth among the remaining oil- that began in 2005 has continued, with GDP importing developing countries actually Figure 1.2 Regional growth trends Percent change in GDP 10 2004 9 2005 2006 8 2007 7 2008 6 5 4 3 2 East Asia and Europe and Latin America Middle East South Asia Sub-Saharan Oil exporters Oil importers the Pacific Central Asia and the and North Africa (excluding Caribbean Africa China) Source: World Bank. 4 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y strengthened and is expected to come in at declines toward 3.5 percent of the labor force. 5 percent for the year as a whole. As a result, the current account surplus should decline to about 3 percent of GDP in 2008. Developing economies to outperform In most developing regions, high oil prices, high-income countries in 2007­08 rising interest rates, and the maturation of the High oil prices are expected to continue to business cycle are expected to restrain growth weigh on growth in industrialized countries. in 2007­08. As a group, however, low- and The slowdown that they and higher interest middle-income countries should again out- rates (working through residential investment perform high-income economies by a wide and household consumption) have already ini- margin--and this strong performance will tiated in the United States is projected to con- continue to be a critical driver of global tinue into the first half of 2007 before an ex- growth. Administrative restrictions on invest- pected relaxation of monetary policy permits ment and slower export growth are expected the economy to pick up. Overall, GDP in the to bring Chinese growth down to a more sus- United States is projected to increase 2.1 per- tainable 8.7 percent by 2008. Higher interest cent in 2007 and 3 percent in 2008. Weaker rates and some further fiscal tightening are ex- domestic demand is expected to be reflected in pected to slow the expansion in India to about slower import growth and should result in a 7.2 percent over the same period, helping to decline in the trade and current account unwind some of the inflationary tensions that deficits of the United States, with the latter have built up in that country. coming in around 5.5 percent of GDP in 2008. Prospects for the remaining oil importers Continued accommodative macroeconomic are varied. Many, particularly in Eastern and policy and pent-up investment demand follow- Central Europe, are overheating and have ing several years of very weak growth should entered a phase of policy tightening. These maintain the pace of the expansion in most countries are expected to decelerate. Others, European countries, without exacerbating including Brazil and Mexico, are projected to underlying inflationary pressures. However, a accelerate or enjoy high but stable growth planned 3 percent increase in the German rates as they continue to benefit from a favor- value-added tax (VAT) is projected to slow able external climate, including low long-term demand in that country in 2007, with knock- real interest rates and interest-rate spreads. on effects elsewhere in the continent. The Overall, developing-country oil importers, ex- higher VAT can also be expected to prompt cluding China and India, are projected to a one-time increase in inflation, although its enjoy broadly stable growth of about 4.8 per- effect should be attenuated by slower growth. cent in 2007­08. Overall, GDP in high-income Europe is pro- For oil exporters (and other large com- jected to slow to about 2.1 percent (1.9 percent modity exporters) strong revenue inflows for the Euro Area) in 2007 and 2008. should continue to fuel robust domestic de- In Japan, vigorous growth in developing mand growth despite lower prices and less East Asia, renewed consumer and business con- rapid increases in global demand for com- fidence, and reduced drag from consolidation modities, resulting in rapid growth of both are positive factors expected to maintain imports and the noncommodity sectors of growth at about 2.5 percent in 2007 and 2008. these economies. Overall, the pace of the ex- The recent return to positive inflation is pro- pansion in developing-country oil exporters jected to persist, allowing short-term interest is expected to decline from 6 percent this year rates to gradually rise to around 2 percent by to 4.9 percent in 2008, as capacity constraints the end of 2008. At the same time, domestic de- slow growth in the resource sector and a rising mand is expected to firm as unemployment share of demand is met by imports. 5 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Regional outlooks Malaysia and the Philippines is also expected More detailed descriptions of economic develop- to come in at around 5.5 percent, while in ments in developing regions, including regional fore- Thailand it is expected to reach only about 4.5 cast summaries and country-specific forecasts, percent, because, despite strong export growth, are available online at http://www.worldbank.org/ globaloutlook.Country-specific forecasts are reported consumption and investment have been de- in the appendix. pressed by high oil prices, rising interest rates, and continued political uncertainty. East Asia and the Pacific2 High oil prices and rapid growth have An emerging growth pole. raised inflation in the region, prompting a The economies of the East Asia and Pacific general tightening of monetary policies during region continued to expand at robust rates in 2005. As a result, both headline and core in- 2006, with regional GDP growth expected to flation are now easing in most of the larger accelerate to 9.2 percent in 2006 from 9 percent economies in the region. Regional equity mar- the year before. In China, continued rapid in- kets were subject to the general correction vestment growth, in conjunction with an unex- affecting many emerging markets during pected surge in exports as new capacity came on May­June 2006. However, spreads on bonds stream, led to a 10.7 percent year-over-year in- remain low, and equity markets began recov- crease in GDP over the first three quarters. The ering in August, suggesting that the earlier overall contribution of the external sector to correction did not represent a reassessment of GDP growth was up because, even though im- the region's economic fundamentals. port growth accelerated, increased output from Growth is expected to moderate only some- China's import-competing sectors prevented what. In China, investment growth and do- import volumes from expanding as rapidly as mestic demand are projected to remain robust. exports. Investment spending has been spurred However, with investment at some 50 percent by growth in credit and the money supply as of GDP, more forceful policy action may be well as strong profits. Concerns about exces- needed to keep credit and investment growth in sive investment creating potential overcapacity check. Moreover, the country's large and per- in specific sectors led the authorities to rein- sistent current account surplus suggests the force administrative measures aimed at con- need, over the longer term, to promote a more taining investment growth. consumption-oriented pattern of growth. In The expansion in the rest of the region re- Indonesia, the ongoing recovery in growth is mains robust. A rebound in global high-tech de- projected to continue, with GDP expanding by mand and stronger import demand from China 6.5 percent in 2008. prompted an acceleration in exports that began Economic pressures for the revaluation of in the second half of 2005 and continued into developing Asian currencies are likely to 2006. Vietnam's growth is expected to reach 8 intensify. In addition to reducing global percent, backed by across-the-board strength in imbalances, revaluation would also reduce in- exports, domestic consumption, and invest- flationary pressures, improve domestic macro- ment. In Indonesia, growth slowed in the first economic management capabilities, steady six months of 2006 following a substantial re- asset markets, and improve living standards duction of fuel subsidies and monetary tighten- for local populations. ing in the latter part of 2005. Activity appears Finally, the region remains susceptible to to be picking up now, with monthly indicators outside risks, including a worsening of the suggesting a strong rebound in domestic con- avian influenza epidemic--either through sumption and investment in the third quarter. wider effects on domestic animals or because For the year as a whole, GDP is expected to in- transmission to (or between) humans becomes crease by about 5.5 percent. Growth in more efficient (World Bank 2006). 6 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y Europe and Central Asia investments, more volatile capital flows are Oil exporters and European Union integration also increasing, and a substantial portion of underpin strong growth. the FDI is going into the banking sector, where Economic activity in the Europe and Cen- it may be more volatile than in other sectors. tral Asia region is estimated to have increased While such flows are likely to remain strong, by 6.4 percent in 2006, up from 6 percent in motivated by investment opportunities associ- 2005. This acceleration comes despite slower ated with European Union (EU) integration, growth in Turkey, where a significant tighten- the real-side disequilibrium that these inflows ing of monetary policy following this spring's are provoking may make these countries sen- financial market turbulence is projected to re- sitive to a change in investor sentiment. In- duce growth from 7.4 percent last year to deed, as events in the spring of 2006 high- 6 percent in 2006. lighted, countries with large current account Faster growth in Europe and low real inter- deficits are particularly vulnerable--especially est rates have helped to maintain growth at those with pegged exchange rates (Hungary high levels elsewhere in the region. Among the and Latvia) and currency boards (Bulgaria, largest economies, growth in the Russian Fed- Estonia, Lithuania)--because sharp reduc- eration is estimated to have picked up to tions in inflows may require very large and 6.8 percent in 2006, supported by high oil disruptive real-side adjustments. In Hungary, prices. Improved incomes and activity in the a budget deficit of close to 10 percent of GDP mining sector boosted growth in Ukraine to an poses further challenges. estimated 6 percent pace in 2006 versus Excess demand has also contributed to in- 2.6 percent in 2005, while rising wages and flationary pressures and a tightening of mone- falling unemployment increased growth in tary policy. Overheating remains a risk both Poland from 3.4 percent in 2005 to an there and in Azerbaijan, the Baltic states, estimated 5.4 percent in 2006. Elsewhere, Belarus, Bulgaria, the Czech Republic, rebounding demand in industrial Europe, cou- Kazakhstan, Romania, Russia, the Slovak pled with rapidly growing demand from re- Republic, Turkey, and Ukraine. Other factors gional oil exporters, notably Russia, bolstered contributing to the slower growth include a exports among smaller oil importers, whose slump in manufacturing activity (especially economies grew 6.1 percent, up from 5.8 per- mining) in Armenia, a marked deceleration of cent in 2005. Higher oil prices and the coming export growth in Latvia, and rising capacity on stream of oil projects lifted the GDP growth constraints combined with declining competi- among oil exporters to 7.3 percent. tiveness in Belarus. Growth is continuing at a The pace of demand growth in many coun- modest pace in the former Yugoslav Republic tries in the region continues to exceed supply of Macedonia (4 percent in 2006), where do- and, as a result, 13 countries have current- mestic demand is recovering slowly following account deficits in excess of 5 percent of GDP, a period of fiscal consolidation and violent and inflation is rising in 12. Strong capital in- conflict in 2001. flows, predominantly in the form of foreign Growth in the region is expected to slow direct investment (FDI), coupled with ex- somewhat over the next two years, coming in tremely rapid domestic credit expansion, are at about 5.5 percent in 2008. Slower growth at the root of excess demand in several coun- in industrialized Europe and higher short-term tries (the Baltic countries, Bulgaria, Hungary, interest rates are expected to cause regional Romania, the Republic of Serbia, and Turkey). export growth to decline for both oil im- Although FDI flows are less easily reversed porters and oil exporters. In the case of the than portfolio and equity investments and are latter, domestic demand growth is expected to more likely to be associated with physical ease but remain strong, because, although oil 7 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 revenues will decline, they will remain high. quarter, and although it slowed in the second Lower prices will contribute to the expected quarter, growth for the year as a whole is slowdown in oil exporters' GDP growth from expected at 3.5 percent. 7.3 percent in 2006 to 6.2 percent in 2008. In In contrast, demand in Argentina and addition, it will also slow demand for exports República Bolivariana de Venezuela, which from regional oil importers, which, in combi- had been expanding at unsustainable rates, nation with weaker export demand from slowed. Nevertheless, demand in each country Germany and the United States in 2007, is remains very strong, and GDP is projected to expected to reduce their GDP growth to about expand by 7.7 and 8.5 percent, respectively, 5.2 percent in 2008. well beyond potential. Unsurprisingly in these The combination of rising inflation and el- conditions, inflation has picked up and now evated current account deficits poses a persis- exceeds 10 percent in both countries. In each tent challenge for regional policy makers. To case, this surge occurred despite price freezes the extent that the contractionary influence of in a number of sectors that are hurting sectoral higher interest rates continues to be offset by investment and supply (inflation of uncon- capital inflows, further fiscal tightening may trolled goods and services is running at 16 per- be unavoidable--even if it means pushing gov- cent in Argentina). The rapid expansion of de- ernment balances into surplus in some coun- mand in República Bolivariana de Venezuela tries. For many countries in the Common- has been fueled by ballooning government wealth of Independent States, future prospects transfers. The growth in supply has been con- will be dependent on continued strong de- centrated in the non-oil sector, as reductions in mand from Russia. Prospects for the poorer investment by the government's oil company countries in the region will also depend on the and by private oil firms (discouraged by high extent to which these countries are able to taxes and royalties and antibusiness policies) strengthen domestic institutions so as to sus- have caused oil production to decline. tain high growth rates. Other countries in the region are also growing relatively rapidly. In Chile, a waning Latin America and the Caribbean investment boom and higher imports have Improved performance but still under- contributed to a slight slowing of growth in performing. 2006. In Central America, growth is expected Economic activity in Latin America and the to accelerate in most countries in 2006, Caribbean has picked up and GDP is esti- boosted by exports and investments associ- mated to have increased by 5 percent in 2006. ated with free trade agreements (Costa Rica, The faster growth reflects favorable interna- the Dominican Republic, El Salvador, tional financial conditions, strong commodity Guatemala, Honduras, and Nicaragua), strong prices, and a relaxation of monetary policy in remittance inflows, increased agricultural Brazil and Mexico, two of the region's largest production due to better weather, and post- economies. hurricane investment spurts (El Salvador). In In Mexico, GDP accelerated sharply in the the Caribbean, growth in Jamaica and the first half of 2006, growing 5.5 and 4.7 percent Dominican Republic has picked up, reflecting (year-over-year) in the first two quarters as foreign investments in the tourism and mining lower interest rates boosted domestic demand sectors (Jamaica) and a growth rebound fol- and construction activity. Stronger sales of cars lowing the 2003 currency crisis (Dominican to the United States and oil exports also con- Republic). Uncertainty over the results of tributed. Brazil, too, benefited from a more re- elections in Nicaragua has hurt investor laxed monetary policy stance, although real in- confidence, partially offsetting the beneficial terest rates remain high at 11 percent. GDP effects of a relatively buoyant agricultural sec- accelerated to about 5.2 percent in the first tor and the writing off of $975 million in debt. 8 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y Electoral uncertainty and concerns about as imports and inflation rise rapidly. Unless the future path of U.S. interest-rate policy con- significant policy restraint is introduced in tributed to volatility in financial markets in the near future, these developments will re- the spring of 2006. The currencies of a num- sult in a deterioration of current account bal- ber of countries depreciated, following earlier ances, leading to an erosion of Argentina's appreciations in some cases (Brazil and current account surplus to about 0.9 percent Colombia). Stock markets also underwent a of GDP and a much-reduced surplus of major correction. However, the improved fis- 7.6 percent in República Bolivariana de cal stance and reduced indebtedness of many Venezuela by 2008. The longer the two coun- countries meant that the region was not par- tries' aggressively expansionary macroeco- ticularly affected by this episode. Risk premia, nomic policies keep demand growing in after rising somewhat, have once again de- excess of supply, the sharper and more dis- clined and currently are just 20 points above ruptive will be the recession required to their historical minimums. reestablish equilibrium. Prospects for countries in the region reflect a number of offsetting influences. The pro- Middle East and North Africa3 jected slowdown in global activity is expected Riding the oil boom. to moderate demand for commodities, result- High oil prices and strong oil demand ing in a modest decline in their prices and continue to be key drivers for the developing slower volume growth. As a result, while rev- economies of the Middle East and North enues from this sector will remain elevated, Africa.4 Overall, these countries' GDP in- they will decline, as will their contribution to creased by an estimated 4.9 percent in 2006, the growth of domestic demand in commodity- the fastest pace in some four years. Among exporting countries. Overall, GDP among developing-country oil exporters, growth is commodity exporters (excluding República expected to reach 4.9 percent, up from last Bolivariana de Venezuela, see below) is pro- year's 4.7 percent. jected to slow to about 3.8 percent in 2008. Reflecting strong investment and remittance Commodity importers also will feel the effect flows from high-income oil exporters and of slower global and U.S. growth. In the case the Euro Area, output among regional oil im- of Mexico, the anticipated cycle in the United porters has strengthened. For the year as a States is expected to be reflected in slower whole, output is expected to come in at 5 per- exports and growth. For most commodity im- cent. Strong Suez Canal revenues in the Arab porters the slowdown is expected to be less Republic of Egypt, better crops following a marked (from 4.6 to 4 percent, excluding drought in the Maghreb, hefty tourism receipts Mexico), in part because many countries have throughout the region, and a pickup in considerable spare capacity. European demand are additional factors ex- As indicated above, unsustainably rapid plaining this relative strength. An exception is growth in Argentina and República Bolivariana Lebanon, where the war and political uncer- de Venezuela, boosted by a dangerously ex- tainty weighed heavily on activity in the first pansionary fiscal and monetary policy, has three quarters. While reconstruction efforts are already strained capacity in these countries. In expected to give a fillip to growth toward the the baseline projection, this unsustainable de- end of the year and into 2007/08, Lebanese mand stimulus is expected to continue, with GDP is expected to contract by about 5.5 per- domestic demand increasing at double-digit cent in 2006. rates. The inability of domestic supply to keep Generous fuel subsidies are pervasive pace means that GDP will grow less quickly, throughout the region. For countries that do declining to 4 percent in Argentina and 5.5 in not benefit from large oil revenues, these sub- República Bolivariana de Venezuela in 2008, sidies have strained fiscal balances. Countries 9 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 such as Egypt, Jordan, Morocco, and Tunisia High oil prices are expected to continue saw fiscal deficits pick up within a range of feeding domestic demand in oil-producing 0.5 to 2 percent of GDP over the course of countries, causing imports to continue rising 2005, in part linked to oil subsidies. Since rapidly, even as growth of export revenue then, Jordan reduced subsidy expenditures by slows. Capacity constraints and strong import 59 percent in the second quarter. In Egypt, growth is projected to slow GDP growth these and other steps have helped reduce among developing oil-exporting countries to the consolidated government deficit from 4.7 percent in 2007 and 4.5 percent in 2008. 9.1 percent in fiscal year 2005 to 6.5 percent Their current account surpluses should decline in 2006. Nevertheless, such subsidies remain from 11 percent of GDP in 2005 to about important in the region and threaten the fis- 5.3 percentofGDPin2008.Intheoil-importing cal sustainability of some countries. They economies, growth is expected to gradually also impede adjustment, although their bal- pick up from 5 percent in 2006 to 5.3 percent ance of payments consequences have been in 2008, reflecting assumed improvements in mitigated by strong remittance and invest- crop conditions, stronger European growth ment flows. and continued robust investment and remit- Rising oil prices during the first eight tance inflows from regional oil exporters. months of 2006 bolstered revenues of the major exporting countries in the region. Oil- South Asia related revenues were up 33 percent in the Rapid growth is pushing against capacity Islamic Republic of Iran and 30 percent in constraints. Algeria, and many governments boosted Despite a tightening of both monetary and spending. Measures included substantial fiscal policy, real interest rates remain low, and investments to augment oil-sector capacity, in- growth in the South Asia region picked up to an frastructure projects, and other non-oil-sector estimated 8.2 percent in 2006. Direct and indi- investments in human and social capital, all rect subsidization of consumer energy prices of which should help boost future supply. have helped contain inflationary pressures but However, a significant share of the additional are keeping government deficits high and con- spending, such as substantial civil service tributing to strong domestic demand. wage increases in several countries and in- With the exception of Nepal, which is only creased spending on fuel subsidies, merely now emerging from political strife, growth stoke demand and may prove difficult to sus- throughout the region was strong in the first tain should oil prices decline further. half of the year. In India, GDP increased by The surge in oil revenue and government 9.3 percent in the first quarter, supported by spending among oil exporters has yet to gen- strong industrial and service-sector produc- erate substantial inflationary pressures. How- tion, while in Pakistan industrial production ever, inflation is up in a number of countries, was up 12 percent. Partly reflecting improved including Egypt, Jordan, Oman, and Tunisia. sales of textiles and clothing after the restric- In the Islamic Republic of Iran, although in- tions on Chinese exports were reintroduced, flation is declining, it still exceeds 10 percent. merchandise exports in the region increased Moreover, regional stock and housing mar- more than 30 percent in the first half of 2006 kets have appreciated enormously throughout (on a year-over-year basis). A good start to the the region. While local markets lost as much monsoon season suggests that agricultural as 25­33 percent of their value in the output (and incomes) will be strong also. May­June 2006 market correction, valua- Overall, regional GDP is projected to increase tions remain high, and there are concerns by 8.2 percent for the year, or 6.4 percent if about increased leverage in private sector bal- the two largest economies (India and Pakistan) ance sheets (IMF 2006). are excluded. In the Maldives a rebound in 10 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y tourism and post-tsunami reconstruction ef- GDP growth should moderate to about 7 per- forts are expected to contribute to a 19 per- cent by 2008. Owing to continued strong cent expansion in GDP, while a new hydro- growth, the region's current account deficit is electric plant may help boost output in Bhutan projected to deteriorate further despite falling by 10 percent. oil and non-oil commodity prices. Notwithstanding robust export demand and a first-quarter current account surplus Sub-Saharan Africa in India, strong domestic demand is expected Another good year. to result in a small further deterioration of GDP in Sub-Saharan Africa expanded by an regional current account deficits in 2006, estimated 5.3 percent in 2006. Oil-exporting particularly in Pakistan, where increased gov- economies are expected to grow 6.9 percent ernment spending (tied to the Kashmir earth- in 2006, about the same as last year. Among oil quake and ongoing military expenditures) is importers (excluding South Africa), the expan- projected to push the current account deficit sion has been sustained, and growth is esti- to 3.9 percent of GDP. In contrast, strong re- mated to have increased 4.7 percent. mittance flows and robust exports are ex- South Africa, the region's largest economy, pected to propel the current account of expanded at growth rates above its potential for Bangladesh toward balance. the third consecutive year. Household expendi- Rapid growth and the relatively expansion- ture has been exceptionally strong, benefiting ary stance of fiscal and monetary policies in from low nominal interest rates, rising real in- the region have provoked a rise in inflation. comes, and wealth effects. As a result, domestic Successive hikes in policy rates in India have demand and output growth in the manufactur- increased interest rates, but higher inflation ing and service sectors have been very strong. means that real rates were negative in August. Despite a large positive terms-of-trade shock as In Pakistan, tighter monetary policy brought metal prices soared, the external sector's contri- inflation down to 6.2 percent in April, but it bution to growth was negative, owing to strong picked up again and was 8.1 percent in Octo- import growth fueled by robust household con- ber. Ample domestic and international liquid- sumption and the stronger rand. The surge in ity has also contributed to substantial in- imports caused the current account deficit to creases in local stock market valuations (up deteriorateto6.2percentofGDPinthefirsthalf about 45 percent in both India and Pakistan). of 2006, which contributed to the sharp depre- Throughout the region, higher international ciation of the rand during May­June. Overall, oil prices have yet to be fully passed through the rand has depreciated 20 percent on a trade- to consumers, placing a strain on government weighted basis since the beginning of the year, accounts and implying significant additional and this has contributed to inflationary pres- inflationary pressure in the pipeline. sures. Nevertheless, consumer confidence and Despite tighter monetary and fiscal policies demand remain at historically high levels. in India and Pakistan, and weaker export In Nigeria, the region's second-largest demand from the United States, low real economy, attacks on oil infrastructure slowed interest rates, strong international capital growth, as oil production fell by 25 percent inflows, and high government deficits are during the first five months of 2006. It has expected to keep domestic demand expanding since picked up but remains down 6.5 percent rapidly. When added to the delayed pass- from a year ago. Nevertheless, the non-oil through of higher oil prices, this should main- economy is expanding rapidly (up 12.8 per- tain inflation pressures in the region and cent in the second quarter) and supplementary sustain rapid import growth. As a result, the budgetary spending is expected to bolster external sector is expected to make a significant growth, perhaps leading to a buildup in infla- negative contribution to growth, and regional tionary pressures. 11 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 The regional expansion is broadly based, Inflation is up in a number of countries be- with a third of the countries experiencing cause of higher international oil prices and sea- growth in excess of 5 percent. Among oil sonal and drought-induced increases in food exporters, growth was particularly strong in prices. In the case of South Africa these factors Angola (16.9 percent), Sudan (11.8 percent), have been exacerbated by excessive domestic and Mauritania (17.9 percent), which began oil demand. Inflation there is projected to exceed production in February. In addition to a strong 6 percent, above the upper limit set by the expansion in oil production, buoyant domestic Reserve Bank. In Nigeria, year-over-year infla- demand is projected to spur rapid growth in the tion remains high but is declining rapidly, non-oil sectors of most oil-exporting countries. owing in part to the appreciation of the naira. The aggregate stability and strength of GDP growth for the region as a whole is growth among the region's oil importers re- projected to remain broadly stable, coming in flects divergent patterns. A number of coun- at about 5.4 percent in 2008, with weaker tries that recently emerged from conflict are growth in South Africa offsetting a pickup experiencing very strong growth rates among oil exporters and stable growth (Burundi, the Democratic Republic of Congo, among smaller oil importers. In South Africa, Liberia, and Sierra Leone). Elsewhere strong higher interest rates are projected to over- international metal and mineral prices are come strong mining and manufacturing generating revenue streams and prompting ad- growth in 2007, before the latter forces gen- ditional investments, which have contributed erate a recovery in 2008. In the baseline pro- to strong growth. However, drought-related jection, emerging electrical shortages due to crop failure, high fuel costs, and energy ra- insufficient generating capacity are expected tioning have resulted in weaker results for to constrain growth in Burundi, Kenya, East African oil-importing countries. In addi- Malawi, Rwanda, Tanzania, Uganda, and tion, although both the numbers and the in- Zambia, but improved rainfall in East and tensity of conflicts in Africa have subsided, the West Africa should help replenish hydroelec- risks associated with political turmoil remain tric dams, thereby improving electrical supply high and are undermining growth in Chad, and manufacturing output. An end to Côte d'Ivoire, the Democratic Republic of drought should also boost agricultural output Congo,Eritrea,Lesotho,theSeychelles,Somalia, and domestic incomes, although weaker agri- Swaziland, and Zimbabwe. cultural prices and high fertilizer prices may Current accounts have come under pres- negatively affect agricultural crops and could sure in several oil-importing economies in the represent a drag on growth. region, although higher commodity prices and increased official and private transfers have helped contain the deterioration. The most Financial markets notable decline in the current account came in Some signs of emerging inflationary South Africa, where an initial appreciation of pressures the rand boosted imports and dampened ex- High oil prices and the rapid pace of global ports, driving the current account deficit to growth have contributed to a gradual increase 6.2 percent of GDP in the first half of 2006. in median inflation among developing coun- Debt relief from Paris Club creditors under the tries, from about 1.7 percent in 2002 to Multilateral Debt Relief Initiative should re- 3.2 percent during the third quarter of 2006 duce debt-servicing cost by substantial mar- (figure 1.3). The acceleration was not consis- gins in several countries, which, in combina- tent across the globe. Inflation has been stable tion with the expected easing in oil prices, is or declining in half of the developing regions projected to provide some relief to current over the past year, falling to an average accounts over the projection period. level of 5.3 in the third quarter. In contrast, in 12 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y Figure 1.3 Inflation has increased Figure 1.4 Inflation is rising in moderately high-income countries End of period, year-over-year Percent change, year-over-year monthly inflation rate (%) 5.0 U.S.: All goods 10 4.5 Sept. 2006 and services 9 4.0 2005 EU: All goods 8 2004 3.5 and services 7 2003 3.0 6 2002 2.5 5 2.0 4 1.5 3 1.0 EU core inflation 2 0.5 U.S. core inflation 1 0 Jan. Jan. Jan. Jan. Jan. 0 2002 2003 2004 2005 2006 Asia and and and Asia Asia East Pacific East AfricaSouth Africa Sources: World Bank; Datastream. the Europe High-income countries America Caribbean Sub-Saharan Central and MiddleNorth Latin the Source: World Bank. Figure 1.5 Signs of overheating in some developing countries high-income countries it rose from 1.3 to 2.7 percent before falling to 1.4 percent in Percent change in consumer prices, year-over-year monthly October as oil prices declined. 16 Botswana Most of the increase appears to reflect the di- 14 rect impact of higher oil prices. Until recently, 12 Argentina core inflation in high-income countries has been Turkey 10 relatively stable (figure 1.4). Core inflation in Bulgaria the United States was rising much of the year, 8 but it has been easing recently and stood at 2.7 6 percent in October 2006. In many developing 4 countries, inflation first picked up in response 2 India to higher oil prices, but it has since declined, re- Thailand South Africa 0 flecting both solid productivity growth and the Jan. Sept. May Jan. Sept. 2004 2004 2005 2006 2006 impact of more credible monetary policies that have helped anchor inflation expectations. Source: World Bank. However, developments in a number of low- and middle-income countries run counter to these general trends. In these countries, infla- rapid growth, high food prices following suc- tion is rising, reflecting the combined influence cessive droughts are playing a role. The con- of several years of above-trend growth and cern is that if an inflationary spiral develops, steep increases in global commodity prices because the credibility of monetary policy is (figure 1.5). Higher inflation would appear to not yet well entrenched, it could have serious reflect overheating in Argentina and several consequences for macroeconomic stability and countries in Europe and Central Asia, the affect the ability of those economies to sustain Middle East, North Africa, and South Asia. the strong growth of the past several years. Inflation has also picked up in Sub-Saharan In countries such as India, regional imbal- Africa. There, in addition to several years of ances in the distribution of growth contribute 13 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 to difficulties because inflationary pressures bonds and nonindexed bonds, which has re- and capacity constraints are concentrated in mained relatively stable at around 2.5 percent- rapidly growing cities and co-exist with con- age points. siderable slack elsewhere in the economy. More volatile financial conditions Rising short-term interest rates for developing countries Higher inflation throughout the developed Despite high short-term interest rates, finan- world has translated into rising short-term cial conditions for developing countries re- interest rates and the gradual removal of the main highly favorable. Several years of very monetary policy stimulus that has character- loose monetary policy, an ample supply of ized the past several years. Although policy global savings (due to aging populations in rates are increasing throughout the developed Europe and rapidly increasing incomes world, the process is most advanced in the among oil exporters), business-sector consoli- United States, while at relatively early stages dation in the United States and Asia, and in Europe and Japan. Many developing coun- high savings rates in the fastest-growing sec- tries also have acted to restrain credit expan- tors of the world economy have combined to sion and contain inflation. Policy rates have buoy global liquidity. This helps explain the risen sharply and appear to be slowing infla- low long-term bond yields and the search for tion in Bulgaria, Indonesia, Thailand, and yield that has boosted the flow of private Turkey. In others (Argentina, India, Pakistan) capital to developing countries. That flow, in the tightening cycle is less advanced and, as a combination with improved fundamentals, result, real interest rates remain low and in- has brought interest spreads down to histori- flation high. cally low levels.5 Long-term interest rates (see figure 1.8) re- However, conditions have become more main low and the yield curve flat, suggesting volatile (figure 1.6). The transition from a slow that markets are confident that the monetary and widely anticipated tightening of U.S. mon- authorities will be successful in containing etary policy toward a more data-driven ap- inflationary expectations--a contention sup- proach increased uncertainty in financial mar- ported by the spread between inflation-indexed kets during the second quarter of 2006. Initially, Figure 1.6 Despite turbulence, financing conditions remain favorable Emerging-market spread as compared with 10-year U.S. T-bill a. Long-term decline b. Recent volatility 1,600 400 Global 1,400 350 Global 1,200 300 1,000 250 Turkey Turkey 800 200 Zoom in 600 150 400 100 200 50 South Africa South Africa 0 0 Jan. Jan. Jan. Jan. Jan. Jan. May Sept. Jan. May Sept. 1998 2000 2002 2004 2006 2005 2005 2005 2006 2006 2006 Sources: Datastream; JP Morgan; World Bank. 14 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y the prospect that dollar-denominated returns some countries, particularly those with large would no longer be rising resulted in a surge of current account deficits (such as South Africa flows into emerging market stocks, commodity and Turkey) and relatively heavy debt bur- markets, and currencies. As values of these dens, the correction was more severe and is assets rose, the market reassessed long-term expected to result in much slower growth in prospects, resulting in a substantial correction. 2006 and 2007, as the real side of these For the most part, this volatility failed to economies adjusts to weaker financial inflows. disrupt growth, and net private capital flows to The combination of several years of low in- developing countries are expected to rival last terest rates has increased global liquidity sub- year's record highs (figure 1.7). However, in stantially (see earlier versions of Global Eco- nomic Prospects and Global Development Finance). Despite the increase in short-term Figure 1.7 Private capital flows to interest rates, the persistence of low long-term developing countries remain strong interest rates, due in part to high savings rates among oil-exporting countries, has kept Total net private capital flows to developing countries global liquidity abundant. The OECD (2006) $ billions Projected estimates that depending on the measure em- 550 500 ployed, high-income liquidity exceeds histori- 450 cal norms by between 15 and 17 percent (fig- 400 ure 1.8). In the baseline, although interest 350 300 rates are projected to rise somewhat, liquidity 250 is projected to remain relatively abundant 200 and continue to be a factor behind strong 150 100 developing-country growth. 50 0 Global imbalances are stabilizing 19901991199219931994199519961997199819992000200120022003200420052006 The imbalances in global spending patterns that have characterized the world economy Source: World Bank. over the past five years began to show signs of Figure 1.8 Ample liquidity keeps long-term interest rates low a. Global liquidity b. Long-term interest rates Percent deviation from long-term trend Percent 20 15 14 U.S. 10-year T-bill 13 15 12 Monetary 11 10 measure 10 9 Credit 8 5 measure 7 6 0 5 4 Germany 10-year 3 5 government bond 2 1 10 0 1980 1985 1990 1995 2000 2005 1980 1985 1990 1995 2000 2005 Sources: OECD (liquidity measures); World Bank. 15 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 An important component of this story has Figure 1.9 Global demand shifts from the been the slowing in U.S. domestic demand United States to Europe and developing and the simultaneous increase in developing- countries country domestic demand (figure 1.9). This Contribution of domestic demand to growth rotation of growth, plus the lagged effect of Percent past depreciations, contributed to a 13 per- 9 cent increase in the volume of U.S. exports in 8 the first half of 2006, almost twice the growth 7 2004 rate for imports (7 percent). Nevertheless, the 6 2005 5 2006 U.S. trade balance declined further, in part 2007 4 because of very high oil prices during the first 2008 3 eight months of the year. The subsequent 2 decline in oil prices should reduce the value of 1 U.S. imports, resulting in an improved trade 0 United States Japan Europe Developing balance during the fourth quarter. Because of countries this strong volume performance and declining Source: World Bank. oil prices, global imbalances are not expected to deteriorate significantly over the projection period--in stark contrast to the recent past, when they deteriorated sharply each year (figure 1.10). stabilizing in 2006. Weaker domestic demand in the United States, the acceleration of eco- nomic activity in Europe, and continued Exchange rates are broadly stable recovery in Japan helped to stabilize the U.S. Despite the substantial financial flows required current account deficit, which is expected to to finance the U.S. current account deficit, the come in at about $850 billion, roughly the dollarhasremainedbroadlystableagainstmajor same share of GDP as in 2005. currencies during 2006--up about 2 percent Figure 1.10 A start to orderly adjustment? Current account balance, 2002, 2006, and 2008 $ billions 400 200 0 200 2002 2006 400 2008 600 800 Oil importers Oil exporters 1000 United Europe Japan Other high- East Asia China Europe and Other High- Low-income States income and the Central developing income exporters countries Pacific Asia importers exporters (excluding China) Source: World Bank. 16 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y surpluses (China and several oil exporters) Figure 1.11 Interest rate spreads support have resisted upward pressure on their curren- the dollar cies with respect to the dollar. 3-month and 10-year bond yield The depreciation against the euro occurred Percent despite a substantial premium currently being 6 offered on both short- and long-term U.S. 10-year U.S. yield bonds (see figure 1.11). With U.S. monetary 5 policy nearing or at the end of its tightening 4 cycle, these differences are expected to nar- row. In the baseline forecast, this narrowing 3 10-year euro yield and slower growth in the United States are 2 projected to cause the dollar to depreciate by 3-month euro yield a further 5 percent against the euro in each of 1 3-month U.S. yield 2007 and 2008, which should further facili- 0 tate the unwinding of global imbalances. Jan. July Jan. July Jan. July Jan. July However, should downward pressures be 2003 2003 2004 2004 2005 2005 2006 2006 more severe, the depreciation could be stronger Sources: World Bank; Datastream. or interest rates in the U.S. may have to rise further (see the section on risks). Currency developments for the remaining against the yen and depreciating by about 5 per- developing countries were dominated by the cent against the euro. In real-effective terms it reemergence of financial market volatility in has lost only 2 percent of its value. In part, this the first half of 2006 (figure 1.12). Downward relative strength is explained because many of pressure on the dollar toward the end of 2005 those countries running large current-account and at the beginning of 2006 saw the currencies Figure 1.12 Turbulence resulted in sharp depreciations for some developing countries Index of euro exchange rates Index Jan. 1, 2006 100 85 Brazil Indonesia 95 105 United States 115 Colombia Depreciation 125 Turkey 135 South Africa 145 Jan. Feb. March April May June July Aug. Sept. Oct. Nov. 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 Source: World Bank. 17 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 of many developing countries appreciate and China increased by 10 and 30 percent, re- strongly. For many countries that appreciation spectively. Trade flows weakened in the sec- was reversed in May and June as investors re- ond quarter but show signs of picking up once assessed their positions. While disruptive, and again in the third quarter. provoking a sharp rise in interest rates in some Over the medium term, growth in mer- countries, the volatility was short-lived and in chandise trade volumes is projected to ease to most cases merely served to unwind earlier about 9 percent, in line with slower global appreciations that had been out of step with GDP growth. The recent relative strength of countries' underlying fundamentals. U.S. export volumes is projected to persist (figure 1.13). Those volumes are projected to rise by more than 9 percent in 2007 and 2008 World trade as the cumulative effect of past and expected S tronger industrial activity was mirrored in future depreciations increase the international world trade. Merchandise trade growth competitiveness of U.S. products. For develop- grew 11 percent during the first eight months ing countries, weaker U.S. import demand of 2006, up from 6 percent the year before. should be partially compensated by stronger Most of the acceleration occurred in China, demand from Europe, but, overall, developing- Japan, and the United States and was concen- country export growth is projected to slow trated in the first quarter. Weaker U.S. con- from an estimated 12.2 percent in 2006 to sumption and investment demand, and grow- 10 percent in 2008, even as countries continue ing domestic demand in the developing world to increase their market share. combined with the lagged effects of past de- preciations to boost U.S. export volumes by Trade outlook--continued expansion an annualized rate of 13 percent in the first Developing-country trade reached a landmark half of 2006, compared with 7 percent in the in 2006. Following 25 years of solid growth, last half of 2005. Measured on the same basis the value of China's exports overtook those and over the same period, exports in Japan of the United States, making China the world's Figure 1.13 Rotation in global trade Merchandise trade volumes Percent change 20 Developing countries Developing countries (excluding China) 15 10 5 Europe 0 5 United States 10 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: World Bank. 18 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y Figure 1.14 China's exports exceed those of the United States Merchandise exports $ millions, log scale 10,000,000 Other developing countries High-income Europe 1,000,000 Brazil China 100,000 10,000 United States India 1,000 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Source: World Bank. second-largest exporter. Increasing exports in becoming privileged destinations for FDI, other developing countries, notably Brazil and both as an export platform for multinational India, have further increased the weight of companies and because they represent the developing countries in world trade (fig- fastest-growing market segment. ure 1.14). Over the long term, as these trends The extent to which other developing continue, the share of developing countries in countries will be able to take advantage of the world trade is projected to reach some 45 per- expected continued strong growth of China cent by 2030 (see chapter 2). and India (see chapter 2) will depend on their While the phenomenal success that China ability to expand exports. This requires has enjoyed in expanding its world market eliminating the anti-export bias in their incen- share since the introduction of market reforms tive framework, reducing costs of produced has increased competitive pressure on both services, and improving customs procedures developing and developed countries (see that undermine competitiveness. It also chapter 4), Chinese imports also have grown requires investments in transport systems to very rapidly (up 477 percent in value terms over reduce transit times (Newfarmer 2005) and in the past decade), and China is a growing desti- other forms of infrastructure, such as electri- nation for the exports of other developing coun- cal generators so as to facilitate the expansion tries (Dimaranan, Ianchovichina, and Martin of capacity. In addition, as discussed in chap- 2006). Sixty-three percent of China's imports ter 4, countries need to reduce rigidities are intermediate goods, 31 percent in the form in product, labor, and financial markets so of parts and components. Overall, 79 percent of that firms can react with agility to new China's imports are sourced from developing opportunities to expand the range of products countries. Partly as a result of China's rapid they produce and sell. increase in imports, the value of other develop- On the multilateral front, the suspension of ing countries' non-oil exports has risen by talks on the Doha Round in July 2006 poses 153 percent, and their global market share has significant challenges. Weakened confidence in increased by 2.3 percentage points. the multilateral system could lead to trade dis- In addition to these direct effects, the ex- putes, rising protectionist sentiment, and trade pansion of developing-country commerce diversion arising from proliferating bilateral means that these countries are increasingly and regional trade agreements. To capitalize on 19 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 progress already made in the Doha Round, gains in dollar terms, up 7 percent by the be- such as the offer to end agricultural export sub- ginning of November, but were broadly stable sidies by 2013, it is important that parties re- if expressed in euros. turn to the negotiating table with the necessary The biggest increases in metals prices were flexibility to conclude an ambitious deal. those of copper (up 64 percent), zinc (up 110 percent), and nickel (up 144 percent). High prices and continued strong demand Commodity markets have prompted significant destocking of S trong global growth, and especially the copper and nickel in China--suggesting that rapid expansion of output in developing prices may remain high even as supply disrup- countries, is largely responsible for the run-up tions ease. Nevertheless, stocks of other prod- in commodity prices over the past several years. ucts such as aluminum, lead, and tin have Improvements in technology and new discover- recovered, and their prices have eased or sta- ies are expected to preclude any major disrup- bilized, suggesting that a peak in the metals tions to growth over the long term (see chapter market may have been reached. 2), but increased demand for energy and other Financial sector activity also played a big natural resources may generate significant en- role in price developments during 2006. Re- vironmental pressures (see chapter 5). cent estimates suggest that more than $19 bil- lion flowed into retail commodity funds dur- Non-oil commodities ing the first eight months of the year, when Metals and minerals take off. prices were rising. More recently, these flows While oil-price increases have received the have reversed sharply. Outflows from ex- bulk of media attention, the rise in the price of change-trade commodities totaled $12 billion metals and minerals during the course of 2006 during the first two weeks of September, when has been much stronger (figure 1.15). Contin- prices were falling (Norman and Shen 2006). ued rapid growth in global output, speculative With global growth projected to slow some- demand, low stocks, and numerous supply what but remain strong, the overall metals and disruptions have pushed metals and minerals minerals index is expected to decline in 2007 prices up by some 48 percent since the begin- and 2008, but remain elevated. ning of 2006. Agricultural prices also posted Moderate gains in agricultural prices. Agricultural prices have risen an estimated 11 percent in 2006 compared with 2005, re- Figure 1.15 Diverging trends in flecting a weaker dollar and the impact of commodity prices higher energy and fertilizer prices. Other fac- Index, Jan. 2003 100 tors, such as crop-specific supply shortfalls 350 and droughts, low carryover stocks, and 300 Metals and minerals strong demand growth contributed to the 250 price increases. Real agricultural prices have increased 35 percent since their cyclical lows 200 in 2001.6 This increase is well below the in- Energy 150 creases in oils and metals, in part because 100 agricultural demand is less sensitive than Agricultural products 50 industrial demand to economic growth, and because agricultural supply responds more 0 2001 2002 2003 2004 2005 2006 quickly to increased demand and prices. High energy prices have contributed directly Source: World Bank. to the surge in the price of some agricultural 20 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y commodities that are either used as energy cost-push factors and because higher energy crops (biofuels) or compete with synthetic prices make biofuel more economically viable, products made from petroleum. The price of generating an additional source of demand sugar, which is being diverted to ethanol pro- for products such as maize and sugar cane. duction for automotive fuel, more than dou- Already, 20 percent of the U.S. maize crop and bled from late 2004 until early 2006, while 50 percent of the Brazilian sugar cane crop are that of natural rubber (a substitute for synthet- used to produce ethanol. Should this trend ics produced from petroleum products) rose continue, demand for other commodities, 60 percent between December and June 2006. especially grains, will increase to substitute for The price of maize, which is used as the feed- crops used for biofuels. stock for ethanol production in the United States, is expected to rise 8 percent in 2006. Oil market High energy costs also have contributed to Rising supply and slow demand cause prices increasing agricultural prices by raising the to ease. cost of fertilizers. This prompts an increase in After showing signs of stabilizing in the fall the cost of production of agricultural goods, of 2005, the price of oil shot up once again in but also induces a reduction in yields because the first half of 2006 (figure 1.16). Prices have farmers use less fertilizer. As a result, energy- since declined and were below $60 in late and fertilizer-intensive crops such as grains are November--bringing the price of oil below expected to show reduced yields and higher the level at which it began the year. Expressed prices in 2006. These factors, plus low carry- in euros, the price has declined 7 percent since over stocks and poor harvests in several im- the beginning of the year and stands at about portant producing areas, are projected to push the same level as before the hurricanes of last wheat prices up 28 percent in 2006 and even summer and fall. higher next year before production increases High prices slowed the growth in demand enough to rebuild stocks. In the case of rice, for oil despite the acceleration in economic higher costs and reduced yields are expected activity in 2006. Oil demand increased by to boost prices by 8 percent. 0.5 million barrels per day (mbpd) in the three In contrast, prices of fats and oils are ex- quarters of 2006, compared with 3.2 mbpd pected to be 5 percent lower, because markets appear to be well supplied. Drought in Kenya is keeping tea prices high (up 14 percent from Figure 1.16 Oil prices continue to rise 2005). Robust coffee prices are expected to av- erage 18 percent higher in 2006, a continuation Price of oil, World Bank average of the price increases that began in 2002. Timber $ and per barrel Nov. 16, 2006 prices are projected to increase 14 percent owing 75 to strong demand, particularly from China, 65 while international supplies remain limited be- cause of controlled logging and export quotas. 55 Dollars per barrel Prospects for agricultural prices in 2007 45 are mixed, with grains and oilseeds higher, Euros per barrel while beverages and raw materials prices will 35 be lower. Overall, agricultural prices are ex- 25 pected to decline by about 1 percent in 2007 and 2.8 percent in 2008 as higher prices begin 15 Jan. July Jan. Aug. Feb. to moderate demand and induce increased 2000 2001 2003 2004 2006 supply. Should oil prices rise further, agricul- Sources: Datastream; World Bank. tural prices could also strengthen because of 21 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Figure 1.17 Higher prices slow oil demand Figure 1.18 A disappointing supply response Change in apparent oil demand Change in global oil deliveries Millions of barrels of oil per day 5 Millions of barrels of oil per day 5 4 World 4 Total 3 3 2 2 1 1 0 0 1 1 2 2 3 2002 2003 2004 2005 2006 2002 2003 2004 2005 2006 Other Other Asia China OECD Other OPEC Former Soviet Union Source: International Energy Agency. Source: International Energy Agency. in 2004 (figure 1.17). Demand among OECD countries actually declined by about Figure 1.19 Spare production capacity remains low 0.5 mbpd, and although demand in develop- ing countries continued to increase by just OPEC spare production capacity under 1 mbpd, this was much slower than Millions of barrels per day the increases recorded in 2003 and 2004. 7 Econometric estimates suggest that had 6 prices remain unchanged, oil demand would 5 have increased by some 2­2.5 mbpd.7 excluding Iraq Notwithstanding some three years of higher 4 Total prices since the recent upward trend in oil prices 3 began in early 2003, and the arrival on stream 2 of new fields in Africa and elsewhere, aggregate 1 non-OPEC (Organization of Petroleum Ex- porting Countries) oil supply was relatively 0 slow to increase (figure 1.18).8 Most recently, Jan. Oct. July April Jan. Oct. July 2002 2002 2003 2004 2005 2005 2006 there has been a pickup in deliveries from the Sources: World Bank; International Energy Agency. former Soviet Union and other non-OPEC sources, with the result that supply rose by 0.8 mbpd during the first nine months of 2006. Despite the limited responsiveness of sup- the world vulnerable to a significant supply ply in the first half of 2006, growth of final oil shock (see the section on risks below). It is demand was even weaker, and as a result in- that vulnerability, plus concerns about future ventories and global spare capacity increased Middle East supplies, that provides the best somewhat (figure 1.19). However, spare explanation for the increase in oil prices capacity remains low (at just 3 mbpd), leaving observed during the spring and early summer. 22 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y Financial market speculation also likely economists at that time, the world plunged played a role, especially in the first half of into the Great Depression. Thus, while much 2006, when a weakening dollar coincided in the current environment is reassuring, a with a significant run-up in emerging-market note of caution is merited. assets (among them the prices of some com- The remainder of this chapter explores four modities), followed by a significant reversal in main risks to the outlook. May and June. An indication of the impor- tance that such forces may have played was Overheating could provoke the 30 percent fall in U.S. wholesale gasoline a sharper slowdown prices in August following the decision of The world economy and, in particular, devel- Goldman Sachs to reduce the share of gasoline oping countries have been expanding at near- in its commodity indexes. record rates over the past few years. So far, the Over the near term, limited spare capacity inflationary effects of fast growth have been and strong global growth suggest that oil largely confined to the markets for global prices will remain volatile. However, high goods, such as commodity markets. The prices should continue to moderate demand inflationary response at the national level has growth, while investments in new capacity al- been remarkably muted. Improved monetary ready in the works are projected to increase policy has succeeded in anchoring inflationary output by about 15 mbpd by 2010, implying a expectations at low levels, while competitive 3 mbpd annual increase--well above expected pressures induced by the entry into the global increases in demand of between 1.5 and 2 mbpd marketplace of the countries of the former annually. As a result, the price of oil is projected Soviet Bloc and China, with their relatively to decline modestly over the next two years, high skills and low wages (see chapter 4), have reaching an average level of $53 in 2008. boosted global productivity and kept the pric- ing power of firms in check. In the baseline pro- jection these factors are assumed to continue to Downside risks predominate hold sway, while tightening of monetary policy A number of factors suggest the soft-landing and slower growth in countries where signs scenario outlined above as the most likely of a pickup of inflation have emerged are as- outcome. Tighter monetary policy in high- sumed to prevent inflation from rising further. income and a number of developing countries However, long-term interest rates are pro- is slowing growth in those countries and jected to remain low and international finan- should alleviate inflationary tensions. Mean- cial conditions relatively loose. As a result, while, still-low long-term interest rates and global growth is expected to remain strong emerging-market spreads are expected to and further inflationary pressures may yet maintain favorable external conditions for de- emerge. In particular, given projected levels of veloping countries, allowing them to grow at global demand, further price hikes in com- a slower but still robust pace of 6 percent. modity markets cannot be ruled out. While the soft landing is the most likely Moreover, should measures to slow growth scenario, the global economy is at a turning in several key developing economies point following several years of very strong (Argentina, China, India, and many European growth--and such periods are fraught with and Central Asian economies) fail, as they risk. Indeed, as described in chapter 2, the last have to varying degrees in recent years, infla- century began under similarly auspicious tion in those countries could pick up. That circumstances, characterized by an extended could lead to a more marked slowdown later period of strong growth buoyed by technolog- on, either because of a much sharper tighten- ical change and ample liquidity. Rather than ing of policy or because of endogenous factors continuing forward as anticipated by leading such as a loss of external competitiveness. 23 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 A housing market crisis could growth observed in the second and third cause a recession quarters, which is projected to continue into Growth in the United States and several other the first half of 2007. high-income countries9 has been bolstered The risk is that the slowdown may be much over the past several years by rapidly rising more severe, either because house prices de- housing prices. In the United States, rising cline more sharply or because the indirect ef- housing prices increased household wealth by fects of the anticipated 9.3 percent decline in 14.6 percent of GDP between 2000 and mid- residential investment has wider impacts on 2006. The pickup in housing valuations was the rest of the economy. A much steeper slow- spurred by low interest rates and the intro- down following a sharp decline in housing duction of new interest-only variable-rate prices11 could accentuate the decline in resi- mortgages. Higher valuations in turn gener- dential investment, driving it down by as ated a boom in home-equity withdrawals, much as 20 percent from its level in mid-2006, which boosted consumer spending and resi- while the reversal in the trend to household dential investment. wealth could cut as much as 1 percent from As short-term interest rates rose, demand growth in personal consumption. On the plus for variable-rate mortgages dried up, and the side, Australia, the Netherlands, and the rate of increase of housing prices cooled sub- United Kingdom have all observed substantial stantially10 (figure 1.20). By the third quarter decelerations and even declines in housing of 2006, the contribution to growth of resi- prices without recession (see OECD 2006 for dential investment had swung from a strong more details). 0.5 percentage points in 2005 to a strongly Such a shock could prompt a recession in the negative 1.1 percentage points. That swing, United States, with growth slowing to as little as plus the end of the additional consumption ­0.2 percent of GDP in 2007 and 2.7 percent in demand generated by home-equity with- 2008. Slower growth would weaken inflation- drawals, underlies the slowdown in U.S. ary pressure in the United States, allowing for lower interest rates in the course of 2007, help- ing to spur a recovery toward the end of 2008. Such a U.S. recession would affect develop- Figure 1.20 After rising rapidly, housing ing economies through three channels: reduced price growth slows sharply exports to the United States, lower commodity Increase in the price of existing U.S. houses and oil prices owing to slower global growth, and more favorable financing conditions. The Percent balance of these forces would vary across re- 25 gions and countries. Regions with the tightest 20 trade ties, such as Latin America and East Asia, would suffer the greatest negative im- Year-over-year growth rate 15 pact. The combination of weaker domestic demand in the United States and less marked 10 slowdowns elsewhere would help to reduce global imbalances. 5 0 A disorderly unwinding of global imbalances remains possible Quarterly growth rate at annual rates 5 The rotation of growth away from the United 1976 1981 1986 1991 1996 2001 2006 States and increased consumption demand in Sources: World Bank; Office of Federal Housing Enterprise Europe and the developing world are welcome Oversight. developments that mark the beginning of an 24 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y orderly adjustment of global imbalances. In countries, both because of its dampening ef- particular, they signal an end to a troubling fect on investment and potential output and trend of rapidly rising U.S. current account because a rapid adjustment would inevitably deficits. While relative stability should reduce result in greater short-term misallocations of financial market concerns, the financing of the resources. gap in the U.S. current account, at 6.5 percent In the baseline scenario, financial sector ad- of GDP, remains a challenge. Such a deficit is justments are assumed to be benign. The ex- not sustainable over the long run. Each year, it pected narrowing of short-term interest-rate augments the net international debtor position differentials is projected to prompt investors of the Unites States and its financing costs. The to continue shifting assets into euros, placing United States has already become the world's downward pressure on the dollar. This should largest net debtor, with the value of foreign- be offset somewhat by a tendency for U.S. owned U.S. assets exceeding that of U.S.-owned long-term rates to rise relative to those in foreign assets by 21 percent of U.S. GDP in Europe. While the relative depreciation of the 2005. In addition, the balance of the interest dollar should be smooth, the dollar could payments on these debts was $8.8 billion weaken quickly if investors were to react ner- during the first three quarters of 2006, the first vously. That would provoke much higher U.S. time in some 30 years that the United States interest rates and a sharper slowdown. Such a paid out more than it received on internation- risk can be reduced by collaborative policy ac- ally held financial assets. Unless savings in the tions to increase public and private savings in United States increase substantially, even as- the United States, strengthen demand in the suming further improvements in the trade bal- rest of the world, and provide for more ance, the net asset position of the United States flexible management of exchange rates. will continue to deteriorate, potentially reach- However adjustment occurs--be it a sharp ing between 65 and 48 percent of GDP by 2015 adjustment led by the financial sector or a more (Higgins, Klitgaard, and Tille 2005). gradual real-side adjustment--the process is As long as the trend toward real-side ad- likely to be relatively short-lived in the context justment (increased savings in the United of the 25-year projections reported in chapter States and increased domestic demand and 2. Although a disorderly adjustment would imports abroad) continues, the resolution of imply up to two years of substantially below- global imbalances should proceed in an or- trend growth for the global economy, this derly manner, even though it may take several would have minimal effects on the average years beyond our medium-term projection long-term growth rates reported there. period (2006­08) before the U.S. current account deficit reaches sustainable levels. An oil-sector supply shock could That said, the medium-term risk to the global disrupt growth economy remains that adjustment will occur With spare production capacity at only not on the real side but on the financial side, 3 mbpd, the world oil market remains vulner- either because investors rapidly lose confi- able to a supply shock. Because no country dence in the dollar--thereby provoking a cur- can easily ramp up production, if output in a rency crisis, much higher U.S. interest rates, producing country were to fall significantly, and financial market turmoil--or because world supply would fall, provoking a decline they increasingly demand higher interest rates in economic activity. on U.S.-denominated assets. While this would Simulations presented in last year's Global help increase U.S. savings and therefore has- Economic Prospects (World Bank 2006) sug- ten adjustment compared with an orderly ad- gest that a negative supply shock of two mil- justment, it would do so at greater cost in lion barrels per day that caused oil prices to terms of growth in high- and low-income double for a period of three months and then 25 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 remained at $80 for nine further months World Bank's East Asia update provides more detailed would cause global output to shrink initially information on recent developments and prospects for by about 1.5 percent of GDP, as compared the East Asia and Pacific region (www.worldbank.org/ with the baseline scenario.12 Inflation would eapupdate/). 3. In addition to the Prospects for the Global pick up rapidly, and on average the current Economy Web site (www.worldbank.org/globaloutlook), account position of oil-importing countries which describes regional developments in more detail, would deteriorate by about 1.1 percent of the World Bank's Middle East and North Africa GDP. The impact would be more severe in Web site, Economic Developments and Prospects large low-income and middle-income coun- (www.worldbank.org/mena) provides an even more tries, both because of higher energy intensities comprehensive discussion of recent economic develop- ments, projections, and policy priorities. and a greater inflationary impact, which re- 4. For the purposes of this report the developing quires a larger contraction to eliminate. countries of the region are Algeria, the Arab Repub- While the impact in terms of GDP for lic of Egypt, Jordan, the Islamic Republic of Iran, current-account-constrained low-income coun- Morocco, Oman, the Syrian Arab Republic, Tunisia, tries is smaller, it is more severe in terms of and the Republic of Yemen. Djibouti, Iraq, Lebanon, domestic consumption and investment. Such and Libya were excluded from the projections be- countries have limited access to international cause of a lack of data. Important regional players such as Bahrain, Kuwait, Qatar, Saudi Arabia, and capital markets and their capacity to pay the United Arab Emirates are included in the high- higher oil prices is limited by their export rev- income aggregate. enues. If these revenues are stable, such coun- 5. As of early November 2006, the credit ratings of tries would be forced to reduce domestic 34 emerging market countries have been upgraded. demand and non-oil imports in order to pay Only 3 have been downgraded. their higher oil bill. As a consequence, when oil 6. Agricultural prices are quoted in U.S. dollars and prices rise, oil consumption remains relatively therefore have been deflated by U.S. inflation. 7. The short-term price elasticity of oil demand constant in terms of volume (being generally is estimated at between 0.01 and 0.2 percent inelastic in the short run), but the oil bill rises. (Burger 2005), implying that immediately following a To compensate, non-oil imports and domestic 100 percent increase in oil prices, such as observed demand tend to decline in unison--leaving between 2002 and 2005, oil demand would be GDP relatively unchanged. For these countries, expected to decelerate by between 1 and 20 percent. the terms-of-trade-shock of the initial increase Long-term elasticities are larger (between 0.2 and in oil prices is estimated at 4.1 percent of their 0.6 percent), implying that the negative effect of higher prices over the past few years will continue to GDP, which would translate into a 2.7 percent be felt. decline in domestic demand, with potentially 8. In the three years following both the 1973 and serious impacts on poverty. 1979 oil price hikes, non-OPEC and non­former Soviet Union oil producers increased their output by some 3.5 million barrels per day. In contrast, since 2002, production from these sources has actually de- Notes clined. OPEC did increase its deliveries during 2004 by 1. Housing prices, which had been rising by 10 per- drawing down its spare capacity, but so far investment cent a year, declined at a 1.2 percent annualized pace to increase that capacity has been limited. in the third quarter of 2006. As a result, increases in 9. Robust increases in residential investment and household wealth slowed, and home-equity with- rising housing prices have been important drivers of drawals, which boosted GDP growth by as much as growth in recent years in Canada, Denmark, France, 1 percentage point during 2000­05, turned negative. Ireland, Spain, and the United States. At the same time, the contribution of residential in- 10. As of September 2006 the median sales price of vestment to GDP growth fell from 0.5 percentage houses in the United States had fallen 1.2 percent (year- points in 2005 to 0.7 and 1.1 percentage points in over-year). This measure, produced by the National As- the second and third quarters of 2006, respectively. sociation of Realtors, differs from data provided by the 2. In addition to the Prospects for the Global OFHEO, which are reproduced in figure 1.20, because it Economy Web site (www.worldbank.org/outlook) the does not control for the quality of the houses being sold. 26 P R O S P E C T S F O R T H E G L O B A L E C O N O M Y 11. Girouard and others (2006) estimate that U.S. OECD Economics Department Working Papers housing prices have a more than 75 percent chance of No. 475. falling if interest rates rise by 100 basis points. Higgins, Mathew, Thomas Klitgaard, and Cedric Tille. 12. Studies suggest that the likelihood of such a dis- 2005. "The Implications of Rising U.S. Interna- ruption occurring over the next several years may be as tional Liabilities." Current Economic Issues (Fed- high as 70 percent (Beccue and Huntington 2005). eral Reserve Bank of New York) 11(5). Institute for International Finance. 2006. Capital Flows to Emerging Markets. References IMF (International Monetary Fund). 2006. World Eco- Beccue, Phillip C., and Hillard G. Huntington. 2005. nomic Outlook: Financial Systems and Economic "An Assessment of Oil Market Disruption Risks." Cycles. Washington, DC: IMF. Final Report of Energy Modelling Forum & SR 8, Newfarmer, Richard. 2005. Trade, Doha, and Devel- Energy Modeling Forum. October. opment: A Window into the Issues. Washington, Burger, Victor. 2005. "House Prices in Developing DC: World Bank. Countries." World Bank, Washington, DC. Norman, John, and Lei Shen. 2006. "How Much Dimaranan, Betina, Elena Ianchovichina, and Will Money Has Left Commodities?" Global Cur- Martin. 2006. "Competing with Giants: Who rency & Commodity Strategy (JP Morgan, New Wins, Who Loses?" In Dancing with Giants: York), September 18. China, India, and the Global Economy, ed. L. Alan OECD (Organisation for Economic Co-operation and Winters and Shahid Yusuf. Washington, DC, and Development). 2006. OECD Economic Outlook Singapore: World Bank and Institute of Policy 80 (December). Studies. World Bank. 2006. Global Development Finance Girouard, N. and others. 2006. "Recent House Price 2006: The Development Potential of Surging Developments: The Role of Fundamentals." Capital Flows. Washington, DC: World Bank. 27 2 The Coming Globalization The emergence of China, India, and the for- in rich and poor countries? What role will de- mer communist-bloc countries implies that veloping countries play--particularly those the greater part of the earth's population is with large populations, such as China and now engaged, at least potentially, in the India? Finally, what forces could accelerate global economy. There are no historical an- growth and globalization--and what could tecedents for this development. derail them? --Ben Bernanke, August 25, 2006 This chapter explores these questions by developing a long-term scenario to 2030. The The last quarter-century, a time of unprece- scenario is anchored in trends already evident dented integration for the global economy, in recent years, and ones unlikely to be re- has witnessed a dramatic rise in standards of versed in the foreseeable future. The results living around the world. The fall in transport describe a world in which the gross domestic and communications costs and in barriers to product (GDP) in high-income countries is trade paved the way for productivity in- slated to nearly double and that of developing creases associated with the integration of countries will more than triple. The progres- emerging economies into global markets. Add sive expansions of China and India, the two to these forces the fall of the Berlin Wall, the largest developing economies and home to subsequent lifting of the Iron Curtain, and the half the people of the developing world, are progressive opening of the Chinese and then projected to drive the process. Their impact Indian economies--and the stage was set for on the global economy will be increasingly a new wave of globalization of production, felt as their exports and energy use, for exam- trade, and finance. While the associated ben- ple, approach the levels of the European efits have been uneven over time and space, Union and the United States. average living standards across the globe have The next 25 years will undoubtedly bring risen markedly. Global income has doubled significant surprises that cause outcomes to since 1980, 450 million have been lifted out deviate from the central scenario in this chap- of extreme poverty since 1990,1 and life ex- ter. Growth in parts of the world may well be pectancy in developing countries is now 65 on more robust than projected in this scenario; average.2 other countries or whole regions may face Can one expect these trends to continue serious setbacks. Many imaginable and even for the next 25 years, and if so, what are the unimaginable shocks are likely. The chapter key forces that will shape the world economy thus includes a discussion of various shocks of tomorrow? If globalization continues, what that could propel growth higher than the does it mean for the allocation of production central scenario--or depress growth with 29 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 impacts devastating for poverty. Developing The evidence of globalization countries are likely to become more important lobalization has been present since the in the global economy. Indeed, if anything, the G dawn of modern humans nearly 50,000 likelihood that developing countries will expe- years ago in Africa (see Wade 2006). The rience higher rises in incomes seems greater Roman Empire stretched from Great Britain than the downside risks. Although outcomes to the Middle East nearly 2,000 years ago and less sanguine than those in the central scenario 500 years ago the age of discoveries led to the are possible, it would take disruptive sea- expansion of European outreach to the west- changes in the structure of the global economy ern hemisphere and East Asia. Two distinct to produce large deviations from, much less periods in more modern times are often cited reversals of, the strong underlying trends as intensified phases of globalization--the toward globalization. 20­30 years before World War I and the years The good performance of developing coun- since World War II. Both witnessed sharp in- tries in recent years, combined with the still creases in trade, international migration, and huge difference in relative incomes between de- flows of finance, accompanied by rapid veloping and developed countries, points to changes in technology--electricity, trains, and strong potential growth across the developing steamships in the first period, and planes, world during the coming decades. The central containers, and telecommunications in the scenario assumes a world growth rate of 2 per- second. While technology was a key factor, cent per capita, slightly faster--by 0.6 points policies were also important--such as the re- per capita--than in 1980­2005. It also as- ductions in trade and financial barriers. This sumes growth in developing countries of section reviews some of the key evidence of 3.1 percent, compared with 1.9 percent in de- the most recent period of globalization hinting veloped countries. There are two main reasons at what trends can be anticipated over the for this. First, policy is far better on average next 25 years. The section will highlight today in developing countries than it was ear- trends in four broad categories that define lier, say in 1980. Second, technological dissem- globalization--trade in goods and services, ination is far faster. Indeed, in the last five years, international migration, capital flows, and growth in developing countries has been sub- technology and information. stantially higher--4.6 percent--than the as- sumption of 3.1 percent in the central scenario. Huge expansion of trade Whatever the scenario, challenges will World trade has exploded since the early abound. Growth and integration will lead to 1960s. World exports have grown from just structural changes, job losses, uneven income under $1 trillion a year (in 2000 dollars) to growth, and other painful transitions. Fast nearly $10 trillion a year, annualized growth growth could lead to ever-increasing competi- of some 5.5 percent per year (figure 2.1).3 tion for scarce resources and put additional They are clearly outpacing global output, strains on the environment. And some regions which increased at some 3.1 percent per year could continue to lag behind, owing to weak over the same period. Between 1970 and institutions, fragile states, and inadequate in- 2004, the share of exports relative to global frastructure. Many of these challenges will be output has more than doubled and is now dealt with at the national level, but some re- over 25 percent. Throughout the early part of quire global leadership. Perhaps one of the this period the export elasticity (the rate of biggest challenges will be shaping a new growth of exports relative to output) was run- global architecture that can take into account ning at about 1.5, but around 1986 the elas- the increasing diversity of countries and inter- ticity picked up substantially, peaking at more ests and allow for peaceful resolution of than 2.5 a decade later. This acceleration came emerging global tensions. on the heels of the collapse of the Iron Curtain 30 T H E C O M I N G G L O B A L I Z A T I O N Figure 2.1 World trade has expanded dramatically... Export to GDP elasticity Export to GDP share (%) 3.0 30 2.5 Elasticity Share 25 2.0 20 1.5 15 1.0 10 0.5 5 0 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Sources: World Bank Development Data Platform (DDP) and staff calculations. Note: Elasticity is calculated as the percent change in real exports relative to the percent change in real GDP. The export share is calculated in nominal dollar terms. Data are smoothed using five-year moving averages. Figure 2.2 ...and become more diversified... Share of developing-country exports, by broad commodity grouping Percent 100 80 60 40 20 0 1962 1967 1972 1977 1982 1987 1992 1997 2002 Agriculture & Food Minerals & Fuel Chemicals Textile & Apparel Machinery & Equipment Other Source: World Bank World Integrated Trade Solution (WITS). and moves by China and India to open their multilateral reductions were also important in economies and pursue an export-led strategy. promoting global trade. Multilateral negotia- Other countries also abandoned inward- tions under the guise of the General Agree- looking strategies and saw their exports jump. ment on Tariffs and Trade (GATT)--now A large part of the opening of domestic the World Trade Organization (WTO)-- economies can be attributed to unilateral deci- undertook stepwise reductions in trade poli- sions, as in China and India, but regional and cies known as rounds, the latest of which, the 31 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Figure 2.3 ...than increase in migrants--in particular toward high-income countries... Millions of migrants % host population 120 Migrants to high-income countries 12 100 10 80 8 Migrants to developing countries, excluding Europe and Central Asia 60 6 40 4 20 2 0 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Sources: World Bank 2006a; World Bank 2006c; staff calculations. Note: Bars are measured on the left axis; lines are measured on the right axis. Figure 2.4 ...and a sharp rise in capital flows. $ billions % GDP 1,200 12 Gross private capital flows 1,000 to developing countries 10 800 8 600 6 Net inflows of FDI to 400 developing countries 4 200 2 0 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Sources: World Bank DDP and staff calculations. Note: Bars are measured on the left axis; lines are measured on the right axis. FDI foreign direct investment. Doha Development Agenda is the ninth in the the Southern Cone Common Market among series. Though initially largely the realm of de- others, with many others in the pipeline. veloped countries,4 with the expansion of Though most of these agreements have tended trade and WTO membership, 149 countries5 to be trade-creating, they can also divert trade are now involved, perhaps complicating the from excluded countries. ability to achieve agreement given the more di- Technological breakthroughs--particularly verse set of objectives. Since 1990 there has in transportation and communications-- also been an explosion in regional trade agree- emerging business practices, capital flows, and ment notifications, many involving the new the growth in a skilled workforce have led to an transition economies, but also including increasing proportion of developing-country expansion of the European Union (EU), the exports in manufactured goods that are more North American Free Trade Agreement, and traditionally the realm of developed countries 32 T H E C O M I N G G L O B A L I Z A T I O N Figure 2.5 Diffusion of traditional technologies has been slow, except in high-growth regions... Fixed phone line subscribers per 100 persons 70 60 50 40 30 20 10 0 High-income East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America countries the Pacific Central Asia North Africa Africa and the Caribbean 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Sources: World Bank DDP and staff calculations. Figure 2.6 ...but the uptake of new technologies has been faster. Mobile phone subscribers per 100 persons 90 80 70 60 50 40 30 20 10 0 High-income East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America countries the Pacific Central Asia North Africa Africa and the Caribbean 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Sources: World Bank DDP and staff calculations. (figure 2.2). Goods as diverse as car parts, air- reduced their dependence on volatile com- planes, semiconductors, and consumer elec- modities, though in some cases the ease of tronics are being sourced in developing coun- moving capital has also induced economic tries. For many developing countries this has volatility, as in apparel manufacturing. 33 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table 2.1 Services exports rise in line with goods exports Levels ($ billions) Growth rate (percent) Percentage of GDP Regions and trade 1984 1994 2004 1984­94 1994­2004 1984 1994 2004 Exports of services World 357.8 978.2 2,009.5 10.6 7.5 3.0 3.7 4.9 High-income countries 303.7 803.9 1,614.2 10.2 7.2 3.3 3.6 4.9 Developing countries 54.1 174.2 395.3 12.4 8.5 2.0 3.8 4.7 East Asia & Pacific 9.1 49.5 115.0 18.5 8.8 1.9 4.7 4.3 South Asia 4.7 9.9 32.3 7.7 12.5 1.8 2.3 3.7 Europe & Central Asia 7.9 47.5 124.6 19.6 10.1 -- 5.1 7.0 Middle East & N. Africa 10.7 19.3 36.7 6.1 6.6 -- 6.7 6.7 Sub-Saharan Africa 4.5 8.4 20.8 6.5 9.5 2.0 3.0 4.0 Latin America & Caribbean 17.2 39.6 65.8 8.7 5.2 2.5 2.5 3.3 External factor income World 330.4 782.4 1,578.3 9.0 7.3 2.8 2.9 3.8 High-income countries 295.2 734.5 1,476.7 9.5 7.2 3.2 3.3 4.5 Developing countries 35.2 47.9 101.5 3.1 7.8 1.3 1.0 1.2 East Asia & Pacific 5.6 15.5 33.7 10.8 8.1 1.2 1.5 1.3 South Asia 0.8 1.3 4.8 4.9 14.0 0.3 0.3 0.5 Europe & Central Asia 1.0 7.3 30.5 21.8 15.3 -- 0.8 1.7 Middle East & N. Africa 15.4 7.1 8.1 7.4 1.3 -- 2.4 1.5 Sub-Saharan Africa 1.4 1.9 4.6 3.2 9.3 0.6 0.7 0.9 Latin America & Caribbean 11.1 14.8 19.9 3.0 3.0 1.6 0.9 1.0 Sources: International Financial Statistics (IFS) and staff calculations. Note: Service exports corresponds to "services credit" from the balance of payments table in IFS (code 78ADDZF). External factor income corresponds to "income credit" from the balance of payments table in IFS (code 78AGDZF). Owing to lack of data, some countries are excluded from regional aggregations. -- not available. Trade in services has been growing at a For developing countries the growth in fac- pace similar to trade in goods at the global tor income from abroad has been much less level (table 2.1).6 Rising from $358 billion in pronounced than the growth in the export of 1984 to $2,000 billion in 2004, the share of services--and as a share of GDP, it has de- services exports in total exports of goods and clined. This is linked to the as yet relatively services has advanced modestly from 16 per- low level of outbound investments by devel- cent to 17.5 percent. For developing countries oping countries. For developed countries, the in aggregate, services exports have risen from expansion of foreign income has been on a par $54 billion in 1984 to nearly $400 billion in with the expansion of service exports driven 2004, raising its share of GDP from 2 percent by rapid investments abroad. to 4.7 percent. The corresponding figure for exports of goods and services is an increase Rapid increase in migration toward from 19.8 percent of GDP in 1984 to 35.1 high-income countries percent in 2004 (with no smoothing in the A second component of the recent globaliza- trend). Though South Asia is often mentioned tion is the rise in international migration, par- as the main source of the growth in trade in ticularly in developed countries. The share of services, the largest contributors to the rise in migrants in developed countries (from both developing-country service exports over the high-income and developing countries) has last two decades have been East Asia and the nearly tripled, going from 4.4 percent in 1960 Pacific and Europe and Central Asia. The lat- to 11.4 percent in 2005--equivalent to an esti- ter region has benefited from its opening up mated 112 million persons out of a total num- to the global economy, its merger with the ber of migrants worldwide of some 190 million European Union, and the rapidly rising share (figure 2.3). It is harder to discern the impacts of services in its economies. of policies on the level of migration. Much of 34 T H E C O M I N G G L O B A L I Z A T I O N the South-to-North migration is predicated on FDI in natural resource sectors does not the huge income differentials between the two, necessarily have the employment and techno- even taking into account differences in the cost logical impacts compared with FDI in the elec- of living. And one would expect that in the ab- tronics sector. A more recent phenomenon has sence of (more or less) tight border controls on been the increase in outward FDI from devel- the movement of people, the number of mi- oping countries from a low base of about grants would increase substantially.7 Pull fac- $2.2 billion in 1990 to $41.1 billion in 2004 tors are also in evidence in developed coun- (World Bank 2006b). tries: slowing or declining labor force growth combined with aging and higher education lev- Faster pace of technological take-up els is giving rise to labor shortages for certain and diffusion skill levels and/or in certain sectors. Migration Technology has been advancing rapidly-- levels in developing countries (excluding the particularly technologies that shrink the countries of the former Soviet Union) have world, easing the flows of goods, capital, and more or less stayed constant over this time technology. The improvements in telecommu- period at about 40­45 million and have de- nications are the most striking example. The clined as a percentage of the population. expansion of computer networking has vastly changed the way large companies organize More integrated financial and capital production and has permitted the introduc- markets tion of production networks that span the The pace of opening of capital markets has globe. These same networks also open up been slower than for trade--even among the market opportunities for small firms that are more homogeneous developed economies. no longer limited to regional markets. Mobile Many countries still maintain restrictions on telephony is having the same impact. As the capital flows but the world has nonetheless costs of developing mobile networks are much seen a huge increase in financial flows both in lower than that of traditional fixed-line net- gross and net terms. Foreign direct investment works, they have expanded rapidly even in the (FDI), which is particularly attractive for de- poorest regions of the world, opening up new veloping countries because it tends to be less market opportunities for once-isolated com- volatile than other capital flows and also has munities.8 Though fixed-line telephony con- other potential externalities such as embodied tinues to be important, at least until wireless technology, has risen both globally and in de- technologies mature, its growth has been lim- veloping countries. From a low initial level of ited by high fixed costs (figure 2.5) except in $22 billion in 1990, FDI toward developing the high-growth countries. Mobile technology, countries is currently running at about $200 by contrast, has taken off (figure 2.6)-- billion a year, some 2.5 percent of developing- perhaps even more sharply than shown by the country GDP (figure 2.4). Developing coun- data, given that the numbers probably vastly tries currently attract about one-third of total understate access since many users are non- global inward FDI, as FDI into developed subscribers. countries is running at some $400 billion a Improvements in transportation technol- year after peaking at over $1,300 billion in ogy have also been impressive. The introduc- 2000 at the end of the dot-com boom. Total tion of the container in the 1950s dropped the private financing of developing countries was cost of loading a ship from $5.83 per ton to nearly $1,000 billion in 2004, over five times 15.8 cents, and even more savings came from the amount in 1990. The aggregate numbers the vast reduction of time ships spend in port fail to show the wide diversity across develop- for loading and unloading (see Levinson ing countries--both in terms of levels (or as 2006). The advent of the jumbo jet airplane in shares of GDP) and externalities. For example, the late 1960s led to the rise of cheaper air 35 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 freight, a key component of the integrated the quality of domestic and international poli- global supply networks. It has enabled farm- cies. Population is expected to add 1.5 billion ers in developing countries to export their people to the planet by 2030, and virtually all time-sensitive produce--such as green beans of the increase will be in developing countries. or flowers--to high-income markets. The im- Moreover, today's high-income countries and provements in transportation and the advent China will become significantly older. Chang- of supply networks and global markets go ing demographics weigh heavily on the results hand in hand with the improvements in influencing the growth of employment, de- telecommunications and networking. mand trends, and changes in savings and in- Looking forward, one would conclude that vestment behavior (and even productivity). many of these forces are likely to provide the While demographic trends are fairly pre- same impetus to globalization as they have in dictable, assumptions about productivity the past--some with diminishing power, and growth are subject to a wider band of possi- others perhaps with more. Trade policies have bilities. There is no agreement on how to in- come a long way toward more integrated mar- terpret recent productivity growth, let alone kets for goods, though tariffs remain high in how to anticipate future patterns. For exam- many developing countries and in some ple, in the view of Gordon (2000), recent sectors--such as agriculture. Other forms of inventions--such as cell phones, the internet, protection are ever present such as unreason- or new drugs--are relatively normal incre- able product standards or ad hoc safeguard mental changes to productivity and are un- measures. The service sectors have also been likely to have the same impact as the new tech- largely untouched by the GATT/WTO disci- nologies at the beginning of the 20th plines, and their reform would likely provide century--electricity, the internal combustion additional impetus to further trade growth. engine, telephones, radio, television, and in- The same could be said for capital flows and door plumbing. Other observers, for example the movement of people. However, the greater David (1990), suggest that it takes time for driver of globalization is likely to be in the new discoveries to have their full impacts-- technological field, because the telecommuni- either because initial costs are too high, or cations revolution is still in its infancy. Adop- because there are network externalities, or be- tion, though rapid, has still bypassed many, cause it takes time for organizations to change and the technology is evolving--with greater their management practices to fully benefit speeds and the broader implementation of from the new technologies. Whether one takes wireless broadband expected. And individuals a sanguine view of new technologies or not, and firms are still learning to adapt to the new large parts of the developing world have yet to technologies and leverage them to open new benefit from "old" technologies. opportunities and increase productivity. The macro assumptions on productivity built into the forecast are largely consistent with the estimates of total factor productivity The world in 2030--the big (TFP) growth from the literature (see, for ex- picture ample, Bosworth and Collins 2003). The world saw a period of very rapid TFP growth Preface: assumptions in the 1960s, followed by a decade of stagna- The central scenario is built up from a number tion coinciding with the energy crisis of the of key driving forces--notably demographic 1970s, recovery to an estimated rate of 0.8 per- trends, savings, and investment behavior, and cent per year in the 1980s and 1990s, and an the role of technological change, and how these acceleration in the 2000s. There have been trends interact with globalization (see box 2.1). large variations across regions and time. The Some of these forces are, in turn, influenced by central scenario assumes a long-term rate of 36 T H E C O M I N G G L O B A L I Z A T I O N Box 2.1 Inside the box--the components of scenario building T he long-term scenarios described in this chapter database (release 6.1; see www.gtap.org) to calibrate are based on the World Bank's Linkage model initial parameters. A scenario is developed by solving with a dynamic core that is essentially a neoclassical for a new equilibrium in each subsequent year through growth model--similar in concept to models used in 2030 with the following key assumptions: other recent scenario work (see Goldman Sachs 2003 The growth in the labor force is driven by and PricewaterhouseCoopers 2006, for example). demographics--essentially given by the growth of the Aggregate growth is predicated on assumptions re- working-age population. Differentiated growth of garding the growth of the labor force, savings/ skilled versus unskilled workers is partly driven by de- investment decisions (and therefore capital mographics and partly driven by changes in education accumulation), and productivity. rates. As education levels rise (in the younger popula- The Linkage model, unlike the aforementioned tions), they eventually drive higher relative growth of models, has considerably more structure--see van skilled workers once they enter the labor force (and der Mensbrugghe (2006a) for a detailed description older unskilled workers retire). of the model and van der Mensbrugghe (2006b) for Savings decisions are partly driven by a summary description of the model and the assump- demographics--rising as youth dependency ratios tions underlying the baseline scenario. fall and falling as elderly dependency ratios rise. In- First, it is multisectoral. This allows for more vestment rates are driven by changes in growth rates complex productivity dynamics including differenti- (the accelerator mechanism) and differential rates of ating productivity growth between agriculture, man- return to capital. Net foreign savings is the difference ufacturing, and services and picking up the changing between domestic savings and investment. structure of demand (and therefore output) as Productivity is derived by a combination of factors, growth in incomes leads to a relative shift into but is also partially judgmental. First, agricultural pro- manufactures and services. ductivity is assumed to be factor-neutral and exoge- Second, it is linked multiregionally, allowing for the nous and is set to estimates from empirical studies (for influence of openness--through trade and finance--on example Martin and Mitra 1999). Productivity in domestic variables such as output and wages. The manufacturing and services is labor-augmenting and a model is also global, with globally clearing markets for constant wedge is imposed between productivity goods and services and balanced financial flows. growth in the two broad sectors with the assumption Third, the Linkage model has a more diverse set that productivity growth is higher in manufacturing of productive factors, including land and natural re- than in services. sources (in the fossil fuel sectors), and a labor split The model assumes that energy efficiency improves between unskilled and skilled. autonomously by 1 percent per year in all regions The Linkage model has a 2001 base year and re- and that international trade costs decline by 1 percent lies on the Global Trade Analysis Project (GTAP) per year. TFP growth in the range of 1.0­1.4 for the Over the last 25 years, the world has seen a high-income countries, somewhat on the high dramatic drop in trade barriers for goods. And end of the Bosworth and Collins estimates. The although they remain high in some countries range for developing countries is somewhat and for some sectors (for example, in agricul- wider--between 0.7 and 2.9 toward 2015 and ture), the dismantling of remaining barriers will declining slowly thereafter as the positive im- not have the same impact as in the past. A pos- pacts of rural-to-urban migration fade. sible exception: dismantling barriers in services The central scenario is also predicated on that remain high could produce significant eco- only modest changes in the policy environment. nomic gains. 37 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 likewise lose about 10 million persons, falling Figure 2.7 World population growth will from 412 million to 402 million. The United be concentrated in developing countries States will see a decline in the population in coming decades growth rate, but fertility is still much higher in Population, billions the United States than in other high-income 4.5 countries--owing in part to immigrants' 4.0 higher fertility. If current trends hold, the U.S. 3.5 population will climb by 45 million to 3.0 345 million in 2030. 2.5 The population growth pattern is more 2.0 highly varied across developing countries. 1.5 Many of the countries in Europe and Central 1.0 Asia are already confronted with declining 0.5 populations--including the Russian Federa- 0 Other China India High-income tion, which is losing population and will con- developing countries tinue to do so unless trends are reversed--at a countries rate of about 0.5 percent each year. In the new 2005 2010 2015 EU accession countries, population declines 2020 2025 2030 average about 0.2­0.3 percent per year through the entire period. At the other end of Sources: UN Population Division; World Bank Development Data Group; staff calculations. the spectrum are Sub-Saharan Africa and the Middle East and North Africa, with popula- tion growth rates currently hovering at about 2 percent, declining toward 1.1­1.4 percent World population will increase per year toward 2030. As noted above, the world will add 1.5 billion The largest contribution to the nearly 1.5 persons to its population between 2005 and billion increase in developing regions can be 2030--going from (about) 6.5 billion to attributed to India, representing 320 million 8 billion (figure 2.7).9 Roughly 12 percent additional persons, and to Sub-Saharan Africa will be living in high-income countries--down excluding Nigeria and South Africa, with a sharply from the 18 percent in 1980 and 14.5 similar increment of 320 million--each percent in 2005. All but 40 million of this contributing 20 percent to the global increase. growth in population will occur in developing Despite China's one-child policy and overall countries. While this represents a substantial aging population, the momentum of the increase in the number of persons--with con- current population will generate 170 million comitant effects on already scarce resources-- additional Chinese by 2030, another 11 per- it also represents a slowing of world popula- cent of the global increase. tion growth that added 2 billion persons between 1980 and 2005. The global popula- The global economy will more tion growth rate, between 1.7 and 1.8 percent than double in the 1980s, will slow to 1 percent by 2015 It is important to keep in mind, turning to the and dip to 0.7 percent toward 2030. economic projections, that these are a combi- High-income countries would start observ- nation of reasoned quantitative analysis and ing actual population declines--Japan by informed judgment and not predicated on 2010 and the EU countries soon thereafter. standard statistically based econometric Japan's population under current projections models, as are the short- and medium-term would fall from about 128 million in 2005 forecasts described in chapter 1. They are to 117 million in 2030. The EU15 would intended to highlight certain key aspects of the 38 T H E C O M I N G G L O B A L I Z A T I O N Figure 2.8 Developing countries will account for a larger portion of world output in coming decades Real GDP, 2001 $ trillions Growth index: 2005 = 100 80 400 Developing countries High-income countries World 70 350 60 300 50 250 40 200 30 150 20 100 10 50 0 0 2005 2010 2015 2020 2025 2030 Source: World Bank simulations using the Linkage model. Note: Bars are measured on the left axis; lines are measured on the right axis. baseline scenario that could be robust to a as a consequence of the combination of im- certain number of alternative assumptions-- proved initial conditions, better policies, and though none that imply highly nonlinear di- the still wide gap in productivity--relative to vergence from current trends. high-income countries. Moreover, developing If growth scenarios obtain, the share of countries have greater capacity and incentives output (in real terms) produced by developing to adapt new technology as communications countries would shift rather steadily. The technology continues to improve, FDI re- global economy would grow from about mains a force in overall development, and $35 trillion in 2005 to $75 trillion in 2030, an education and skill levels improve. If one overall increase of some 2.1 times (fig- decomposes the last 25 years in two ure 2.8).10 The developing-country share periods--1980­2000 and 2000­2005-- would jump from $8 trillion to $24.3 average growth in developing countries trillion--effectively tripling its output between jumped from 3.2 percent per year in the first 2005 and 2030 and increasing its global share period to 5 percent per year in the second. of output from 23 percent to 33 percent. This recent acceleration has not been shared This represents a modest acceleration of by all countries--nor is it exclusively a China what was observed between 1980 and 2005. and India phenomenon. The global economy has increased by a factor Perhaps somewhat surprisingly, these differ- of 2.1. For high-income countries the projec- entiated growth rates will have only relatively tion represents a slight decrease (from 2 to modest impacts on the ranking of countries/ 1.9) but a more significant acceleration for regions based on the volume of output.11 The developing countries (from 2.4 to 3.1). Part rankings of the top six countries/regions would of this acceleration is due to compositional remain identical to those of today led by the factors--higher-growth developing countries United States, the European Union, Japan, have higher weights today than back in 1980. China, the newly industrializing economies However, it is mostly based on the chapter (NIEs), and Latin America (excluding Brazil authors' judgment that many developing and Mexico). India would jump three spots countries are on an accelerated growth path from its current 10th ranking, essentially 39 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 swapping spots with Canada. Other countries/ Looking behind again, one sees that there regions moving up include the rest of East Asia have been some spectacular jumps in the past aggregate, Indonesia, and Iran. Sub-Saharan 25 years as well as some spectacular Africa, with its assumed more modest growth declines--most reflected by the fall of the Iron rates, would fall further behind with the rest of Curtain (table 2.2). The clear winners have Sub-Saharan Africa aggregate losing an addi- been Ireland, Singapore, Sri Lanka, Costa tional three spots by 2030. Rica, El Salvador, Equatorial Guinea, and Table 2.2 Country rankings--1980­2005 Country 1980 2005 Change Country 1980 2005 Change Country 1980 2005 Change United States 1 1 0 Hungary 52 45 7 Jamaica 96 90 6 Japan 2 2 0 Philippines 40 46 6 Bolivia 98 91 7 Germany 3 3 0 New Zealand 49 47 2 Azerbaijan 80 92 12 United Kingdom 5 4 1 Algeria 34 48 14 Ghana 89 93 4 France 4 5 1 Nigeria 33 49 16 Albania 102 94 8 China 10 6 4 Peru 51 50 1 Botswana 117 95 22 Italy 6 7 1 Romania 37 51 14 Paraguay 84 96 12 Canada 8 8 0 Bangladesh 56 52 4 Honduras 97 97 0 Spain 12 9 3 Ukraine 30 53 23 Ethiopia 83 98 15 Mexico 13 10 3 Kuwait 50 54 4 Uganda 103 99 4 Korea, Rep. of 23 11 12 Morocco 60 55 5 Senegal 100 100 0 India 11 12 1 Vietnam 35 56 21 Nepal 101 101 0 Brazil 9 13 4 Kazakhstan 54 57 3 Gabon 90 102 12 Australia 14 14 0 Slovak Republic 61 58 3 Mauritius 115 103 12 Netherlands 16 15 1 Croatia 57 59 2 Madagascar 92 104 12 Russian Federation 7 16 9 Slovenia 64 60 4 Namibia 106 105 1 Switzerland 19 17 2 Ecuador 63 61 2 Nicaragua 99 106 7 Taiwan, China 32 18 14 Oman 73 62 11 Burkina Faso 108 107 1 Sweden 18 19 1 Guatemala 67 63 4 Mali 111 108 3 Austria 24 20 4 Tunisia 68 64 4 Congo, Rep. of 104 109 5 Turkey 27 21 6 Syrian Arab Republic 58 65 7 Georgia 82 110 28 Saudi Arabia 15 22 7 Bulgaria 55 66 11 Benin 113 111 2 Indonesia 20 23 3 Dominican Republic 69 67 2 Guinea 110 112 2 Norway 28 24 4 Sri Lanka 86 68 18 Chad 119 113 6 Poland 26 25 1 Sudan 66 69 3 Armenia 109 114 5 Denmark 29 26 3 Belarus 62 70 8 Niger 105 115 10 Greece 36 27 9 Costa Rica 94 71 23 Kyrgyz Republic 107 116 9 South Africa 22 28 6 Lithuania 72 72 0 Malawi 114 117 3 Argentina 21 29 8 Kenya 78 73 5 Swaziland 123 118 5 Hong Kong, China 42 30 12 El Salvador 93 74 19 Togo 120 119 1 Finland 31 31 0 Uruguay 70 75 5 Rwanda 112 120 8 Ireland 53 32 21 Angola 81 76 5 Central African 122 121 1 Iran, Islamic 17 33 16 Côte d'Ivoire 71 77 6 Republic Rep. Of Panama 88 78 10 Sierra Leone 116 122 6 Portugal 45 34 11 Cameroon 76 79 3 Lesotho 124 123 1 Thailand 38 35 3 Trinidad and Tobago 77 80 3 Mauritania 121 124 3 Israel 48 36 12 Yemen, Republic of 79 81 2 Belize 126 125 1 Venezuela, R. B. de 25 37 12 Zimbabwe 74 82 8 Burundi 118 126 8 Malaysia 43 38 5 Latvia 75 83 8 Seychelles 127 127 0 Singapore 59 39 20 Bahrain 91 84 7 Gambia, The 125 128 3 Colombia 39 40 1 Equatorial Guinea 129 85 44 Guinea-Bissau 128 129 1 Czech Republic 41 41 0 Tanzania 65 86 21 Vanuatu 130 130 0 Egypt, Arab Rep. of 46 42 4 Iceland 95 87 8 Pakistan 44 43 1 Jordan 87 88 1 Chile 47 44 3 Estonia 85 89 4 Source: World Development Indicators, World Bank. Note: Based on five-year moving average centered on 1982 and 2003, respectively. Rankings based on GDP in current dollars. 40 T H E C O M I N G G L O B A L I Z A T I O N Botswana.12 More modest, but still substan- Per capita income growth is what matters tial improvements include the Republic of Economic size and ranking have their impor- Korea, Taiwan (China), Hong Kong (China), tance, not least in terms of determining power Israel, Oman, Panama, Portugal, and relations, be it at the global, regional, or bilat- Mauritius. There is little obvious commonality eral level. But from a welfare point of view, across these economies with the exception that what really matters is income per capita, not none (save Equatorial Guinea) is an oil pro- the overall size of an economy.13 Using the ducer or a transition economy. China's GDP in- market dollar exchange rate of an economy crease has been fast, but it has only moved provides a biased estimate of individual well- from 10th place to 6th over the 25 years and being because prices differ substantially across India has lost a spot, with both Mexico and economies--particularly for nontraded goods Korea jumping over India in the rankings. such as personal and housing services. For this Many of the countries, having lost ground reason, it is more appropriate to use the PPP in the global ranking, are concentrated among exchange rates, which take into account these oil producers and transition countries includ- differences in prices.14 Even using PPP ex- ing the Russian Federation, Saudi Arabia, change rates, the speed of convergence be- Indonesia, the Islamic Republic of Iran, tween developing- and developed-country República Bolivariana de Venezuela, Nigeria, incomes would be modest under this scenario. Romania, Bulgaria, Gabon, and Georgia. At today's income in PPP terms, the average However, other countries have also fared developing-country resident receives about poorly--for example Brazil, Argentina, and 16 percent of the average income of high- Uruguay in Latin America, and Ethiopia, income countries--$4,800 versus $29,700 Tanzania, and Zimbabwe in Sub-Saharan (figure 2.9). This ratio would rise to 23 percent Africa--though it is generally the case that in 25 years' time, representing an average countries that have avoided conflict have developing-country income of $12,200 versus managed to maintain their ranking. $54,000 for high-income countries. Figure 2.9 In some developing regions, per capita incomes will begin to converge with those in high-income countries Index: high-income countries = 100 40 35 30 25 20 15 10 5 0 East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America Low- and the Pacific Central Asia North Africa Africa and the Caribbean middle-income countries 2005 2010 2015 2020 2025 2030 Source: World Bank simulations using the Linkage model. Note: Ratio of PPP-adjusted per capita incomes relative to high-income average. PPP is fixed at base year (2001) level. 41 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 There is, perhaps needless to say, great vari- growing rapidly in most countries owing to ance across countries. Chinese incomes would the large number of births over the last two rise from 19 percent of the average high- decades and most are only seeing modest in- income level to 42 percent, a significant nar- creases in the share of the elderly in the popu- rowing of the gap and would achieve an aver- lation because rising life expectancy largely age income close to the lower range of today's impacts current (and larger) generations poorest high-income countries. There would rather than past. be a further falling behind in Sub-Saharan For developed economies, the standard Africa with its modest per capita growth economic impacts of slowing population below the high-income average, and Latin growth and aging suggest that aggregate America would see little if any convergence on savings will decline, all else being equal, as average. As the previous 25 years have shown, aging populations tend to dis-save or consume there is plenty of scope for surprises and coun- out of existing assets. This would tend to de- tries doing significantly better, even compared crease the amount of savings available for to countries with similar initial conditions.15 developed countries. The evidence for this dis- The rather modest level of convergence saving is mixed and other factors--such as overall nevertheless obscures the fact that current levels of public and/or international market opportunities for both developed and indebtedness--may influence the long-term developing countries will increase dramati- patterns of savings and investment. On the cally as the sheer size of the population of de- other hand, lower rates of employment veloping countries ensures the growth of a growth could have mixed impacts on invest- very significant middle and upper class likely ment. Lower labor supply could lessen the to rival the purchasing power of today's high- need for investment in sectors where labor and income consumer (see chapter 3). Thus, capital are close complements.16 But more in- notwithstanding the challenge that poverty tense investment may counteract this effect in will continue to hold on the global commu- sectors where labor and capital are substitutes nity, the wider spread of wealth globally will and labor-saving technology is an option.17 also provide greater means to deal more sub- Aging populations can have other conse- stantively with poverty and other global con- quences. Productivity growth could be higher cerns such as the environment and health. in economies with rapid increases in the num- The next sections delve more in depth into ber of youth joining the labor force. They can the underlying assumptions of this central also be associated with changes in consumer scenario and some of the policy implications behavior with less demand for food and educa- that can be derived from them and their tional services and more demand for leisure potential alternatives. and health services (McKibbin 2005; Bryant 2004; Helliwell 2004; Tyers and Shi 2005). Demographics are central to the There could also be fiscal implications as growth scenario promises to earlier generations in terms of so- Two significant demographic changes are oc- cial welfare benefits prove hard to finance with curring at the moment. Developed economies a lower tax base. This eventually may involve a have seen a huge decline in fertility rates (well combination of lower benefits and delay of re- below replacement rate), a stable labor force tirement age or other forms of higher labor- that will begin to decline, and rapidly aging force participation rates by the elderly. populations. Developing countries--some ear- For developing countries, some of these im- lier than others--are now also seeing signifi- pacts are reversed. With a lower proportion of cant declines in fertility rates and a substantial youth to care for--including provisions for reduction in the number of youths relative to housing, education, and nourishment--more those in the labor force. Labor forces are still can be saved and invested, particularly because 42 T H E C O M I N G G L O B A L I Z A T I O N many countries still have a low proportion of between 2020 and 2025--somewhat later than elderly. To the extent that available savings even for the NIEs of East Asia. from developed countries decline, the higher Labor force growth is still rapid in devel- savings in developing countries would tend to oping countries--though on a declining trend offset the decline. throughout the period. Currently, developing Starting with employment, developed- countries need to increase employment by country employment growth, though positive nearly 50 million jobs per year to keep up through 2010 at about 1.2 million new jobs per with working-age population growth under year, becomes negative thereafter, with an aver- the proviso of no change to the labor force age loss of about 700,000 jobs between 2010 participation rate including females. This and 2015, jumping to an annual average loss of latter assumption may be dubious in light of over 3.2 million between 2025 and 2030 (fig- the fact that fertility is declining rapidly in de- ure 2.10).18 This latter number represents a veloping countries. The largest needs are in decline of about 1 percent per year. Among the largest countries, with China and India other things, this negative employment needing to create 8­10 million jobs each year. growth implies--through standard growth This may be easier in these rapidly growing accounting--that combined capital accumula- economies. Countries in Sub-Saharan Africa tion and productivity will have to accelerate to also need to create close to 10 million jobs compensate if aggregate growth of 2­3 percent each year. With their lower economic growth per year is to be maintained. The start of the rates and relatively small urban populations, decline in the labor force varies across coun- the task appears to be much harder.19 tries, already (potentially) observable in Japan, The trends for China also show the im- beginning in the European Union shortly after pacts of its decades-long population policies 2010, and delayed in the United States (and limiting births. In a relatively near future, Australia and New Zealand) until sometime employment growth will decline precipitously Figure 2.10 Labor force growth is slowing Average annual change in employment, millions of persons 16 14 12 10 8 6 4 2 0 2 4 6 High-income East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America countries the Pacific Central Asia North Africa Africa and the Caribbean 2005­10 2010­15 2015­20 2020­25 2025­30 Sources: UN Population Division; World Bank Development Data Group; Global Trade Analysis Project (GTAP) Database; World Bank simulations using the Linkage model. 43 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 from over 5 million between 2010 and 2015 many high-income countries, where fertility to under 500,000 between 2015 and 2025 has dropped to between 1 and 1.5 births per and will turn negative thereafter. The only woman.20 other region affected by negative employment Over the longer term, the increase in growth is Europe and Central Asia, whose developed-country fertility would lead to a population is more similar in structure to the slight rise in the share of the population aged European Union than to the average develop- 15 and below, and even an absolute rise, ing country. sometime after 2020 (figure 2.11). In summary, employment growth initially The growth in the number of youths in de- will provide significant stimulus to economic veloping countries will stay more or less con- growth, but its share will decline rapidly in stant on average over the entire time horizon-- most developing regions as the current gener- though again highly variegated across regions, ation of youth join the workforce and leave with large declines in East Asia and the Pacific behind a smaller pool of potential workers as and Europe and Central Asia offset by positive fertility continues to fall. if declining growth rates in Sub-Saharan Africa The demographic projections for youths and to a lesser extent South Asia. But even in and the elderly are distinctly different in na- Sub-Saharan Africa, the assumption of re- ture mainly because the future elderly are all placement-level fertility will lead to a sharp de- alive today and thus the projection is based on cline in births. Between 2005 and 2030, the changes to mortality rates that tend to be share of youths in Sub-Saharan Africa will easier to gauge than changes to fertility rates. drop from 44 percent to 35 percent. Despite In fact, the UN population forecasts--the this leveling, the pressure to educate (and pro- basis of the World Bank's forecasts--are pred- vide health services) for the young in Sub- icated on all countries converging toward Saharan Africa will be a challenge in a region population replacement levels of fertility by that is significantly off-track in terms of 2050. This implies an increase in fertility in achieving the Millennium Development Goals Figure 2.11 Due to the demographic dividend, fewer resources will be needed for a declining youth population Youth as a share of total population (%) 50 45 40 35 30 25 20 15 10 5 0 High-income East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America countries the Pacific Central Asia North Africa Africa and the Caribbean 2005 2010 2015 2020 2025 2030 Sources: UN Population Division; World Bank Development Data Group. Note: Youths include those who are 0­14 years old. 44 T H E C O M I N G G L O B A L I Z A T I O N (MDGs) by 2015. The number of young people developing countries. These ratios will drop, in Sub-Saharan Africa will jump by 100 million but even in 2030 will remain at 60 or above. from 300 million currently, so it is not simply a The number of elderly will more or less question of building new classrooms and train- double over the next 25 years--from 464 mil- ing new teachers for today's population, but lion in 2005 to about 910 million in 2030. also taking into account the bulge in the By and large, future population aging is student-age population as one looks forward. a developed-country phenomenon--though Potentially, the resources to take care of the only one in three elderly currently lives in de- youth will improve as the number of workers veloped countries and another one in three grows more rapidly than the number of lives in China or India. The number of elderly youths in developing countries. The depen- per 100 workers in developed countries dency ratio--defined as the number of youths would rise from 30 to 53 between 2005 and per 100 workers--will drop pretty steadily be- 2030 and reach 63 in Japan and 59 in the EU tween 2005 and 2030, starting at a level of 60 (figure 2.12). Even in the United States the and falling to 47. Even with the sharp drop in rate could nearly double from today's low of the youth dependency ratio in developing 23 to 44 in 2030. This will undoubtedly ne- countries, they will still have an average ratio cessitate forceful policy changes because exist- considerably higher than the average in high- ing unfunded promises to future elderly would income countries centered at around 35.21 require unprecedented taxes on workers. These ratios will reach developed-country For developing countries, aging populations levels in East Asia and the Pacific and in (as defined by the number of elderly per 100 Europe and Central Asia--pretty rapidly in workers) will rise only slowly from current both cases--by about 2015. Both Middle East levels through about 2020, but will start accel- and North Africa and Sub-Saharan Africa erating modestly afterwards to reach a level stand out as having particularly high youth of nearly 19 starting from 12 in 2005. This is dependency rates--85 and 93 (per 100 work- still well below the developed-country average ers), respectively, well above the average for of 30 today and differs widely across regions. Figure 2.12 More resources will be needed to take care of a growing elderly population Elderly persons per 100 workers 60 50 40 30 20 10 0 High-income East Asia and South Asia Europe and Middle East and Sub-Saharan Latin America countries the Pacific Central Asia North Africa Africa and the Caribbean 2005 2010 2015 2020 2025 2030 Sources: UN Population Division; World Bank Development Data Group. Note: Elderly include those who are 65 or older. 45 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 China will see a sharper rise in its elderly dropped dramatically since 1980, they still re- dependency rate, moving from 12 currently main stubbornly high in some sectors, for ex- to 25 by 2030. This could be contrasted with ample in agriculture and services, or in some India, which has a level similar to China's at 11, countries. Protection can also take other but rising to only 16 by 2030. As mentioned forms, for example antidumping, questionable earlier, population-wise Europe and Central standards, or variable levies (as bound tariffs Asia is more similar to high-income regions and are typically well above applied tariffs). the elderly dependency ratio will hit 34 by Progress in opening markets has stalled at the 2030. What moderates this to some extent in multilateral level, but countries continue to Europe and Central Asia is the inclusion of pursue liberalization either unilaterally or Turkey with its relatively young and large pop- through bilateral and regional agreements. ulation and, more unfortunately, the precipi- While the standard theory of trade has fo- tous decline in life expectancy in some parts of cused on comparative advantage, new trade Europe and Central Asia. theory places much more emphasis on the role These demographic trends provide a signif- of specialization. Specialization is manifested icant opportunity for many developing coun- in two ways. The first is consumers' desire for tries22 that will be able to devote less resources greater varieties of the same categories of to their youth and that do not yet have to de- goods. Whereas 25 years ago consumers had a vote significant resources to their elderly-- relatively modest selection of automobiles or although they would be well advised to avoid fashion, today's range of consumer goods is making some of the choices that developed huge. This love of variety has provided pro- countries have made regarding long-term ducers from a diverse set of countries with op- commitments to their elderly without ade- portunities to export. A second form of spe- quately making provisions for them. cialization is represented by production networks that allow for the breaking up of the production process across multiple firms The four channels of globalization and/or countries. The growth in production W hat follows is a discussion of the four key networks has been predicated on many tech- channels of globalization and how they nological advances--both physical, as in interact with the development process--trade telecommunications and transport, and man- in goods and services, movement of persons, fi- agement processes, such as supply chain logis- nancial flows, and technological diffusion. tics. There is little evidence that these factors will subside anytime soon. Trade integration will accelerate Under the central scenario, the level of ex- The trade dimension of globalization has per- ports would more than triple--from about haps been the most prominent, especially with $9 trillion in 2005 to over $27 trillion in 2030-- the emergence of Asia and the transition with a concomitant rise in the world export-to- economies over the last two decades. Growth output ratio, jumping to 34 percent from in trade has outpaced growth in output by a 25 percent currently. For developing countries factor of two or more and the causes behind exports will increase from about $3 trillion to this phenomenon are in place to sustain it over over $12 trillion, reflecting in part these coun- the next two decades. tries' greater output growth. These baseline Income growth, changing comparative numbers are predicated on the assumption of advantage, and the push toward greater no change to current trade policies.23 Under a openness--the impasse in the Doha Round broad reform scenario whereby all countries re- negotiations notwithstanding--will continue duce tariffs on merchandise goods (and domes- to lead to expanding global trade over the tic agricultural protection) by three-quarters, next two decades. Though import tariffs have exports by developing countries would increase 46 T H E C O M I N G G L O B A L I Z A T I O N by an additional $2 trillion in 2030, a jump of restriction on the exchange of goods and ser- some 18 percent over the baseline. vices has an economic cost and migration is no different. Global Economic Prospects 2006 The push and pull factors driving (World Bank 2006a) explored in depth the international migration will persist main impacts of such migration, illustrating in International migration has risen substantially particular that the greatest beneficiaries are the recently--though the lack of reliable data, par- migrants themselves, though through remit- ticularly of irregular migration, makes it diffi- tances, the sending countries could also gain cult to assess the actual number of migrants in substantially. The aggregate impacts for the re- either developed or developing countries. Cur- ceiving countries are also on balance positive, rent estimates are that 11.4 percent of devel- though they could have negative distributional oped countries' population are foreign-born, consequences. up from 6.2 percent in 1980. South-South mi- This chapter's central scenario uses the un- gration is also an important phenomenon, but derlying UN methodology and projections for data prove even less reliable. the growth in country population and makes While developing countries on average will no additional assumptions as regards interna- see improvements in living standards relative tional migration. Though migration can make to high-income countries, the forces underly- a significant impact for the migrants them- ing South-to-North migration will continue to selves, in the context of a 25-year scenario, in- have a strong impact. First, there is the exist- ternational migration is unlikely to have large ing, considerable wage gap (even taking into macroeconomic impacts save perhaps for a consideration differences in purchasing handful of smaller economies or countries that power) that will shrink, but will still be sub- receive high levels of remittances. The United stantial well into the future. Second, the com- Nations forecasts the net number of migrants bination of existing migrant stocks (and the to developed countries to increase by 98 mil- push to reunite family and friends) with po- lion between 2005 and 2050, or about tential reductions in migration costs will pro- 2.2 million annually. This is expected to more vide ongoing impetus for South-to-North mi- or less offset the net natural population de- gration. Third, the slowing growth of the crease in developed countries (that is, the ex- workforce in developed countries and the cess of deaths over births). For developing aging of populations will be a pull factor in in- countries, this represents only 4 percent of creasing migration over the next two decades. total incremental population between 2005 However, unlike the trade in goods and ser- and 2050 (and thus a small fraction of the vices, or the flow of capital, migration is total population) (see UN 2004). subject to considerable regulation and control and is also fraught with many additional con- Financial integration will intensify siderations. Sending countries are concerned Savings, investment, and finance. The global with the social and family aspects of outward financial system is likely to change dramatically migration, or in some cases with brain drain. over the course of the next 25 years, as The receiving countries may also be concerned technological innovations and even greater with the social implications of migration and integration of markets expand the reach of the economic and fiscal consequences, partic- global financial intermediaries. Some of these ularly for those whose populations compete changes are impossible to anticipate. For directly with the migrants. example, it is not clear whether the future Notwithstanding these legitimate concerns, communications technologies will favor a in a global context, the economic impacts of continued concentration of financial inter- increasing South-to-North migration can be mediation, or encourage the growth of global highly beneficial. Any form of economic banks and other financial institutions in a 47 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 wide range of markets, or lead to even greater considerations.24 On balance, some decline in decentralization as smaller investments are savings can be expected as elderly dependency required to obtain the information necessary to ratios increase. carry out financial transactions. Other changes This simple theory of individual behavior, can be partially anticipated. For example, as in conjunction with demographic trends set developing countries take up a greater share of off by the baby boom after World War II and global output, it is likely that their importance the impressive increases in longevity in the de- in financial markets will continue to grow. veloping world, has dramatic implications for Some decline is already apparent in the the global economy. For Europe, Japan, and dominance of the dollar as a currency of East Asia, which have relatively high saving lending and reserves (World Bank 2006b), but rates and supply a large share of global finan- whether currencies from developing countries cial flows, the rise in dependency ratios should will play a major role in global financial lead to a decline in savings.25 By contrast, the markets is not yet apparent. very low saving rates in Sub-Saharan Africa One major issue facing developing countries may be at least partially explained by the re- over the next quarter-century is the impact of gion's very high youth dependency ratios. Sav- demographic trends on the countries' access to ing rates should rise as these young people external savings. The rise in old-age depen- move into the workforce, boosting investment dency ratios in industrial countries, and in and growth. some developing countries, is likely to be asso- The decline in saving rates is not expected ciated with a decline in saving, a rise in interest to follow a smooth trend over the next rates, and a fall in their current account sur- 25 years. In industrial countries, saving rates plus. All else being equal, the elderly tend to should rise in the near future, as the bulk of save less or even dis-save, as they live off of sav- the baby boom generation remains in the ings earned during their working years. While workforce during peak saving years. However, forecasts of saving rates are uncertain, and es- over the next 20 years this generation will re- timations of the relationship between aging tire, and saving rates will go down. Russia and and saving rates vary widely, the prospect of some of the other countries of the former So- reduced global saving over the coming decades viet Union are likely to see a decline in the needs to be considered seriously. labor force, and thus savings, owing to rising elderly dependency ratios shortly after the in- The coming savings decline. The life-cycle dustrial countries, followed closely by China theory of consumption argues that saving rates and some other parts of East Asia. Latin are low during young adulthood to provide for America and South Asia may see some effect children, rise as individuals save for retirement, of rising elderly dependency by the end of the and then fall as retirees live off of their forecast period. By contrast, Sub-Saharan accumulated assets. This theory is subject to Africa and the Middle East and North Africa significant qualifications, as individuals also have relatively young populations and should save to provide a bequest for their children and see increasing labor force participation and to maintain a stock of wealth to deal with savings through 2030. Overall, the forecast adverse shocks. Saving rates also are influenced drop in global saving is quite substantial, from by a host of macroeconomic factors, including 21.6 percent of income in the first half decade growth, interest rates, inflation, borrowing of this century to 19.9 percent by 2030. constraints, fiscal policy, pension systems, and Demographic influences also imply a de- income distribution (Loayza, Schmidt-Hebbel, cline in investment demand, as fewer workers and Servén 2000). Econometric estimates are available for each unit of investment.26 have provided mixed support for the view Other aspects of aging may boost investment that savings behavior is governed by life-cycle demand. Aging may spur investments in 48 T H E C O M I N G G L O B A L I Z A T I O N human capital to compensate for reduced remain relatively risky investments. If risk numbers of workers (Fougère and Mérette aversion rises with age, the aging of the rich 1997). The decline in the labor force is likely countries will imply a greater premium for to lead to higher wages, thus increasing in- risk, and thus less willingness to lend to vestments that save on labor, either in produc- developing countries. Higher risk premia tive processes or in the supply of services at could result either because older individuals the household level. Similarly, aging may ac- control a greater share of investment funds, or celerate technical progress by increasing in- because the share of pensions and other centives to innovate. Cutler and others (1990) institutional investors in financial systems estimate that a decline of 1 percentage point in increases.28 In the United States, the number labor force growth in 29 countries for of Americans aged 65 and over will double 1960­85 was associated with a 0.5 percentage between 2000 and 2030, so the asset holdings point increase in TFP growth. On the other of the elderly are likely to grow substantially hand, older workers may be less innovative, compared with total holdings (Bellante and reducing technical progress (Börsch-Supan Green 2004). 1996). It is likely that risk aversion does rise with On balance, it is likely that investment will age. Older individuals have less time to make decline in regions where elderly dependency up any shortfall in savings owing to the high ratios are rising, but not by as much as sav- volatility of investment returns. In the stan- ings, leading to a decline in these countries' dard models of portfolio choice, the only current account surplus (or a rise in their factor that explains age-related differences in deficit), along with a rise in global interest portfolio allocation is differences in risk aver- rates. This is roughly consistent with findings sion (Poterba and Samwick 1997). Bodie, in Helliwell (2004), where half the impact of Merton, and Samuelson (1992) show that demographic change was matched by a corre- older individuals will place a smaller share of sponding change in investment, and half their portfolio in risky assets than will showed up in the current account. Estimates younger individuals if the latter can vary their of reduced form relationships between demo- supply of labor to offset volatility in asset graphic ratios and current account balances returns. Kimball (1993) argues that facing (Bryant 2004), cross-country time-series increasing risks in general, for example higher analysis (Lührmann 2003), and forecasting medical risks, should make individuals less models based on estimations from historical willing to bear other risks, for example finan- relationships (IMF 2004; Turner and others cial risks. Samuelson (1989) concludes that, 1998; Higgins 1998) find that countries with assuming that one must ensure a minimum dependency ratios that are rising relative to level of wealth (to ensure subsistence) at re- other countries' tend to experience a weaken- tirement, younger individuals will be more ing of current account balances.27 willing to take risks than older. Despite this theoretical support, empirical Implications for developing countries' access estimates of the relationship between age and to finance. According to this chapter's risk aversion are inconclusive (Ameriks and simulations, the high-income countries' Zeldes 2004).29 Measuring whether aging is current account surplus is likely to deteriorate associated with a shift to less risky assets is by $800 billion by 2030, or 1.7 percent of fraught with difficulty because it is hard to GDP. The decline in capital flows to distinguish the impact on portfolio allocation developing countries and the rise in interest of age, of the person's date of birth (different rates on developing countries' loans may be age cohorts may behave differently), and of greater than anticipated by this demographic the date of observation.30 Moreover, the data model. Developing countries are likely to on household allocation, even in the United 49 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 States, are incomplete and subject to measure- savings in high-income countries. They will ment error. need to improve creditworthiness through One way that developing countries could sound fiscal and monetary policies, mainte- adjust to account for increased risk aversion nance of an appropriate exchange rate, open in financial markets is a greater use of securi- trade policies, and institutional reform to im- tization, particularly of future receivables, prove the efficiency of investment. Relatively such as export revenues, remittance receipts, youthful countries in Sub-Saharan Africa and and diversified payment rights (DPRs). Securi- the Middle East and North Africa can benefit tization or structured finance techniques in from the coming rise in savings in their developing countries are designed to enhance economies, but only if an appropriate invest- the credit ratings of debt issued by borrowers, ment climate provides secure financial instru- typically to an investment grade status. This ments for keeping savings, and efficiently allo- can allow sub­investment grade borrowers to cates saving to productive investment. As pierce the sovereign "rating ceiling," which capital becomes scarcer in the global economy, often constrains the access of subsovereign many developing countries will have the op- entities in developing countries to interna- portunity to improve financial returns and di- tional capital markets (Ketkar and Ratha versification through capital outflows, while 2001). Securitization usually results in re- low-income countries in particular could ben- duced spreads and longer maturities for efit from South-South flows. South-South cap- emerging market debt issues, compared to ital flows have risen greatly over the past conventional or unstructured debt. While tra- decade (World Bank 2006b), and demo- ditional items such as oil and gas and mining graphic trends will provide a further impetus receivables were among the first to be securi- over the next quarter-century. In short, policy tized, other assets (such as remittances and reform and a strengthening of the institutional DPRs) have increasingly taken their place in environment should enable developing coun- recent years (figure 2.13). tries to maintain their access to the savings re- Fundamentally, however, developing coun- quired for growth in the face of a decline in tries will need more than innovative financing external finance from industrial countries and techniques to deal with the coming decline in a rise in global interest rates. The transition to the medium term. Figure 2.13 Future-flow securitizations in Expectations of the decline in industrial developing countries, 1990­2004 countries' savings over the medium term Telephone receivables Airline ticket receipts may have important implications for short- 3% 3% term instability in financial markets. The Export receivables: other Credit card sustainability of the U.S. current account deficit 10% receivables 20% is an important vulnerability. As outlined in previous editions of Global Economic Prospects, there is a danger that investors will lose confidence in the ability of the United States to finance continued, large deficits, Export leading to a sharp decline in external finance receivables: Remittances oil and gas and DPRs and thus some combination of large increases in 37% 19% interest rates and a sharp depreciation of the Export receivables: dollar. Anticipated demographic trends would mining 8% exacerbate the shortfall between the existing level of the U.S. current account deficit and Source: Ketkar and Ratha 2005. what foreigners are willing to finance. 50 T H E C O M I N G G L O B A L I Z A T I O N The basic issue is that the baby boom gener- Trade, FDI, foreign travel and education, ation in the United States is now passing and improvements in mass communication through what should be its period of highest have all played a significant role in the past saving--if baby boomers are to ensure that they 25 years to enhance productivity in many have adequate financial resources to sustain parts of the world--and a reinforcement of themselves through retirement. While high these trends is likely to continue. immigration rates (relative to those in other in- More than ever before, all countries have dustrial countries) should continue to support access to a large share of the world's most labor force growth for the next decade, the U.S. advanced technology through improvements in labor force is forecast to slow to 0.5 percent communications technology and access to the from 2005 to 2015 (compared with 1.3 percent World Wide Web.32 The capacity to harness from 1995 to 2005) and should actually decline these technologies has enabled countries such beginning about 2020. At the same time, U.S. as China and Thailand to quickly advance up personal saving rates are at their lowest point the technology ladder and evolve from export- since the government began compiling consis- ing natural resource­ and/or low-skilled, labor- tent statistics in 1959.31 If saving rates do not based goods toward exporting advance rise in the near term, the country will be in a technology­laden goods such as microproces- very poor position to face rising dependency sors and flat panel displays. Given the greater ratios, a declining labor force, and hence an availability of information and technology, impetus for further declines in savings. what will differentiate countries is their ability More generally, unfunded pension liabilities to adopt these technologies--the skill level of combined with anticipated demographic trends their workforce, the appropriate capital and in- pose a considerable challenge to industrial- frastructure, openness to trade and FDI, and country policy makers that could imply more generally the investment climate. slow growth, economic instability, or both. Some technologies actually allow firms and Industrial-country governments may impose even individuals to overcome these obstacles. higher taxes to cover unfunded pension costs, Mobile telephony and access to the Internet eroding incentives to work and invest. Alterna- have the potential to transform and raise in- tively, governments may accommodate the formation sharing to unprecedented levels, conflicting demands of pensioners and current particularly for the poorest and most isolated workers through monetary expansion, leading in the global economy. For example, Sub- to inflation and a more pronounced economic Saharan Africa has long lagged most develop- cycle. In any event, an inability to appropri- ing countries in telecommunications infra- ately deal with the challenges posed by the structure. Mobile telephony penetration has demographic transition would have serious been impressive (figure 2.6)--and as noted consequences for the global economy. above, the number of subscribers most likely largely understates the actual access because Technological diffusion: productivity, small-scale mobile service firms have made information, and knowledge service available to a much broader share of It has long been recognized in the economic the population through the selling of access to literature that higher incomes are produced in small time slices of mobile phone service. the long run primarily through productivity A significant portion of productivity growth growth rather than factor accumulation. With can be captured by technology embodied in declining labor forces in some countries and imported inputs and capital, or through learn- declining labor force growth in all, productiv- ing by doing or imitating. At the same time the ity will play a more prominent role in main- larger and more diverse developing economies taining economic growth over the next have built considerable, if yet infant capacity in 25 years. research and development. But one particular 51 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 challenge for the global community will be to would have fewer resources to tackle them improve the research and development poten- and be more reluctant to compromise in un- tial for underfunded regions and sectors--for dertaking multilateral action. Faster growth example to jumpstart a green revolution in Sub- would likely ease distributional concerns and Saharan Africa or in medical research to allevi- labor market adjustments, but increase pres- ate the scourge of tropical diseases.33 sures on the global environment. The bright side of faster growth for the environment is that an accelerated pace of technological What will happen if growth is changes and investments in capital stock slower--or faster--in the next means that abatement technologies can be adopted sooner and at lower costs than with 25 years? slower growth. H istory has shown that past trends are not immutable over time. In fact, the only A slow-growth scenario thing certain about the future is that surprises The last quarter-century has shown a diver- will occur. However, even if growth rates turn sity of growth trends across regions, but out to be faster or slower than in the central trends have been less volatile in the global ag- scenario, the demographic and globalization- gregates. It is hard to identify in the global related strains in the global economy iden- aggregates well-known systemic crises such as tified in that scenario are likely to persist--if the Latin America debt crisis, the fall of the in somewhat different form. If developing Berlin Wall, or the more recent Asian finan- countries grow by only by 1.5­2 percent per cial crisis and its aftermath--with the one no- capita over the next 25 years, a glum scenario table exception of the energy crisis of the from any point of view, globalization-related 1970s. Figures 2.14a­c show the evolution of problems would remain--including the issues long-term per capita growth rates over the examined in subsequent chapters, such as period 1970­2005. The growth rates reflect income distribution, labor market adjustments, the 10-year period average growth rate for and the environment. Slower growth is likely to each year. That is, the 1970 number reflects heighten all of these problems, as countries the average annual growth rate between 1960 Figure 2.14a Past global growth... 10-year per capita income annual growth rate, percent per year, 1970­2005 4.5 4.0 3.5 High-income countries Developing countries 3.0 2.5 2.0 World 1.5 1.0 0.5 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Source: World Bank DDP. 52 T H E C O M I N G G L O B A L I Z A T I O N Figure 2.14b ...has been around 2 percent per capita for high-income regions... 10-year per capita income growth rates, percent per year, 1970­2005 10 Japan 9 8 Korea, Rep. of 7 6 5 High-income countries 4 3 2 United States 1 EU15 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Source: World Bank DDP. Figure 2.14c ...and much more volatile in developing countries 10-year per capita income growth rates, percent per year, 1970­2005 8 6 East Asia and the Pacific 4 South Asia 2 All developing countries 0 Latin America Sub-Saharan Africa 2 Middle East and North Africa Europe and Central Asia 4 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Source: World Bank DDP. and 1970 in per capita terms. By 2005, the would be about 25 percent lower per capita. growth rate for developing countries had This translates into a reduction of GDP in accelerated to 3.4 percent. 2030 by some $5.5 trillion, nearly $800 per If instead of the 3.1 percent growth in per person. This would be disappointing, and it capita incomes assumed for developing coun- underscores the importance of competent tries in the central scenario, developing coun- collective global economic management--and tries were to grow at their average for the en- well-conceived domestic policies. tire 25-year period of 1.9 percent (with world Still, it would take a sharp set of shocks to growth a meager 1.4 percent), their incomes depress growth rates to this level. And only if 53 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 these shocks were to occur in tandem, in more 1930s, with many more actors having a much than one region, and with some adverse policy greater stake in an open global economy. But feedbacks would rates likely be depressed sub- in many countries, domestic pressures to re- stantially below this level. Even then, the re- verse the trends toward greater openness are versal in the growth and global integration ever present and one can never be too com- process worldwide would likely be relatively placent about the strength of existing interna- short-lived. One reason is that one sees much tional institutions. greater stability on average for the three major A key downside risk for high-income coun- economies of the world--Japan, the European tries may come from the transition from a Union (EU15), and the United States (fig- regime of steady economic growth and rela- ure 2.14b). Together, they make up more than tively stable labor force to one with a declin- two-thirds of the global economy. The Euro- ing labor force and a rising and dependent pean Union, and even more so Japan, benefited population of elderly. The long stagnation of after the end of World War II from catching up Japan through the 1990s and early part of this to the United States and rebuilding after the decade may be an indication of the pressures devastation of the war. The oil crisis of the high-income countries will face in the next 1970s made a dent in the long-term growth decades. The pressures are already being felt rate in the early 1980s, but after a period of in Europe and the United States as economic adjustment, long-term growth was fairly policy makers attempt to deal with the im- steady throughout much of the remaining pending "transfer" crisis--the benefits period. The exception is Japan, which had a promised to aging baby boomers will translate long period of adjustment during the 1990s, into ever-higher tax rates on ever-smaller perhaps in part related to its changing workforces unless benefits are modified. The demographics--occurring earlier than else- way out will most likely involve a package of where. Korea, which in 1980 was not yet con- steps, in order to minimize the overall costs, sidered a high-income country, continued to but there is no guarantee that these steps will show the process of catch-up that is only now be politically acceptable. beginning to show signs of fading. The figure The history of the 20th century, if not ear- suggests that it would take a really major lier, has also shown the danger to the global event to shove the high-income countries off well-being from the competition for ever- their relatively steady rate of 2 percent growth. scarcer resources, for example energy or The early energy crisis was such an event for a minerals. The overall outlook for resources, at few years, but the long-term stagnation in least through 2030, suggests the ability to Japan has not had the same impact. cope with a growing global economy. It is always possible that nonlinear distur- Smoothly functioning markets should be able bances may cause a break in trends. The to allocate resources and/or provide the right downside risks are also potentially consider- signals for developing and supplying alterna- able. As history has shown, countries could tives. Nonetheless, interference with markets backtrack on their commitment to openness. that lead to substantial market disturbances Failure to address the negative consequences could lead to a rise in international tensions of a more integrated global economy could and pressures to use military force. Over the generate domestic pressures to reverse the last 50 years, conflicts that have arisen have process of opening. International tensions been relatively contained, but in a changing might degenerate into tit-for-tat tariff escala- global environment where economic and tion or competitive devaluations. This was political objectives do not necessarily align, certainly an important factor in driving the the chance for miscalculations could lead to world into recession in the 1930s. The world broader-based conflict with significant global is probably more integrated today than in the implications (see box 2.2). 54 T H E C O M I N G G L O B A L I Z A T I O N Box 2.2 Challenge of geopolitical shifts for long-term economic forecasts: lessons of history T he economy does not exist in a vacuum; it is af- fected in myriad ways by the political, historical, Predicted G5 GDP, 1990 International and social context in which economic agents act, and $ millions (ln) history shows that to ignore this geopolitical context Prediction using ARMA(1,1) with time trend can lead the economic forecaster awry. For example, Log GDP at the beginning of the 20th century the prevailing Great Depression mood in Europe and its offshoots (such as the United 31 WWI WWII States) was one of optimism and confidence. Per capita growth in Europe had accelerated to an un- Upper bound precedented 1.5 percent between 1896 and 1913 on 30 Central the heel of 1.1 percent growth between 1820 and prediction 1896 coming after three centuries of near-zero growth. This growth was sustained by a relatively 29 peaceful geopolitical environment thanks to the Pax Lower bound Britannica and a stable balance of power in Europe, rapid technological change brought about by the first 28 (steam) and second (electricity) industrial revolu- tions, and policy changes that enhanced openness, Actual GDP such as Britain's decision to reduce protectionism. 27 However, this rapid economic growth was accompa- nied by important political, social, ideological, and 1860 1880 1900 1920 1940 1960 military changes, and by 1913 there was a growing Sources: Maddison 2001; staff estimates. sense that war was somehow inevitable. Based on the historical growth of the 30 years to 1900 and using standard econometric estimates, one would have predicted a fairly optimistic GDP trend By 1913 the political environment had deteriorated for the "G-5" (France, Germany, Japan, the United and led to outbreaks of conflict that ultimately esca- Kingdom, and the United States) for the first half of lated into World War I. In its aftermath, the Great De- the 20th century. The box figure shows the central pression arose from a severe real economic shock that prediction and the upper and lower bounds. The was exacerbated and propagated worldwide owing to prediction is calculated by assuming yearly shocks poor economic management, an unprepared financial similar in nature to those observed between 1870 system, and weak institutions. In the long run, however, and 1899 and implicitly on the strength of eco- the forces pushing worldwide economic integration for- nomic dynamics generated by continued globaliza- ward dominated the adverse impact of political shocks tion and technological progress. With the geopoliti- and globalization recovered powerfully after World cal shocks of the early 20th century, however, War II. Policy was decisive in facilitating postwar resur- economic development followed a very different gence: to keep "history at bay" in the words of a noted path. In fact, by 1949, actual GDP (of the G-5) was historian, domestic and international institutions were almost $300 billion (in 1990 international dollars) built that have allowed globalization to flourish in an below the lower-bound prediction, an amount environment of relative global peace. equivalent to 13 percent of actual output, highlight- ing the substantial and persistent effects of adverse geopolitical events. Source: Fardoust and Goldberg 2006. 55 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 The ability of the planet to carry a growing debt levels (at least for developing countries on population with rapidly growing demand for average), more diversified economies with less goods and services may be put to a severe test reliance on volatile commodities, a much as the world moves forward. And even if cat- greater role for services (which tend to be less astrophes are largely avoided in the years to volatile), much improved production manage- 2030, there is growing evidence that action ment with lower inventories (which tended to needs to be taken soon, if not immediately, to be a major factor in past business cycles), and avoid catastrophe in some future not far away. better macroeconomic management, particu- Rising incomes provide an opportunity--and larly monetary policy. the desire--to deal with many environmental The past 25 years have had numerous set- issues, but this is no guarantee that the right backs afflicting growth in the developing coun- decisions will be taken. Major changes in the tries. Four of the six regions have suffered from environment, such as higher-than-predicted very long bouts of stagnant or even negative temperatures and/or dramatic changes in growth--Sub-Saharan Africa, the Middle East weather patterns could seriously impact re- and North Africa, Latin America, and Europe gional economies, if not the global economy, and Central Asia. They each had specific rea- with lower productivity, or worse yet, sickness sons for these periods of depressed growth and deaths. ranging from Latin America's debt crisis in the Deviations from the central scenario are 1980s, the Middle East and North Africa's more likely to be in the form of extended peri- (and, to a lesser extent, Africa's) energy de- ods of very rapid growth in some countries and cline, and Europe and Central Asia's emer- regions and extended periods of stagnation in gence from its transition toward market-based others, such as those witnessed over the past economies. Nonetheless, growth has been 25 years. A cataclysmic event that affects the much improved overall since 1998, in almost entire globe for an extended period has a low all regions, significant crises notwithstanding, probability--though from a geopolitical and the decade-long run to 2005 produced in- point of view, the world is likely in a period of creases per capita of some 4.6 percent annually. transition. The end of the Cold War has shifted Therefore the upside potential is even the world's major stress point and was, to a higher for developing countries, not only for large extent, a conflict among the industrial technological and policy reasons, but also be- countries--even if it had global spin-offs. New cause the current momentum in the global tensions are more likely to arise between the economy remains strong. Developing-country traditional industrial powers and developing growth could exceed 3 percent in 2020 and be countries--those that are rapidly rising and down to 2.2 percent by 2030 in the central will ask for an increased voice in global discus- scenario (figure 2.15). sions and decision making and those from The upside potential for the high-income failed states and/or regions. The already exten- countries, however, is much more muted. This sive integration of many countries in a global scenario therefore implies greater income con- economy raises the stakes for all, but also pro- vergence between developing countries and vides an incentive to find a resolution through developed countries. It would also imply peaceful methods rather than through violence. much greater weight for developing countries in the global economy. Instead of rising to What if the world grows faster? 37 percent from an initial share of 21 percent in The global economy is benefiting from an- 2001, the developing-country share in global other period of sustained and broad-based output would be 43 percent. More remarkably, growth. Among reasons are improved macro- China's share would climb to 15 percent, economic conditions (such as less inflation and almost on a par with Europe, from its share of inflationary expectations), more sustainable 3.7 percent in 2001, whereas the share for the 56 T H E C O M I N G G L O B A L I Z A T I O N Figure 2.15 More acceleration in growth is possible Average annual growth in per capita incomes between 2010 and 2030 (%) 8 7 Baseline 6 High 5 4 3 2 1 0 rld and Asia India and and and Wo China Africa Europe Japan States middle- Asia AsiaPacific South East Africa High-income countries and countries United Europe Eastthe America Caribbean Low- CentralMiddleNorth the income Sub-Saharan Latin Source: World Bank simulations using the Linkage model. United States would drop to less than 25 per- the Doha Development Agenda--without even cent from an initial position of 32 percent. describing the considerable distributional At the global level, the difference in the two impacts of removing protection in the most dis- scenarios--the central and the high-growth torted sectors, such as agriculture, could have in scenario--translates into an additional $11 many developing countries.34 trillion in 2030, of which $10 trillion is addi- There are many other choices facing policy tional income for the developing countries-- makers--most going well beyond the ability that is, an overall increase of 44 percent. Thus, of this chapter's analytical framework to cap- much is at stake. This difference in outcome is ture, such as improving institutions and the in- about the same as one would have calculated vestment climate, deepening infrastructure, in 1980 using the previous 10-year growth to and implementing policies to achieve or sur- predict where developing countries would be pass the MDGs and make for a healthier and in 2005. more productive labor force. Cumulatively, A number of discrete policy changes can making the right policy choices could dramat- have significant impacts on long-term growth ically change the growth prospects for a large even if they do not individually imply large number of countries. There are a number of deviations from baseline growth rates-- countries that have demonstrated the ability particularly as seen from a regional or global to achieve very high growth rates for very long level. The aforementioned trade liberalization periods even with very different initial condi- scenario that is limited to merchandise goods tions in terms of endowments--human and only would raise the average developing-coun- natural--and institutions. In the same vein, try growth rate by 0.2 percentage points and as getting "everything" right may not necessarily high as 0.5 percentage points for some regions. lead to the kind of high growth rates that Cumulatively over a 25-year period, this addi- many countries in East Asia have been able to tional growth would end up as significantly generate. higher incomes for many. This illustrates the im- The potential nonlinearities could also sur- portance of making progress in the current prise on the upside. This chapter may be seri- round of multilateral negotiations known as ously underestimating the potential for further 57 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 technological change because the world is still returns to endowments could also be influ- at the dawn of the information age, the enced by sectoral changes such that a migrant biotechnology revolution, and other innova- moving from rural to urban areas and from a tions. The chapter also assumes a rather be- low-wage country to high-wage country may nign policy environment even though barriers benefit from higher wages even given the same to trade in some instances, for example agri- intrinsic endowment (that is, skill level), albeit culture and services, are still prohibitively perhaps with a correction in welfare due to high, and many countries can still make vast changes in prices. improvements in domestic policies that could The central scenario includes labor market improve the investment climate and lead to segmentation between agricultural and nona- capital deepening. Cumulatively some of these gricultural sectors (for unskilled labor alone). changes could have the same impact as that The segmentation is only partial because agri- witnessed in East Asian economies starting cultural workers seek higher-paying jobs in with Japan in the 1950s and 1960s, followed urban areas. All else equal, this would tend to by the newly industrializing economies, and raise wages in rural areas and cap wage rises in now more recently by China. urban areas (compared with a no-migration scenario). The scenario suggests that the rural exodus could be a significant factor in the years Challenges of the coming ahead with the share of agricultural workers globalization dropping from about 51 percent currently to T he discussion suggests that an abrupt re- less than 35 percent in 2030. Owing to popu- versal in current trends--particularly to- lation increase, this leads to only a small de- ward a global economic collapse--has a very crease in the agricultural labor force, but to an low probability. The central scenario--and any increase of over 1 billion in urban workers. reasonable upward or downward deviation The outward movement of agricultural work- around it--will generate mostly positive conse- ers is highest in East Asia and the Pacific and quences, but not exclusively. Globalization and leads to a decline in the urban wage premium. growth will have uneven impacts leading to Another critical dimension in determining structural change, job losses in some sectors distributional outcomes is the change in the so- and regions, and the risk of some being left be- called skill premium, that is, the ratio of skilled hind. And virtually any growth scenario will wages relative to unskilled wages. According to put stress on natural resources in the absence the estimates (and definitions) of the authors of of corrective action. this study, the share of skilled workers is ap- proximately 32 percent in developed countries Income distribution and jobs and less than 10 percent in developing coun- As mentioned, incomes rise rapidly in devel- tries. The scenario assumes an acceleration of oping countries--increasing an average 3.6 skilled workers relative to unskilled workers. percent per capita across all regions, with the Despite the acceleration in numbers, the skill fastest growth in East and South Asia.35 This premium tends to increase in most regions is some 1 to 1.2 points more than income under this scenario. This reflects the assump- growth in developed countries, so there is tion that skilled labor is a complement to capi- some overall convergence in incomes--with tal, so demand for it increases more rapidly some important exceptions, for example Sub- than supply. The skill premium increases most Saharan Africa and Latin America. rapidly in those countries with a high invest- The distributional gains across households ment rate. A second factor is the relative glut of will largely reflect household endowments--of unskilled workers as the rural exodus--largely capital, land, and skills--and changes to the an unskilled phenomenon--continues. A third underlying returns to these endowments. The factor is the relatively higher concentration of 58 T H E C O M I N G G L O B A L I Z A T I O N skilled workers in high-income elastic sectors, in 2015 from 20.2 percent in 2003 (table 2.3). notably services. The MDG target is just under 14 percent.36 The In summary, the increase in value added for poverty MDG is met broadly in all regions with all developing countries can be decomposed the glaring exception of Sub-Saharan Africa, into volume and price effects and further which will miss by a wide margin. By 2030, the differentiated by factor of production. The percent of poor living on $1 a day or less will be average annual increase is 4.0 percent over near 8 percent of the developing-country popu- the entire period. There is a rotation in value lation, or roughly 550 million persons.37 Even added toward skilled workers, their total with the longer term horizon, it is unlikely that share increasing from 11 percent to 17 percent, the 2015 poverty MDG will be met in Sub- largely taken from capital's share that declines Saharan Africa without an acceleration in to 47 percent from 59 percent in 2005. Thirty growth and more targeted interventions. percent of the increase is determined by the in- crease in unskilled wages and 15 percent by the The global environment will come under increase in skilled wages (figure 2.16). The for- increasing stress mer is more important because of the relative While incomes, inequality, and poverty are at weight of unskilled workers in total value the heart of the debate on globalization and its added. Combined, wage increases account for impacts, energy and more specifically environ- 44 percent of the total increase in value added. mental impacts are lurking not far behind. The next-largest segment is the increase in the Though the energy issue had been somewhat capital stock, representing 44 percent of the relegated to a less prominent position during growth in value added (with changes in the re- most of the 1990s and early 2000s, the recent turn to capital not a factor). The results suggest run-up in fossil fuel prices and the more alarm- a modest improvement for workers in develop- ing evidence of global warming have returned ing countries over 25 years relative to owners energy and environment to the front pages. of capital, but with a somewhat better outcome With world growth in the central scenario for skilled workers. running at about 3 percent per year on average, Under the baseline scenario, the poverty primary energy demand (coal, oil, and natural MDG is reached in 2015 at the global level, gas) runs at about 2 percent per year. Under with the headcount index falling to 11.8 percent standard assumptions, this would not generate any significant tension on energy markets, with prices rising at about 1.4 percent per year in real terms from base year levels. Demand for Figure 2.16 Wages outpace profit income natural gas would tend to outpace demand for Capital oil and coal as policies and technology tend to favor this relatively cleaner fuel. Renewable Skilled workers Volumes and nuclear energy would tend to see some- Unskilled workers what higher growth rates (see chapter 5), but from a low base and therefore making only a Return to capital modest impact. To accelerate their adaptation Skilled workers Prices requires a more significant push from policies to increase investment in these technologies Unskilled workers and taxes on conventional fuels to induce 0 10 20 30 40 50 greater substitution. Percent Stagnant or declining production in high- Source: World Bank simulations using the Linkage model. income countries and high growth in some Note: Decomposition of growth of value added for developing countries would lead to some (per- developing countries, 2005­30. haps dramatic) changes in net trade in fossil 59 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table 2.3 Regional breakdown of poverty in developing countries Millions of persons living on less than $1 per day less than $2 per day Region 1990 2003 2015 2030 1990 2003 2015 2030 East Asia and the Pacific 472 213 57 18 1,116 745 317 148 China 375 179 50 16 825 531 229 108 Rest of East Asia and the Pacific 97 34 7 2 292 213 88 40 Europe and Central Asia 2 9 5 3 23 71 40 26 Latin America and the Caribbean 49 49 38 30 125 134 118 103 Middle East and North Africa 6 5 3 1 51 62 45 31 South Asia 462 472 273 159 958 1,131 1,017 902 Sub-Saharan Africa 227 320 345 337 382 530 613 653 Low- and middle-income countries 1218 1,068 721 547 2,654 2,671 2,150 1,863 Excluding China 844 889 671 531 1,829 2,140 1,921 1,755 Percent of the population living on less than $1 per day less than $2 per day Region 1990 2003 2015 2030 1990 2003 2015 2030 East Asia and the Pacific 29.6 11.5 2.8 0.8 69.9 40.2 15.5 6.7 China 33.0 13.9 3.6 1.1 72.6 41.2 16.5 7.3 Rest of East Asia and the Pacific 21.1 6.0 1.1 0.2 63.2 37.7 13.5 5.4 Europe and Central Asia 0.5 1.9 1.0 0.6 4.9 15.0 8.4 5.5 Latin America and the Caribbean 11.3 9.1 6.1 4.1 28.4 24.9 18.8 14.2 Middle East and North Africa 2.3 1.7 0.7 0.2 21.4 21.0 12.3 6.5 South Asia 41.3 33.2 16.2 8.1 85.5 79.5 60.2 46.0 Sub-Saharan Africa 44.6 45.0 37.4 29.9 75.0 74.5 66.5 58.0 Low- and middle-income countries 27.9 20.2 11.8 7.8 60.8 50.5 35.1 26.7 Excluding China 26.1 22.2 14.2 9.7 56.6 53.5 40.6 31.9 Source: World Bank. fuels. The high-income countries may be sub- impacts of rising greenhouse gas concentra- ject to an increase in their energy imbalance tions are accelerating--at least in some parts amounting to some $400 billion in 2030 (in of the world, notably at the two poles. Even 2001 dollars), more than a doubling from accelerated penetration of clean energy is $175 billion in 2001. China's small deficit likely to leave the world largely fossil fuel could balloon to over $100 billion and India's dependent--at least over the next two to $50 billion. The positive counterparts on a decades--thus technologies need to be devel- regional basis would be the Middle East and oped that limit the damage from burning North Africa, Sub-Saharan Africa, Russia, conventional fuels, such as carbon sequestra- and Latin America. Dutch-disease-type effects tion. Such technologies combined with poli- combined with the political economy of nat- cies to accelerate the use of renewable fuels ural resource­rich countries may make it and improve energy efficiency would form difficult for some to diversify their economies the basis of a package to deal more forcefully and prepare for a post-energy future. with greenhouse gas emissions. These are Relatively benign economic impacts as re- developed further in chapter 5. gards energy do not imply that the negative externalities associated with continued depen- These three problems will require policy dence on carbon-based fossil fuels will not responses lead to severe environmental consequences-- The purpose of this chapter has been to out- if not immediately, at some point in the line a plausible evolution of the global econ- future. There is mounting evidence that the omy and to highlight some of the key findings 60 T H E C O M I N G G L O B A L I Z A T I O N from the forward-looking exercise. Irrespec- unit value (MUV) index. This index is set to 1 in the tive of whether growth rates exceed or fall base and all subsequent years. Thus future values, for short of the central scenario, it has exposed example of GDP, do not integrate the normal secular increase in prices that are generated by changes in several problems that require further analysis money supply. Technically the price of manufactured and policy response. Three of the most impor- exports of high-income countries is fixed and only tant problems--income distribution, tensions changes in relative prices matter. Say for example that in labor markets, and environmental risks that world GDP is 100 in 2001 and 200 in 2030 in real require multilateral response--are the subjects terms, that is, the volume of output has doubled. Say of the next three chapters. that relative to the numéraire, the price of GDP is unchanged so that the value is also 200. If instead there were a steady increase in nominal inflation, say prices would have increased by an average of 2.5 percent per Notes year, the price of GDP would have more or less dou- bled between 2001 and 2030 and nominal GDP would 1. Measured as the difference in the number of be 400, not 200. There would of course be no change poor in 2002 using the 1990 poverty incidence (head- in volume growth, and assuming super-neutrality, all count) and the actual number of poor (at the $1/day relative prices would also be identical to assuming zero poverty line). nominal inflation. 2. World Development Indicators 2006, table 2.19. 11. The rankings refer to the model-based aggrega- 3. Unless otherwise stated, historical growth rates tion, not at the actual country level. are calculated using a log-linear regression growth 12. To attenuate the problems with exchange rate model. movements, the rankings are based on a five-year mov- 4. The first five rounds, concluding with the Dillon ing average of dollar-based GDP centered on 1982 and Round in 1961, involved 13­38 countries at most. 2003. 5. As of December 11, 2005 (www.wto.org). 13. Here welfare is equated with income per 6. The comparisons with goods trade are not nec- capita, but of course the authors recognize that there essarily straightforward. First, there are no price in- are many other variables that affect individual well- dexes for trade in services, so they are measured only being--such as health, family and friends, and so on. in current dollar terms. Second, it is more difficult to 14. One such commonly used index is the so-called evaluate trade in services, so their level is likely to be Big Mac index popularized by the Economist. This underestimated. index compares the cost of purchasing a Big Mac in a 7. There is recent evidence that migration from the variety of cities across the world. While McDonald's new EU member countries is higher than analysts had prides itself on selling a well-recognized and largely ho- anticipated. See John Kay, "How the Migration Esti- mogeneous product across the world, the raw inputs in mates Turned Out So Wrong," Financial Times, Sep- a Big Mac--perhaps priced the same everywhere if one tember 5, 2006. assumes that they are perfectly traded on world 8. Anecdotal evidence is provided in Sharon markets--only represent a small portion of the cost of LaFraniere, "Cellphones Catapult Rural Africa to 21st the final product. A large portion of the Big Mac will Century," New York Times, August 25, 2005; Rodrique be composed of the wages paid to local workers and Ngowi, "Africa's Cellphone Explosion Changes Eco- managers and the rent on land and buildings. These are nomics, Society," Associated Press, October 16, 2005; likely to be much lower in many developing countries and Kevin Sullivan, "For India's Traditional Fisherman, and hence represent an approximation of the PPP ex- Cellphones Deliver a Sea Change," Washington Post, change rate. Thus if a Big Mac sells for an average of October 15, 2006. $4 in the United States but $1 in China, the PPP ex- 9. To avoid repetition, unless stated otherwise, all change rate would be 4. When this calculation is scaled incremental values represent changes between 2005 up, if China's GDP is evaluated at 8 trillion renminbi and 2030--either absolute levels or average annual and converted to U.S. dollars at a rate of 8 renminbi compound rate changes. per $1, its GDP in dollar terms (at the market exchange 10. All prices are at 2001 levels--the base year of rate) is $1 trillion. Using a PPP exchange rate of 4, the the scenario. Volume growth will therefore reflect 2001 Chinese economy would then evaluate to $4 trillion. weights. Prices and values reflect changes with respect However, this chapter argues that this conversion is to the model numéraire, which is a price index of man- largely valid to make intercountry welfare comparisons ufactured exports from the high-income countries-- and not for making judgments about the relative size of similar in concept to the World Bank's manufactured the respective economies. 61 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 15. There is recent theoretical and empirical work the macroeconomic literature refers to the youth and el- on what makes countries grow that could provide derly dependency ratios, it could also be true that there more practical implications for policy makers than the are macroeconomic dimensions to the gross dependency widely discussed cross-country panel regressions. Both ratio, that is, the ratio of all nonworkers to workers. of these strands rely on complementarities across poli- Countries that have had high births in the recent past cies and other development-related necessities such as and that have relatively low labor force participation infrastructure needs. In the theoretical literature this rates will tend to have higher total dependency ratios, has been referred to as the O-ring theory of growth and all else being equal. Sub-Saharan Africa, for example, comes from the analogy with the U.S. space shuttle dis- has relatively greater declines in the total dependency aster. In that disaster, it was a simple and cheap O-ring ratio than the Middle East and North Africa despite its that failed. So despite the billions of parts and scientific more rapid population growth because the labor force know-how that goes into putting the shuttle in space, participation rate is higher. In fact, on this score, Sub- the failure of any part, no matter how simple or Saharan Africa has a lower dependency ratio than Latin inexpensive, is enough to bring it down. In growth the- America, which has lower workforce participation and ory, the same can be applied. A country can have, say, a more rapidly aging population. Altering assumptions 9 out of 10 growth-related necessities absolutely per- on labor force participation--for example a rise in fect, but if the 10th is a failure, there will be no take- female and/or elderly participation--could affect these off because of the complementarities across these ne- total dependency ratios. cessities. Recent empirical work by Hausman, Rodrik, 23. The baseline does track changes in policies be- and Velasco (2004) has used so-called growth diagnos- tween the base year of 2001 and 2005, notably China's tic tools to assist in finding the bottlenecks to growth commitments following its accession to the WTO, EU and providing a road map for the appropriate se- expansion, and the removal of the textile and apparel quencing of policies to overcome the bottlenecks. quotas. 16. For example, fewer office workers could re- 24. The literature on empirical studies of the life- duce the need for office space, computers, furniture, cycle theory is voluminous. Studies of macroeconomic and the like. data in industrial countries have found significant rela- 17. Car manufacturing in Japan is much less labor tionships between dependency ratios and saving rates intensive than in other countries owing to the scarcity (see for example Masson and Tryon 1990; Meredith (and hence the price of) labor. Agriculture in the United 1995; IMF 2004; Higgins 1998; Masson, Bayoumi, States could become more mechanized in the absence and Samiei 1998; Lee and Kim 2005). By contrast, of abundant cheap labor ("The Worker Next Door" by household survey evidence typically finds only weak, Barry Chiswick, New York Times, June 3, 2006). or even positive effects, of dependency ratios on sav- 18. These numbers are based purely on the growth ings (Turner and others 1998). The difference is likely rate of the working-age population, defined as the pop- due to weaknesses in the survey data (for example, the ulation aged between 15 and 65. In effect it currently surveys often do not include pension data), the failure assumes that labor force participation rates are zero for to adequately assess the disproportionate impact of the the rest of the population and that the participation wealthy on aggregate savings, differences in age cohort rates for the 15­65 group are fixed. Further, it makes behavior, and the failure to consider interactions no explicit assumption regarding migration, though among households, firms, and government. one could infer that the proportion of migrants as a 25. This would reverse the trend of past decades. share of the population remains constant. Bloom and Canning (2004) estimate that one-third of 19. Much of the recent turnaround in Sub-Saharan the East Asian economic miracle may be accounted for Africa is also largely based on increased global demand by a demographically induced rise in savings and in- for natural resources--whose labor intensity in most vestment. Higgins and Williamson (1996) find that cases is relatively low, save for agriculture. much of the rise in Asian saving rates since the 1960s 20. The standard replacement rate, that is, the rate is due to a decline in youth dependency ratios. of fertility that would keep population at a steady-state 26. On the other hand, Börsch-Supan, Ludwig, and level is 2.1 births per female, to take into account Winter (2001) claim that the elasticity of substitution be- average mortality rates and other factors. tween capital and labor is close to one in industrial coun- 21. Of course, this represents a relatively narrow tries, meaning that production processes can be modified definition of dependency because an increasing number easily to substitute labor for capital, which would limit of youths pursue education well beyond high school the impact of demographic change on investment. and often with parental support. 27. OECD (2005) notes that the model-based sim- 22. Several additional points are worth highlighting ulations typically assume that labor is immobile, that regarding the demographic scenario. Though most of capital is perfectly mobile, and that investors have 62 T H E C O M I N G G L O B A L I Z A T I O N perfect foresight. Allowing for immigration, capital ac- higher in the base year, respectively 1,068 and 20.2 per- count restrictions, and risk aversion would reduce the cent compared with 1,011 and 19.3 percent. (A printing magnitude of capital flow shifts in response to aging. error appeared in last year's report regarding the 2002 28. The share of institutions, defined as pension numbers for the $1-a-day poverty indicator.) Principally funds, insurance corporations, and mutual funds in this is due to three factors. (1) Population figures have household portfolios in OECD countries has risen been revised. (2) The new estimates for the base year in- from 17 percent in 1970 to 38 percent in 2003 (OECD corporate 38 new surveys covering the 2003/04 period. 2005). World Bank (1997) shows that regulations also Many of the new surveys include large countries such as limit the share of investments by institutional investors Argentina, Brazil, Mexico, and the República Bolivari- in foreign assets. ana de Venezuela in Latin America; Pakistan in South 29. Morin and Suarez (1983) and Bakshi and Chen Asia; and Nigeria in Sub-Saharan Africa. (3) Adjust- (1994) find evidence that risk aversion rises with age; ments to a common base year incorporate the latest in- Riley and Chow (1992) and Halek and Eisenhauer formation on price inflation and growth in private con- (2001) find that risk aversion declines with age until sumption. These same factors also impact the forecast to 65, but then increases; while Bellante and Saba (1986) 2015, since the new surveys will be associated with new and Jianakoplos and Bernasek (1998) find that risk estimates of the poverty-to-growth elasticity and the re- aversion falls with age. Bellante and Green (2004) find vised population forecasts will impact the estimate of the that risk aversion tends to increase with age, for any absolute number of poor. One additional factor has been given level of wealth. a revision to the Chinese estimate of the poverty elastic- 30. For example, Poterba and Samwick (1997) ity with respect to changes in the income distribution (as find significant differences in asset ownership of differ- measured by the Gini coefficient). The estimate of this ent birth cohorts. Older households today are more elasticity has been revised upward so that more poverty likely to hold relatively risky stock, and less likely to is associated with a rise in inequality. While for most hold tax-exempt bonds, than younger households. But countries, the 2015 forecast assumes distribution neu- this says nothing about how the current, older- trality, in the case of China, both the rural and urban generation attitudes will change as they age. Gini coefficient is assumed to deteriorate by 10 percent. 31. Considerable controversy exists concerning the Thus, a rise in this elasticity leads to a worsening of appropriate measure of personal savings in the United poverty, all else remaining constant, though counter- States, the determinants of the steady decline in saving acted, nonetheless, by relatively high income growth. rates, and the extent to which low saving represents 37. The 2030 poverty forecast does not use the household dependence on the rise in housing assets or same methodology as the poverty forecast for 2015 reliance on pensions. Whatever the actual level of sav- and instead is based on a straight poverty and growth ing, there is little doubt that appropriately measured, elasticity approach. The 2015 forecast uses all of the it has fallen significantly over the past two decades. available household information at the country level 32. Not all technologies are freely available, of combined with a country-specific forecast of per capita course. Patents and other forms of intellectual property consumption growth. The 2030 forecast uses the elas- protection imply that still significant portions of tech- ticity approach at the regional level. It combines the nological transfers will come through joint ventures implicit growth elasticities from the 2003/2015 with firms holding the rights to those technologies. poverty forecast with the regional per capita growth 33. The philanthropic efforts of the Bill and rate between 2015 and 2030 as anticipated in the cen- Melinda Gates Foundation and others provide some tral scenario. The projection is meant to be broadly in- reason for optimism that progress can be made in some dicative of potential improvements and to highlight of these areas. regional differences. 34. Significantly more detail on the potential im- pacts of various Doha scenarios is available in Ander- son and Martin (2006). 35. Per capita growth is somewhat higher when References measured at PPP exchange rates owing to different Ameriks, John, and Stephen P. Zeldes. 2004. "How Do weights--notably for China. Household Portfolio Shares Vary with Age." 36. The 2015 forecast represents some changes from Working Paper, Columbia Business School, New last year's poverty forecast published in Global Eco- York, NY. nomic Prospects 2006. The first notable difference is the Anderson, Kym, and Will Martin, eds. 2006. Agricul- change in the base year for the forecast--now 2003 in- tural Trade Reform and the Doha Development stead of 2002. Despite the more recent year, both the Agenda. Washington, DC: Palgrave Macmillan number and percentage of poor (at the $1-a-day level) is and the World Bank. 63 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Bakshi, Gurdip, and Zhiwu Chen. 1994. "Baby Boom, `Rearview Mirror.'" World Bank, Washington, Population Aging and Capital Markets." Journal DC, August. of Business 67 (2): 165­202. Fougère, Maxine, and Marcel Mérette. 1997. "Popula- Bellante, Don, and Carole A. Green. 2004. "Relative tion Aging and Economic Growth in Seven OECD Risk Aversion among the Elderly." Review of Countries." Economic Modeling 16: 411­27. Financial Economics 13: 269­81. Goldman Sachs. 2003. "Dreaming with BRICs: The Bellante, Don, and Richard P. Saba. 1986. "Human Path to 2050." Global Economics Paper 99 Capital and Life-Cycle Effects on Risk Aversion." (October). Journal of Financial Research 9 (1): 41­51. Gordon, Robert J. 2000. "Does the `New Economy' Bernanke, Ben S. 2006. Remarks made at the Federal Measure Up to the Great Inventions of the Past?" Reserve Bank of Kansas City's Thirtieth Annual Journal of Economic Perspectives 14 (4) (Fall): Economic Symposium, Jackson Hole, WY, 49­74. August 25. Halek, Martin, and Joseph G. Eisenhauer. 2001. "De- Bloom, David E., and David Canning. 2004. "Global mography of Risk Aversion." Journal of Risk and Demographic Change: Dimensions and Economic Insurance 68 (1): 1­24. Significance." In Global Demographic Change: Hausman, Ricardo, Dani Rodrik, and Andrés Velasco. Economic Impacts and Policy Challenges, pro- 2004. "Growth Diagnostics." Kennedy School of ceedings of a symposium, pp. 9­56, sponsored by Government, Harvard University, Cambridge, the Federal Reserve Bank of Kansas City, Jackson MA. Hole, Wyoming, August 26­28, 2004. Helliwell, John F. 2004. "Demographic Changes and Bodie, Zvi, Robert C. Merton, and William F. Samuel- International Factor Mobility." NBER Working son. 1992. "Labor Supply Flexibility and Portfo- Paper No. 10945, National Bureau of Economic lio Choice in a Life Cycle Model." Journal of Research, Cambridge, MA. Economic Dynamics and Control 16: 427­49. Higgins, Matthew. 1998. "Demography, National Sav- Börsch-Supan, Axel. 1996. "The Impact of Population ings, and International Capital Flows." Interna- Ageing on Savings, Investment and Growth in tional Economic Review 39 (2): 343­69. OECD Area." In Future Global Capital Shortages: Higgins, Matthew, and Jeffrey G. Williamson. 1996. Real Threat or Pure Fiction? Paris: Organisation "Asian Demography and Foreign Capital Depen- for Economic Co-operation and Development. dence." NBER Working Paper No. 5560, National Börsch-Supan, Axel, Alexander Ludwig, and Joachim Bureau of Economic Research, Cambridge, MA. Winter. 2001. "Aging and International Capital IMF. 2004. World Economic Outlook. Washington, Flows." NBER Working Paper No. 8553, Na- DC: International Monetary Fund. tional Bureau of Economic Research, Cambridge, Jianakoplos, Nancy Ammon, and Alexandra Bernasek. MA. 1998. "Are Women More Risk-Averse?" Eco- Bosworth, Barry P., and Susan M. Collins. 2003. "The nomic Inquiry 36: 620­30. Empirics of Growth: An Update." Brookings Ketkar, Suhas, and Dilip Ratha. 2001. "Development Paper on Economic Activity 2: 113­206. Financing during a Crisis: Securitization of Future Bryant, Ralph C. 2004. "Cross-Border Macroeco- Receivables." Policy Research Working Paper No. nomic Implications of Demographic Change." In 2582, World Bank, Washington, DC. Global Demographic Change: Economic Impacts ------. 2005. "Recent Advances in Future-Flow Secu- and Policy Challenges, proceedings of a sympo- ritization." The Financier 11/12: 1­14. sium pp. 9­56, sponsored by the Federal Reserve Kimball, Miles S. 1993. "Standard Risk Aversion." Bank of Kansas City, Jackson Hole, Wyoming, Econometrica 61 (May): 589­611. August 26­28, 2004. Lee, Jong­Wha, and Soyoung Kim. 2005. "Global Im- Cutler, David M., James M. Poterba, Louise M. plications of Demographic Changes." Korea Uni- Sheiner, Lawrence H. Summers, and George A. versity. Paper prepared for the keynote speech at Akerlof. 1990. "An Aging Society: Opportunity the International Symposium on Aging Problems or Challenge?" Brookings Papers on Economic in Northeast Asia, August 25­26, Fudan Univer- Activity 1: 1­73. sity, China. David, Paul A. 1990. "The Dynamo and the Com- Levinson, Marc. 2006. The Box: How the Shipping puter: An Historical Perspective on the Modern Container Made the World Smaller and the World Productivity Paradox." The American Economic Economy Bigger. Princeton, NJ: Princeton Uni- Review 80 (2) (May): 355­62. versity Press. Fardoust, Shahrokh, and Jonathan E. Goldberg. 2006. Loayza, Norman, Klaus Schmidt-Hebbel, and Luis "Predicting the Future, Looking into the Servén. 2000. "Saving in Developing Countries: 64 T H E C O M I N G G L O B A L I Z A T I O N An Overview." The World Bank Economic Re- PricewaterhouseCoopers. 2006. "The World in 2050: view 14 (3): 391­414. How Big Will the Emerging Economies Get and Lührmann, Melanie. 2003. "Demographic Change, How Can the OECD Compete?" March. Foresight and International Capital Flows." MEA Riley, William B., and K. Victor Chow. 1992. "Asset Discussion Paper 38-03, Mannheim Research In- Allocation and Individual Risk Aversion." Finan- stitute for the Economics of Aging (MEA), Uni- cial Analysis Journal 48 (6): 32­37. versität Mannheim, Germany. Samuelson, Paul A. 1989. "A Case At Last for Age-Phased Maddison, Angus. 2001. The World Economy: A Mil- Reduction in Equity." Proceedings of the National lennial Perspective. Paris: Development Centre Academy of Sciences (November): 9048­51. Studies, OECD. Turner, David, Claude Giorno, Alain De Serres, Ann Martin, Will, and Devashish Mitra. 1999. "Productivity Vourc'h, and Pete Richardson. 1998. "The Growth and Convergence in Agriculture and Man- Macroeconomic Implications of Ageing in a ufacturing." World Bank Policy Research Working Global Context." Economic Department Work- Paper No. 2171, Washington, DC. August. ing Papers No. 193, Organisation for Economic Masson, Paul, Tamim Bayoumi, and Hossein Samiei. Co-operation and Development, Paris. 1998. "International Evidence on the Determi- Tyers, Rod, and Qun Shi. 2005. "Global Demographic nants of Private Saving." World Bank Economic Change, Labour Force Growth and Economic Review 12 (3): 483­501. Performance." Faculty of Economics and Com- Masson, Paul, and Ralph W. Tryon. 1990. "Macroeco- merce, Australian National University, June. nomic Effects of Projected Population Aging in UN (United Nations). 2004. World Population Industrial Countries." IMF Staff Papers 37 Prospects: The 2004 Revision. New York: (September): 453­85. Department of Economic and Social Affairs, McKibbin, Warwick J. 2005. "The Global Macroeco- Population Division. Available at www.un. nomic Consequences of a Demographic Transi- org/esa/population/publications/WPP2004/ tion." Centre for Applied Macroeconomic Analy- 2004EnglishES.pdf. sis, Australian National University, August. van der Mensbrugghe, Dominique. 2006a. "Linkage Meredith, Guy. 1995. "Demographic Change and Technical Reference Document: Version 6.0." Household Saving in Japan." In "Saving Behavior World Bank, Washington, DC. and the Asset Price Bubble in Japan," IMF Occa- ------. 2006b. "Scenario Building through 2030 with sional Paper No. 124, International Monetary the Linkage Model." World Bank, Washington, DC. Fund, Washington, DC. Wade, Nicholas. 2006. Before the Dawn: Recovering Morin, Roger A., and Antonio Fernandez Suarez. the Lost History of Our Ancestors. New York: 1983. "Risk Aversion Revisited." The Journal of Penguin. Finance 38 (4): 1201­16. World Bank. 1997. Private Capital Flows to Develop- OECD. 2005. Financial Market Trends: Ageing and ing Countries: The Road to Financial Integration. Pension System Reform, Implications for Finan- New York: Oxford University Press. cial Markets and Economic Policies, Vol. 2005, ------. 2006a. Global Economic Prospects: Economic Supplement 1. Paris: Organisation for Economic Implications of Remittances and Migration. Co-operation and Development. Washington, DC: World Bank. Poterba, James M., and Andrew A. Samwick. 1997. ------. 2006b. Global Development Finance. Wash- "Household Portfolio Allocation over the Life ington, DC: World Bank. Cycle." NBER Working Paper 6185, National ------. 2006c. "South-South Migration and Remit- Bureau of Economic Research, Cambridge, MA. tances." World Bank, Washington, DC. 65 3 Income Distribution, Inequality, and Those Left Behind Over the past 20 years, the global distribution over the coming decades are likely to be of income has undergone significant struc- much more dramatic than those suggested tural shifts. While aggregate measures of by the averages--and other households are global inequality have changed little between likely to benefit less than average. This chap- the 1980s and today, the relative positions of ter explores future trends in income distribu- countries and the welfare of millions of the tion to identify households positioned to world's citizens have experienced much more benefit most and least, and suggests policy in- dramatic transformations. The sustained high terventions to help spread the benefits of the growth rates of China and India (and to a anticipated growth over the next several lesser extent, those of other Asian nations) decades. Building on the demographic and lifted millions out of poverty, while the stag- educational trends described in the previous nation in many African countries caused them chapter, it explores whether incomes are to fall behind. In comparing the world income likely to become more equal across and distribution in 1980 with that in 2002, one within countries. It also examines the role of study notes that the poorest country in 2002 globalization in producing these outcomes, had a lower income per capita than the poor- through the lens of microanalysis at the est country in 1980 (Bourguignon, Levin, and household level (see box 3.1; see also Bussolo Rosenblatt 2004). The same is true for the en- and others [forthcoming], available at tire bottom 6 percent of the world income dis- www.worldbank.org/prospects/gep2007, for tribution. Are these trends likely to continue methodological details). in the future? Who will be the poor and the Findings for any specific country or region rich of 2030 under the global scenarios devel- should be taken with a grain of salt. First, mi- oped in the previous chapter? crosimulation techniques used here mimic Average incomes of people in developing markets' adjustments and agents' responses countries are expected in chapter 2's baseline only imperfectly. Furthermore, potentially scenario to converge slowly toward levels in large measurement errors and comparability high-income countries. But for households in issues affect the income and consumption particular countries and particular social data used in the microsimulation model. Sec- groups, improvements in living standards ond, the focus on income (or consumption) inequality deals with inequality of outcomes and not inequality of opportunities. This is because it is less difficult to measure income For details on the methods used to project the world in- come distribution in 2030 please visit www.worldbank. inequality than to measure inequality of org/prospects/gep2007. opportunities. 67 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 3.1 Changes in demographic structure, occupational choices, and factor rewards determine the authors' hypothetical 2030 world income distribution T his chapter's forward-looking exercise is based applied to each worker in the sample depending on on methodologies developed in recent literature, their education and sector of employment. The number including Bourguignon and Pereira da Silva (2003); of workers changing sectoral occupation and the differ- Ferreira and Leite (2003, 2004); Chen and ential growth rates in wage rewards used to "shock" Ravallion (2003); and Bussolo, Lay, and van der the study's micro-data are consistent with the results of Mensbrugghe (2006). The objective of the exercise is the global computable general equilibrium (CGE) to create a hypothetical income distribution for all model described in the previous chapter. (Note that the countries of the world in 2030. The starting point is outcomes of the CGE model are also influenced by the global income distribution in 2000, assembled using same demographic changes described above.) data from household surveys for 84 countries and The sequential changes described above reshape na- data on income groups (usually vintiles) for the re- tional income distribution under a set of strong assump- maining countries (see this book's Web site for a full tions. In particular, income inequality within population detailed list). The hypothetical 2030 distribution is subgroups formed by age, skills, and sector of employ- then obtained by applying three main exogenous ment is assumed to be constant over the period. More- changes to the initial distribution: (a) demographic over, data limitations affect estimates of the initial in- changes, including aging and shifts in the skill com- equality and its evolution. In particular, consumption position of the population; (b) shifts in the sectoral data are not available for all countries' surveys, so, to composition of employment; and (c) economic get a global picture, the study had to include countries growth, including changes in relative wages across for which only income data were available. Consump- skills and sectors. tion expenditure is a more reliable welfare measure than In reality these changes take place simultaneously, income, and its distribution is normally more equal than but in this chapter's simplified framework they are ac- the distribution of income. Finally, measurement errors commodated in a sequential fashion. In the first step, implicit in purchasing power parity (PPP) exchange total population in each country is expanded until it rates, which have been used to convert local currency reaches the World Bank's projections for 2030. The units, also affect comparability across countries. structure of the population is also changed; for exam- The resulting income distribution should thus not ple, as fertility rates decrease and life expectancy in- be seen as a forecast of what the future distribution creases, older age cohorts will become larger in many might look like; instead it should be interpreted as the countries. To accommodate these changes in the sur- result of an exercise that captures the ceteris paribus veys data, larger weights have been assigned to older distributional effect of demographic, sectoral, and eco- people than have been assigned to younger individu- nomic changes. Although the results of this exercise als. In the next step, workers move from traditional provide a good starting point for debating potential agricultural sectors to more dynamic industrial and policy trade-offs, they should not be used as the basis service sectors, and new incomes are estimated for for detailed policy blueprints. these movers. Finally, consistent with an overall growth rate of real income per capita, changes in labor remuneration by skill level and sector are Note: For details see www.worldbank.org/prospects/gep2007. Therefore policy conclusions based on in- reflection of efficient incentive structures, even come inequality scenarios in this chapter though excessive levels of inequality are should be considered with caution. For exam- often associated with market distortions and ple, some degree of inequality can be the protection of vested interests. Moreover, the 68 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D redistribution of opportunities has to include concentrated in Sub-Saharan Africa. At "deep" institutional reform often accompanied present, almost half of the poorest tenth of or "financed" by some redistribution of out- the world's people live in South Asia; by comes. These concerns notwithstanding, indi- 2030 this could be reduced to just one- viduals' economic status can shape the oppor- fifth. By contrast, Africa, now home to tunities they have to improve their situation one-third of the world's poorest people, (World Bank 2005), so income levels are often may see its share double by 2030. The like- correlated with access to better education and lihood of this outcome is high, even if fa- health, which in turn are key determinants of vorable developments in Africa continue. future earnings. To reiterate a central message · In a given growth context, individuals of the 2006 World Development Report (World will realize most of their income gains by Bank 2005: 10), "equity-enhancing redistribu- moving from one social group to an- tions can often be efficiency-increasing." other. The income gains that people With these limitations in mind, the chap- achieve by migrating out of agriculture ter's exploration of the distributional effects of into manufacturing and services or by at- future scenarios in the global economy raises taining higher skill levels surpass by far some broad policy issues. It highlights five key the gradual increases of those who do messages: not move. Consequently, and conditional on higher sustainable growth rates, poli- · For a large number of people in developing cies that reduce restrictions to mobility countries, the convergence to Organisation across sectors and that provide broader for Economic co-operation and Develop- access to education are key to spreading ment (OECD) income levels will come the benefits of growth. much faster than the average numbers sug- · Although general indicators of global in- gest. In 2030, 16.1 percent of the world come distribution will probably change population will belong to what can be little, growth will generate pressure to- called a "global middle class," up from ward increasing inequality within a num- 7.6 percent in 2000. That is, in 2030 more ber of developing countries, calling for than a billion people in developing coun- policy interventions to offset these tries will buy cars, engage in international forces. Trade integration, a key aspect of tourism, demand world-class products, globalization and important for effi- and require international standards for ciency, does not seem to systematically higher education. Compare that with only increase inequality. As average incomes 400 million people in developing countries rise, the number of poor will shrink and who had access to these kinds of living the tax base will grow, making effective standards in 2005. Assuming faster income assistance easier to provide and social convergence in a scenario where develop- safety nets a viable remedy for increasing ing countries continue for the next 25 years inequality. the sustained pace of growth in recent · Although investing in education may years, the share of the global middle class in not by itself be enough to spur growth, the world population will rise even further, improved access to education at any to 19.4 percent. This large middle class will given level of growth can limit the rise create rapidly growing markets for interna- in income inequality and reduce poverty tional products and services--and become by facilitating the movement of poor a new force in domestic politics. people from low-paying jobs in agricul- · Poverty will decline worldwide, but the ture to higher-paying jobs in industry remaining poor are likely to be more and services. 69 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 The global distribution of income increases in inequality within countries; therefore by this measure, global in- Assessment of past inequality trends equality has remained roughly constant is contentious since the late 1980s. Assessing what has happened to global in- come distribution in the last two decades-- Even though these three methodologies can and what will happen in the next 25 years-- yield quite different pictures of past and future presents challenges. Part of the difficulty lies trends, and none is clearly preferable to the with choosing an appropriate measure of in- others (Ravallion 2004), it is worth elaborat- equality. The literature identifies three main ing on some general trends.4 approaches to measuring income inequality, Intercountry measurements of inequality all of which have strengths, but each of which suggest that the last five decades of develop- measures a slightly different thing.1 ment have done little to bring the average in- comes of developing countries closer to those · Intercountry inequality is a concept fa- of OECD countries. For example, Quah (1996, vored by macroeconomists. It measures rel- 1997) finds "emerging twin peaks" in the ative movements of per capita incomes global distribution, supporting the argument across countries and gives each country an that the relative distance between the top and equal weight in the world distribution (that the bottom of the global income distribution is, population size does not matter). This has increased since the 1950s. More generally, literature tends to conclude that in the last Pritchett (1997) has concluded that a "big two decades, income distribution has time" divergence in incomes occurred between become more unequal. 1870 and 1990, evidenced by a doubling of the · International inequality takes into account gap between the per capita incomes of the rich the relative sizes of countries (that is, re- and poor countries.5 Underlying this general sults are population-weighted). Its propo- pattern is a large degree of variation in individ- nents (such as Theil and Seale 1994) point ual country performance, with growth peaks out that failing to use population weights and valleys across various regional groupings will cause, for example, the fast growth of and time periods. However, the overall trend is China to be exactly offset by the anemic of an increasing distance between countries in growth rates of Malawi or Honduras, even different income brackets, although Pritchett though the number of Chinese citizens (1997) also shows evidence of convergence at who experienced improvements in their the top of the distribution (that is, among the incomes far exceeds the populations of ei- group of today's high-income countries). ther of the other two countries.2 The Once different weights are assigned to broad consensus in this literature is that in- countries based on their population (using the come inequality has decreased, although international inequality approach), the global this finding is mostly driven by the fast income distribution appears to have imp- growth in China and India.3 roved. For example, Bourguignon, Levin, and · Global inequality, which compares indi- Rosenblatt (2004) demonstrate a decrease in vidual incomes regardless of country world income inequality between 1980 and of citizenship, is a fairly recent concept 2002, as long as the relevant inequality mea- (Milanovic 2002). It takes into account sures are not too sensitive to the distance of within-country inequality, which is ig- mean income from the bottom.6 A similar de- nored by the international inequality ap- crease is observed by Atkinson and Brandolini proach, where each individual is deemed (2004).7 However, these approaches do not to earn the country's average income. To take into account inequality within countries a large extent, fast growth in the large (see box 3.2 for the importance of accounting emerging economies tends to offset the for within-country distributional changes), 70 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Box 3.2 Aggregate economic performance: distribution matters I n measuring social welfare, economists have strug- Brazil, Mexico, Panama, and República Bolivariana gled to provide simple statistics that reflect changes de Venezuela showed strong positive growth. in both aggregate income (that is, the gross national However, because of their worsening income distribu- product--GNP) and distribution. This box presents tions, when equal weights (0.2 for each quintile) or a graphic approach--the growth incidence curve poverty weights (0.6 for the poorer 40 percent, 0.3 (GIC)--that, by jointly measuring size and distribu- for the next 40 percent, and 0.1 for the richest quin- tion effects, provides an intuitive evaluation of tile) are used, these countries display much lower wel- welfare changes. fare increases. Conversely, countries enjoying improv- The basic idea behind the GIC was already present ing income distribution during the 1960s and 1970s, more than 30 years ago in a well-known study enti- such as Colombia, El Salvador, Sri Lanka, and tled "Redistribution with Growth." In this study, Taiwan (China), scored better when their perfor- Chenery and others (1974) proposed to use the mance was measured with indicators that gave more weighted sum of the growth of all income groups as a weight to poorer individuals. summary measure for changes in social welfare. In a This weighting idea underlies the GIC, originally typical developing country the top two quintiles--the proposed by Ravallion and Chen (2003). The GIC is richest 40 percent of the population--would normally a graphical representation of the growth rate in in- account for about three-quarters of total GNP. There- come or consumption at each percentile of the distri- fore the GNP growth rate, the most commonly used bution. It can summarize the distributional effects of index of performance, measures the income growth of income growth by plotting the cumulative share of the richer minority and "is not much affected by what the population (the x-axis) against the income happens to the income of the remaining 60 percent of growth rate of the nth percentile of the distribution the population" (Chenery and others 1974: 40). The (the y-axis) when the population percentiles are trends observed in aggregate economic performance ranked in ascending order of income. Ravallion and will differ according to the weights associated to the Chen (2003) show that a measure of pro-poor various income groups. Chenery and others (1974) growth can be obtained by integrating under the found that when using GNP growth rates, where the GIC. However, a simple comparison of the growth weights are income shares of the initial distribution, rate of the poorest percentiles against the mean Growth incidence curves Mexico Brazil % per capita income growth % per capita income growth 0 15 Doha rural FTAA 1 urban FTAA all 10 Doha urban Doha all FTAA 2 5 rural 3 0 0 20 40 60 80 100 0 20 40 60 80 100 % population, ranked by per capita income % population, ranked by per capita income Source: Bussolo and Medvedev 2006. Note: FTAA Free Trade Area of the Americas. (continued) 71 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 3.2 (continued) aggregate gross domestic product (GDP), poor Global growth incidence curve: base Mexicans gained much less than poor Brazilians. In other words, the GIC for Mexico shows that % per capita income growth trade reform can be somewhat regressive, whereas 200 Growth incidence strong progressivity is observed for Brazil. This im- plies that focusing exclusively on changes in macro 150 variables cannot convey the full amount of informa- tion needed to evaluate different policy alternatives. 100 Now consider the global GIC in the microsimula- Average income gain tion here, obtained by comparing the initial situation in 2000 with a final distribution in 2030. It shows 50 that for 81 percent of the world's population, per 0 20 40 60 80 100 capita income will rise faster than the global average. % population, ranked by per capita income The growth incidence curve shows that the pat- tern of expected growth is not clearly pro-poor, Source: Authors' calculations. since the poorest 2 percent of households gain less growth rate of the entire distribution already demon- than half of the global average. Instead, future strates whether income growth is biased for or changes in the global economy are likely to particu- against the poor. For example, consider the effects of larly benefit the households in the third, fourth, and trade liberalization on the income distributions of fifth world income deciles. Although these changes Mexico and Brazil (obtained from Bussolo and do not favor the extremely poor (because the bene- Medvedev 2006). While both reforms (the Free fits are not concentrated at the bottom of the in- Trade Area for the Americas in Mexico and the come distribution), the poor and the middle class, Doha Round in Brazil) produced similar gains in taken together, benefit much more than the rich. which has been steadily increasing since the late mobility argument. For them, the fact that 1980s (World Bank 2005). Nonetheless, the ex- some world citizens lost (for example, in Sub- tent to which increases in inequality within Saharan Africa or the Former Soviet Union) is countries have offset the decreases in inequality not necessarily compensated by the fact that between them is a hotly debated subject.8 others, initially poorer, in China or India have Therefore the overall direction of change in gained. The initial income position matters and global inequality since the 1980s is not clear.9 the social cost of falling incomes is not com- Bourguignon, Levin, and Rosenblatt (2004) pensated by the social gain of increasing in- offer a "mobility" argument to reconcile the comes, even if these changes take place in the seemingly divergent strands of the literature on same income range" (Bourguignon, Levin, and intercountry and international inequality. Rosenblatt 2004: 21). Most of the improvement in global income dis- Using the global inequality approach tribution since the mid-1980s has been driven (which takes into account within-country by increases in the incomes of millions of peo- inequality), Milanovic and Yitzhaki (2002) ple in East and South Asia. So the individuals at proposed disaggregating world income distrib- the bottom of the income distribution today ution into three categories irrespective of coun- are not the same as the poor of 20 years ago. try of citizenship--the poor, the middle class, Therefore, "those who insist upon equal- and the rich, where the middle class is defined weights inequality and corresponding worsen- as individuals earning an income falling be- ing of the distribution have in mind the implicit tween the per capita income of Brazil and the 72 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D per capita income of Italy. They then showed The results of table 3.1 are based on an that, in 1993, the resulting middle-class group absolute definition of the "global middle accounted for 8 percent of global population class": the per capita income thresholds are and 12 percent of global income, and that approximately equal to $4,000 and $17,000 income differences between the rich, the poor, (in 2000 international dollars) and remain and the middle class captured 90 percent of the same in 2030.13 Since an average middle- inequality between countries and almost class family from a developing country has 70 percent of total global inequality.10 4.3 household members, these income bound- The next section turns to the future and aries imply annual household earnings of uses the concept of global inequality and three $16,800 to $72,000 in PPP terms. This ab- global classes to identify the characteristics of solute definition implies that today (as of those whose fortunes are likely to improve-- 2000) many of the relatively rich in develop- the new global middle class--and of those ing countries are in the global middle class, who risk falling behind. while the vast majority of the absolutely rich (per capita incomes above $17,000) live in The future: an emerging global middle OECD countries. Since the study projections class contain only positive growth rates for all While the global middle class' share in the pop- countries in the world, there is some "nat- ulation remained largely the same from 1993 ural" expansion in the absolute size of the to 2000, its income share rose from 12 percent middle class. However, since these growth to 14 percent (table 3.1). By 2030, the size of rates represent growth in real incomes, it is this group is projected to surpass one billion, not appropriate to eliminate this "natural" making it the fastest-growing segment of the expansion by setting higher thresholds for world's population.11 Meanwhile its income 2030 relative to the thresholds of 2000.14 share will remain largely unchanged, indicat- The study's definition of the global middle ing that inequality between countries is falling. class is based on real purchasing power, Today 56 percent of the members of the middle which remains constant throughout the class reside in developing countries; in 2030 model horizon and is therefore equally this share should reach 92 percent.12 relevant in 2000 and 2030.15 Table 3.1 The global middle class is growing, its composition changing Percentage shares 1993 2000 2030 Pop. Income Pop. Income Pop. Income Poor (per capita income below the average of Brazil) 76 29 82.0 28.7 63.0 17.0 Middle class (per capita income between Brazil and Italy) 8 12 7.6 13.8 16.1 14.0 High-income country nationals 3.4 6.8 1.2 1.0 Low- and middle-income country nationals, of which: 4.2 7.0 14.9 12.9 East Asia and the Pacific 1.3 2.0 7.3 6.4 Eastern Europe and Central Asia 0.8 1.3 2.2 1.9 Latin America and the Caribbean 1.5 2.7 2.6 2.2 Middle East and North Africa 0.4 0.6 0.8 0.7 South Asia 0.1 0.1 1.6 1.3 Sub-Saharan Africa 0.2 0.3 0.5 0.4 Rich (per capita income at or above the average of Italy) 16 58 10.5 57.5 20.9 69.0 Total 100 100 100.0 100.0 100.0 100.0 Source: Authors' calculations. Note: Totals may not sum to 100 because of rounding. Estimates for 1993 are from Milanovic (2002). Thresholds of Brazil and Italy are annual per capita incomes (2000 PPP) of US$3,914 and US$16,746. 73 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 There are several reasons behind the dra- cent of earners within their own countries. By matic increase projected in the size of the mid- contrast, only 10 percent of global middle- dle class and the major shift in composition in class members occupy the lower seven deciles favor of the low- and middle-income countries. of their national income distributions. Thus, Faster population growth in the developing for many nations, the correspondence be- world is responsible for some of the change in tween the global middle class and the within- the composition. Thus regions with population country middle class is quite low. growth above the world average (for example, The situation will change quite dramatically South Asia and Sub-Saharan Africa) will in- by 2030. A full 42 percent of developing- crease their share in the global middle class. country members of the global middle class will The main determinant of joining the middle be earning incomes in the seventh decile or class ranks, however, is not population growth lower at the national level. Consider the exam- but income growth. Although East Asia's pop- ple of China, where 56 million people belonged ulation grows more slowly than the world av- to the global middle class in 2000--each of erage, this region is projected to increase its them earning more than 90 percent of all Chi- share of residents in the global middle class by nese citizens. By 2030, there will be 361 million a factor of five, compared with a doubling for Chinese in the global middle class, and their Africa. The difference is due to the fact that an- earnings will range from the sixth to the ninth nual per capita income growth in Asia is fore- decile of the Chinese national income distribu- cast to be more than twice the growth in Sub- tion.16 They will no longer be among the rich- Saharan Africa, easily offsetting the decline in est Chinese citizens but will probably be con- the former's population share. sidered upper middle class. Another example is Another determinant of the changing com- Brazil, a country that grows one-third as fast as position of the middle class is the (unequal) China in per capita terms. Even with slower shape of the initial income distribution by re- growth, the number of Brazilians in the global gion. South Asia, which could see a dramatic middle class will expand by more than one- increase (87-fold) in the share of its residents in third by 2030. The compositional change is also the global middle class, is currently the least important. In 2000, the Brazilians in the global unequal region in the world. This means that middle class were split evenly across the eighth the benefits of its projected per capita growth and ninth income deciles of their national dis- of 3.9 percent per year (roughly equal to that of tribution. By 2030, 75 percent of the members East Asia) are distributed across the population of the global middle class will earn the incomes much more equally than in other regions. Sub- of the sixth and seventh deciles in Brazil, and no Saharan Africa, by contrast, has an initial in- member of that class will earn more than equality level that is twice as high. Therefore 80 percent of the country's population. the same amount of growth would be much Consistent with these data, by 2030 the less effective at moving large numbers of peo- middle class, together with the rich, will ac- ple up the ladder of income distribution. count for a larger share of the population in Most developing-country members of a greater number of countries. In 2000, the today's (as of 2000) global middle class earn middle class and the rich exceeded 40 percent incomes far above the averages of their own of the population in just six developing coun- countries of residence. In other words, being tries, and these countries were home to classified as middle class at the global level is 0.7 percent of the population of the develop- equivalent to being at the top of the distribu- ing world. By 2030, the middle class and the tion in many low-income countries. For exam- rich will exceed 40 percent of population in ple, in our sample, as of 2000, 165 million (out of the total 30 countries, and these countries will account 231 million) developing-country citizens in for 36 percent of the world's developing- the global middle class are in the top 20 per- country population. Therefore, although the 74 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Figure 3.1 Middle-class expansion is sensitive to growth assumptions 40 Middle-class income range Global income 30 distribution in 2000 Global income distribution 20 in 2030 (high growth) Global income distribution in 2030 (baseline) 10 Rich (84% OECD) 0 0 7 55 400 3,000 22,000 Monthly household per capita income (1993 PPP) Source: Authors' calculations. ability of the global middle class (together allowing an additional 235 million people to with the rich) to influence policy in many gain access to middle-class standards of living. low- and middle-income countries is initially In addition to growth assumptions, policy limited by its small size, this group is likely to intervention at the global and national become a much stronger political force at levels--such as trade liberalization--can also both the global and national levels by 2030. affect the rate of middle-class expansion. The The increase in developing-country nationals effects of policy reforms are considered in the in the global middle class may also policy section at the end of this chapter. strengthen developing countries in the global policy arena. The growth of the global middle class may It is important to emphasize that the pro- have far-reaching consequences jected expansion in the global middle class is The ascent of hundreds of millions of not a formal forecast. Alternative assumptions developing-country nationals into the global about income and population growth, as well middle class will produce a large group of peo- as effects of policy interventions, can have a ple in the developing world who can afford, significant impact on the estimates of table 3.1. and will demand access to, the standards of Figure 3.1 illustrates some of these possibilities living that were previously reserved mainly for by plotting the income distribution of the the residents of high-income countries. This world in 2000 and in 2030 under different growth assumptions.17 The size of the global has two major implications: the demand for international goods and services will rise, and middle class is represented by the area under pressures for policies that favor global inte- the distribution curve between the two middle- gration will increase. class boundaries. Faster growth shifts the peak of the distribution closer to the middle-class Goods and services. Much of the effect of the threshold, although even the optimistic sce- middle-class expansion on the world economy nario here--which increases growth to 1.6 per- will be realized through a changing demand for cent above the baseline growth rates--falls goods. The fact that the middle class will be short of moving the thickest part of the distrib- growing twice as fast as the overall population ution into middle-class territory. Still, under implies that multinational enterprises will be the high-growth scenario the global population able to market their products to a much larger share of the middle class rises to 19.4 percent, audience in 2030 than they do today. 75 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Furthermore, the rules of this new global The rise of the global middle class is also marketplace will be increasingly determined by likely to increase the demand for international the tastes and preferences of the developing tourism services. Already in 2004, 20 percent of world, particularly the desires of consumers in all outbound tourism came from East and South East and South Asia. Therefore, while most of Asia, with an additional 6 percent from Africa the world's purchasing power will continue to and the Middle East (World Tourism Organiza- be concentrated in the OECD countries, the tion 2006). By 2020 the overall number of global economic influence of those countries tourist arrivals is expected to double to will vastly diminish. By 2030 marketing to the 1.5 billion, with a growing share coming from developing world will be a much more developing regions (figure 3.2). important strategy for multinationals than it is today. Integration policies. A significantly larger The rise of the global middle class will also global middle class composed mainly of affect demand for services. For example, given developing-country nationals will exert a the strong correlation between education levels stronger influence on international and and income, the growing middle class is likely domestic policy making. As shown above, by to demand more and better education. The 2030 these middle-class members will constitute share of the global middle class in developing a significant share of their home country countries with less than a secondary school cer- populations, allowing them to have a greater tificate is projected to decline from 47 percent say in the policy process. in 2000 to 38 percent in 2030. This is roughly Some evidence points to a correlation be- comparable to the mean education levels tween rising incomes and a shift in demand among rich individuals in 2000, when 32 per- toward more globalization-supportive policies. cent of the working-age population had not Recent literature has found that pro-trade pref- completed secondary school. Furthermore, by erences are significantly correlated with an in- 2030 the likelihood of completing at least pri- dividual's skill level and the relative abundance mary school will be virtually the same for the of skilled labor in a given country (Scheve and rich and the middle class.18 The increased em- Slaughter 1998; O'Rourke 2003). These re- phasis on education among the middle class sults link pro-trade attitudes to the predictions will help establish the foundations for contin- of the Stolper-Samuelson trade theorem, which ued growth in the developing countries, as ris- states that wage rates for skilled workers rise ing educational attainments and growing de- (relative to the returns of other factors) in skill- mand for schooling deepen the human capital abundant countries as international trade in- stocks across the developing world. creases. Mayda and Rodrik (2005) confirm Demand for health services is also likely to these findings, while showing that individuals' rise with the growth in the global middle class. relative economic and social status is highly The ability to afford better care is a major de- correlated with pro-globalization preferences. terminant of health outcomes: the World Bank Therefore, not only will the new global middle (2005) estimates that eliminating within- class possess the means to purchase products country differences in infant health would pre- previously targeted mainly toward consumers vent 3.1 million infant deaths in developing in the OECD countries, but their demand for countries--more than three-quarters of the these products is likely to become a major dri- total reduction that could be achieved by low- ver of calls for further openness. ering mortality to the OECD averages. How- The literature on the political economy of ever, the increasing demand for education and trade policy proposes that the direction of health is likely to put pressure on the budgets of policy is determined by the preferences of the developing-country governments and will re- median voter (Mayer 1984).19 Today the me- quire heightened policy attention in the future. dian voter in most developing countries is 76 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Figure 3.2 World tourism is expected to double between 2004 and 2020 Millions of outbound tourists, 1950­2020, by region Millions Actuals Forecasts 1,600 1.6 billion 1,400 1 billion 1,200 1,000 694 million 800 600 400 200 0 1950 1960 1970 1980 1990 2000 2010 2020 South Asia Middle East Africa East Asia and the Pacific Americas Europe Source: World Tourism Organization. unlikely to be a member of the middle class, employment competition from developing- which may help explain why some studies find country nationals entering the global middle a negative relationship between pro-market class. Therefore understanding and managing policies of the incumbent party and its perfor- the effects of globalization on within-country mance at the ballot box (Olivera and Lora distribution of income are likely to become 2005). However, the near-tripling of the more important in the future; this point will global middle class by 2030 increases the like- be revisited later in this chapter. lihood that the median voter in many coun- Other policy goals--among them improved tries will have a pro-openness stance. transparency, intensified anticorruption These changes are likely to have an impact efforts, and demand for a more open society not only on the domestic policy arena (for ex- and cleaner environment--are also likely to ample, increased pressure for unilateral lower- move to the forefront of the policy agenda with ing of tariffs) but also on negotiations in the expansion in the size of the middle class. Al- multilateral forums such as the World Trade though most of these issues are usually more Organization (WTO). Countries with a easily addressed by domestic policy, multilat- rapidly growing middle class could emerge as eral efforts can assist the progress. For exam- strong proponents of improved dialogue and ple, Bonaglia, Braga de Macedo, and Bussolo faster progress on multilateral liberalization of (2001) found a strong link between increased trade in goods and services. trade openness and lower corruption in a large However, as calls to remove trade restraints sample of countries between 1980 and 1998. become stronger in some countries, they may Other challenges, such as improving the qual- turn weaker in others. Liberalization of trade ity of the environment, require at least as much may also lead to an antiglobalization backlash cooperation on the multilateral front as they from lower-income citizens of industrial coun- do in domestic policy circles. (See chapter 5 for tries, who will experience increased wage and a discussion of these issues.) 77 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Africa may fall behind standards in Sub-Saharan Africa compared to other regions. 20 Even though a rising share of the global There are three main factors driving population will have access to living standards Africa's decline: high initial income inequality, currently reserved mainly for OECD relatively high population growth, and the nationals, more than half the world in 2030 lowest per capita income growth among will continue to earn less than middle-class developing-country regions. The second and incomes. Although the share of people whose third reasons imply that more and more living standards fall below those of the middle Africans are falling behind the rest of the class will decline from 82 percent in 2000 to world, while the first compounds the problem 66.5 percent in 2030, those left behind are by limiting the ability of the poor to enjoy the likely to become increasingly concentrated in growth benefits equally. Similar mechanisms Sub-Saharan Africa, revealing geographic operate in Latin America, which also is ex- polarization in the lower ranges of the global pected to increase its share in the bottom income distribution. By 2030 Sub-Saharan decile. Slower growth of income per capita rel- Africa alone could be home to almost ative to other regions means that the share of 55 percent of the poorest decile of the world Latin America in the bottom decile could rise income distribution--an 80 percent increase by 50 percent in 2030--a much slower in- from its initial share in 2000 (figure 3.3). In crease than that of Sub-Saharan Africa, but other words, in 25 years the likelihood that a significant nonetheless. This underscores the random person in the bottom decile will live in universal importance of growth and growth- Africa may increase twofold, indicating a oriented policies, which are equally relevant significant deterioration of relative living for low- and middle-income regions. The bleak outlook for Sub-Saharan Africa (and to a lesser extent Latin America) is not foreordained or immutable. Policies that raise Figure 3.3 The world's poor may be growth rates, both international and domes- concentrated in Africa tic, as well as policies aiming at efficiency- enhancing redistributions, can lead to differ- Population share of each region in the bottom decile of global income distribution ent outcomes. Consider the third column of figure 3.3, which represents the high-growth Percent 1.9 4.1 5.5 100 scenario described in chapter 2. In this 2.0 1.6 1.0 90 6.2 9.2 13.7 scenario, Sub-Saharan Africa performs slightly 80 17.8 6.3 7.0 better because it experiences a larger-than- 70 average increase in per capita growth. By con- 60 29.5 55.3 53.1 50 trast, Latin America falls further behind. It is 40 important to keep in mind that figure 3.3 sum- 30 marizes a relative measure of performance and 20 42.6 23.5 10 19.7 that everyone's living standards improve 0 under high growth relative to the baseline. 2000 2030 (baseline) 2030 (high growth) However, the important point is that while growth is effective in raising living standards, Europe and Middle East and closing the income gap with wealthier coun- Central Asia North Africa Latin America and East Asia and tries requires faster-than-average growth-- the Caribbean the Pacific which is successfully achieved in South Asia Sub-Saharan Africa South Asia but not in Sub-Saharan Africa, even under this chapter's optimistic growth scenario. Simi- Source: Authors' calculations. larly, maintaining one's relative standard of 78 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Figure 3.4 By 2030, East and South Asia are likely to move up the global income distribution ladder, while other regions will lag Population of each region by deciles of global income distribution Percent 100 80 60 40 20 0 2000 2030 2000 2030 2000 2030 2000 2030 2000 2030 2000 2030 2000 2030 Sub-Saharan South Asia Middle East and Latin America Europe and East Asia and High-income Africa North Africa and the Caribbean Central Asia the Pacific countries Deciles of global income distribution: Low (1­3) Medium (4­7) High (8­10) Source: Authors' calculations. living is also conditional on growing at least as most of South Asia's poor will earn incomes in fast as the global average. the second and third deciles in 2030. The The relative stagnation in Africa and Latin growth effect is exactly the opposite in Sub- America is not limited to the poorest 10 per- Saharan Africa, where an average African is cent of the world. Virtually all Africans are at 30 percent more likely to be in the three risk of underperforming their counterparts bottom deciles in 2030 than in 2000. from other regions. For example, in 2000, Moving away from geographic regions, it is 59 percent of the population of Sub-Saharan possible to identify alternative typologies of Africa and 25 percent of the population of countries whose citizens could fail to improve East Asia were in the bottom third of the or even lose their position in the world income world income distribution (figure 3.4). By distribution. One group includes low- and 2030, more than three-quarters of the popula- middle-income energy exporters, defined as tion of Sub-Saharan Africa is likely to be countries whose exports of oil or natural gas among the world's poorest, while only 16 per- exceed 20 percent of their total value of ex- cent of East Asia's residents will remain in the ports.21 In 2000 citizens of energy-exporting bottom third. This contrasting performance is countries made up 15 percent of the first (bot- largely a function of the difference in per tom) decile of the global income distribution. capita income growth rates. South Asia will By 2030, the population share of energy ex- continue to be the largest group in the three porters in the poorest decile could rise to bottom deciles in 2030 owing to the very high 27 percent. Similarly, agricultural exporters initial poverty rates. But its citizens are mov- may fall behind by 2030.22 While in 2000 ing up through the ranks of the global distrib- their citizens accounted for just one-tenth of ution at a fast pace owing to high per capita the poorest global decile, that share could rise growth and, unlike in the initial situation, to 23 percent in 30 years. Although everyone 79 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 in the above countries will be better off in determinants have found that income inequal- 2030 than they are today in absolute terms, ity within countries shows no time trend. Li, these developments imply a large deteriora- Squire, and Zou (1998), using the Gini coeffi- tion in the relative living standards of a large cients for 47 developing and developed coun- share of the population.23 tries covering the period 1974­94, found no The outlook for Sub-Saharan Africa significant time trend. Bruno, Ravallion, and underlines the importance of international Squire (1998) found very few countries that efforts to reduce poverty. International devel- had recorded discernible long-term changes in opment policy is already focused on the prob- inequality in either direction. lems facing Sub-Saharan Africa, but still more More recently, however, this view of con- attention is needed. One avenue for improving stant income inequality has been challenged the lot of countries left behind will be the in- by some new evidence. Focusing on the OECD creased demand for multilateral trade liberal- countries between the 1970s and 1995, ization. Another mechanism of global income Osberg (2003) concluded that inequality redistribution that has the potential to changed relatively little in Canada, Sweden, help the poor is represented by international and Germany, but that income distribution in aid. (These two global policies are discussed the United Kingdom and the United States saw in more detail in the final section of the substantial increases in polarization.24 Similar chapter.) conclusions were reached by Atkinson (2003). In the developing world, inequality has generally increased in many, if not most, coun- Within-country inequality tries since 1980, even though a sizable minor- and poverty reduction ity of countries have exhibited the opposite T he moderately sanguine conclusions about trends toward greater equality. In East Asia, the expansion of the middle-class popula- inequality has increased significantly over the tion and the increasing access of developing- last several decades--and more so during the country residents to living standards currently recent period of high growth in China and reserved to OECD nationals are only one part Vietnam than in the earlier years of growth of of the global income-distribution story. the East Asian "tigers" (World Bank 2005). Changes in the distribution of income within However, Ravallion and Chen (2004) caution countries are no less important. Worsening against drawing a causal relationship between inequality can mute the positive effects of growth and inequality in China, since inequal- growth on poverty reduction in both the ity increased fastest during periods of short and long run, increase the risk of social slow growth. In South Asian countries, the alienation of people at the bottom of the evolution of inequality in India is difficult to income distribution, and perhaps produce ascertain owing to data problems, but other counterproductive backlashes against further countries in the region experienced very large integration with the global economy. increases in inequality during the 1990s (World Bank 2005). For the countries of Latin On balance, past trends of inequality America, de Ferranti and others (2004) show are mixed that inequality increased almost uniformly When one looks backward, clear trends of ris- during the 1980s (a period of volatile and low ing or falling inequality are difficult to iden- growth coupled with high inflation), but that tify, but recent evidence casts doubt on the in the 1990s (a period of improved macroeco- view of unchanging inequality. Some empirical nomic stability) the deterioration was less studies concerned with the intertemporal pronounced and limited to approximately half evolution of inequality and its possible the countries in the region. 80 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Is trade a cause of changes in inequality? major avenue for individuals looking to escape One potential determinant of inequality is the poverty. The size of the migration-related increasing integration of developing countries reduction of poverty depends on the initial into the global economy, which, while raising poverty rate in agriculture, and the income dif- overall incomes, may also increase the return ferential between households whose heads are to more mobile factors of production such as employed in agriculture and those whose heads capital and highly skilled workers. But the im- are employed in nonagricultural activities. For pact of trade (one channel of globalization) on all developing countries in the sample, the income inequality shows no consistent pat- headcount poverty ratio falls by 2 percentage tern. Another source of the past decade's in- points (calculated as an unweighted average) crease in inequality could be increases in the when 10 percent of the agricultural population premium for skills generated by technological moves to industry or services. In Sub-Saharan change. The effects of trade are difficult to iso- Africa, where agricultural households account late from technological diffusion and foreign for 75 percent of national poverty and investment, and the combination may raise agriculture-related incomes are only 47 percent the relative wages of skilled workers and of incomes earned in the other sectors of the widen the distribution of income (for more economy, the equivalent reduction in poverty details see the "Policy Implications" section is 4 percent. This reduction could be larger, below and chapter 4). but not all migrants are poor, and not all of the poor who migrate escape poverty.26 Demography and social mobility affect Although migration does not lead to a large equality and poverty reduction in poverty at the national level, the Another determinant of inequality is demo- improvement in welfare of individuals migrat- graphic change. The aging of the world's pop- ing from agriculture can be quite large. Even ulation may increase inequality, as older for impoverished migrants who fail to escape workers often earn higher salaries (Deaton poverty, an increase in income from migration and Paxson 1997) and inequality tends to be can reduce the poverty gap. Other long-term higher among older age cohorts (Jenkins effects can also be attributable to the migration 1995; Mookherjee and Shorroks 1982).25 The process. By reducing the labor supply in the mixed rise in inequality in developing coun- agricultural sector, wages of nonmovers in this tries has been accompanied by a fall in sector tend to rise, exerting a direct positive poverty, largely driven by high growth rates in impact on relatively poor households. East and South Asia. Nevertheless, rising in- equality will hamper further poverty reduc- Education facilitates mobility tion, particularly in Africa, where poverty is To help the poor take advantage of new rising and inequality remains high. This section growth opportunities, governments can imple- and the next one assume circumstances of ment measures ranging from expanding rele- healthy growth in the modern sectors, which vant infrastructure to increasing poor people's give rise to new jobs in industry and services. access to credit and insurance. This section fo- The role that intersectoral mobility can play in cuses on how education can improve the pro- reducing poverty is then considered, as well as poor effects of the described employment shift how policies can help the poor move between out of agriculture.27 To the extent that the occupations and take advantage of the new poor lack access to education, intersectoral opportunities offered by growth. migration may be limited. The unequal access Moving from low-paying jobs in agricul- to education in many developing countries ture, where poverty rates are often high, to is documented in World Bank (2005) and higher-paying jobs in industry or services is a works cited therein. Among the relevant 81 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 findings: family members of households headed the head of the household and the average level by women and of rural households have of education of the other members of the house- marked disadvantages in attaining higher levels hold was observed to be very high. The average of education. Additionally, parents' initial value for this correlation in the developing- wealth and education greatly influence their country sample is almost 0.4. children's educational achievements and their The influence of education on the poverty- expected earnings, thereby contributing to fu- reducing impact of intersectoral migration is ture income inequalities. For many countries, illustrated in figure 3.5 for a sample of devel- the correlation between the education level of oping countries. For the majority of these Figure 3.5 Migration out of agriculture reduces poverty more when education is more equally distributed Poverty reduction due to intersectoral migration, expressed as percentage of initial poverty Thailand Turkey Vietnam Indonesia Nicaragua Cambodia Guatemala Peru Yemen, Rep. of Mexico Egypt, Arab Rep. of Jamaica India El Salvador Bangladesh Brazil Pakistan Paraguay Honduras Burkina Faso Venezuela, R. B. de Madagascar Dominican Republic Cameroon Chile Rwanda Bolivia Sierra Leone Haiti Uganda Colombia South Africa Burundi Guyana Zambia Argentina Mozambique Panama 8 7 6 5 4 3 2 1 0 1 2 8 7 6 5 4 3 2 1 0 1 2 Percent Percent Biased selection Simulated selection Unbiased selection Source: Authors' estimates based on country-specific household surveys. Note: The three bars in the figure represent the percentage reduction of poverty headcount ratios due to intersectoral employment migration (equivalent to a 10 percent decrease in initial agricultural employment and a corresponding increase in nonagricultural employment). The reductions depicted by the teal bars (biased selection) occur when workers switching occupation are selected according to their personal characteristics (the selection bias is estimated using a logit model). White bars represent a similar estimation, except that the effect of education is netted out (simulated selection). In the case of the gray bars (unbiased selection), movers are chosen randomly. 82 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D countries, heads of households who are more infrastructure, and materials required for a educated, younger, and already in urban areas useful educational experience, not just en- are more likely to migrate from agriculture-- rolling everyone in school. and less likely to be poor.28 Thus poverty reduction through intersectoral mobility is By 2030, inequality within countries may limited, reflecting a phenomenon known as rise, leaving the unskilled poor farther biased selection, represented by the teal bars behind of the figure.29 By contrast, if heads of house- More than two-thirds of low- and middle- holds were randomly selected to move out of income countries in the study sample, compris- agriculture, poverty reduction would be ing 86 percent of the population in the greater, as a larger share of the poor would developing world, are projected to experience move (shown in the gray bars of the figure). a rise in inequality by 2030. For some countries One way the government could improve the increase is quite significant (figure 3.6). the poverty-reducing impact of intersectoral Rising inequality is worrisome because mobility would be to increase access to educa- there is an inverse relationship between in- tion. Consider this thought experiment. If equality and poverty reduction.30 Even if every individual initially employed in agricul- growth is distribution-neutral (that is, if the ture were given the same level of education, incomes of the poor rise by the same amount poverty reduction would rise closer to the ran- as average incomes), inequality can still ham- dom selection case (referred to as "simulated per the ability of growth to reduce poverty. selection" and represented by the white bars This point is illustrated in the left panel of of figure 3.5). For example, in Burkina Faso, a figure 3.7, which plots the relationship be- migration out of agriculture of 10 percent of tween the partial (neutral) poverty elasticity of those employed in agriculture reduces poverty growth and the Gini coefficient for a sample by 5.5 percent in the best-case scenario, that of 84 developing countries (see also World is, when the poor and nonpoor have the same Bank 2005). This elasticity has been calcu- chances of moving. Poverty decreases by a lated by simulating a counterfactual income smaller amount, 4 percent, when movers are distribution, where the income of each person selected according to their characteristics and in a given country rises by 1 percent, and cal- the nonpoor have a greater chance to move. If culating the resulting percentage change in the Burkinabe policy makers were able to grant poverty headcount. The results show that the same education level to all citizens em- there is a robust positive relationship between ployed in agriculture, the intersectoral migra- the level of initial income inequality and the tion considered here would approach the absolute value of the poverty elasticity. At low outcome of the best-case scenario: poverty levels of income inequality, a 1 percent in- reduction would be 5.4 percent. crease in per capita growth generates a more- It is important to reiterate that complemen- than-proportional change in the poverty head- tary policies are necessary to exploit fully the count. However, as inequality rises to the high poverty-reducing impact of expanding access levels of Lesotho or Haiti, the ability of to education. As already said, in many coun- growth to reduce poverty approaches zero. A tries a poor investment climate limits the abil- similar relationship is observed for the total ity of the economy to absorb newly educated elasticity of growth (the right panel of figure workers. Improvements in economic policies 3.7), which is calculated using observed in- and institutions are often critical to encourag- come growth rates and therefore allows in- ing the higher investment required to employ equality to vary within the sample period. graduates. And care must be taken to main- These results show that, while growth is the tain quality standards. Raising access to edu- major vehicle of lifting individuals out of cation means providing the trained teachers, poverty (see Dollar and Kraay 2002), it is 83 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Figure 3.6 Changes in inequality are mainly due to economic shifts Change in Gini coefficient for individual countries Gini coefficient 0.08 Total change 0.06 Change due to Change due to demographic shift economic shift 0.04 0.02 0.01 0.03 0.05 iti a a r r a s r r de ionani bia xico a a a sh sia d a f. of ia ia d n a al ire n a a o as HaBoliviaaicBrazilblic ma entinlvado PeruguayChileadour guayRicadova key blic ric gary biquezania ani piaInd lgar stan mbia scaanda T a menia namAf n p. oep.Nige undiLanka Jordan Jam RepuPana a Re R Ethio Bu PolanBeni Za NigeriNepd'Ivo Ug Albani Rw anda FaGuine egal Arg Sa LithuaniaUru Ecu sta lomMe GuyanaLeone temalabodiiet V Indone Sen Cameroo BurSri El ParCo MolNicaragu HonduraR.B.Repu FederatRomCo Ar Hunnglade am Ta Thailan Paki Sierra Gu Cam South Ba Maurit Moz men,Arab CôteMadaga Burkin minican Kyrgyzian Ye pt, Do Venezuela, ss Ru Egy Less Total change in income inequality More Sources: Authors' calculations based on survey data and microsimulation results. Note: Total change represents the Gini level for income distribution in 2030 minus the level in 2000. Figure 3.7 Inequality hampers the potential of growth to reduce poverty Partial elasticity vs. Gini (headcount ratio, US$2 PPP) Total elasticity vs. Gini (headcount ratio, US$2 PPP) Gini coefficient Gini coefficient .2 .3 .4 .5 .6 .7 .2 .3 .4 .5 .6 .7 0 0 MNG-R MNG-U MNG-R MNG-U VEN DOM HTI NER SLE NGA HTI MDG VNM PNG SLE TZA RWA ZMB .2 NGA VNM LKA ZWE LSO PNG NIC EST MDG CHN-U PHL NER MYS KAZ TZA ZMB CHN-R KHM LAOBDI ZAF MWI LSO IND GHA UZB ZWE BRA .5 BGD NPL YEM IRN IND KHM LKA RWA IDN GTM AZE LTU CHN-R LAO BOL THA MEX COLSLV GTM .4 MDA CHN-U BGD BDI UZBPHL MYS TUN HND BOL NIC GHA MWI ZAF PAK TUR NPL BFA ARG YEM IRN SLV TJK MRT KEN PRY BRA THA BLR ETH EGY SEN LVA ECU IDN TUR UGA KAZ ARM SEN CHL UKR YUG GEO TKM COL PAK AZE MEX ROM JAM EGY LTU ARG HND BGR ALB .6 TUN CMR MDA KGZ PER CRI PAN JAM PRY RUS BFA MAR MRT KEN BLR TJK 1 PER CIV ETH YUG ECU JOR ARM LVA TKM CHL UKR GEO CMR VEN .8 CIV KGZ BGR ALB CRI PAN ROM EST DOM RUS MAR 1 JOR 1.5 Poverty elasticity of growth Poverty elasticity of growth Source: Authors' calculations. 84 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D more likely to be pro-poor when initial in- is that countries with relatively large public equality is low. sectors and relatively high education levels Figure 3.7 thus demonstrates the long-term (such as countries in Eastern Europe and the benefits of reducing income inequality: in addi- Commonwealth of Independent States) tend tion to a contemporaneous reduction in to have more egalitarian distribution of in- poverty that may be expected from lowering in- come among skilled workers, possibly because equality, policies that promote a more equal their governments and other bureaucracies distribution of income are likely to enable the have more compressed wage structures. economy to realize greater poverty reduction Hence, changes in the demographic structure from future growth. The projected rise in in- work to reduce income inequality. By contrast, equality would imply that in 2030 poverty elas- many countries in Latin America and Sub- ticities will be lower and, with more unequal in- Saharan Africa experience an increase in in- come distribution in 2030, countries will need equality as the shares of older and more higher growth rates than they need today to skilled workers rise, since wage dispersion achieve a given reduction in poverty. If higher within these groups tends to be high. growth rates cannot be achieved, the countries Although aging and the accumulation of will need more active redistribution policies. human capital imply important changes in the demographic structure of many countries, the Within-country inequality in 2030: two main overall effect of demographic changes on in- drivers equality varies within a narrow band (figure In each country, income distribution is affected 3.6). On the other hand, widening gaps in fac- by two sets of factors: shifts in the demo- tor rewards, and particularly in the premium graphic structure of the population, in terms of paid for higher skills, tend to produce larger aging and education attainment, and changes changes in inequality and generally determine in rewards for individuals' characteristics, such the overall direction of the effect. This is as their education level, experience, sector of shown in figure 3.6, where for large changes employment, and so on. Although in the real in inequality, the distance between the black world these demographic and economic and teal marks--that is, the change in in- shocks occur simultaneously and jointly deter- equality attributable to changes in economic mine inequality changes, this analysis applies factors--increases, a sign that economic each of them sequentially and decomposes the factors are the most important determinant total change into various components. for the final level of inequality.33 This study's view of the demographic struc- The initial skill premia and the pattern of ture of the world in 2030 is based on the growth experienced by each country deter- World Bank's population projections by age mine the consequences for inequality of the group and a simple model of human capital economic factors. Those consequences are ob- accumulation that assumes a continuation of tained by applying the changes in the factor the educational trends observed over the rewards of the model in chapter 2 to the in- 1980­2000 period. Controlling for other fac- come sources of individual households. For tors, both the level and dispersion (inequality) example, countries in Latin America are of household income tend to increase with the characterized by high initial income inequality age and education of the household head.31 and relatively slow growth rates. This implies Therefore as the population shares of groups a slower transition to a service-oriented econ- with more income inequality rise, one may ex- omy and lower rates of capital deepening-- pect to see higher inequality.32 However, as both of which dampen the growth of the shown by teal tick marks in figure 3.6, there is wages of skilled workers, whose labor is a no clear pattern in changes in inequality dri- complement to capital and is highly demanded ven by demographic forces. One explanation in the service sectors. 85 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Since initial wage gaps are high--the per inequality (the incomes of unskilled-headed capita income of a household headed by an un- households are 52 percent of the skilled-headed skilled worker in Brazil is only 27 percent of incomes) and an average per capita growth of that of a household headed by a skilled more than 4 percent, which leads to a sub- worker--and growth is relatively unskilled- stantial rise in the skill premium. The rise in intensive, unskilled wages rise faster than inequality is somewhat mitigated by conver- skilled incomes and inequality tends to fall (fig- gence between farm and nonfarm incomes, ure 3.6).34 The reduction in inequality is com- but this effect is quite small because growth is pounded by the diminishing rural-urban wage concentrated in the nonagriculture sectors. differentials in countries with a comparative ad- In sum, changes in income inequality over vantage in agriculture, which tends to be rela- the next 30 years are likely to be driven mainly tively unskilled-intensive. For example, farm by changes in the rewards for individual char- workers in Brazil--a country with one of the acteristics and investment in education--rather largest decreases in the Gini coefficient--earned than globalization in isolation. Countries with 40 percent of the average manufacturing wage low initial inequality and fast growth are likely in 2000 but will likely earn more than 72 per- to experience a worsening distribution of in- cent of the average industrial wage in 2030. come, while countries with slower growth rates By contrast, countries in East and South and greater initial inequality in income are Asia will experience increasing inequality, dri- likely to see inequality fall. The results therefore ven by low initial skill premia and high per illustrate a "convergence" of income distribu- capita growth rates. Faster income growth tions across countries, which can be interpreted generates more demand for skill-intensive as a manifestation of the Kuznets hypothesis or products and requires higher rates of invest- as a consequence of the globalization-induced ment, both of which increase the returns to equalization of factor prices. skilled labor. For example, one of the largest It must be borne in mind that these trends increases in inequality in the sample is ob- are driven by the assumptions of the baseline served in India--a country with low initial scenario and are far from inevitable. Figure 3.8 Figure 3.8 Restricting intersectoral mobility can lead to large increases in inequality Change in Gini coefficient for individual countries Gini coefficient 0.15 Change attributable to low 0.10 mobility scenario 0.05 0 Change attributable Total change from to high-growth scenario baseline scenario 0.05 iti a a r y s a a a a f. r n a r n a ay va de ionan bia ico ia ia la a sh sia d of ia d al ire so HaBoliviaaicBrazilblic ma entinlvado aniaPerugu Chileadorur gua key Ricado blic ric gary bique ani piaInd mbia andaFa egal T a men bodietnamAf p. oep.Nige lgariastanolanBeni scaandaroo undiLanka a Jam n RepuPana El sta lomMexGuyanLeone tema a R Ethio Bu P Za NigeriNepd'Ivo Ug Albani Rw Guinea Jordan Arg Sa Lithu Uru Ecu Vi ParCo MolNicaragua HonduraR.B.RepuderatRomCo Ar Hun Tanzaniailan Paki Sen Fe Sierra BurSri Gu Cam Indone zam ThMaurit, Re Came South BangladeMo men Arab CôteMadaga Burkin minica Venezuela, ss Kyrgyzian Ye pt, Do Ru Egy Less Total change in income inequality More Sources: Authors' calculations based on survey data and microsimulation results. 86 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D illustrates the inequality consequences of two young men who live in rural areas of Senegal alternative scenarios: the high-growth sce- and Yemen, respectively. Rahmane and Ali nario introduced earlier (gray marks) and a have not completed primary education, work low-labor-mobility scenario (teal marks), in agriculture, and belong to families whose where unskilled workers are not allowed to per capita income is in the poorest decile. After move from farm to nonfarm activities. By in- completing his primary education, Rahmane's creasing rural incomes, the high-growth sce- probability of remaining in poverty would be nario reduces within-country inequality, al- reduced by more than 13 percentage points. though the overall magnitude of the changes is This is explained, to a great extent, by the not very large. On the other hand, limiting the 40 percent increase in his income produced by intersectoral mobility of workers markedly in- completing his primary school education. Ali's creases income inequality for the majority of efforts to combine his hard work in the field countries. For example, India experiences an with elementary studies will not be met with as 11-point increase in the Gini coefficient, great a reward. Once he gets his primary which makes its level of income inequality school degree, his probability of escaping approximately the same as that of República poverty will fall by less than 1 percentage Bolivariana de Venezuela. point, because his income will increase only The inability of workers to take up jobs in 6 percent. the urban sector counteracts the natural This example illustrates the large variation processes of growth and urbanization, apply- in the welfare effects of education among ing upward pressure on nonfarm wages while different countries in different geographical re- depressing earnings in agriculture. Even in such gions. In Europe and Central Asia, for exam- countries as Brazil, which has a comparative ple, completing primary school reduces the advantage in agriculture, labor-market rigidi- probability of being in the lowest income decile ties in the low-mobility scenario result in a sig- by 11 percentage points and increases income nificant increase in inequality. Because distor- by less than 3 percent (table 3.2). By contrast, tions can have severe effects on inequality, in Sub-Saharan Africa, completing primary ed- policy makers must be careful not to erect bar- ucation reduces the probability of being in the riers to labor mobility.35 On the other hand, lowest income decile by 7.2 percentage points as is argued below, public intervention can and increases income by more than a third. counteract the tendencies toward rising in- Even among countries in the same region, there equality by creating new opportunities that is heterogeneity. For example, in the Middle benefit low-income groups. East and North Africa, as mentioned, a Yemeni who obtains a primary education is only Who is left behind: the face of the poor in slightly less likely to end up in the lowest in- 2030. As is true today, in 2030 most people come decile (a difference in probability of less in the lowest income decile will be without than 1 percentage point), whereas Egyptians primary school education, will work in with a primary education improve their agricultural sectors, and will live in rural areas. chances of escaping the bottom decile by more Lack of education appears to be the single than 10 percentage points. Nevertheless, there most important characteristic common to is a strong negative correlation within all re- people at the bottom of the distribution. gions between the returns to education and the Completing primary education reduces the marginal effect of primary school education on probability of being in the lowest income decile the probability of being in the lowest decile: in every developing country in the forecast. where the return to education is high, the prob- However, the magnitude of this effect varies ability of remaining poor is low. dramatically across countries. Consider, for Additional variables, such as the number example, the cases of Rahmane and Ali, two of elders in the household and the gender and 87 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table 3.2 Where the return to education is high, its poverty-reducing impact is also high Poverty and income effects due to completing primary education (regional averages) Within-region correlation Marginal effect on Marginal effect on between effect of primary school probability of being in the returns to primary completion on poverty and return Region lowest decile schooling to primary schooling Sub-Saharan Africa ­0.072 0.340 ­0.674 East Asia and the Pacific ­0.079 0.242 ­0.733 Europe and Central Asia ­0.111 0.264 ­0.743 Latin America and the Caribbean ­0.066 0.431 ­0.464 Middle East and North Africa ­0.056 0.229 ­0.996 South Asia ­0.068 0.257 ­0.946 Source: Authors' estimates based on country-specific household surveys. Note: For each country, the income of a head of household and the probability of the head of household being in the bottom income decile depend on individual and household-specific characteristics--among them education, age, gender, and sector of employment. As simple averages for all the countries within the six developing regions, the first two columns represent the marginal effect of completing primary school on the probability of being in the bottom decile and on income, respectively. Table 3.3 Some factors affect the probability of being in the lowest income decile more than others--and the differences are changing over time Poverty effects of specific characteristics (developing-country averages) Marginal effects on Marginal effects on probability of being probability of being in the lowest decile in the lowest decile Difference Factor (2000) (2030) (2000­30) Primary school ­0.066 ­0.081 0.016 Secondary school ­0.110 ­0.100 ­0.011 Gender (women 1) 0.020 0.017 0.006 Age 0.002 0.002 0.000 Age squared 0.000 0.000 0.000 Number of elderly in the household 0.021 0.020 0.003 Industry effects Mining ­0.028 0.011 ­0.038 Manufacturing ­0.066 ­0.013 ­0.054 Public services ­0.066 0.008 ­0.071 Construction ­0.060 ­0.007 ­0.057 Retail, Hotels ­0.076 ­0.025 ­0.051 Transport communications ­0.065 ­0.023 ­0.050 Finance services ­0.065 ­0.014 ­0.047 Other services ­0.067 ­0.018 ­0.052 Others not well specified ­0.011 0.020 ­0.026 Source: Authors' estimates based on country-specific household surveys and microsimulation results. Note: For each country, the probability (estimated with a probit model) of the head of household being in the bottom income decile depends on individual and household-specific characteristics--among them education, age, gender, and sector of employ- ment. As simple averages for all the developing countries, the first two columns represent the marginal effect of each independent variable estimated at the initial and final years, respectively. For each country, the difference between the marginal effects between the two years has been calculated for each factor and the factor's average across all countries is shown in the last column. the sector of employment of its head, among lowest income decile than are households others, affect the likelihood of being poor headed by a man. A similar difference is ob- (table 3.3). Everything else being equal, served between workers in agricultural sectors households headed by a woman are more and those in nonagricultural sectors. Working likely (by 2 percentage points) to be in the in the former increases the probability of being 88 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D in the lowest income decile by 5 percentage Global policies: is development assistance points. a useful instrument to reduce inequality? The correlations between poverty and indi- The improving fortunes of the developing vidual characteristics change over the forecast world raise the question of whether official period. Owing to a slightly increasing skill development assistance (ODA) is still neces- premium, completing primary education re- sary. It is--for the following reasons. duces the probability of a person being in the Aid flows can have a significant impact on lowest income decile by 6.6 percentage points the global distribution of income when they in 2000; it could reduce that probability by raise the incomes of the poorest countries. To 8.1 percentage points by 2030. Hence lack of be sure, aid has to be well invested and rela- education is likely to become a more impor- tively free of corruption to be fully effective. tant determinant of who is left behind in the Bourguignon, Levin, and Rosenblatt (2006) next 25 years. By contrast, the gender of the show that as long as aid is distributed head of household will become less impor- equally--that is, it does not change the na- tant. As just noted, in 2000, households tional income distribution--in recipient coun- whose head was a woman were 2 percentage tries, its effect can be particularly beneficial to points more likely to be found in the lowest the poor: 41 percent of all aid accrues to the income decile than were male-headed house- bottom decile of the global income distribution holds. That difference could shrink to 1.7 per- and another 25 percent to the second decile.37 centage points in 2030. Finally, as agricultural Furthermore, empirical evidence demonstrates incomes approach those generated in other that aid can lead to additional growth in the re- sectors,36 disparities in the probability of cipient countries, although some studies reach poverty of workers in the agricultural sector the opposite conclusion (Easterly, Levine, and and those in other industries may be less in Roodman 2004), and others show that aid can 2030 than they were in 2000. enhance growth only in the presence of good institutional characteristics (Burnside and Dollar 2000; Collier and Dollar 2002). Policy implications To illustrate the potential effect of aid on- T hese forecasts of growth, demographic incomes by region, consider a simple exercise shifts, and trends in inequality point to that adopts the same methodology that significant challenges--and opportunities. Bourguignon, Levin, and Rosenblatt (2006) Developing countries' growing participation used to estimate the global redistribution in the global middle class will represent a sub- effects of aid flows. This chapter calculates stantial improvement in welfare for hundreds growth rates without aid using the empirical of millions of people, increase the weight of relationship between annual income growth developing countries in the global economy and ODA estimated by Collier and Dollar and in international policy, and possibly even (2002).38 It is further assumed that the share increase support for open economic policies of aid in developing countries' GDP does not that could further improve growth rates. change between 2000 and 2030, that institu- While poverty will fall quite sharply, hundreds tional quality39 remains constant, and that aid of millions of people, concentrated in Africa, is distributed equally within recipient coun- will continue to live on less than $1 a day. De- tries.40 By removing all aid, the forecast mographic shifts, coupled with unequal access growth rate for Sub-Saharan Africa would fall to both wealth and services, are likely to by more than 0.5 percentage points a year, or increase inequality within countries, thus almost one-third of projected per capita in- further hampering the potential for overall come growth. By contrast, the complete cessa- growth to reduce poverty. Policy can help tion of aid flows would have small effects on lessen the effects of these tendencies. growth in East Asia and Latin America. 89 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 enable larger reductions in poverty than those Figure 3.9 Ending aid would hurt the poor anticipated in this study's forecasts. On the Percent income loss, by centiles of global distribution other hand, efforts to reduce poverty could be if ODA were to cease (expressed as difference from baseline scenario of chapter 2) hampered by conflict, macroeconomic insta- bility, or high levels of corruption that afflict % population, ranked by per capita income 0 20 40 60 80 100 many of the poorest countries.42 0 A final limitation of the approach followed here is that it does not consider the general 4 equilibrium effects of removing aid. In fact, the 8 different without-aid growth rates may have implications for global trade (among other 12 effects) and thus may affect relative prices of Growth incidence Average 16 income goods and factors: these second-order effects loss are not considered here. However, even with 20 these limitations, the conclusion that aid can Percent be powerfully pro-poor, combined with the Source: Authors' calculations. worsening outlook for Sub-Saharan Africa in the baseline, underscores the importance of fo- cusing the aid flows on Africa and improving Figure 3.9 extends this approach to the full its effectiveness. set of countries in the microsimulation model.41 For the world as a whole, per capita Global policies: further liberalization income gains are 2 percent lower than in the of trade stands to benefit everyone baseline. Distributional effects are much more The lowering of trade barriers around the pronounced, with 87 percent of the world ex- world would benefit all segments of the world periencing greater-than-average losses, al- population, including the poor. Previous esti- though no one ends worse off in 2030 than mates showed that full multilateral trade re- they were in 2000. The bottom 1 percent of form could lift roughly 100 million people out the income distribution experiences an income of extreme poverty (defined as living on less loss of 17 percent relative to the baseline. Ex- than $2 a day)--see Anderson, Martin, and pressed positively, the poorest 1 percent will van der Mensbrugghe (2006). Increased prefer- see their incomes rise by 37 percent between ence for free trade, combined with greater visi- 2000 and 2030 if aid levels remain un- bility of the plight of the poor, may help imple- changed, versus a 20 percent real income gain ment the global reforms that can be effective in if their countries receive no aid. elevating the living standards of the poor. These results are only illustrative. The ex- Unfortunately, as illustrated by the impasse ercise assumes that the allocation of aid and in the multilateral Doha negotiations, the the effectiveness of aid in promoting growth progress toward freer trade is currently follow their historical patterns. The growth stymied and will take a major effort among penalty of aid removal is thus constant the rich and poor countries together to realize through the forecast period--and such fast- even its limited progress. growing countries as China and India appear This section illustrates the potential effects to be penalized for not receiving developmen- of a successful global trade reform by imple- tal assistance, even though they may require menting a scenario with a 75 percent cut in significantly less of it by 2030. In reality, im- tariffs and domestic support in all countries by provements in the allocation of aid to the 2025,43 thus projecting into the future the lib- poorest countries and to countries with good eralization trend of the past few decades. The policies could boost aid effectiveness and resulting income gains, which include the 90 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D both inequality and poverty in any specific Figure 3.10 Global trade reform can be country. Such policies need not interfere with pro-poor sustainable long-term growth. In fact, as Percent income gain from trade liberalization, by clearly shown by World Bank (2005: 10), the centiles of global income distribution (difference from baseline scenario of chapter 2) "dichotomy between policies for growth and policies specifically aimed at equity is false," Percent and governments should be able to design 24 equity-enhancing policies that can also in- 20 Growth crease efficiency. Even so, potential trade-offs incidence 16 may arise. Raising direct taxes to excessive lev- 12 els to finance social services, such as education, targeted to the poor may create disincentives 8 Average income gain and even curb investment. However, in the 4 long run, once access to education has become 0 more equitable, a larger share of the popula- 0 20 40 60 80 100 tion will be educated; growth should also be % population, ranked by per capita income higher. These long-term benefits of redistribu- Source: Authors' calculations. tion should be considered when assessing trade-offs between equity and efficiency. In addition, specific policies that may boost positive effects of increased trade openness on growth, such as trade liberalization, may in productivity, are quite modest: average per some cases negatively affect the poor. In many capita income (in PPP terms) in the final year cases, the solution consists of designing com- rises by 7 percent relative to the baseline. The plementary policies that mitigate the adverse distributional consequences of trade reform poverty consequences of reform rather than are summarized in figure 3.10, which for each abandoning or modifying the pro-growth centile of the global distribution shows the in- policy, either of which may have even worse come gains experienced over and above the consequences for equity. In the trade-liberal- baseline improvements in income.44 Despite ization example, mitigation policies may the modest overall gain, trade reform is decid- range from investing in access roads to im- edly pro-poor in the sample because the poor- prove access by the poor to markets, to setting est households experience income gains above up or improving safety nets, and to better the global average. Furthermore, the bottom labor-market policies and institutions. 30 percent gain slightly more than the four The design and successful implementation middle deciles, and more than twice the in- of a development strategy that positively rein- crease in incomes experienced by the top forces growth and equity objectives is highly 30 percent of the world.45 In absolute terms, country-specific. It will depend, among other these income gains translate into a 13 percent things, on countries' initial conditions in terms increase in the size of the global middle class of equity, institutions, and economic struc- and reduce the number of people earning less tures. Yet from recent literature, and through than middle-class incomes by 231 million one's consideration of the scenarios described relative to the baseline. in this and the previous chapters, some policy lessons emerge that may be relevant for a large Domestic policies: powerful instruments number of countries. to attain mutually reinforcing growth A first lesson can be inferred by observing and equity objectives that with increasing incomes and the expan- Well-designed domestic policies are likely to sion of the middle class, governments should be the most powerful instruments to reduce be able to collect larger revenues and thus 91 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 gain fiscal space for redistributive spending. that "a good deal of the success of these Furthermore, the composition of tax sources programs stems from their system of targeting also changes with a shift toward more direct [. . .] On average 71 percent of conditional taxes and fewer indirect taxes. Unfortunately [for education] cash transfer program benefits the distributional impact of such a shift cannot go to the bottom 40 per cent of families." be tested directly on this study's household sur- Evidence on the educational impact of these vey data, which do not report tax payments, programs is sparser, and given their relative but the available literature is not overly opti- recent implementation, very little is known mistic. For example, ex post studies for Chile, about the long-term earnings benefits accru- a country with one of the most effective tax sys- ing to the recipients. However, the existing tems in Latin America, estimated that in 1996 evidence is strongly positive: most reviewed the after-tax Gini coefficient was 0.496-- programs have achieved their objectives of slightly higher than the before-tax index of increasing enrollment rates among their 0.488.46 Lopez and others (2006) show that in- targeted population.47 come inequality in European countries is Increasing educational levels must be ac- barely affected by taxes and social security companied by a strong investment climate to contributions, indicating that the overall distri- ensure productive jobs for the newly educated, butional effect of taxation is almost equivalent and the quality of education needs to be main- to that of a proportional (flat-rate) tax. The tained. The shift from agriculture should be same study also shows that redistribution takes undertaken within a wider context of improv- place mainly through transfers rather than ing agricultural productivity and expanding through taxes: in most European countries opportunities in modern sectors, not through transfers seem to be almost equally distributed policies that discriminate against agriculture. across the population, thus contributing to a Labor-market policies are important in aiding substantial reduction in income inequality be- worker adjustment and enhancing mobility, fore tax and transfers. but too often such policies end up raising This evidence suggests that although taxa- the cost of hiring labor and push workers into tion can be redistributive, at least in principle, informal sectors or unemployment. World Bank transfers and expenditures (for education, for (2005) emphasizes the role of labor unions in example) may be governments' preferred improving worker conditions, but cautions that levers of redistribution. This chapter has em- product markets must be competitive to prevent phasized the critical role that education can the unions from commanding rents at the ex- play in reducing poverty. Improving access to pense of consumers. Success stories include education can reduce poverty both by increas- unionizing landless export agriculture workers ing individual productivity and by facilitating in northeastern Brazil, which resulted not only the movement of poor people from low-paying in better worker protections but also enhanced jobs in agriculture to higher-paying jobs in in- productivity and quality of output. dustry and services. Even more important, Government intervention is another mech- public spending on education (as well as on anism for improving working conditions. health and other human capacity), when tar- Cambodia was able to significantly limit labor geted toward the poor, can produce a double abuse in the textiles sector through a monitor- dividend, reducing poverty in the short run ing program designed to improve labor and increasing the chances for poor children to standards in exchange for higher U.S. import access formal jobs and thus break free from quotas. The Slovak Republic lowered employ- the intergenerational poverty trap. Empirical ment taxes and increased labor-market evidence of the double advantages of targeted flexibility through concentrated efforts by a education programs has been emphasized by reform-minded government seeking to join the other studies. Morley and Coady (2003) state European Union. 92 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Figure 3.11 The inequality effects of trade liberalization are not large and depend on the structure of initial protection Change in Gini coefficient for individual countries Gini coefficient 0.06 Change due to trade liberalization scenario 0.04 Total change from 0.02 baseline scenario 0.00 0.02 0.04 iti a a r y s ionani bia ico a a a e a f. r n a r ay de ic ia la sh sia d ia of ia d al ire so al HaBoliviaaicBrazilblic mantinlvado Perugu Chileadorur key Ricadova ric gary biqu an piaInd scaanda andaFa T aguata men bodietnamAf p. oep.Nige lgariastanolanBeni undiLanka a Jam RepuPana lomMexGuyanaLeone tema R P Zambia NigeriNepd'Ivo Ug Albania GuineanegJordan ArgeSa LithuaniaUru Ecu s MolNicaragua Vi Ethio Bu Bur Rw HonduraR.B.Republ deratRomCo Ar Cam Hunnglade am TanzaniMaurit Indone Thailan , Re Paki CameroonSri Se El ParCo Fe Sierra Gua South Ba Moz men Arab CôteMadaga Burkin minican Kyrgyzian Ye pt, Do Venezuela, ss Ru Egy Less Total change in income inequality More Sources: Authors' calculations based on survey data and microsimulation results. Domestic policy reforms can also strongly well-designed additional policy interventions, influence the final effects on inequality and especially those that improve education and in- poverty of multilateral trade reforms. As frastructure and address other "behind the shown above, multilateral trade liberalization border" investment climate reforms, can mili- has a discernible impact on global income tate against the inequality changes that may distribution; however, the "pure" trade policy result from trade liberalization. effect on within-country inequality does not Such policies are likely to play an increas- seem to be very large (figure 3.11). Within- ingly important role in the future, not least be- country inequality will change according to the cause the coming globalization will include initial pattern of protection, the evolution of two new challenges--the integration of large global prices, and the sectoral and factor- emerging economies such as China and India, specific productivity impacts (see Winters, and the global sourcing of services. While the McCulloch, and McKay 2004 and Bussolo and above scenario explores the impact of reduc- Round 2005). For example, in a scenario where tions in tariffs on goods, the global sourcing of global trade barriers are eliminated and inter- services, enabled by new advances in tech- national prices for agricultural goods increase, nology, is leading to an increasing number of Brazil, which currently protects skill- (and services-related tasks--and increasingly higher- capital-) intensive industries more than it does skilled tasks--being undertaken in developing agriculture, will likely experience a reduction countries. This will bring new implications for of within-country inequality. Conversely, India wage distribution along the skill spectrum, or Mexico, countries with tariff structures that and most likely change the inequality results protect unskilled workers (especially in agricul- shown in figure 3.11. The implications of this ture), will probably have to face pressure to- combination of technological advance and ward increasing inequality. Because they as- trade integration for workers in developing sume that other factors will remain equal, these and developed countries are discussed in more have been labeled "pure" trade effects. Clearly, detail in chapter 4. 93 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Notes 11. The size of the middle class doubles between 2000 and 2030. In comparison, the number of the rich 1. In this discussion the authors have adopted the increases by 75 percent and the number of the poor naming conventions of World Bank (2005). Milanovic decreases by 19 percent. (2005) refers to the following different measurements 12. The number of developing-country citizens in as inequality concepts 1, 2, and 3. the global middle class increases 3.25 times between 2. Bourguignon, Levin, and Rosenblatt (2004) point 2000 and 2030. The share of low- and middle-income out that using the intercountry concept may represent an country nationals among the rich rises even faster implicit welfare judgment, whereby the rising incomes (4.7 times), but their influence in this group is likely of more populous countries cannot offset the losses of to be moderate, as OECD citizens will still constitute smaller countries when their incomes are falling. one-half of the category. 3. The influence of China and India is so large that 13. Note that this is not true for the first column of omitting these two countries would reverse the conclu- table 3.1, which is based on 1993 international dollars. sion: international inequality excluding China and India 14. This study's qualitative conclusions about the has increased in the past two decades (World Bank 2005). composition of the middle class hold even if the au- 4. It should also be noted that measurement of in- thors adopt a relative definition of the middle class equality is sensitive to both the precise indicators used and confine their attention to the fifth decile of to measure it and the time horizon chosen. the world's income distribution--that is, the "median" 5. The ratio of per capita incomes of the richest individuals. and poorest country in the sample has grown by a 15. The authors' global middle class concept is in factor of more than five. this sense similar to a poverty line, which is the amount 6. Bourguignon, Levin, and Rosenblatt (2004) of real income required to buy a fixed amount of calo- show that it is possible to produce rising inequality ries. Poverty lines do not move through the periods of statistics if, for example, the sensitivity of the Atkinson growth and decline, since the latter do not affect inequality index to deviations from mean income at the caloric intake requirements. Similarly, the study's defi- bottom of the distribution is set sufficiently high nition of the middle class is based on the ability to af- (over five). ford a certain basket of goods and services, and anyone 7. Atkinson and Brandolini (2004) use the Gini who can purchase this basket (whether in 2000 or in coefficient, the Theil index, and mean logarithmic 2030) is a member of the middle class. deviation to show that income inequality declined 16. Note that the simulation design for China differs between 1970 and 2000. from the majority of countries in this study's sample. Be- 8. Bourguignon and Morrison (2002) argue that cause the authors do not have information on individual inequality between countries has been responsible for earnings, they cannot pass the changes in skill premia most of the time-series variation in global inequality. from the CGE model to the microsimulation. All other See also Milanovic (2002), who shows that in 1993, steps, including demographic change and growth in per inequality between countries accounted for three- capita incomes, remain fully consistent with the stan- quarters of global inequality. dard simulation approach. See box 3.1 and Bussolo and 9. Some of the studies examining global inequality others (forthcoming) for more details. have relied on parameterized Lorenz curves to add the 17. The distribution is plotted as a kernel density within-country dimension to the analysis: see for function of the household per capita incomes. example, Sala-i-Martin (2002a, 2002b), and Bhalla 18. In other words, the study results show that (2002). Others, such as Milanovic (2002) and World 13 percent of the middle-class population in 2030 will Bank (2005), have built up the global distribution from have completed less than a full cycle of primary educa- household surveys. tion. The relevant population share for those earning 10. The between-group decomposition is accom- more than middle class incomes is 12 percent. plished by giving each person within the group that 19. As cited in Mayda and Rodrik (2002). group's average income. As a result, differences between 20. Note that the Africans of 2030 will be better the average incomes of the rich, the poor, and the middle off in absolute terms than in 2000, since the study fore- class account for 68 percent of total world inequality. casts non-negative real growth rates even at the bottom On the other hand, if every person in the world is as- of the distribution. signed the average income of his or her country of resi- 21. There are 12 developing energy exporters in dence, income differences between countries account for the study's 114-country database. These include three 76 percent of global inequality. Thus, income differen- countries in the Middle East and North Africa, two in tials between the rich, the poor, and the middle class are Sub-Saharan Africa, three in Latin America and the responsible for the bulk of global variation in incomes. Caribbean, and four in Europe and Central Asia. 94 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D 22. There are 15 developing-country agriculture 27. By facilitating access to higher-paying jobs, ed- exporters in the study sample. A country is defined as ucation contributes to reducing poverty even for those an agriculture exporter if its exports of any one agri- workers who do not move across sectors. cultural commodity exceed 20 percent of total exports. 28. This finding is informed by country-specific re- 23. This is another case showing that this chapter's gression analysis focusing on the determinants of em- results should not be considered "forecasts" but ceteris ployment in farm and nonfarm sectors (probit models). paribus scenarios; a forecast should include at least Although this pattern is true for most countries, in some countries with negative performances. some cases migrants tend to have different characteris- 24. Osberg (2003) measures polarization by the tics. For example, in a recent study for Brazil, Bussolo, shares of population earning less than 50 percent and Lay, and van der Mensbrugghe (2006) used a more more than 150 percent of the median income. In both complex behavioral model to show that, with only a the United States and the United Kingdom, the shares few exceptions, poorer individuals are more likely to of low and high earners increased substantially over migrate to nonfarm occupations. the sample period. 29. The authors of this study assume that, once the 25. These effects are somewhat ameliorated be- migrants are selected, in one way or another, they find cause with population aging, older and more experi- a job in the rest of the economy and earn the modern enced workers tend to become less scarce, reducing their sector's higher wage adjusted to take into account their wage premium (Higgins and Williamson 1999). Also, in personal characteristics. 45 percent of the developing countries in the study sam- 30. See, for example, World Bank (2005) and ple, younger household heads tend to earn more than Lopez and others (2006). older ones, owing in part to higher education. 31. The relationship between household income 26. The maximum poverty elasticity (assuming and age of the household head is positive in approxi- that all migrants are poor and all poor migrants in- mately 70 percent of the sample countries, while crease their incomes sufficiently to escape poverty) for the age-income profile is positive in 60 percent of the developing countries is about 2--see table below, col- countries. umn (a). However, on average, of every 100 migrants 32. The literature on the evolution of income in- only 20 are below the poverty line (column b). These equality identifies three channels that determine the ef- results are based on the case where movers are selected fects of demographic change: first, given an upward- according to their characteristics (that is, with a logit sloping age-earnings (incomes) profile, aging will selection model). Furthermore, of the average 20 poor increase inequality between old and young groups migrants among the movers, 7 remain in poverty in (Deaton and Paxson 1997); second, different age their new occupation (column c). This results in an groups are characterized by different within-group in- observed poverty elasticity of 0.2 (column d). equality, and inequality tends to be higher among older Moving from agricultural to nonagricultural occupations reduces poverty in some regions more than others Migration-poverty elasticity when 10 percent of the population employed in agriculture migrates Poverty among Poverty among Observed Maximum migrants migrants after poverty poverty elasticity before moving moving elasticity Region (a) (b) (c) (d) Sub-Saharan Africa 1.60 0.35 0.14 0.36 East Asia and the Pacific 3.78 0.11 0.09 0.48 Europe and Central Asia 4.45 0.03 0.01 0.03 Latin America and the Caribbean 1.74 0.19 0.08 0.09 Middle East and North Africa 1.18 0.26 0.01 -- South Asia 1.81 0.23 0.07 0.15 Average in the developing world 2.03 0.19 0.07 0.20 Source: Authors' estimates from household surveys. Note: Column (a) shows how much poverty could be reduced in percent terms with respect to the initial poverty headcount if all migrants are initially poor and all escape poverty after the move. Column (b) represents the actual poverty headcount of movers (accounting for the fact that many nonpoor migrate). Column (c) shows the poverty headcount among movers calculated after they are assigned the income of the new occupation. Column (d) is the actual migration-poverty elasticity (for a 10 percent migration rate). -- not available. 95 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 age cohorts (see Deaton and Paxson 1997; Jenkins PPP terms, which account for lower prices of nontrad- 1995; and Mookherjee and Shorroks 1982). With able goods in the recipient countries. everything else remaining constant, when older cohorts 38. This is given by the equation: become more populous, as is the case with lower pop- Gi {Set of variables not related to aid} 0.54 ulation growth rates, aggregate inequality increases. * (ODAi/GDPi) 0.02 * (ODAi/GDPi)2 These two channels affect aggregate inequality without 0.31 * (CPIAi * ODAi/GDPi) any change in the age premium, that is, with a fixed 39. Institutional quality is captured by the World age-earnings profile; however, the third channel con- Bank's Country Performance and Institutional Assess- siders changes in inequality due to changes of the life- ment (CPIA) ratings. For International Development cycle income profile. As the population ages, older Association (IDA) member countries, CPIA scores are high-wage and more experienced workers tend to be- available online starting with 2005. come less scarce and the wage premium they initially 40. The assumption of equal distribution of ODA receive will be reduced (Higgins and Williamson implies that the removal of aid flows does not change 1999). This third channel works through the labor the income distribution within countries. On the one market and contributes to attenuating the inequality hand, assuming equal distribution may be too increases brought about by the first two channels. This optimistic--aid may not reach the desired recipients channel is explored in more detail as part of the owing to a host of factors including corruption and lack discussion on price-wage adjustments. of access to infrastructure. On the other hand, it may be 33. Some of the changes in inequality shown in fig- too pessimistic, since the rich in the recipient countries ure 3.6 may seem implausible when compared with are assumed to derive some benefits from the aid that some ex post evidence; however, the aim of this figure the donors never intended for them to obtain. is not to present forecasts of income inequality, but 41. In estimating this effect, this study's methodol- rather to show what may happen, ceteris paribus, to in- ogy and the approach of Bourguignon, Levin, and equality in a specific scenario for the evolution of the Rosenblatt (2006) differ in two important ways. First, global economy. this study uses growth rates generated by the model of 34. The simulated reduction of the gap between chapter 2 rather than historical growth rates. Second, skilled versus unskilled workers' wages for Latin America while Bourguignon, Levin, and Rosenblatt (2006) dis- is plausible and in line with recent evidence. Manacorda, regard the within-country distribution of income (by Sanchez-Paramo, and Schady (2005), using micro-data assigning every individual within a country that for five Latin American countries, show that the relative country's per capita gross national income--GNI), this rewards of workers who completed tertiary school have study's approach explicitly takes into account both increased but, apart from Mexico, relative wages of between- and within-country distributions. workers who completed secondary school have de- 42. In 2004 only 10 of 66 low-income aid recipient creased. In this study's micro-simulation the authors de- countries received a "good enough" rating of their fine a worker as skilled when his or her level of education budget system according to the CPIA indicators (World is, at least, completed secondary. In Latin America, about Bank 2006). In the presence of bad policies, conven- 25 percent of heads of household have secondary educa- tional aid delivery methods are unlikely to benefit the tion (without tertiary) compared with 12 percent of heads intended recipients, even if the aid is targeted toward with tertiary schooling. Therefore, even with an increase human development­intensive sectors such as educa- of the tertiary-educated workers' wages, the average tion and health (World Bank 1998). This is because aid wage for the group defined in the study as skilled would is often fungible and can be easily reallocated away still be reduced. from target activities once it enters the public budget. 35. Notice that this is not the same as encouraging At the same time, even the most distorted policy envi- mobility by means of "forced urbanization," which is ronments are likely to have "pockets of reform," which known to generate negative consequences. The can become the focal point of donors' efforts to improve focus here is on removing distortions rather than the overall policy environment. For example, efforts to adding them. improve public procurement--the mechanisms through 36. These income dynamics--that is, the changes which governments purchase goods and services--lie at of the premia received by agricultural workers versus the heart of the ability of aid to deliver the desired nonagricultural ones, as well as those obtained by outcomes (World Bank 2006). skilled versus unskilled--are consistent with the CGE 43. Notice that this implies a larger absolute cut in results of chapter 2. protection for developing countries, whose initial tariff 37. Bourguingon, Levin, and Rosenblatt (2006) also levels are significantly above the high-income average. show that although aid is often viewed as a zero-sum 44. In other words, figure 3.10 represents the dif- game, that is not the case if aid flows are measured in ference between the growth incidence curve of the 96 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D trade reform scenario and the growth incidence curve ------. 2006. "Global Redistribution of Income." of the baseline scenario. The horizontal line is the World Bank Policy Research Working Paper increase in the global average per capita income. 3961, Washington, DC, July. 45. Figure 3.10 shows the dynamic gains from trade Bourguignon, François, and Christian Morrison. 2002. reform, which allow for feedback from increases in "Inequality among World Citizens: 1890­1992." trade openness (exports-to-output ratio) to total factor American Economic Review 92 (4): 727­44. productivity. Since low-income countries tend to have Bourguignon, François, and Luiz Pereira da Silva, eds. higher trade barriers, the trade reform scenario results 2003. The Impact of Economic Policies on in larger absolute tariff cuts in these countries and there- Poverty and Income Distribution: Evaluation fore greater increases in trade flows. The CGE model Techniques and Tools. Washington, DC: World used to simulate the trade reform scenario does not cap- Bank and Oxford University Press. ture the possibility of imperfect pass-through of price Bruno, Michael, Martin Ravallion, and Lyn Squire. shocks to different individuals (because they are in re- 1998. "Equity and Growth in Developing Coun- mote areas, involved in subsistence activities and the tries: Old and New Perspectives on the Policy like), and accounting for these imperfections would Issues." In Income Distribution and High- dampen the pro-poor potential of trade liberalization. Quality Growth, ed. Vito Tanzi and Ke-young 46. Engle, Galetoviv, and Raddatz (1998), cited in Chu. Cambridge, MA: MIT Press. Lopez and others (2006). Burnside, Craig, and David Dollar. 2000. "Aid, Poli- 47. Morley and Coady (2003) even attempt to es- cies, and Growth." American Economic Review timate the future earnings of poor children receiving 90 (4) (September): 847­68. transfers under programs in Mexico and Nicaragua. In Bussolo, Maurizio, Rafael De Hoyos, Denis Medvedev, their words: "under the reasonable assumption that the and Victor Sulla. Forthcoming. "Demographic wage structure of the future labor force will be the Change, Economic Growth, and Income Distrib- same as it was in the year of the most recent survey, we ution: An Empirical Analysis Using Ex-Ante estimate that the additional education would add Microsimulations." Background paper for about 8 per cent to the average earnings of the poor in Global Economic Prospects 2007: Confronting Mexico and about 9 per cent in Nicaragua." Challenges of the Coming Globalization. Bussolo, Maurizio, Jann Lay, and Dominique van der Mensbrugghe. 2006. "Structural Change and References Poverty Reduction in Brazil: The Impact of the Anderson, Kym, Will Martin, and Dominique van der DohaRound."InPovertyandtheWTO:Impactsof Mensbrugghe. 2006. "Doha Merchandise Trade the Doha Development Agenda, ed. Thomas Hertel Reform: What Is at Stake for Developing Coun- and Alan Winters. Washington, DC: World Bank. tries?" World Bank Economic Review 20: 169­95. Bussolo, Maurizio, and Denis Medvedev. 2006. "What Is Atkinson, Anthony B. 2003. "Income Inequality in the Role of Labor Markets in Making Trade Reform OECD Countries: Data and Explanations." Pro-Poor in Latin America?" World Bank Policy Munich: CESifo Working Paper Series 881. Research Working Paper, Washington, DC. Draft. Atkinson, Anthony B., and Andrea Brandolini. 2004. Bussolo, Maurizio, and Jeffery Round, eds. 2005. "Global World Inequality: Absolute, Relative or In- Globalisation and Poverty: Channels and Policy termediate." Paper prepared for the 28th Confer- Responses. London: Routledge/Warwick Studies ence of the International Association for Research in in Globalisation, Routledge. Income and Wealth, Cork, Ireland, August 22­28. Chen, Shaohua, and Martin Ravallion. 2003. "House- Bhalla, Surjit. 2002. Imagine There's No Country: hold Welfare Impacts of China's Accession to the Poverty, Inequality, and Growth in the Era of World Trade Organization." Policy Research Work- Globalization. Washington, DC: Institute for ing Paper 3040, World Bank, Washington, DC. International Economics. Chenery, Hollis, Montek Ahluwalia, C. Bell, John Bonaglia, Federico, Jorge Braga de Macedo, and Duloy,andRichardJolly.1974.Redistributionwith Maurizio Bussolo. 2001. "How Globalization Im- Growth. New York: Oxford University Press. proves Governance." CEPR Discussion Paper 2992, Collier, Paul, and David Dollar. 2002. "Aid Allocation Centre for Economic Policy Research, London. and Poverty Reduction." European Economic Bourguignon, François, Victoria Levin, and David Review 46 (8): 1475­1500. Rosenblatt. 2004. "Declining International de Ferranti, David, Guillermo E. Perry, Francisco H. G. Inequality and Economic Divergence: Reviewing Ferreira, and Michael Walton. 2004. Inequality the Evidence through Different Lenses." in Latin America: Breaking with History? Économie Internationale 100: 13­25. Washington, DC: World Bank. 97 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Deaton, Angus, and Christina Paxson. 1997. "The Education in Latin America: The Role of Demand Effects of Economic and Population Growth on and Supply of Skills." CEP Discussion Paper National Saving and Inequality." Demography dp0712, Centre for Economic Performance, 34 (1) (February): 97­114. London School of Economics. Dollar, David, and Aart Kraay. 2002. "Growth Is Mayda, Anna Maria, and Dani Rodrik. 2005. "Why Good for the Poor." Journal of Economic Are Some People (and Countries) More Protec- Growth 7 (3): 195­225. tionist than Others?" European Economic Easterly, William, Ross Levine, and David Roodman. Review 49 (6) (August): 1393­1430. 2004. "Aid, Policies, and Growth: Comment." Mayer, W., 1984. "Endogenous Tariff Formation." American Economic Review 94 (3) (June 1): American Economic Review 74: 970­85. 774­80. Milanovic, Branko. 2002. "True World Income Distrib- Engle, E., A. Galetoviv, and C. Raddatz. 1998. "Taxa- ution, 1988 and 1993: First Calculation Based on tion and Income Distribution in Chile: Some Un- Household Surveys Alone." Economic Journal pleasant Redistributive Arithmetic." NBER (Royal Economic Society) 112 (476) (January): Working Paper 6828, National Bureau of Eco- 51­92. nomic Research, Cambridge, MA. ------. 2005. Worlds Apart: Global and International Ferreira, Francisco H. G., and Phillippe G. Leite. 2003. Inequality 1950­2000. Princeton, NJ: Princeton "Meeting the Millennium Development Goals in University Press. Brazil: Can Microsimulations Help?" Economía Milanovic, Branko, and Shlomo Yitzhaki. 2002. 3 (2): 235­79. "Decomposing World Income Distribution: Does ------. 2004. "Educational Expansion and Income the World Have a Middle Class?" Review of Distribution: A Microsimulation for Ceará." In Income and Wealth 48 (2): 155­78. Growth, Inequality and Poverty, ed. Anthony Mookherjee, Dilip, and Anthony Shorrocks. 1982. Shorrocks and Rolph van der Hoeven. London: "A Decomposition Analysis of the Trend in UK In- Oxford University Press. come Inequality." Economic Journal 92: 886­902. Higgins, Matthew, and Jeffrey G. Williamson. 1999. Morley, Samuel A., and David Coady. 2003. From So- "Explaining Inequality the World Round: Cohort cial Assistance to Social Development: Targeted Size, Kuznets Curves, and Openness." Staff Re- Education Subsidies in Developing Countries. ports 79, Federal Reserve Bank of New York. Washington, DC: Center for Global Development Jenkins, Stephen P. 1995. "Accounting for Inequality and the International Food Policy Research Trends: Decomposition Analyses for the UK, Institute. 1971­86." Economica [London School of Eco- O'Rourke, Kevin. 2003. "Hecksher-Ohlin Theory and nomics and Political Science] 62 (245) (February): Individual Attitudes towards Globalization." 29­63. CEPR Discussion Paper 4018, Centre for Li, Hongyi, Lyn Squire, and Heng-fu Zou. 1998. "Ex- Economic Policy Research, London. plaining International and Intertemporal Varia- Olivera, Mauricio, and Eduardo Lora. 2005. "The tions in Income Inequality." Economic Journal Electoral Consequences of the Washington 108 (446): 26­43. Consensus." Inter-American Development Bank Lopez, J. Humberto, William F. Maloney, Guillermo E. Working Paper 530, Washington, DC. Perry, Omar S. Arias, and Luis Servén. 2006. Osberg, Lars. 2003. "Long Run Trends in Income In- Poverty Reduction and Growth: Virtuous and equality in the United States, UK, Sweden, Vicious Circles. Washington, DC: World Bank. Germany and Canada: A Birth Cohort View." Lutz, Wolfgang, and Anne Goujon. 2001. "The Eastern Economic Journal 29 (1): 121­42. World's Changing Human Capital Stock: Multi- Pritchett, Lant. 1997. "Divergence, Big Time." Journal State Population Projections by Educational At- of Economic Perspectives 11 (3): 3­17. tainment." Population and Development Review Quah, Danny T. 1996. "Twin Peaks: Growth and Con- 27 (2): 323­39. vergence in Models of Distribution Dynamics." Lutz, Wolfgang, Anne Goujon, Gabriele Doblhammer- Economic Journal [Royal Economic Society] 106 Reiter. 1998. "Demographic Dimensions in (437) (July): 1045­55. Forecasting: Adding Education to Age and Sex." ------. 1997. "Empirics for Growth and Distribution: Population and Development Review 24 (Supple- Stratification, Polarization, and Convergence ment: Frontiers of Population Forecasting): Clubs."JournalofEconomicGrowth2 (1):27­59. 42­58. Ravallion, Martin. 2004. "Competing Concepts of In- Manacorda, M., Carolina Sanchez-Paramo, and equality in the Globalization Debate." In Brook- Norbert Schady. 2005. "Changes in Returns to ings Trade Forum 2004, ed. Susan Margaret 98 I N C O M E D I S T R I B U T I O N , I N E Q U A L I T Y , A N D T H O S E L E F T B E H I N D Collins and Carol Graham. Washington, DC: Theil, Henri, and James Seale. 1994. "The Geographic Brookings Institution. Distribution of World Income, 1950­1990." De Ravallion, Martin, and Shaohua Chen. 2003. "Mea- Economist 142 (4): 387­419. suring Pro-Poor Growth." Economics Letters Winters, L. A., N. McCulloch, and A. McKay. 2004. 78 (1): 93­99. "Trade Liberalization and Poverty: The Evidence ------. 2004. "China's (Uneven) Progress against So Far." Journal of Economic Literature 42 (1): Poverty." World Bank Policy Research Working 72­115. Paper Series 3408, Washington, DC. World Bank. 1998. Assessing Aid: What Works, What Sala-i-Martin, Xavier. 2002a. "The Disturbing `Rise' Doesn't, and Why. New York: Oxford University of Global Income Inequality." NBER Working Press. Paper 8904, National Bureau of Economic ------. 2005. World Development Report 2006: Research, Cambridge, MA. April. Equity and Development. Washington, DC: ------. 2002b. "The World Distribution of Income (Es- World Bank and New York: Oxford University timated from Individual Country Distributions)." Press. NBER Working Paper 8933, National Bureau of ------. 2006. Global Monitoring Report 2006: Economic Research, Cambridge, MA. May. Millennium Development Goals--Strengthening Scheve, Kenneth, and Matthew Slaughter. 1998. Mutual Accountability, Aid, Trade, and Gover- "What Determines Individual Trade Policy nance. World Bank: Washington, DC. Preferences?" NBER Working Paper 6531, World Tourism Organization. 2006. "Facts and National Bureau of Economic Research, Figures." Available at http://www.unwto.org/ Cambridge, MA. facts/menu.html. 99 4 New Pressures in Labor Markets: Integrating Large Emerging Economies and the Global Sourcing of Services Rapid technological progress, trade in goods, for the emergence of the Asian tigers, India and and international sourcing of services come to- China's sheer size raises the specter of surging gether to put new pressures in labor markets, new export competition. Many developing pressures that will only become more acute in countries fear that exports from these large the next 25 years. Through these channels, new players may swamp their domestic mar- globalization is creating a progressively more kets, squeeze them out of the global market, integrated global market for labor. The impact foreclose avenues of diversification in manu- is tempered by differences in the skills, tech- factures as a road to higher growth, and gobble nology, and know-how available to workers. up all the investment flows. And high-income Globalization offers opportunities for ex- countries worry that if the large emerging port growth and access to a wider range of economies can readily acquire and master the cheaper imported products that can fuel pro- newest technologies, their exports may soon ductivity growth and rising average living take over high-tech markets. standards. But globalization also imposes Global sourcing of services exerts pres- adjustment costs on certain groups within sures in the same direction. The transfer of countries, primarily through labor markets by relatively skilled service activities to firms in influencing wages and job security and by developing countries is putting new pressures demanding retraining, and the upheaval of on white-collar employment in both the high- moving between jobs. The unskilled have seen income countries and advanced developing their wages worsen relative to skilled workers countries. This puts higher-paying and higher- and their jobs become less secure. This is true skill jobs at risk in both high- and middle- even in developing countries--contrary to ex- income countries. Unlike displacement in pectations that the unskilled benefit relative low-skilled manufactures trade, services off- to the skilled as labor-intensive manufacturing shoring has the potential to destroy the previ- moves to low-wage countries. The projections ous investments of white-collar workers in in this report offer little reason to believe that firm-specific knowledge. this will change in the coming decade. The analysis here suggests that three fac- Two challenges are particularly demanding: tors are likely to mitigate these effects in the one is the rise of China, India, and other emerg- medium and long term. ing economies as manufacturing powerhouses, and the other is the emergence of global sourc- · First, the growth of the Chinese, Indian, ing of services. While the qualitative implica- and other emerging markets offers enor- tions of increasing exports of manufactured mous offsetting opportunities for other products from India and China are the same as developing and developed countries to 101 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 increase exports. As China and India in- and know-how from abroad and seize the op- crease their exports, they will have to portunities created by rising demand from-- increase imports of intermediate inputs, and production shifts in--India and China. energy, technology, and investment But openness will not foster integration in the goods. Driven by China, Asia was the absence of an attractive investment climate, principal destination for accelerated ex- one with sound institutions and policies that ports from Africa and Latin America in allow labor, capital, and knowledge to flow the late 1990s and the early 2000s. from low-return to high-return sectors. Devel- · Second, accompanying the rising value of oping knowledge-intensive activities as future exports and domestic living standards in drivers of growth will require investing in the emerging economies will be rising wages. institutions and policy frameworks that foster This--together with the inevitable ex- innovation, and in education and lifelong change rate adjustment to the rise in learning for all workers. Developing countries global demand for these countries' prod- with wages currently higher than those in ucts and services--will create space for China and India will have to place greater at- low-income countries to move into the tention on their institutions and on education lowest-skill activities vacated by produc- policies to create a climate for greater innova- ers in the large emerging countries. tion and skill enhancement. · Third, developing the social institutions Social policies should focus on protecting that support a dynamic market economy workers rather than protecting jobs. Even in the in China and India will take time, most propitious policy and institutional envi- providing an opportunity for smaller, ronments, rapid growth, globalization, and more flexible countries to progress faster labor-market flexibility are likely to quicken the in institutional development--and for pace of job creation and job destruction. This rich countries to continue to lead in demands policies to cushion the adjustment productivity-enhancing innovation. The costs associated with increased volatility and in- flow of services activities from rich to voluntary dislocation. The returns to skilled poor countries, which entails some trans- labor will continue to increase faster than those fer of know-how, will be slowed to the to unskilled labor, perpetuating a natural wage- extent that institutional frameworks dis- widening tendency in many (if not most) coun- courage foreign direct investment (FDI) tries and underscoring the need for measures to and in particular fail to protect the own- support workers at the low end of the scale. ership of such assets. Rising wage inequality, together with volatile labor markets, are heightening insecurity The policies that countries adopt will among workers throughout the world. determine whether they will be able to take advantage of these new opportunities. Effec- tive policy responses will need to position The impact of globalization: countries to harness the opportunities from globalization while also addressing the adjust- the story so far G ment tensions that inevitably arise from the lobalization, coupled with technological unprecedented magnitude and speed of change, has driven growth in the world change in labor markets. economy, bringing new employment opportu- Policies to embrace, rather than resist, global nities and enabling millions of people to escape integration will lay the foundations for future absolute poverty. That said, impacts have growth and job creation. Openness to trade varied across and within countries--and not all and FDI will become ever more critical if the workers have benefited equally. While many poorest countries are to absorb technologies countries have seized the opportunities offered 102 N E W P R E S S U R E S I N L A B O R M A R K E T S by greater integration of markets in goods and This increased two-way trade reflects the services, others, especially in Sub-Saharan growth of outsourcing and global production Africa, have remained marginalized. Mean- chains (see Global Economic Prospects 2003). while demand for skilled labor has increased in Enabled by falling barriers to trade and FDI, both developed and developing countries and lower transport and communication costs, and greater global competition has become associ- new technologies, global chains break down ated with a growing sense of insecurity for goods into their constituent parts, each pro- many workers worldwide. duced where it can be done most efficiently and at least cost, whether by an affiliate or by an in- Product markets are rapidly integrating . . . dependent supplier. Ghose (2003) sees a corre- with a geographical redistribution of lation between countries' share of world mer- manufacturing chandise exports and their share of FDI inflows: Developing countries' trade has accelerated between 1982 and 1999, foreign affiliates of over the last few decades. In the markets of transnational corporations increased their share developed countries, the share of developing of world exports from 31 percent to 45 percent. countries in imports of manufactured prod- At the same time, manufacturing employ- ucts grew from barely 14 percent in 1973 to ment has been redistributed between developed nearly 40 percent in 2003 (figure 4.1).1 and developing countries. While the precise Imports of developing countries have grown numbers are debated, the gain in the latter has just as quickly as their exports to the rest of the been much larger than the loss in the former world (Ghose 2003). Developing-country im- (Sapir 2005; Ghose 2003).2 Overall, employ- ports grew at about 2 percent per year during ment in manufacturing in developed countries the 1980s, accelerating to 9.5 percent per year declined from 28.7 percent in 1995 to 24.8 per- during the 1990s (Bhorat and Lundall 2004). cent in 2005, while most developing regions saw gains (table 4.1). Not all developing countries have experi- Figure 4.1 Developed countries' imports enced gains in manufacturing employment, of manufactures increasingly come from however. Consider, for example, the striking dif- developing countries ferences between East Asia and Latin America. Share of high-income countries' imports of Over the 1990s, employment in manufacturing manufactures originating in developing countries increased in China (by just under 15 percent Percent cumulatively), India (by about 38 percent), 45 Malaysia (40 percent), and Thailand (about 40 49 percent). In Latin America, however, aggre- 35 gate manufacturing employment declined over 30 the 1990s; increases in Chile (about 10 percent) 25 20 were more than offset by declines in Brazil 15 (about 50 percent) and Argentina (14 percent) 10 (Bhorat and Lundall 2004).3 5 Services employment has increased in 0 both developed countries and all developing 1973 1983 1993 2003 regions, except the Middle East and Africa Europe and Central Asia North Africa, where it has remained the same Middle East China (table 4.1). Over the 1990s, large increases Latin America Other Asia in services employment were seen in Brazil Sources: Data from World Bank World Integrated Trade (57 percent) and Mexico (62 percent), with Solution (WITS) and staff calculations. smaller increases in China (about 13 percent) Note: Excludes trade among EU15 countries. and India (25 percent) (Bhorat and Lundall 103 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table 4.1 Employment in developing countries has shifted out of agriculture into manufactures and services Trends in sectoral shares in employment, 1995­2005 (percent) Agriculture Industry Services World region 1995 2005a Change (%) 1995 2005a Change (%) 1995 2005a Change (%) World 44.4 40.1 9.7 21.1 21 0.5 34.5 38.9 12.8 East Asia 54.4 49.5 9.0 25.9 26.1 0.8 19.7 24.4 23.9 South East Asia and 55.3 43.3 21.7 15.4 20.7 34.4 29.3 36 22.9 the Pacific South Asia 64.1 61.2 4.5 13.4 14.1 5.2 22.5 24.6 9.3 Latin America and the 23.4 17.1 26.9 20.2 20.3 0.5 56.4 62.5 10.8 Caribbean Middle East and 30.8 26.3 14.6 20.3 25 23.2 48.9 48.7 0.4 North Africa Sub-Saharan Africa 70.1 63.6 9.3 8.2 8.9 8.5 21.7 27.5 26.7 Developed economies and 5.1 3.7 27.5 28.7 24.8 13.6 66.1 71.4 8.0 the European Union Central and Eastern 27.9 22.7 18.6 27.5 27.4 0.4 44.6 49.9 11.9 Europe and CIS Source: ILO 2006; Bank staff calculations. CIS: Commonwealth of Independent States. a. Indicative. 2004). Also worthy of note are the relatively the investment climate--if opening the econ- high levels of female employment in the sector. omy does not attract FDI, potential short- To some extent, this increase may reflect term wage losses from opening the economy changes in business organization, where func- may not be offset (World Bank 2002). tions once performed inside manufacturing While average wages rise more rapidly in companies are now outsourced to other firms open economies than in closed ones, increasing on a contract basis, resulting in their reclassifi- relative demand for skilled labor is widening cation as services. This change in business the wage gap between skilled and unskilled organization has also crossed borders, with workers in both developed and developing multinational companies sourcing activities countries.4 The latter is contrary to expecta- from subsidiaries or external firms around the tions based on traditional trade theory that globe. (The global sourcing of services will be globalization will increase the relative return revisited below.) to abundant unskilled labor in poor countries. While available evidence attributes wage Globalization has generally been widening primarily to technology, trade is also associated with rising average wages--but important. The relative impacts are hard to not all workers are benefiting equally . . . disentangle since technology can lead to trade, While an economy's openness to trade and in- and technological innovation in turn can be a vestment is in general associated with faster response to increased competition from trade growth of average wages over the longer (box 4.1). term, short-term impacts can vary. Although A widening wage gap between skilled and the initial impact of trade liberalization on unskilled workers is particularly evident in the wages may be negative in some countries, it United States, where lighter labor-market becomes significantly positive over time. For regulation permits faster adjustments in FDI, the picture is reversed: an initial positive wages. In Europe, where labor markets are effect on wages is reduced to nothing after more tightly regulated, the outcome of rising five years. This highlights the importance of relative demand for skilled labor has been 104 N E W P R E S S U R E S I N L A B O R M A R K E T S Box 4.1 What causes the gap between skilled and unskilled labor--technology or trade? T raditional theory expects trade with low-wage than trade, is the most important factor in wage countries to result in a shift in the composition inequality--for example in Chile (Reinecke and of employment toward skilled labor between sectors Torres 2001) and South Africa (Edwards 1999). (as industries expand or contract in response to for- Trade plays an important role in disseminating eign competition). In both developed and developing new technologies, however. Robbins (1997), for ex- countries, however, labor composition within sectors ample, finds that the amount of capital equipment has moved toward skilled labor (also reflected in a imported into a subset of developing countries is a dramatic increase in their relative wages), suggesting significant factor in raising the demand for tertiary- that technological change has been the major force at educated workers relative to demand for those com- work. Moreover, the sectors shifting toward skilled pleting only primary school. Moreover, technological labor in developing countries in the 1980s had done upgrading can itself be a response to trade competi- so in the United States in the 1960s, suggesting a tion; there is substantial evidence that firms improve migration of technological change from developed productivity following competition from imports to middle-income countries. (Hoekman and Winters 2005). Companies in high- Technological change is generally viewed as the wage countries facing import competition from most important force in terms of the rising demand lower-cost developing-country suppliers may engage for skilled labor (Krugman 1995), as evidenced by the in "defensive innovation," moving up the value positive correlations between technology and growth chain and into more capital-intensive production. of employment of skilled workers within industries, They may also respond by offshoring more produc- and by the fact that increases in the relative wages tion to reduce costs--generally the low-skilled activi- (cost) of skilled workers have been accompanied by ties, with the high-skilled activities remaining in the an increase in their relative demand (Helpman 2004). home market (Anderton, Brenton, and Whalley Trade, by contrast, is a less important force-- 2006).a This can occur in both high-skill-intensive although estimates vary. Feenstra and Hanson (2003) and low-skill-intensive sectors so that trade with conclude that the offshoring of manufacturing ac- developing countries can thus potentially have an counts for 15­24 percent of the shift toward more impact on a wide range of sectors--and even within skilled labor. Anderton and Brenton (1999), on the low-skill-intensive sectors, the higher-skilled activities other hand, find that, when only offshoring to low- could still expand. In this sense, offshoring can have wage countries is included, trade may actually account the same effect as technology in reducing the relative for about 40 percent of the rise in the wage bill share demand for unskilled labor within an industry of skilled workers and approximately one-third of the (Feenstra and Hanson 2003), and reallocating away increase in their employment in the U.K. textiles sec- from high-wage economies to low-wage economies. tor. The OECD (2005d) finds that the average decline in employment in 15 Organisation for Economic Co- operation and Development (OECD) countries was a 27 percent in industries characterized by high interna- Currency and exchange rate movements can also prompt a shake-out that leaves only higher technology firms in a sector. tional competition compared with 16 percent in total Disproportionate increases in offshoring can be seen during large manufacturing. Evidence from developing-country exchange rate appreciations and the costs and difficulty of rever- studies also suggests that technology and FDI, rather sal may see offshoring remain after the currency has stabilized. reflected in higher unemployment among the Hence, while average U.S. real wages did not unskilled.5 In the United States, increased de- change significantly between the late 1970s mand for skilled workers since the mid-1980s and the mid-1990s, real wages of high-wage has led to a relative increase in their employ- workers increased and those of low-wage ment and wages (Katz and Autor 1999). workers declined (Helpman 2004)--despite 105 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 women's skills tended to be lower than average Figure 4.2 In many developed countries and hence their entry in the 1990s depressed the gap between high- and low-income earnings to lower-skilled workers (although earners has widened overall it raised the income of poor households) Percent change in gap between 90th- and (de Ferranti and others 2002). 10th-percentile earners, late 1970s to mid-1990s Second, demand for skilled labor increased Percent (World Bank 2002). The share of skilled work- 30 ers in total employment and the relative de- 25 mand for these workers increased between the 20 early and late 1990s in a range of countries-- 15 including Brazil, Chile, Malaysia, the Philip- 10 pines, Singapore, and Thailand, with Mexico 5 experiencing a slight decline. In all cases, in- 0 cluding Mexico, the growth rate of relative wages of skilled workers exceeded that of 5 unskilled workers. 10 d e y The demand for skilled labor depends on the Italy States lands Kingdom Zealan Canada Franc her Norwa skill intensity of the export sector. In many coun- Germany United Net New United tries in Latin America, increased exports based on abundant natural resources raised demand Source: Katz and Autor 1999. for complementary skilled labor and capital (Perry and Olarreaga 2006). Furthermore, ac- tivities considered relatively low skill in devel- the fact that the relative supply of skilled oped countries may nonetheless be relatively workers grew over the same period.6 skilled in developing countries, especially in The same decline in the relative wages of manufacturing. Transfer of activities considered low-skilled workers is found in other devel- relatively low skill in developed countries to de- oped countries, although to a lesser extent. veloping countries raises the relative demand for While the gap increased by 29 percent in the and relative earnings of high-skilled workers in United States and 27 percent in the United the latter (Feenstra and Hanson 2003). Kingdom (figure 4.2), it was only 15 percent Additionally, enhanced competition from in New Zealand, 14 percent in Italy, and trade can affect the relative demand for 9 percent in Canada (Katz and Autor 1999). skilled labor by reallocating resources within The relative wages of skilled workers in- sectors away from less productive, unskilled- creased in developing countries in the late labor-intensive firms toward more productive 1990s, along with a rise in their relative firms using more skilled labor. Trade also facil- employment levels (Majid 2004; Bhorat and itates the transmission of skill-biased techno- Lundall 2004). There are several possible logical change. Protection in many developing reasons for this. countries favored unskilled-labor-intensive sec- First,insomedevelopingcountries,notablyin tors, hence liberalization led to expansion of Africa, increased demand for low-skilled labor skilled-labor-intensive industries (Perry and did not lead to wage increases because there was Olarreaga 2006). Foreign direct investment can a large surplus of labor to be absorbed (Ghose also increase the demand for skilled labor. 2003; Wood 1997; Fox and others 2004). Stud- In Eastern Europe, for example, privatization ies in Latin America and the Caribbean also sug- and FDI helped bias employment composition gest that the increased participation of women and relative wages significantly toward skilled in the labor market may have contributed to a labor as production facilities were upgraded widening wage distribution. In that region, with new technologies.7 106 N E W P R E S S U R E S I N L A B O R M A R K E T S Even outside the traded sector, globalization mobility in the form of migration may also have may have an impact on low-skilled workers in an impact on low-skilled workers, although other ways. In all countries, wages in the non- there is considerable debate on this issue. Most traded sector may be affected by wages and em- studies focus on developed countries and tend to ployment in the traded sector if there is mobility find small overall impacts--or indeed positive between the two. In addition, international impacts--frommigrationonwages(seebox4.2). Box 4.2 Workers in the nontraded sector--the role of migration E ven where workers do not experience increased decline in the relative wages of high school dropouts competition from international trade, they may over 1980­95 (depending on wage elasticity). The be affected by an increase in the supply of workers effect of immigration is diffused throughout the econ- resulting from migration, although this impact may omy, as natives move in response to immigration. be positive. As immigration increases labor supply, Large immigrant flows to one region may discourage and wages are reduced, more capital may be attracted flows to that region of native workers, but may encour- and more jobs created. Moreover, consumption by age flows of capital. Ottaviano and Peri (2005) find migrants also increases the overall demand for native that migration increases total employment by 10 per- labor and capital. That said, many studies identify cent, and increases U.S.-born workers' wages by redistributional effects, with the impact of immigra- 3­4 percentage points, largely because U.S.- and tion concentrated on the lowest-skilled workers. The foreign-born workers are not perfectly substitutable, impact depends critically on the extent to which even when they have similar observable skills. migrants and natives are substitutes for one another; College graduates, high school graduates, and college that is, whether they are competing for the same jobs dropouts all gain (about 2.4 percent real wage in- or operating in segmented markets (for a fuller crease), but the low-skilled lose (by the same amount). discussion, see Global Economic Prospects 2006). Cortes (2005) also sees sizeable wage effects on the Borjas (2003) finds negative impacts on workers up low-skilled, but these are concentrated on other immi- to some college level, with immigration harming the grants, as low-skilled immigrants and low-skilled employment prospects and lowering the wage of com- natives are far from perfect substitutes. A 10 percent peting native workers (a 10 percent increase in supply increase in the number of low-skilled immigrants in reduces wages by 3­4 percent). The lowest-skilled a city reduced the wages of low-skilled natives by were hardest hit: wages fell by 8.9 percent for high 0.6 percent and of low-skilled immigrants by 8 per- school dropouts, 4.9 percent for college graduates, cent. But migration also reduces the prices of non- and 2.6 percent for high school graduates while barely traded goods and services. A 10 percent increase in changing for workers with "some college." However, the average city's share of low-skilled immigrants in this analysis ignores, among other things, the long-run the labor force decreases the price of immigrant- capital adjustments induced by immigration. If the intensive services such as housekeeping and gardening capital stock does adjust, overall wages are unaffected by 1.3 percent, with about 50­80 percent of this net and the loss of wages to high school dropouts is cut to effect caused by reduction in wages. below 5 percent (The Economist, "Economics Focus: Card (2005) argues that the wages of natives with Myths and Migration," April 8, 2006). less than a high school education relative to native high Borjas, Freeman, and Katz (1997) argue that the ef- school graduates have remained nearly constant since fect of immigrant-induced increases in relative labor 1980. This is despite immigrant inflows that have in- supply are strongly concentrated on U.S. workers with creased the supply of workers with less than high less than 12 years of schooling, many of whom work in school education and despite the growing wage gap be- the nontraded sectors. Migration increased the supply tween other education groups. Most of the absorption of workers with less than high school education by 15 to of unskilled workers occurs in the form of city-specific, 20 percent over 1980­95, leading to a 27­55 percent within-industry increases in low-skilled intensity. 107 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 . . . and workers are feeling less secure job prospects to developing countries, it can Individuals are concerned not only about the also bring greater volatility and insecurity. level but also the security of their earnings. Bergin, Feenstra, and Hanson (2006) find that Greater global competition, along with more while offshoring of production from the rapid technological change and diffusion, can United States to Mexico has been an impor- increase wage and employment volatility,8 tant source of growth in Mexico, there is a although separating the effect of trade from high degree of volatility in these activities. technological change in volatility is difficult. Domestic demand shocks in the United States Labor turnover is high in many countries, are amplified when they are transmitted to the fueling individuals' perceptions of economic offshored activities in Mexico. In this way, off- insecurity. Available data show gross sec- shoring has led the United States to export to toral rates of job creation and destruction of Mexico some of the employment fluctuations between 5 and 20 percent, adding up to an an- that it experiences over the business cycle. nual job turnover of up to 40 percent in some In Latin America, overall labor turnover is countries. A significant part of that turnover higher than in OECD countries. However, (often 30­50 percent) can be traced to the turnover depends on education, per capita entry and exit of firms, important for output income, and other demographic and growth and productivity growth. About 20 percent of variables. For example, young workers change firms are created and destroyed each year in jobs more frequently than older workers and many countries, involving 10­20 percent of lower levels of education can imply lower lev- the workforce (World Bank 2005). While els of firm-specific capital and hence a higher evidence is not uniform, some studies of incidence of voluntary separation. Adjusting OECD countries find that increased trade ex- for these factors, the region does not show posure is associated with more labor churning conditionally higher turnover. There is only (Hoekman and Winters 2005). Overall it is mixed evidence that either greater trade liber- estimated that 3 to 5 percent of the OECD alization or exposure to technological change workforce experiences involuntary layoff in leads to greater overall turnover in the region; any given year (Kuhn 2002). however, to the degree that trade liberalization In the United States, more than 7 million has expanded the share of tradables in total jobs have been destroyed on average every output, it may have led to more churning in quarter over the last decade as a result of the the job market (de Ferranti and others 2000). normal functioning of the economy (OECD To the extent that it reflects labor-market 2005b), matched for the most part by equal or flexibility and the reallocation of resources to greater job creation. Among the unemployed, more productive sectors, increased turnover about 45 percent were laid off, 10­15 percent can be a sign of healthy adjustment, linked are persons voluntarily between jobs, and the to further growth and job creation. However, remainder are persons entering or reentering churning can also be negative--for example, the labor market as new job-seekers (Kletzer where job creation lags well behind job 2001). Voluntary attrition may account for up destruction, where high turnover lowers to two-thirds of employment reduction in the workers' and employers' incentives to invest United States (OECD 2005d). in education and training (thus ultimately Falling transport and communication costs reducing productivity in a sector), or where are creating new opportunities for developing churning results in labor moving into less countries to participate in global production productive sectors. In the absence of social chains by providing specific activities and tasks safety nets, workers in developing countries (see box 4.7). However, discrete activities are may be unable to finance job searches likely to be more footloose than whole sectors, and may be forced into the informal sector so that while globalization can bring better (where productivity is generally lower) or 108 N E W P R E S S U R E S I N L A B O R M A R K E T S into low-productivity, relatively low-growth developed and developing countries, and sectors such as agriculture. increasing wage gaps between low- and high- The extent and nature of churning depend skilled workers worldwide. What will be the on the policy environment. Onerous labor- impact of the key challenges now facing global market regulations and restrictions on entry labor markets--namely absorption of large and exit of firms can discourage firms from emerging economies and the global sourcing hiring new regular workers, and workers from of services? searching for jobs in the formal sector. They can also limit the movement of resources out of low-productivity sectors. Overly restrictive New challenge I--absorbing employment protection (such as restrictions large emerging economies into on hiring and firing) tends to have the effect of the global market protecting only some workers (insiders, usu- ally prime-age males) at the expense of others By 2030 China and India together will (outsiders, usually youth, women, and low- account for about 40 percent of the skilled workers). Strict employment protection world's workforce, which will remain is associated with higher income disparities predominantly unskilled and a greater incidence of informal work By 2030 the world's labor force will number (World Bank 2005). It also raises the costs of some 4.1 billion workers, 90 percent of whom workforce reorganizations, thereby reducing will live in the developing world. The global incentives for innovation and implementation labor force is predicted to grow by about of new technologies (Arias and others 2005).9 1 percent per year over 2001­30, with higher The precise impact of strict employment growth in developing countries offset by some protection on job creation depends on who contraction in developed countries (table 4.2). bears the cost: where wages absorb less of the East Asia, the Pacific, and South Asia together cost than firms, the disincentives to create em- will account for just over half the world's ployment are greater. In Latin America, firms workforce, with China and India alone repre- can bear up to 50 percent of the cost of non- senting 40 percent--although China's labor wage benefits, resulting in reduced wages, force will grow far more slowly than that of greater informality, or both (World Bank India. Sub-Saharan Africa will experience the 2005). In the OECD countries, partial reforms highest rate of growth (about 2.4 percent per have tended to reinforce labor-market in- year) and will be the third-largest developing equality, with temporary contracts for new region. entrants (youth or women) but only limited Worldwide, the supply of skilled workers is access to more permanent jobs. Strict employ- likely to grow faster than that of unskilled ment protection is also associated with a workers, but the vast majority of the world's greater feeling of insecurity, perhaps because workforce will remain unskilled in 2030.11 In workers realize that their chance of long-term the developing world, rates of growth in the unemployment is higher (OECD 2005d, number of skilled workers will be highest in 2004).10 A benefit of globalization is the pres- Sub-Saharan Africa, South Asia, and the sure it exerts on institutions that cramp pro- Middle East and North Africa. Given the large ductivity growth and on governments to de- pool of unskilled labor, however, these in- velop efficient safety nets that cushion creases will raise the share of skilled workers workers from the worst aspects of economic in developing countries' workforces only insecurity, while preserving job creation and slightly (from 9.6 percent to 11.3 percent). flexibility. There will be significant regional variations Today's global labor market is characterized in the developing world. The Middle East by volatility, shifts in employment between and North Africa, Latin America and the 109 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table 4.2 In 2030 most workers will be in developing countries and unskilled Growth in the global labor force 2001­30 All workers (millions) Unskilled workers (millions) Skilled workers (millions) Growth Growth Growth (% per (% per (% per World region 2001 2030 year) 2001 2030 year) 2001 2030 year) World total 3,077 4,144 1.03 2,674 3,545 0.98 403 598 1.37 High-income countries 481 459 0.16 327 276 0.58 154 183 0.60 Developing countries 2,596 3,684 1.21 2,347 3,269 1.15 249 415 1.78 East Asia & the Pacific 1,060 1,279 0.65 988 1,163 0.56 71 117 1.70 China 773 870 0.41 740 816 0.34 33 54 1.72 South Asia 632 1,005 1.62 589 925 1.56 42 81 2.27 India 473 712 1.42 441 653 1.36 32 59 2.10 Europe & Central Asia 236 233 0.04 195 192 0.06 41 41 0.02 Middle East & North Africa 119 205 1.88 87 144 1.74 32 61 2.25 Sub-Saharan Africa 313 617 2.36 293 573 2.33 20 44 2.74 Latin America & the Caribbean 236 345 1.32 194 273 1.19 42 72 1.85 Source: World Bank staff calculations. Caribbean, and Europe and Central Asia con- two driven by large changes in China (67 to tinue to have relatively high rates of skilled 42 percent) and India (54 to 34 percent). workers (30, 21, and 18 percent, respectively), Moreover, while average incomes will con- compared to East Asia and the Pacific (9 per- tinue to increase with new opportunities for cent), South Asia (8 percent), and Sub- growth, the skill premium--the ratio of skilled Saharan Africa (7 percent). But in absolute wages to unskilled wages--will also increase. numbers, India and China each have more Projections from the model developed in chap- skilled workers than Europe and Central Asia ter 2 suggest that the skill premium in develop- or Sub-Saharan Africa, and almost as many as ing countries will rise from 3.5 on average in the Middle East and North Africa. Overall, 2001 to 4.2 in 2030. In India the premium rises developing countries have more than twice as from 4.3 to 4.9 in 2030 while in China the in- many skilled workers as developed countries, crease is even larger, from 5.4 to 7.7. Develop- even though the proportion of skilled workers ments in Sub-Saharan Africa are similar, with a in the workforce is four times higher in the rise from 5.1 to 6.8. The Middle East and North developed world. Africa sees only a modest increase in the skill Agricultural workers will constitute a premium from 1.3 to 1.5, while the premium shrinking share of the world's labor force, remains constant at 2.2 in Latin America. declining from about 43 percent in 2001 to about 30 percent in 2030. While the share of Pressures on unskilled workers will agricultural workers will fall by about half in intensify in both developed and developed countries, the stark decline is from developing countries . . . an already low base (from 4 to 2.6 percent). Between 1995 and 2005 the global labor force The more significant change will occur in (employed and unemployed) grew by some developing countries, where agricultural 438 million workers, or 16.8 percent (ILO workers will shift from about 50 percent of 2006). However, the effective increase in the the workforce in 2001 to 34 percent in 2030. global labor market is considerably larger, The most notable shifts will occur in Sub- because many workers in the emerging Saharan Africa (61 to 47 percent), East Asia economies were previously only weakly con- and the Pacific (62 to 39 percent), and South nected to the global economy. Freeman (2005) Asia (55 to 35 percent)--with the latter calculates that the integration of China, India, 110 N E W P R E S S U R E S I N L A B O R M A R K E T S and the former Soviet Union has led to a in China and India risk losing ground follow- "great doubling" of the global labor force. ing the entry of these countries into global The increasingly competitive global market commerce. The sheer size of China and India for labor may be the most important issue may also preclude the diversification of the facing workers worldwide. Freeman (2005) poorest countries into manufactures and so argues that because the workers in these coun- close off a route to growth and development tries brought little capital with them into the (Cline 2006). In rich countries, low-skilled global labor force there has been a massive labor is expected to lose as well, and future drop in the overall global ratio of capital to growth opportunities will depend on whether labor. In response to the huge amounts of new the rich countries' comparative advantage in low-wage labor, therefore, capital should high-technology sectors can be maintained. hemorrhage from rich countries and flow to According to this view, it is the quantity of China, India, and the ex-Soviet bloc. At the new entrants from China and India that risks margin, new investment should take place in swamping the global market, undermining the China and India, where returns should be prospects of unskilled workers in all other highest. countries, both rich and poor. This may not be The prognosis from this view is that devel- the case, however; competition is not always oping countries with wages higher than those what it appears (box 4.3). Box 4.3 Is the world flat . . . or just smaller? I n The World Is Flat, Thomas Friedman (2005) ex- Shanghai. Does everyone now live in a world in amines the rise of China and India in global supply which distance--physical, linguistic and cultural--no chains for both goods and services, describing the longer isolates jobs from competition? Is the world increasing pace and intensity of competition across flat or are jobs protected from competition by rela- skilled activities as the "flattening" of the globe. tionships and geography? But, as Leamer (2006) asks, is flatness the right metaphor? What if the world is not flat, but just Competition is not always what it appears . . . smaller? Smallness may confer a larger market without In the past, geography--physical, cultural, and generating many new competitors in sectors where informational--had limited competition by creating there are highly localized economies of scale, cost-advantaged relationships between proximate agglomeration (or cluster) effects, and first-mover sellers and buyers. Three revolutionary forces are advantages (consider the success of Hollywood in now driving a smaller world: (a) the presence of the global market for cultural products). Where more unskilled workers in the global labor market you are still matters. Economic activity is dispersing resulting from liberalizations in China, India, the around the globe, but with very strong clustering to Russian Federation, and Latin America; (b) new benefit from agglomeration effects. Commerce still equipment for knowledge workers (the Internet, declines dramatically with distance (although cul- computers) that has raised productivity, emphasized tural or linguistic forms of closeness can compen- talent, and reduced the need for helpers; and sate for physical distance), and trade remains a (c) communications innovations that extend the neighborhood phenomenon, close to home both geographic reach of suppliers and the competition geographically and organizationally. Consumer for routine work and standardized products. In a preferences and trust contribute to this pattern-- smaller world, exchanges are more contested and U.S. Web surfers still favor foreign sites close to relationships between buyers and sellers weaker. the United States, particularly when financial In a small world, wages in Los Angeles are set in transactions are involved. (continued) 111 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 4.3 (continued) Moreover, competition in knowledge products is two-thirds, and the pace of innovation in the 21st not necessarily a win-lose proposition: knowledge century will be unlike anything previously seen. products have their value enhanced by the existence of other products (software is an example). And Education, infrastructure, and safety nets are not all work is commoditized and sold in global essential, but technology guarantees that markets. Most exchanges still rely on long-term inequality will persist relationships between buyer and seller--relationships Global sourcing of services presents issues similar that create the language needed to communicate, to those posed by manufacturing. The lessons are that establish the trust needed to carry out the clear--make the education and infrastructure invest- exchange, that allow ongoing servicing of implicit ments needed to keep high-paying, noncontestable, or explicit guarantees, and that monitor and creative jobs at home, and argue for strong protec- enforce the truthfulness of both parties. This is tion of intellectual property rights (IPRs) to preserve the difference between negotiated rather than con- the value of knowledge goods sold abroad. But it is testable exchanges. Reliability--and liability--form important to recognize that technology can accentu- limits to the contestability of high-skilled jobs. To ate inequality by magnifying the importance of tal- date, global sourcing of intellectual work has been ent and enabling it to reach a much larger customer a small drop in a very large bucket, and the devel- base. Education may help to remedy the income- oped countries remain extremely well-positioned to inequality problems caused by technology, but there compete in the Internet-based segment of the are limits. First, if training is more effective for the economy. talented, they are likely to receive more of it--and the amount of training needed to equalize incomes . . . but competition--or the threat of it-- may be enormous and a great social waste. (How matters for routine tasks much training does it take to turn a World Bank Global competition is tight for standard tasks for economist into a Pavarotti? And is this a good use which global markets exist, both in manufacturing of resources?) Second, many jobs involve job- and services. Movement of jobs is not the only specific tacit knowledge gained only through on-the- indicator of global competition--contestability job experience. But will workers invest in acquiring may be reflected in a deterioration of wages and these skills if the job is likely to disappear? Will working conditions, rather than the movement of the incentives for skill acquisition also disappear? jobs. That is, the possibility of factor mobility Policies are needed to facilitate the formation of creates competitive pressure even in the absence of long-term relationships between workers and actual movement. Once this is factored in, the real employers and so instill the confidence to make effect of contestability--of global competition--is relationship-specific investments from which great hard to assess. returns can flow. Metaphors matter Innovation is key, but innovation moves The landscape of global competition is not flat, at around the world, and its pace is quickening least not much of it. The flat plains of open compe- Ideas stowaway with goods. As manufacturing work tition for mundane tasks certainly exist, but much of moves to China, so naturally do process innovations-- the landscape is hills and mountains--where endow- as those closest to production are best placed to ments, human capital, and policy matter. That land- work out how to do it better. But will product inno- scape is also constantly changing--today's hill might vation also move? The Internet has increased the be tomorrow's plain, creating new opportunities and speed and reduced the cost of distributing ideas obstacles and demanding continual adaptation. (subject to the constraints of infrastructure and literacy). Add the integration of former outsiders that has increased the size of the global brain by Source: Leamer 2006. 112 N E W P R E S S U R E S I N L A B O R M A R K E T S Productivity differences matter. Firms in rich countriescombineunskilledworkersinpro- Figure 4.3 Average wages in China have increased more than in other countries duction with more and better capital and techni- cal know-how than do firms in poor countries. Internationally comparable average wage rates, indexed, 1998 100 What matters is whether the wage gap is greater 230 than the difference in productivity--and China whether productivity differentials can be main- 180 tained.Similarly,theleastdevelopedcountriesin Mexico Africa that have lower wages than China and 130 India South Africa India will be able to compete in the global market--but only if their levels of productivity are close to those in India and China. The 80 Philippines sources of productivity differences across coun- tries will be discussed in more detail below. 30 There is another problem with the view 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 that the global market will be swamped by Sources: China Statistical Yearbook 2005, People's Bank of products from China and India. The law of China, International Labour Organization (Philippines, South comparative advantage implies that there will Africa), IBGE (Brazil), Banco de Mexico, Ministry of Statistics and Programme Implementation (India); exchange always be opportunities for other countries to rates from IMF International Financial Statistics. Wages are average wages for China, the Philippines, and South Africa, export, even though China and India will average private sector wages in Brazil, and manufacturing come to dominate certain sectors. In general, wages for India and Mexico. as the global demand for Chinese manufac- Note: 1998­2000 wages for the Philippines have been estimated using observed wages from 2001 and projecting tured products increases, dollar-denominated them backward using GDP per capita growth rates. wages in China will tend to increase, in re- sponse to higher wage demands from Chinese workers (especially if the rural and urban infrastructure (Eifert, Gelb, and Ramachandran labor markets remain partially segmented) 2005). The poor business environment leads and from the inevitable additional upward to lower returns to labor in production, pressure on the yuan. depressing labor demand and real wages. There is evidence that this process is Even within sectors where China is already underway (figure 4.3). In 2004, real expected to dominate world trade, there are wages in China were 2.11 times the level of examples of growing exports of other devel- 1989, and the rate of wage increase acceler- oping countries. The removal of quotas in the ated in 2004­05, especially in the coastal re- United States and the European Union on gions (Yusuf, Nabeshima, and Perkins 2006). imports of textiles and clothing products from In 2005 alone, according to the People's Bank China and India was expected by some to of China, average wages for Chinese workers decimate exports of these products from other rose by 14.8 percent (China Daily, "Worker developing countries. For example, it was sug- Shortage Drives Salary Rise," May 27, 2006). gested that one million jobs would be lost in Thus China's development should not keep Bangladesh and that half the factories in the the poorest countries from being able to industry in Sri Lanka would close down export low-skill-intensive products, as long as (Oxfam 2004). However, exports of clothing these countries can manage to create and sus- from both of these countries to the United tain a business climate that supports invest- States have increased since the quotas were ment and trade. In Africa, competitiveness dismantled. Sri Lankan exports in 2005 were based on low-cost labor is undermined by 6 percent higher than in 2004 (with a further high indirect costs, with the main barriers growth of 3 percent over the first six months being corruption, crime, and inadequate of 2006 relative to the same period in 2005). 113 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Exports of clothing from Bangladesh to the as the challenges it poses for developed and de- United States increased by 21 percent in 2005 veloping countries. The large markets of China and by a further 28 percent over the first six and India have changed the dynamic of South- months of 2006. South trade and offer developing countries a Nevertheless, the growth of Chinese route to decreased dependence on rich countries, exports of textiles and clothing has had nega- whose demand for products produced in the tive impacts on other countries. Many jobs poorestcountrieshasbeenrelativelystagnantfor have been lost in Mexico's maquiladoras years. Demand in Asia, and primarily in China because activities in sectors such as clothing and India, has been the main source of the accel- have been unable to compete with China in eration in African exports since 1990. Asia also the U.S. market. Clothing exports from hasbeenakeysourceofrecentexportgrowthfor African countries have declined substantially Latin America. Overall, China's share of the since 2004, amid reports of substantial loss world'snon-oilimportsgrewfrom1.8percentin of jobs in the sector. It is clear, therefore, that 1990 to 6.5 percent in 2004, implying substan- the emergence of China and India as major tial opportunities for its trading partners to ex- exporters will entail significant adjustment in pand exports and create jobs (figure 4.4). some sectors in some countries. The adjust- As a result of Asia's increasing demand for ment costs are likely to be higher in countries resources there is an increasing correlation that offer a less favorable climate for business between growth in China and India and and investment and that suffer from more growth in developing countries that have a rigidities in product and labor markets. comparative advantage in natural-resource- It should not be forgotten that trade and intensive products (Lederman, Olarreaga, FDI have contributed to unparalleled reduc- and Soloaga 2006). Even resource-abundant tions in poverty in China and can continue to countries that have not increased exports do so. The poverty rate (people living on less to Asia--such as Bolivia, Colombia, and than $1 a day) in China fell from almost 60 percent in 1980 to 17 percent in 2003. While lifting more than 400 million people Figure 4.4 China's imports from out of poverty is a remarkable achievement, developing countries have surged over close to 200 million people still live on less the last two decades than $1 a day, many of whom stand to bene- fit from China's continued trading strength. Non-oil imports, $ billions 140 For other countries the impact of the inte- gration of the large emerging economies 120 should not be qualitatively different from the 100 pressures that globalization has exerted on 80 labor markets over the past 30 years, as sum- marized above. Unskilled workers in both rich 60 and developing countries are likely to face 40 greater volatility of employment and continu- 20 ing downward pressure on relative wages. 0 The following section will discuss how policy 1985 1990 1995 2000 2005 makers can help to ameliorate these costs. Africa Latin America South Asia Europe and Central Asia . . . but opportunities for export and Middle East East Asia and the Pacific growth will remain for all countries and North Africa The entry of large economic entities into the Sources: WITS and World Bank staff calculations. global market offers opportunities as enormous 114 N E W P R E S S U R E S I N L A B O R M A R K E T S Ecuador--have seen benefits from higher export-diversification efforts. Markets in Asia world prices for their exports. Lederman, are very large, but key products for developing Olarreaga, and Soloaga (2006) also find countries face high tariff barriers. For exam- fairly strong complementarity rather than ple, cocoa beans face applied tariffs of 30 per- substitutability between the exports of China cent in India and 8 percent in China (in and Latin America to third markets. They contrast to zero protection in developed coun- attribute this complementarity to the growing tries). Second, for traditional commodities, importance of production networks and the even if the reduction in tariffs in Asia does not ability of Latin American firms to join lead to new exports for a specific developing them, the impact of cheaper imports of inter- country, there will be a positive effect through mediate inputs on export competitiveness, and the impact on world prices. Third, it is im- learning by exporting larger amounts to portant to consider the tariffs on the final China. Nevertheless, they suggest that if Latin products that use resource-intensive inputs American and Caribbean countries were to exported from developing countries. Reducing refrain from protectionist policies that prevent such protection will expand the demand for them from using cheap inputs from China those inputs. It would also reduce tariff esca- and India and were to invest more in skills, lation and ease one of the constraints that research and development (R&D), and insti- limit higher value­added activities from being tutions, they would be able to further exploit undertaken in developing countries. opportunities in the new global economy. Moreover, both China and India have The surging demand in Asia for minerals become significant sources of FDI for both has been the primary driver of growing South- developing and developed countries. India's South trade. But China and India offer huge outward FDI stock grew from $0.6 billion potential for a range of other products as well, in 1996 to $5.1 billion in 2003. China and including agricultural products. However, India now occupy positions 54 and 72 (out of trade restrictions keep many developing coun- 132 economies) in terms of outward FDI per- tries from gaining market access for many formance (UNCTAD 2005).13 About two- such products.12 For most developing coun- thirds of cumulative Indian FDI has gone to tries multilateral trade negotiations are poten- other developing countries, but developed tially a major route to better access to growing countries (in particular the United States) are markets in Asia and to better prices for tradi- important markets at the moment. The lead- tional exports. The key feature of access to ing developing country is Mauritius, which markets in Asia for developing countries in attracts about 10 percent of Indian investment Africa and Latin America is that access flows. In the information technology sector, should occur on a most-favored-nation Indian firms' success in global sourcing ex- basis--that is, each country should be entitled posed them to new knowledge and business to the best trade terms an importer offers to methods from developed-country companies any nation. Therefore market access is best and induced outward FDI through demonstra- addressed through multilateral trade negotia- tion and spillover effects. Liberalization of tions, so that tariff concessions made by the Indian government's policies on outward China and India are immediately available to FDI since 2000 also proved critical. Restric- all developing countries, regardless of their tions on maximum overseas investments as a size and global importance. percentage of net worth have been removed, Lower duties in Asia can buoy the export as have the requirement to obtain prior ap- prospects of other developing countries in proval for investments from the Reserve Bank three ways. The first is through lower tariffs of India and prohibitions against overseas on products currently exported by these devel- investments in the same activity as the com- oping countries or that are the focus of pany's core activity in India (UNCTAD 2004). 115 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Fears that China and India will quickly Innovations matter. To understand differ- dominate high-technology sectors are ences in levels of income across countries and misplaced differences in rates of growth of per capita Some worry about the impact of the rising income, it is necessary to explain the sources numbers of skilled workers in China and of variations in productivity. Innovation is India.14 Freeman (2006) suggests that the at the heart of such explanations. In recent increase in these numbers together with in- models of endogenous growth, innovations creased capacity for technological advance- lead to new products and processes that are to ment will undermine the advantage that rich some extent protected by patents and other countries have in high-tech, high-productivity institutional mechanisms that return profit to activities. Trefler (2005) has put this issue the innovator and bolster the incentive to concerning long-run comparative advantage invest. Where protection of the innovation in the following way: will China and India is less than full, a certain amount of dominate high-tech goods and services to the "disembodied" knowledge becomes accessible west, leaving, for example, "Americans to to other innovators and so adds to the stock of mend the socks of Chinese business execu- knowledge available to all, reducing the costs tives"? The prognosis that China and India of future research and development (see will swamp the global market not only with Helpman 2004 for a survey). low-skilled-intensive products but also skill- Some of a country's R&D effort may thus intensive high-tech products is based on the be accessed by other countries, even as it assumption that success in high-tech sectors augments the national stock of knowledge. The depends on the absolute number of scientists main conduits for such technology transfer and engineers rather than the relative number are FDI and trade. In this way the innovative of such workers in the overall workforce. This efforts of rich countries push out the global view also fails to take account of a well-estab- technology frontier and support the growth lished literature that identifies the critical im- of their total factor productivity. Developing portance of domestic institutions in driving countries, which invest little in R&D, can and sustaining innovation-based growth.15 achieve long-run productivity growth through Much evidence supports the view that a process of continually catching up to the growth and income levels do not depend technology frontier. Policies that attract FDI solely on the physical amounts of capital and from rich countries, openness to technology- labor that are available in a country; instead, intensive imports, and learning by exporting they depend on how those factors are com- into the most demanding markets are crucial bined in production. Cross-country varia- for this catching up. This learning can also be tions in per capita income cannot be ac- enhanced by temporary movement of people.16 counted for by differences in endowments of Multinational firms exhibit the highest capital and labor, but by variations in pro- levels of total factor productivity and create ductivity. For example, in 1988 output per more knowledge inputs than other types of Chinese worker was about 6 percent of that firms. Criscuolo, Haskel, and Slaughter (2004) of the typical U.S. worker. Most of that dif- find that globally engaged firms generate more ference was due to lower productivity in ideas than their purely domestic counterparts, China rather than lower capital per worker not only because they employ more re- or lower levels of human capital. If produc- searchers, but also because they have access to tivity levels had been the same, output per a wider pool of knowledge. That pool is deep- worker in China would have been more than ened by contacts with suppliers and customers 50 percent of that in the United States (Hall and, for multinationals, by the intrafirm and Jones 1999). stock of ideas. Others (Bernard, Knetter, and 116 N E W P R E S S U R E S I N L A B O R M A R K E T S Slaughter 2004, cited in Criscuolo, Haskel, advantage and so influence the commodity and Slaughter 2004) find that the parents of structure of trade. Nunn (2005) shows that U.S.-based multinationals perform about two- countries with a good institutional environ- thirds of all private R&D in the United States ment for contract enforcement will tend to have but are a small fraction of 1 percent of the a comparative advantage in producing and ex- total number of U.S. firms. Thus, it appears porting goods that require relationship-specific that openness to trade is important not only investments.17 Countries with poor contract for poor countries to absorb new technologies enforcement will suffer from underinvestment created by firms in developed countries, but and thus higher costs of production for also for wealthier countries to stimulate in- goods that require relationship-specific invest- vestment and productivity growth. ment. Such investments are more likely to be necessary in industries in which firms have So do institutions. Even after taking into some form of firm-specific asset, which in turn account innovation efforts, a substantial are more likely to be high-technology and amount of the variation in per capita income innovation-intensive industries. The same is levels and growth rates across countries equally true for services. remains unexplained. What accounts for the The structure of all countries' exports is rest of the variance? Institutions and insti- thus influenced by the nature of domestic in- tutional quality are now accepted as the reason stitutions. The growth of exports from China why some countries have higher productivity and India in some products and services that than others and why growth rates have require relationship-specific investment has differed across countries, even when factor been facilitated by having good institutions endowments and rates of innovation are in particular enclaves of the economy such as similar (Helpman 2004). For example, Hall special economic zones. The ability of these and Jones (1999) conclude that "a country's countries to substantially increase exports of long-run economic performance is determined these goods and services further will depend primarily by the institutions and government on the ability to engender economywide policies that make up the economic institutional change, something that will be environment within which individuals and much harder to achieve and will occur more firms make investment, create and transfer slowly. ideas, and produce goods and services." Continual technological innovation and To compete with the United States, the changes in demand make comparative European Union, and Japan in innovation and advantage a dynamic concept. It is very diffi- high-tech products, China and India will cult to predict in which sectors and tasks require institutions similar to those of the countries will be efficient producers. Thirty OECD countries. The two countries are a long years ago, who could have predicted the way from having such institutions at present. emergence of the iPod or known how the Moreover, building them takes a long time value added in production and the return and is unlikely to occur within 25 years to knowledge would be distributed across (Trefler 2005). Thus the United States leads in countries (box 4.4)? innovation-based growth not because it has The opportunities of global production more scientists and engineers, but because it chains will encourage the upgrading of has an institutional framework that allows domestic institutions, as countries compete companies such as Microsoft, Apple, and on quality and efficiency as well as price--just Yahoo to exploit new ideas. as they have for another set of domestic insti- Recent research has highlighted how institu- tutions related to labor standards. Rather tional quality can determine comparative than a race to the bottom, with declining 117 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 4.4 Global production and the iPod T ake just one component of the iPod nano, the It is suggested that overtime is compulsory. Never- central microchip provided by the U.S. company theless, wages are higher than the average of the PortalPlayer. The core technology of the chip is region in which the assembly plants are located and licensed from British firm ARM and is modified by allow for substantial transfers to rural areas and PortalPlayer's programmers in California, Washington hence contribute to declining rural poverty. State, and Hyderabad. PortalPlayer then works with PortalPlayer was only established in 1999 but microchip design companies in California that send had revenues in excess of $225 million in 2005. the finished design to a "foundry" in Taiwan (China) PortalPlayer's chief executive officer has argued that that produces "wafers" (thin metal disks) imprinted the outsourcing to countries such as India and with hundreds of thousands of chips. The capital Taiwan (China) of "non-critical aspects of your busi- costs of these foundries can be more than $2.5 mil- ness" has been crucial to the development of the firm lion. These wafers are then cut up into individual and its innovation: "it allows you to become nimbler disks and sent elsewhere in Taiwan (China) where and spend R&D dollars on core strengths." each one is tested. The chips are then encased in Since 2003, soon after the iPod was launched, the plastic and readied for assembly by Silicon-Ware share price of Apple, the company that produces and in Taiwan (China) and Amkor in the Republic of sells the iPod, has risen from just over $6 to over Korea. The finished microchip is then warehoused $60. Those who own shares in Apple have benefited in Hong Kong (China) before being transported to from the globalization of the iPod. mainland China where the iPod is assembled. Working conditions and wages in China are low Sources: C. Joseph, "The iPod's Incredible Journey," Mail on relative to Western standards and levels. Many Sunday, July 15, 2006; "Meet the iPods's `Intel,'" Business workers live in dormitories and work long hours. Trends 32(4)(April), 2006. wages and standards as countries compete on of U.S.-owned patents with at least one inventor trade and investment, globalization is encour- who is a resident of China or India. Thailand aging gradually higher labor standards, both and Mexico have not witnessed such growth. directly in terms of attracting FDI and indi- But Cravino, Lederman, and Olarreaga rectly, through higher growth (box 4.5).18 (2006) find no evidence that FDI by Western firms in China and India is displacing FDI Implications for middle-income countries. in Latin America. In a detailed econometric While China and India are unlikely to threaten exercise, Bravo-Ortega and Lederman (2006) Western dominance of "big idea" innovation, find no statistical evidence that current patent- Puga and Trefler (2005) suggest that the great ing activity by China and India has had an capacity of these countries for lower-level impact on the number of patents of Latin incremental innovations may have important American countries. They do find some evi- implications for middle-income developing dence, however, that the stock of patents to countries. The presence of many well-trained which China and India have contributed scientists and engineers in China and India is feeding the innovation process in Latin means that Western firms looking to invest will America. In other words, innovators in Latin tend to be attracted to these countries--for their America can learn from innovations under- greater capacity to assist the firm in incremental taken in China and India. innovation--rather than to other countries, Thus both economic theory and the avail- such as Thailand and Mexico. Puga and able evidence suggest that while the sheer size Trefler refer to the rapidly increasing number of new entrants into the global economy poses 118 N E W P R E S S U R E S I N L A B O R M A R K E T S Box 4.5 Does globalization lead to a race to the bottom on labor standards? L ooking at the relationship between core labor around 50 million persons worldwide. That said, the standards (freedom of association and collective majority of EPZs are covered by national labor laws, bargaining; and elimination of forced labor, child and physical conditions and wages tend to be better labor, and discrimination in employment) and trade, than in the rest of the economy (ILO 1998). EPZs the OECD (1996) finds no empirical support for with poor working conditions do not attract long term the view that low-standards countries will enjoy investment--"smart" EPZs have introduced measures gains in export-market shares to the detriment of to continuously upgrade labor (OECD 2000b). high-standards countries. There is also no evidence that low core labor standards are associated with Why not a race to the top? low unit labor costs: real wages actually grew faster Globalization may be forcing a race to the top, as it than productivity growth in a number of low- places a new emphasis on speed, efficiency, and qual- standards countries from the mid-1980s to the ity as well as cost, shifting the focus from cheap mid-1990s. While core labor standards will not labor to productive labor (ILO 1998; Aggarwal necessarily affect comparative advantage negatively 1995). Countries can gain an advantage by improv- and indeed may have a positive affect, noncore or ing labor standards. Strengthened core labor stan- economic standards such as working time and mini- dards can increase economic growth and efficiency mum wages may affect trade performance negatively by raising skill levels, thereby creating an environ- (OECD 2000a).a However, the picture is not clear; ment that encourages innovation and higher produc- Dehejia and Samy (2002) find no clear link between tivity (Stiglitz 2000; OECD 2000a). At the same labor standards and a country's competitiveness. time, in the sectors with a reputation for poor labor Rodrik (1996) finds that labor standards are a signif- standards (clothing, footwear, and sporting goods) icant determinant of labor costs when one controls consumers are increasingly demanding products pro- for productivity, but not of comparative advantage, duced under acceptable working conditions, with which is mostly determined by factor endowments. monitoring and certification, often by trusted non- Evidence on FDI also suggests that firms are governmental organizations. attracted to countries with higher, not lower, labor Indeed, efforts to promote labor standards at a standards (OECD 2000b; Aggarwal 1995; Rodrik global level have been increasing: examples include 1996; Brown, Deardorff, and Stern 2002).b Multina- the 1998 International Labour Organization (ILO) tionals invest principally in the largest, richest, and Declaration on Fundamental Principles and Rights at most dynamic markets; with the significant exception Work (under which monitoring and reporting on of China, countries where core labor standards are not core labor standards is extended to all members) and respected receive a very small share of global flows. development cooperation programs to reduce child Even in China, the average foreign affiliate pays wages labor. More controversially, trade agreements have 30 percent higher than the average in state-owned en- been used to promote compliance with labor stan- terprises and has higher occupational safety and dards. The United States suspends access under the health standards than Chinese-owned firms (Lardy Generalized System of Preferences (GSP) in the event 2004). Overall, multinational firms provide incentives of noncompliance, while the European Union grants to improve, rather than worsen working conditions; additional access for compliance. Labor provisions pay higher wages than alternative employment; and or side agreements figure in U.S. free trade agree- tend to promote, rather than repress worker rights ments. Links to the World Trade Organization (Brown, Deardorff, and Stern 2002). (WTO) have faced strong resistance from developing In some countries, labor regulations do not apply countries.c Other market-based mechanisms, such as and a range of labor standards issues still arise in labeling schemes or codes of conduct for firms (at export processing zones (EPZs), which now employ the OECD, ILO, and firm level) have expanded, with (continued) 119 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 4.5 (continued) most U.S. Fortune 500 companies now embracing (for example, minimum wages), which will vary depending such codes (see OECD 2000a; Stern 2003).d on the level of income in a given society. b Labor standards rise with income (Stern 2003; Data on freedom of association rights in 75 countries that OECD 2000a), and the path to higher growth for represent virtually all of world trade and all inward and out- ward FDI show no significant deterioration in these rights in developing countries lies in seizing the opportunities any of the 75 countries between 1980 and 1999 (the period of global production networks in goods and services. during which competition for FDI heated up). Data show But this in turn requires efforts to raise productivity significant improvement in those rights in 17 countries and create a stable and attractive environment for (OECD 2000b). FDI. And the evidence suggests that improving core cMoran (2004) provides a persuasive analysis of the labor standards and creating frameworks for sound practical problems of using dispute settlement under trade agreements to enforce labor standards, given the lack of inter- and stable labor relations can contribute to both of national agreement on exactly what core labor standards mean these goals, with the potential to create a virtuous and what is required for adequate implementation and the circle of rising wages and standards for workers in reliance on incomplete, nonrepresentative, noncomparable, developing countries. and potentially biased sources of information. dMore recently, attention has shifted beyond core labor standards to the concept of decent work (work that is freely aBates (2000) distinguishes between core labor standards, chosen, provides an income sufficient to satisfy basic economic which are viewed as fundamental human rights and can create and family needs, respect for rights and representation, basic the framework conditions for the economy to operate security through some form of social protection, and adequate efficiently, and developmental or economic labor standards conditions) (ILO 2004). a number of challenges to other countries, both of the services value chain to be performed in developed and developing, there are enormous different locations around the globe--a phe- opportunities. To grasp these opportunities re- nomenon that has come to be known as "out- quires that countries have in place a policy en- sourcing" or "offshoring," but could perhaps vironment that allows competitive advantages be most accurately termed "global sourcing to be exploited and the key sources of growth of services."19 Global sourcing has increased to flourish, while ensuring that those workers competition in services markets for a wide va- adversely affected are assisted in adjusting by riety of activities, from low-skilled functions moving to new sectors and/or by augmenting such as data entry, word processing, and call their particular skill set. In other words, while centers to higher-skilled activities such as soft- aggregate gains are available to all countries, ware development, consultancy, medical ser- some industries, firms, and workers will incur vices, and R&D. A range of services previously some pain. Appropriate policy responses are thought to be nontradable are now being pro- discussed further below. vided electronically over large distances. Global sourcing allows firms to benefit from around-the-clock production (for just-in- New challenge II--global sourcing time delivery of both goods and services) and of services lower wage costs. Estimates of the total cost savings from global sourcing vary across a Workers in previously sheltered services wide range--for example, from 15­30 percent face international competition (Atkinson 2004) to 30­60 percent (industry The global competition in goods that has been estimates cited in Kirkegaard 2005). under way for decades is now visible in ser- While absolute numbers to date are not vices, as falling telecommunications costs and large, growth rates have been high, and global greater openness to FDI enable different parts sourcing of services is expected to grow by 120 N E W P R E S S U R E S I N L A B O R M A R K E T S The number of service jobs that migrate Figure 4.5 Developing-country exports of from rich to poor countries could be large business services are growing rapidly Estimates to date of the scale of potential job Growth in export of business services, 1994­2003 movements from global sourcing of services vary widely according to the definitions and India methodology used. A great deal of attention Estonia was initially given to a report by Forrester Romania China Research (2002) that estimated that 3.3 mil- Israel lion service jobs would move offshore from the Brazil United States by 2015. However, when put Argentina into perspective--the U.S. economy creates Mauritius about 30 million jobs per year--that number is EU15 quite small. Even for the job categories deemed United States to be vulnerable to outsourcing, including Canada management and computer operations, the Barbados predicted impact amounts to just over 0.5 per- Dominicaa cent of existing employment (see Bhagwati, Australia Panagariya, and Srinivasan 2004). Ghana Japan Blinder (2006) asserts that a much broader range of services will be liable to global sourc- 0 100 200 300 400 500 600 700 Percent ing (almost all of those activities that do not require direct personal delivery) as communi- Source: Data from IMF Balance of Payment Statistics. cations costs decline further and technology Note: "Business services" are defined as total services minus transportation, travel, and government services. continues to advance. His rough estimate is a. 1994­2002. that between 28 and 42 million jobs in the United States could move overseas as a result of global sourcing. OECD (2005c) concludes 30 percent per year over 2003­08.20 While de- that close to 20 percent of total employment veloped countries still dominate the trade, some could potentially be affected by information developing countries experienced fast growth in and communications technology­enabled off- exports of business services between 1994 and shoring of services. 2003: nearly 700 percent for India; more than The potential size of the future market for 200 percent for China, Brazil, and Argentina; global sourcing of services remains a matter of and more than 100 percent for Mauritius, debate, reflecting uncertainties over how the Barbados, and Dominica (figure 4.5). dividing line between tradable and nontrad- Sourcing locations expand and change over able services will shift over time (see box 4.6). time, and technology advances are likely to Even if these higher predictions come to allow more services to be provided offshore. As fruition they do not imply a corresponding costs in Ireland rose, activities moved to India lower level of employment. The impact of and the Philippines. Now, as costs in India rise, global sourcing of services on the overall level other locations, including some in Eastern Eu- of unemployment in rich countries may be rope, are becoming popular (Atkinson 2004). small, especially in countries that are effective Language patterns influence location deci- in generating new jobs. sions, but these are not immutable.21 Against this backdrop, countries across all regions and Global sourcing will benefit both levels of development--from Senegal to Sri developed and developing countries Lanka, Argentina to Zambia--are now seeking The shift in jobs resulting from global sourc- to become sites for services sourcing. ing of services is unlikely to be a zero-sum 121 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 4.6 The number of services jobs liable to be moved abroad: large or small? T o identify the tradability of industries and occu- second bar adds those jobs that could potentially be pations, Jensen and Kletzer (2005) use indicators lost if all jobs in all tradable sectors (including those of regional concentration of production in the United where production is only relatively geographically States to group industries into three categories of concentrated) could be sourced globally. This tradability, leaving a similar number of industries in changes the picture dramatically; more than half the each category (the more the geographical concentra- jobs in nonpersonal services could be affected by tion, the higher the degree of tradability). They then globalization (about 30 million jobs). use this degree of tradability of the sector to estimate In Jensen and Kletzer's analysis there is a distinct the number of jobs that are prone to global sourcing. difference in the vulnerability to global sourcing be- Results are very sensitive to assumptions about tween wholesale and retail activities, which are less what is considered tradable. The first figure below concentrated and therefore deemed less prone to presents possible effects for the U.S. economy global sourcing (about 36 percent of jobs in these (excluding public administration), distinguishing sectors are at risk), and professional services (where between the agriculture and manufacturing sector, 71 percent of jobs are at risk). However, technologi- nonpersonal services, and personal services. For cal change may make global sourcing more relevant each category, the first bar shows the number of to the wholesale and retail activities. The top part of workers currently employed, while the second bar the second columns shows that an additional 9.7 mil- shows the number of jobs that may potentially be lion jobs could be lost if the sensitivity to global globally sourced under three possible scenarios. The sourcing for professional sectors were equal to that lowest part considers only the highly concentrated of the other nonpersonal services (the light gray part industries as tradable; here fewer than 4.5 million in the second column for nonpersonal services). The jobs in the nonpersonal services sector could be second figure below shows the same analysis for the lost overseas. In contrast, the middle part of each European Union (EU15). The figure suggests that the Jobs and potential for outsourcing, United States Jobs and potential for outsourcing, EU15 Millions of jobs Millions of jobs 70 70 62 60 57 57 55 60 50 50 40 40 36 9.7 10.3 30 30 23 20 20 21.8 24.4 25.2 3.6 13.5 2.4 10 10 0.8 0.3 6.8 10.1 4.3 4.4 0 0 Agriculture Nonpersonal Personal Agriculture Nonpersonal Personal and services services and services services manufacturing manufacturing Sources: Jensen and Kletzer 2005; World Bank staff calculations. Note: The first, black bar for each sector represents the number of workers currently employed in that sector. The second bar indicates the number of jobs potentially globally sourced. The black section represents only workers in highly geographically concentrated industries subject to global sourcing; the other two sections represent workers in less geographically concentrated industries (see text). 122 N E W P R E S S U R E S I N L A B O R M A R K E T S Box 4.6 (continued) potential future adjustment for the EU15 is some- sourced. For these reasons it is preferable to identify what skewed toward the agricultural and manufac- tasks, rather than sectors, as tradable or nontradable turing sectors, where currently a higher share of total (see box 4.7). Jensen and Kletzer make a crude at- employment can be found than in the United States. tempt at this by repeating their exercise for broad A further important caveat to these estimates is occupational groups. They find that 11 percent of total that only certain elements of production in tradable employment is represented by tradable occupations in sectors can be sourced overseas. Equally, there could industries that are classified as nontradable. Similarly, be activities in the production process of largely non- about 22 percent of the total workforce is found in tradable services sectors that could be globally nontradable occupations in tradable industries. game, owing to the presence of significant There is also little evidence to date that offsetting factors. While workers whose jobs tradable service activities have lower employ- are liable to offshoring will face lower labor ment growth than other service activities, demand and downward pressure on their rel- or that net outward investment or imports of ative wages, workers who are complemen- business services are associated with signifi- tary to the offshored activities will see a rise cant declines in the share of employment po- in their productivity and an increase in rela- tentially affected by outsourcing (Jensen and tive wages. In addition, global sourcing will Kletzer 2005; OECD 2005a; Amiti and Wei augment the productivity of firms that utilize 2004). However, growth is lower at the lower the opportunities presented by lower labor end of the skill distribution--although this costs overseas. These firms are more likely to may also indicate that these jobs are most expand than other firms, increasing their readily substituted by technology. Worker dis- demand for labor, some of which will be placement rates are higher in tradable ser- for local tasks that can be fulfilled by the vices, but affected workers have higher skills type of worker affected by offshoring, thus and higher predisplacement earnings than offsetting, to some extent, the impact of displaced manufacturing workers (Jensen and offshoring on wages (Grossman and Rossi- Kletzer 2005). Hansberg 2006). The key labor-market issues raised in rich Moreover, demand is not inelastic--lower countries by the global sourcing of services are wages for software workers in developing the nature of the new jobs that will replace countries raise global demand for software, those transferred overseas, and the difficulties benefiting all countries. Some OECD coun- that firms and workers may face in adjusting tries are experiencing a net inflow of service to this new facet of globalization. Workers jobs from outsourcing (Amiti and Wei previously sheltered from global competition 2004); investment by foreign companies in are facing greater job insecurity, downward the United States, for example, exceeded pressure on their wages, and potential costs of investment by U.S. companies in foreign coun- adjustment in moving from one job to another tries every year over 1996­2001 (Atkinson or in upgrading their skills to obtain new 2004),22 and several OECD countries have employment following displacement. These experienced double-digit growth in exports of issues are explored in more detail below, fol- business services. For example, exports grew lowing a brief discussion of an appropriate at 11 percent in both the United States and framework in which to assess the nature and Australia (OECD 2005a).23 impacts of global sourcing. 123 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Global sourcing of services offers extremely adverse location, one of the highest opportunities as well as challenges population densities in the world, and a high Global sourcing of services is creating consid- population growth rate. While increasing the erable opportunities for development in poor, quality and quantity of exports of traditional low-wage countries, through export possibili- agricultural exports (coffee) and minerals is ties and through access to cheaper service in- crucial to increases in incomes for the poor in puts that raise productivity when used in other the short to medium term, the government of sectors. Global sourcing is providing impor- Rwanda has identified the provision of IT- tant new employment--in India, employment intensive services, both locally and abroad, as in the information technology (IT) sector is a base for growth in the long run, to provide now three million, although this is concen- for employment and turn the country's large, trated in five or six urban centers (Yusuf, but very young, population into a driver of Nabeshima, and Perkins 2006). Employment development rather than a constraint. creation is at a wide range of skill levels, re- The important new opportunities for devel- flecting the range of activities open to global oping countries are accompanied by consider- sourcing. In the relatively low-value segments able challenges related to the provision of such as call centers, wage costs are important necessary infrastructure, the design and imple- determinants of location (along with language mentation of appropriate regulation, better skills), and competition is fierce among devel- education to increase the supply of human oping countries. At the high end, global sourc- capital, and the creation of strong marketing ing of services may be reducing incentives profiles and reputations for reliability.24 Ac- for skilled migration by creating new opportu- cess to relatively cheap and reliable electricity, nities at home. A large number of those em- a critical problem for many poor countries, ployed as a result of global sourcing are will be necessary. High-quality telecommuni- women, offering a different route to develop- cations infrastructure must be accompanied ment than those based on the growth of agri- by a competitive framework for the provision culture and manufacturing. of telecommunications services. Liberalization While India and China are likely to come of the trade and investment regime, comple- to dominate the market for global sourcing mented by an appropriate and effective regu- of services, comparative advantage will en- latory environment, can help ensure the sure that there are opportunities for many efficient and competitive provision of the developing countries. Small island economies telecommunications backbone services. in the Caribbean, for example, have been Many developing countries could assist their able to attract certain back office activities nascent IT sectors by joining the Information from the United States, such as data entry. Technology Agreement (ITA) of the WTO. The services revolution and global sourcing The agreement covers the main categories are offering opportunities for new exports of IT products, computers, telecommunica- and for attracting services-related foreign tions equipment, semiconductors, semicon- investment for a range of poor countries. IT ductor manufacturing equipment, software, and global sourcing offer new and alterna- and scientific instruments, and commits mem- tive drivers of development that circumvent bers to bind tariffs at zero on these items. some of the key constraints to growth driven Joining the ITA can provide a strong signal to by the expansion of exports of agricultural investors, both domestic and foreign, of a and manufactured goods. country's commitment to an open IT environ- This is most apparent for landlocked coun- ment by ensuring access to necessary equip- tries and small (often island) economies that ment at world prices. The ITA has 43 mem- face very high transport costs. For example, bers (with the European Union treated as development in Rwanda has to confront an one), among them industrial countries and 124 N E W P R E S S U R E S I N L A B O R M A R K E T S some large and small developing countries, rise, but displaced workers face some costs of such as China, the Arab Republic of Egypt, El adjustment. Salvador, India, Mauritius, Moldova, and For Trefler (2005), by contrast, the fact that Morocco. As yet, however, none of the least skilled workers lose their jobs when services are developed countries is a member. sourced from low-wage countries has impor- tant economic implications that do not arise But does the global sourcing of services when low-skilled jobs disappear. The loss of have different implications for labor in relatively high-wage jobs and the pressure on developed and developing markets? the wages of high-skilled workers may reduce Trade in services that use skilled labor inten- economic incentives to invest in and to acquire sively is not new. In the standard analysis skills. In addition, in knowledge-intensive ser- of multinational firms, parent companies in vice activities skilled workers are more likely to developed countries are seen as exporting a have obtained some industry- and firm-specific range of services such as design, management knowledge that is lost when the job is lost. This and engineering consultancy, marketing, and may have a direct negative impact on produc- finance to their overseas subsidiaries in poorer tivity, especially if the knowledge is comple- countries. What is new is trade in the opposite mentary to other skills or factors. In developing direction, as services both within multination- countries the opposite will tend to occur. The als and through arm's-length trade flow from transfer of know-how, the increasing demand low-wage countries to richer markets. for skilled workers, and the upward pressure An immediate implication of this new on skilled wages will tend to increase the in- development is that the standard factor- centives to acquire skills. This will increase endowments model of trade (countries export demands on the education system in develop- goods and services that make intensive use of ing countries, which in many cases is likely to factors abundant in their country) cannot ex- become a constraint on this process. plain why skilled-labor-intensive services are The rapid pace of change and flexibility de- being exported from countries with very manded by competitive global markets, along scarce skilled labor. A common explanation with new trends such as global sourcing of ser- involves the absence in developing countries vices, will lead to potentially rising adjustment of the knowledge-based assets that are com- costs falling on a wider range of--more highly plementary to skilled labor. The lack of these skilled--workers. These trends all argue for assets limits the use of skilled labor at home countries to review their domestic policy and and keeps such workers cheap even though institutional frameworks to ensure that their they are relatively more scarce than in rich advantages can be exploited and that affected countries. Globalization in the form of the workers are supported when they incur ad- transfer of know-how to complement cheap justment costs. skilled labor in poor countries leads to trade in skilled-labor-intensive services. Policies to confront the labor There is much discussion of whether the global sourcing of service activities to low- market challenges of globalization F wage countries presents features and issues dif- ocusing on factors that determine the ferent from those associated with global trade growth of productivity will be key to con- in goods. Bhagwati, Panagariya, and Srini- fronting the challenges of globalization with- vasan (2004), for example, argue that global out neutralizing its opportunities. This will re- sourcing of services has effects that are not quire a change of mindset by policy makers, qualitatively different from those emanating who must grasp and internalize the fundamen- from the sourcing of goods. In both cases, there tal changes in the nature of international pro- are gains from trade and national incomes duction and trade (see box 4.7). 125 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Box 4.7 Trading goods and services or trading tasks? I n the classic conception, international trade is the likely for complex tasks that require tacit informa- exchange of complete goods and services across tion. The latter often require relationships and are national boundaries. Countries gain from specializa- often best performed in clusters of individuals. Your tion in particular sectors of the economy, such as neighbors matter. Even for some mundane tasks, textiles and steel. Within firms, gains are had from such as mowing the grass, physical presence is higher productivity, that is, by allowing workers to required. specialize in particular tasks. In the past, effective co- In this new global environment, interventions that ordination of these efforts, and the combination of target particular sectors will be ineffective relative to tasks to produce a product, required proximity. initiatives to provide an environment that supports Communication required physical presence and the activities and tasks. This entails greater emphasis on transportation of intermediate inputs was slow and a business environment that facilitates the entry and costly. Specialization led to geographic concentration exit of firms across all sectors and policies and infra- of production. International trade occurred if con- structure and regulations that support the free flow sumers lived in another country. and low cost of imported inputs (whether physical or However, the nature of production has changed. information) to which domestic workers can con- Revolutions in transport and communications tech- tribute their tasks. What matters is the quality of nologies have led to enormous reductions in cost, roads, ports, telecommunications, and electricity to- allowing tasks to be separated in time and space, and gether with relatively low tariffs on imported inputs weakening the link between specialization and geo- and effective regulation of key backbone services. graphic concentration. Instructions and information Finally, the increasing importance of trade in can be effectively conveyed over long distances and tasks creates a challenge for the measurement of intermediate inputs can be transported quickly and international trade flows. Currently imports of goods much more cheaply than before. Thus, increasingly it are recorded according to their invoice value as they is tasks in addition to final goods and services that are cross the border and the whole value of the import is exchanged across national boundaries, resulting in attributed to the country in which the last substantial global production networks of activity in a wide transformation occurred. There is no system by range of sectors. In this new global economy there are which the countries that contributed value added to additional gains from specialization, as firms take the product are identified. Thus, for example, the advantage of differences in the cost of labor and skills value of the iPod discussed in box 4.4, when im- across countries to allocate tasks internationally. ported into the United States, is attributed to China, Some tasks can be offshored more easily than oth- where it is assembled. Yet most of the value of the ers. What matters is the extent to which a particular iPod is added by tasks undertaken in other countries. task is contested globally (Leamer 2006). This is more likely for standard, mundane tasks that can be coordinated through codifiable information and less Sources: Grossman and Rossi-Hansberg 2006; Leamer 2006. In the new environment, productivity investments in human capital, and reductions growth requires openness to new ideas and the of barriers to the flow of knowledge, capital, ability to exploit new technologies and oppor- and labor. This process is not without adjust- tunities. Economies need to be sufficiently ment costs, and complementary policies are flexible to enable resources to move from low- needed to ensure that particular groups in productivity to high-productivity tasks and society do not bear a disproportionate share activities, which policy makers cannot identify of the pain. beforehand. This places a premium on institu- The appropriate policy mix that provides a tions and policies that encourage innovation, framework for productivity growth will vary 126 N E W P R E S S U R E S I N L A B O R M A R K E T S over time and according to country character- investment, and growth. Domestic policy istics, such as level of development and size. reforms, underpinned where necessary by in- Nevertheless, key ingredients will be openness creased "aid for trade" from the international to trade and FDI and an attractive climate for community, will be essential in helping the investment and for innovation, investing in ed- poorest countries benefit from the opportuni- ucation, and repositioning labor-market poli- ties of new global markets. cies to focus on protecting workers, not jobs. In addition, policies that affect innovation Rich countries have a particular responsibility and access to technology are crucial. For the to maintain and, indeed, increase the openness least developed countries, moving up the tech- of their markets to goods and services pro- nology ladder by acquiring technological duced in poor countries. A related issue is the know-how from overseas through trade and impact that globalization and openness may FDI will be a key driver of growth over the have on a country's capacity to raise tax rev- next 20 to 30 years. Innovation and learning enues to fund infrastructure for trade or train- will continue to play essential roles in raising ing for affected workers. productivity and sustaining growth in rich countries and increasingly in the middle- Supporting open access to markets, income countries, placing emphasis on the innovation, and a strong business climate institutions that frame incentives to invest in Openness to trade in goods, services, and R&D and in the acquisition and application ideas provides a critical stimulus to innovation of knowledge.26 In the middle-income coun- and productivity growth, both for countries tries there will be opportunities to be had from at the global technology frontier and those incremental innovations to processes that catching up. But because trade and technology improve the tasks undertaken for foreign can lead to lower relative wages and greater firms and to products that can be tailored for employment volatility for some workers, pol- growing domestic markets. In the richer coun- icy makers are often tempted to meet the chal- tries it is innovation and learning creating new lenges of globalization by increasing trade goods, services, and new processes for pro- protection. Doing so compromises a key ducing them, that will be of greater impor- source of growth. As a former finance minis- tance. The key elements of a policy framework ter of a developing country that undertook to support innovation and learning will differ successful reforms stated, "Trade shocks are between countries according to level of devel- better dealt with through more, rather than opment as well as size but to varying degrees less, trade."25 will include the following: Trade policies interact critically with other elements of the business and investment · Investing in human capital to overcome climate. Reaping the benefits of globalization shortages of skilled labor, including due requires not only openness to trade and in- to migration. Learning-by-doing in firms vestment but also physical infrastructure increases with its workers' human capital. and a policy framework that enables actual · Supporting public research through and potential exporters to effectively exploit universities and research centers, and their advantages. High costs of clearing facilitating interaction with private busi- customs, poor port infrastructure, weak nesses to ensure dissemination of "basic" telecommunications services, and poor regula- knowledge that stimulates research for tion, for example, raise costs and hamper commercially exploitable innovations. competitiveness. These major challenges for · Defining--and enforcing--adequate intel- developing countries, particularly the least lectual property rights to encourage do- developed, must be addressed if trade liberal- mestic investment in innovation and ac- ization is to be effective in stimulating trade, quisition of technology through FDI.27 127 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 · Promoting access to finance, especially could place greater pressure on the education for new entrants and small and medium system to provide the industry-specific skills enterprises (SMEs), which are more previously provided by firm-level training; likely to be innovative (Geroski 1990). while continuing volatility could also place a · Careful review of fiscal incentives to premium on providing workers with the gen- stimulate R&D and innovation, taking eral skills that enable continuous adaptation. into account that evidence of effective- In sum, education systems will be expected ness is scarce (de Ferranti and others to provide more people with more opportuni- 2002) and that there could be crowding ties to learn across a broader menu of educa- out or in of private investment (Jaumotte tional and skill-development options at more and Pain 2005). stages of their lives than ever before. This will require a new model of education and train- Providing more people with lifelong ing, as well as ongoing reform of traditional learning methods, providers, and financing of educa- In all countries investment in education will tion (box 4.8). become an ever more critical determinant of labor-market performance in the context of Protecting workers, not jobs greater global competition and increasing While globalization offers new opportuni- rewards to skills. Higher-skilled workers are ties for workers, it can also entail greater better at dealing with changes, including movement--for example, between jobs, sec- adoption of new technologies, new workforce tors, or regions--and this brings with it addi- organizations, and ongoing pressures for ad- tional risk. This calls for policies that shift the justment and shocks, and also support the cre- emphasis from measures designed to protect ation of well-functioning institutions (World those in employment--which, as discussed Bank 2006; Hoekman and Javorcik 2006).28 earlier, can discourage job creation--to mech- Countries need to focus not only on enroll- anisms aimed at ameliorating the potentially ment in education but also on quality and rel- negative effects of greater labor movement evance, a fact underlined by the prevalence of through targeted labor-market policies and youth unemployment in both developed and social safety nets. While the precise combina- developing countries. tion of labor-market programs and income Education systems everywhere face new support measures will need to be determined challenges, however. In the face of rapid at the national level--taking account of local changes in technology and business organiza- circumstances and involving all relevant stake- tion, these systems struggle to keep pace with holders--and there can be important differ- demand for new skills--a trend likely to be ex- ences in the types of programs that are most acerbated by global sourcing of services. For effective in developing and developed coun- individual workers, this means rapid changes tries, some general lessons can be drawn. in the value of their skills, demanding constant In all countries, income support programs retraining and skill upgrading. But workers will remain the core of worker assistance. In facing more rapid obsolescence or devaluation OECD countries, the redistributive impact of of skills may have lower incentives for skill the tax-transfer system increased in the late acquisition. Moreover, firms already faced 1980s and 1990s (Brenton 2006). In develop- with competitive cost pressures and concerns ing countries, the design of such programs that they will not benefit from their invest- raises specific challenges. ment in training as their workers leave for other firms, will increasingly have access to a · Unemployment benefits can ease adjust- global pool of workers with the desired skills ment and maintain public support for to substitute for the existing workforce. This structural change, but if set too high 128 N E W P R E S S U R E S I N L A B O R M A R K E T S Box 4.8 Key challenges for education systems in the new global economy W hile some of the key challenges relate to prob- continuously upgrade how they produce in whatever lems that education systems have traditionally sector they might be employed (de Ferranti and oth- faced, such as increasing access to and quality of ed- ers 2002). This means moving away from a model in ucation, others relate to revisiting the nature, type, which the teacher is the source of knowledge to a and purpose of educational offerings to equip a glob- system where educators function as guides to multi- ally competitive workforce. In key respects, tradi- ple sources of knowledge. tional educational methods are ill-suited to providing Include a range of providers. Including private the lifelong learning that is necessary in the new sector as well as public sector providers can pro- global economy (World Bank 2003). mote greater access to education and greater variety Increase access. In Sub-Saharan Africa and South in educational offerings. Additionally, foreign insti- Asia, more than 40 percent of those aged 25 and over tutions can help upgrade standards, although sound in 2000 had not completed any formal education. In regulation is needed to ensure quality and access, developing countries, public funding--directed and to provide clear and nondiscriminatory condi- through public educational institutions or to individu- tions for investors. Here again, government mea- als (loans or vouchers)--can help expand access sures to ensure broad access may be necessary. (World Bank 2005). Recent policies in Brazil that ad- Strengthen the links between education and dressed supply-side constraints in the education system work. The mismatch between graduates' skills by establishing a minimum spending level per student and labor-market needs in many developing have proven successful in increasing enrollment rates countries argues for greater links between the substantially (de Mello and Hoppe 2005). While the private and public sectors. In the Middle East and central government transfers funds to the local govern- North Africa, skills geared toward public sector jobs ments in case these are unable to finance the prescribed are ill-suited to the needs of industry, while the tra- spending levels, demand for education is increased by ditional focus on law, philosophy, and theology in using school attendance as a requirement for certain education in much of Latin America and the types of income transfers to low-income households. Caribbean is argued to have slowed the development Provide access at all ages. Preschool and early of natural resource sectors (de Ferranti and others childhood programs establish a solid basis for subse- 2002). Postsecondary education should be balanced quent learninga while primary and secondary educa- between academic and technical-vocational training tion give workers the basic skills that enable them to (OECD 1996), with the latter assessed, certified, learn new skills required by technology-induced and formally recognized. changes (OECD 1996).b Lifelong learning helps work- ers to adjust, but government support (for instance, aIn developing countries, early childhood and preschool pro- via a training levy) may be needed (World Bank 2005). grams show returns of $2­5 for every $1 invested (World Bank In many emerging economies, improving the access to 2006). b secondary and postsecondary education will be critical For a full discussion of ensuring access and quality in edu- cation, see World Bank (2006). For a discussion of issues in the in view of the rising skill premium. delivery of basic education services, see World Development Improve quality.c Efficient increased spending Report 2004: Making Services Work for Poor People. should be combined with strengthened incentives to cChildren in Argentina, Chile, and Mexico perform about two teach and learn, and with improved accountability standard deviations below children in Greece, one of the poorest- (World Bank 2006). Quality assurance mechanisms performing countries in the OECD. In reading competence (based (including regionally) and national qualifications on the OECD's Programme for International Student Assessment [PISA] 2001), the average Indonesian student performed at the frameworks raise standards and facilitate interna- level of a French student at the seventh percentile (World Bank tional recognition. 2006). More than 20 percent of firms in many developing coun- Focus on learning to learn, equipping workers tries rate inadequate skills and education of workers as a major or to learn throughout their working lives, and severe obstacle to their operations (World Bank 2005). 129 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 and given for too long, can slow down periods of unemployment or more frequent adjustment. In developing countries, in- job change argues for new mechanisms to en- formality makes targeting of benefits sure that they are not left without access to es- difficult as the unemployed may also sential health services. Moreover, health care have jobs in the informal sector, and the benefits provided by firms are a burden on registered unemployed may be middle- globally contestable jobs and make employers rather than low-income workers (Hoek- wary of forming long-term relationships with man and Winters 2005).29 However, prospective employees (Leamer 2006). In de- unemployment insurance may also be an veloping countries, extending universal basic alternative source of credit for self- medical care not linked to other aspects of for- employment. Individual savings ac- mality could help to reach the poorest, but counts or similar types of unemploy- risks increasing incentives for informality ment insurance may be a better solution (Arias and others 2005). One option is general for developing countries, although there health provision, funded by tax revenue rather is a risk that workers have insufficient than attached to employment; a more modest resources. alternative is the creation of a health insurance · Mandatory severance pay is the most subsidy for displaced workers, as proposed by common income support program in Kletzer (2001).31 developing countries, as compliance is Active labor-market programs can be effec- complaint-driven and an expensive tive in keeping workers in the labor market and bureaucracy is not required. However, if upgrading their skills, but experience has been overly large, severance pay may discour- mixed and programs need to be designed to age hiring or reforms as costs become suit the conditions in developing countries.32 unmanageable. While there is considerable experience in the · One-off compensation programs have OECD countries,33 less is known about poli- also been used in both developing (public cies in developing countries.34 For the latter, sector downsizing) and developed (re- the key issues are the size of the informal sec- structuring of declining industries) coun- tor, limited administrative capacity, and the ab- tries. While supporting relatively well-off sence of broader social safety nets. Leakage workers (the previous beneficiaries of risks are higher, as the unemployed may also rents), they are often seen as politically have informal jobs and may not be low- necessary for reform--although experi- income. Policies need to improve conditions in ence suggests that they have often not suc- the informal sector, but avoid creating addi- ceeded in attaining their stated goals.30 tional incentives for informality. · Wage insurance gives workers a propor- In both developed and developing countries tion of their former wage for a set period successful interventions are comprehensive, of time, conditional upon their finding oriented to labor demand, linked to real work- new employment. This eases adjustment, places, and carefully targeted (box 4.9). Inter- provides an incentive to take a new, ventions are also more effective when the albeit lower-paying, job and (in effect) economy is growing. But longer-term assess- subsidizes on-the-job retraining--the ments are needed (most cover one to two most effective kind (Kletzer 2001). years) and a range of effects--deadweight (im- pact would have been achieved in the absence Global competition and movement of of the program), substitution (participants workers argues for separating health care from substitute for nonparticipants in the labor employment status. The possibility that in fu- market and the employment effect is zero), ture more types of workers could experience and displacement (firms with subsidized 130 N E W P R E S S U R E S I N L A B O R M A R K E T S Box 4.9 Overview of the impact of active labor-market programs Employment services. Generally have a positive Training for youth. Less successful than earlier in- impact on employment and earnings and are cost- vestments in the education system. Some success in effective. But they are of limited use where struc- Latin America with programs that integrate training tural unemployment is high and labor demand low. with remedial education, job search assistance, and May be less effective in developing countries social services. For example, the "Jovenes" programs where informality is high and implementation of Argentina, Chile, Peru and Uruguay are targeted capacity limited. at disadvantaged youth. They combine training and Public works. Effective, including for informal work experience with other services, include the pri- workers, as a short-term safety net in developing vate sector, and are financed by tripartite levy-grant countries but do not improve future labor-market schemes or the government. They have substantial prospects, especially where a stigma is attached to and positive impacts on employment and earnings, participation. Wages need to be sufficiently low to but can be small scale, may not be cost-effective, and target those with low incomes and few job participants can displace other workers. prospects, and projects should also target poor Employment subsidies. Mostly for disadvantaged areas. Can be most redistributive, but require groups, although some countries (Belgium, France, the government expenditure. Netherlands) provide for all low-paid work. A signifi- Training. Can result in higher employment rates, cant share of overall active labor-market program if not earnings. Programs work best with active em- (ALMP) spending in several OECD countries. Most ployer involvement (on-the-job training), and lim- do not have a positive impact and have substantial ited evidence from developing and transition deadweight (workers would have been employed economies suggests better results for women than without the subsidy) and substitution (worker dis- men. But firms are reluctant to train lower-skilled places a nonsubsidized worker) costs. workers; in the OECD countries the least qualified Microenterprise development and self-employment are only a quarter to a third as likely as the highly assistance. Some evidence of positive impacts for qualified to participate in job-related training. A older and better-educated workers, but take-up is growing number of countries fund enterprise-based low and business failure rate is high. training via compulsory levies (usually 1 percent of payroll), with reimbursement based on training pro- vided in some cases (examples include Singapore Sources: Betcherman, Olivas, and Dar 2004; World Bank 2005; and Mauritius). Rama 2003; Arias and others 2005; OECD 2005a, 2005d; Heckman and Pagés 2000. Retraining after mass layoffs. No positive impact, except with a comprehensive package of employment Note: Betcherman, Olivas, and Dar (2004) build on an earlier World Bank study of 72 scientific (that is, using a control services, and expensive. Workers are often geograph- group) evaluations of ALMPs by Dar and Tzannatos (1999) by ically concentrated with industry-specific skills. adding 87 new studies, 39 of which cover programs in develop- Best results have been achieved with longer programs ing and transition economies. Similar conclusions on a number that include some worker contribution to costs. of points have been made in OECD reviews (see Martin 2000). workers displace those without)--need to be Globalization may undermine funding taken into account. Given their mixed record, for programs to support labor and the challenges of appropriate design, gov- While integrating into the world economy ernments need to be realistic about what ac- requires that import taxes be kept low and tive labor-market policies can achieve. relatively uniform, for the least developed 131 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 While feelings of greater economic insecu- Figure 4.6 Low-income countries depend rity among workers may lead to greater de- heavily on import duties for tax revenues mands for social insurance, it has been argued Percent of GDP that globalization is limiting the capacity of 7 VAT, sales, or governments to fund such protection--and to 6 turnover taxes support the productivity-enhancing measures Import 5 duties discussed above. The fear is that globalization 4 and greater mobility of capital and wealthy 3 workers will undermine the tax base, as these 2 factors move to the lowest tax locations, com- 1 promising social welfare programs for those 0 bearing the burden of adjustment. There is little evidence of this process in developed countries (OECD 2000b), however, and the preferences income per-middle- income Low-income High-income Lower-middle- Up of wealthy individuals for education, health, Source: World Bank staff calculations. law and order, and social welfare suggest that fears of large numbers of skilled workers emi- grating to low tax havens are unlikely to be jus- countries they are a key source of revenue rel- tified. Moreover, much of the knowledge of ative to value-added tax (VAT) and sales taxes these workers is gained and creates value from (figure 4.6). High-income countries are able to interactions and synergies with clusters of other recover revenues lost from trade liberalization similar workers. As Leamer (2006) says, it from other sources: on average, middle- matters who your neighbors are. income countries recover 45­60 percent of lost tariff revenues while least developed countries The international community--working recover less than 30 percent (Baunsgaard and together--can help realize the potential Keen 2005). of globalization But many countries collect far less in tariff The rise of China, India, and other emerging revenue than the applied tax rates would sug- economies amounts to a huge increase in the gest, owing to the widespread (discretionary) supply of unskilled labor on a global scale. granting of exemptions. Exemptions make the This could heighten the existing--and grow- tax regime opaque and difficult to administer ing--inequality between skilled and unskilled and can lead to a distorted incentive structure workers in both developed and developing that discriminates against small firms with less countries that has resulted from a mix of tech- influence. Further, there is little evidence that nology and trade effects. There are fears that exemptions have a significant impact on in- there will be no space for other countries, par- vestment, their primary justification. Many ticularly developing countries, to compete in countries could substantially reduce applied low-wage exports or in attracting global capi- tariffs while maintaining or even increasing tal flows. In addition, because China and revenue if exemptions were removed and col- India are rapidly upgrading the skills of their lection improved. However, it is still necessary workforce, and services activities along the to address the development challenge high- value chain are being globally sourced, skilled lighted in figure 4.6 of moving from easy- workers everywhere are increasingly facing to-collect trade taxes to harder-to-collect competitive pressure. There are mitigating consumption and income taxes.35 Simply forces, however; China and India offer huge implementing a VAT is not sufficient; a high markets for the exports of other countries, degree of collection efficiency (the ratio of their own wages are bound to grow rapidly, actual to potential revenues) is needed.36 and developing the full range of institutions 132 N E W P R E S S U R E S I N L A B O R M A R K E T S needed to underpin a modern, dynamic mar- simultaneously and most were in manufacturing, al- ket economy will take time. though there was a significant increase in shifts of This new global climate poses challenges-- white-collar-services jobs to India. 3. Ghose (2003) identifies a group of manufacturing- but also offers considerable opportunities--for exporting developing countries that have shown im- all countries. The countries best placed to ad- pressive growth in employment in the sector, from dress the challenges will be those best able to about 50.9 million in 1980 to 82.8 million in 1997 (or seize the opportunities and generate the new from 79 to 88.7 percent of total employment over the sources of growth and wealth needed to fi- same period). In this group are China, the Arab nance additional investments in social safety Republic of Egypt, India, Indonesia, Israel, the nets and education. In many developing coun- Republic of Korea, Malaysia, Mauritius, Morocco, the Philippines, Singapore, Sri Lanka, Taiwan (China), tries, domestic reforms to reduce rigidities in Thailand, and Turkey. However, other manufacturing- the labor market and improve the climate for exporting countries have witnessed declines of 13.5 to business and innovation will be critical. All 10.6 million (21 to 11.3 percent) over the same period-- countries will need better mechanisms to cush- this group consists of Argentina, Brazil, Hong Kong ion adjustment costs and distribute the benefits (China), Malta, Mexico, Pakistan, and South Africa. of growth to offset the tendencies toward in- Note that Mexico's figures do not include the equality and volatility. But collective action by maquiladora; if they did, the country would have appeared in the first group. the international community will also be 4. A further question is the distribution of gains be- needed in two important respects. First, as the tween labor and capital. However, this issue has pressures of globalization increase the calls for proven difficult to analyze and beyond the immediate beggar-thy-neighbor protectionism, the inter- scope of this chapter, which focuses on the distribution national community will need to band together within labor markets between skilled and unskilled to preserve and extend the open markets that workers. have underpinned recent advances in growth 5. Whether the impacts of greater openness operate more or less through wages as opposed to employment and poverty reduction. In the absence of open depends on labor-market institutions, the efficiency of global markets, many of the new opportunities capital markets, and social policies. Hence in the United from the coming globalization will disappear. States, the more flexible labor market and more effi- Second, the international community needs to cient financial sector mean that wages bear a greater provide the financial and technical support-- share of shocks than in the European Union. In devel- the "aid for trade"--to enable the poorest oping countries, it also appears that wage responses are countries to overcome the infrastructure and greater than employment impacts, suggestive both of labor-market rigidities and industry rents engendered capacity constraints that limit their ability to by trade policy (Hoekman and Winters 2005). take advantage of new trade opportunities. 6. For full-time workers in the United States be- tween 1979 and 1995, real wages of those with 12 years of education fell by 13.4 percent, and real wages of those Notes with less than 12 years of education fell by 20.2 percent. 1. Trade among developing countries is also grow- During the same period, real wages of workers with 16 ing significantly; South-South trade now constitutes or more years of education rose by 3.4 percent so that one-quarter of developing-country exports, and this the wage gap between these groups grew significantly trade is growing 50 percent faster than world trade. (Feenstra and Hanson 2003). 2. A recent U.S. study (Brofenbrenner and Luce 7. Some of the overall increases in the skill pre- 2004) concludes that the Bureau of Labor Statistics mium could also be related to the artificially low prices (BLS) grossly underestimates the total number of jobs for skilled labor prior to opening. Modest increases are lost to global production shifts. While the BLS re- also found in China and Vietnam (World Bank 2002). ported 4,633 private sector workers in establishments 8. Economic literature has mainly focused on the with 50 or more employees who lost their jobs be- impact of globalization on labor demand elasticities. Au- cause of global outsourcing from January to March thors such as Rodrik (1997) argue that globalization 2004, the authors, drawing on media reports, find has led to an increase in the labor demand elasticity, with evidence of a minimum of 25,000 jobs lost over that the result that changes in product prices now have period. Moves were often to several destinations magnified impacts on wages and employment. The 133 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 empirical support for this link is mixed, although esti- that engineering graduates from Chinese universities mating elasticities is prone to difficulties. Rodrik (1997) numbered only 351,000 per year as opposed to previ- finds that the interaction between trade openness and ous estimates of over 600,000. Note that the new num- variation in a country's terms of trade is positively linked ber is still two and a half times as many graduates as in with volatility of growth and government expenditures. the United States; however, China's population is four The latter, Rodrik argues, reflects the increasing demand times as large. for social protection as globalization increases insecu- 15. It is interesting to note that the Soviet Union rity. On the other hand, Iversen and Cusack (2000) overtook the United States in the number of research argue that what is required is to show that volatility workers (Nolting and Feshbach 1980) but did not suc- from international markets is greater than that in do- ceed in achieving strong and sustained innovation- mestic markets; they find that in developed countries driven growth. there is no correlation between trade openness and out- 16. Migration of skilled workers can be positive for put, earnings, or employment volatility. Others have the country of origin when those workers come back. In tried to link trade liberalization to changes in labor de- Morocco, high-skilled labor has started to return, mand elasticities. Slaughter (2001) finds that labour de- bringing substantial know-how and technological mand elasticities for low-skilled workers in the United knowledge into the country, increasing productivity and States have increased over time but with no clear link to boosting innovation in terms of improved business trade variables. Fajnzylber and Maloney (2005), for a practices. set of Latin American countries, and Krishna, Mitra, 17. The value of a relationship-specific investment and Chinoy (2001) for Turkey do not find strong sup- is significantly higher within a buyer-seller relationship port for this link. On the other hand, Hasan, Mitra, and than outside it. An example is where suppliers or sub- Ramaswamy (2003) find a strong link between trade re- contractors to a car producer make investments in de- form and wage and employment volatility in India. sign modifications that improve the fit or ease of as- Nevertheless, even though a positive link between glob- sembly with other parts but which are not relevant to alization and observed measures of volatility has not the production process of other car makers. Such been found, globalization may still have contributed to investments tend to be associated with longer-term greater risks and to heightened economic insecurity contractual commitments between producers and (Scheve and Slaughter 2002). their suppliers and less repeated bargaining (Joskow 9. Data for 19 developed and developing (1987)). Spencer and Qui (2001) find that such rela- economies suggest that flexible hiring and firing rules tionships in the Japanese car industry tend to limit the are positively associated with higher rates of entry of range of imports to less important parts and that it is new firms, which are often better at harnessing new possible that no parts are imported despite lower technologies (World Bank 2005). production costs overseas. 10. Alternatively, of course, it could be argued that 18. Looking at U.S. imports from 10 major devel- strict employment protection is a response to the oping countries (which together accounted for 26.5 higher level of anxiety of workers in these countries. If percent of U.S. imports at the time of the cited study), that is the case, however, one would also have to Aggarwal (1995) noted that sectors with egregious conclude that protection has not been very effective in labor conditions were not a primary share of these reducing that anxiety. countries' exports; that standards were often lower in 11. Skilled workers are considered to be those with less export-oriented or nontraded sectors; and that, some secondary education, plus those with secondary within the export-oriented sector, labor conditions in education or above. This selection does not take into firms more involved in exporting were either similar to account the quality of the education received or or better than those in other firms. Raynauld and Vidal comparability among countries. (1998) showed that, since 1980, countries with low 12. This is notwithstanding the fact that trade lib- standards had not increased their share of global ex- eralization in India and especially China has advanced ports and two-thirds of 39 countries with low labor greatly over the last 15 years, in China driven in part standards had seen their international competitiveness by World Trade Organization (WTO) accession and in (as measured by unit labor costs) stagnate or decline India by unilateral reforms. (decline reflects either a decline in labor productivity 13. It should be noted that Hong Kong (China) is relative to the nominal cost of labor or a rise in the the world's third-largest outward investor, with flows nominal cost of labor relative to its productivity) while of about $40 billion in 2004 (UNCTAD 2005). 14 of the 18 high-standards countries had increased 14. There is controversy around the exact number their international competitiveness. of graduates from Chinese and Indian institutions. For 19. "Outsourcing" or "offshoring" have example, a recent study by Duke University indicated both been used to refer to the global sourcing of 134 N E W P R E S S U R E S I N L A B O R M A R K E T S services--technically "outsourcing" refers to the sourc- 27. However, IPR protection must be balanced ing of an activity outside a company (such as the con- against the need to avoid stifling competition. tracting out of billing services), which can also take place 28. Education can also promote access to technol- within the domestic market, while "offshoring" is the ogy and development of new sectors: in 21 countries in movement of production of a service outside a country. Latin America and the Caribbean, an increase of five For firms, offshoring need not be outsourcing when the years in the average level of education in those above activity stays within a foreign affiliate; from the perspec- 15 was associated with an increase in FDI of 3 percent tive of national labor markets, it is the movement of of GDP (de Ferranti and others 2002). production to another territory that is the focus of inter- 29. While the informal sector can also be a way of est. Strictly speaking, not all FDI is offshoring, as in cases managing risk, there are limits to its role as a safety where a foreign affiliate is built to service the local net, as it often generates most of the flows into unem- market in the host country (Kirkegaard 2005). ployment (about 60 percent in Argentina, Brazil, and 20. Note that balance of payments (BOP) statistics Mexico) and is ineffective in cases of multiple/covariate imperfectly measure the full extent of global sourcing shocks (Arias and others 2005). of services because of classification and data limita- 30. In the United States, the structure of the politi- tions; figures should be taken as an underestimate. cal system (including, for example, passage of Trade 21. Atkinson (2004) notes that a concerted effort Promotion Authority in Congress) and nature of pres- by the Chinese government to expand acquisition of sures have led to the development of trade-specific English could see China moving beyond non-language- adjustment measures (Brenton 2006). Under the U.S. based services to other business services. Trade Adjustment Assistance (TAA) program, qualified 22. Care needs to be taken with comparisons be- workers can receive an additional 52 weeks of unem- tween outsourcing and inward FDI, as the foreign es- ployment insurance provided they are enrolled in an ap- tablishment may be created primarily to serve the do- proved training program; a similar program was created mestic market (see note 19). That said, the comparison for the North American Free Trade Agreement may be more relevant in terms of the contribution of (NAFTA) in 1993. TAA and NAFTA payments are foreign companies to job creation within the United about $300 million annually (Kletzer 2001). The Euro- States, to offset the movement of jobs overseas. pean Union is also now considering a Globalization 23. For every $1 of call-center work offshored by Fund of half a billion euros to help retrain and relocate U.S. firms, an estimated $1.43 is reinvested in the U.S. 35,000­50,000 workers a year whose jobs are lost to economy; the amounts are $1.33 and $1.42 for infor- global sourcing and trade. Money would be available in mation technology services and high-end knowledge the case of layoffs of at least 1,000 people in regions services (such as equity research, tax preparation, and with a population of at least 800,000 where the unem- risk management), respectively. The net benefit to the ployment rate was already higher than the European or U.S. economy of shifting $1 previously spent in the national average; or where several companies in a sector United States to India could be as high as 12­14 cents laid off at least 1,000 workers over six months in regions per dollar (McKinsey Quarterly, October 2003). with a population of up to three million and where the 24. The absence of international standards for job losses added up to 1 percent of total employees in many service sector activities reinforces the importance that sector. While the fund will cover a relatively small of reputation in attracting new clients. number of workers, it is seen as important in resisting 25. Nicolás Eyzaguirre, former Minister of Finance growing calls for protection (Kanter 2006). It is not from Chile, comparing the experiences of Chile and clear, however, that there is an equity argument for dis- Argentina, in a presentation to the Mauritius High tinguishing between trade-affected and other workers, Level seminar in September 2006. such as those affected by technological change. In some 26. Rates of return can also be influenced by the circumstances--such as mass layoffs--trade-specific as- lack of competition. If incumbents are able to extract sistance may be more cost-effective. However, it should large rents that are not endangered, the incentive to be used sparingly, aimed at orderly adjustment, be time- innovate is severely restrained because the returns limited, and include both services and manufacturing will replace some of the rents they are actually workers as well as workers who have lost their jobs from collecting, reducing the net value of innovation. Evi- both import and export competition (OECD 2005d). dence indicates that monopolies are not particularly 31. Under Kletzer's proposal for the United States, innovative and that small firms stimulate innovation all full-time displaced workers would be eligible to re- (Geroski 1990). Opening markets by reducing exter- ceive a health insurance subsidy for up to six months, nal barriers and creating better regulatory frame- or until they found a new job, whichever is earlier. works for natural monopolies is hence likely to raise 32. ALMPs include employment services, training, rates of innovation. public works (which offer short-term employment on 135 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 community projects in sectors such as construction, Anderton, Robert, and Paul Brenton. 1999. "Out- rural development, and community services), wage and sourcing and Low-Skilled Workers in the UK." employment subsidies, and self-employment assistance. Bulletin of Economic Research 51: 267­86. 33. Over 1990­2002, average national expendi- Anderton, Robert, Paul Brenton, and John Whalley. ture on ALMPs in OECD countries remained relatively 2006. Globalisation and the Labour Market: constant at about 0.75 percent of gross domestic prod- Trade, Technology and Less-Skilled Workers in uct (GDP). This average masks wide differences, how- Europe and the United States. London and New ever, with some European countries spending over York: Routledge. 1 percent of GDP while the United States, Japan, Arias, Omar, Andreas Blom, Mariano Bosch, Wendy Korea, and the United Kingdom spent under 0.4 per- Cunningham, Ariel Fiszbein, Gladys Lopez cent. Training accounted for the bulk of spending Acevedo, William Maloney, Jaime Saavedra, (36 percent), followed by public employment services Carolina Sanchez-Paramo, Mauricio Santamaria, (24.5 percent) and job subsidies (19.5 percent). Transi- and Lucas Siga. 2005. "Pending Issues in Protec- tion economies show similar (but lower) patterns of tion, Productivity Growth, and Poverty Reduc- spending (Betcherman, Olivas, and Dar 2004). tion." World Bank Policy Research Working 34. There is some evidence that Latin American Paper No. 3799, Washington, DC. December. countries have been investing significantly in youth Atkinson, Robert. 2004. "Understanding the Offshoring training and public works programs, but in Africa Challenge." Progressive Policy Institute Policy there is very little active programming on any signifi- Report, May, available at www.ppionline.org. cant scale (Betcherman, Olivas, and Dar 2004). Bates, Jenny. 2000. "International Trade and Labor 35. For many resource-rich developing countries Standards." Progressive Policy Institute (PPI) facing rising demand from China and India and Policy Report, available at www.ppionline.org. higher world prices, a crucial opportunity that has to Baunsgaard, Thomas, and Michael Keen. 2005. "Tax be addressed is translating higher revenues into in- Revenue and (or?) Trade Liberalization." Interna- vestments in social and educational programs that en- tional Monetary Fund (IMF) Working Paper hance competitiveness and support diversification. For No. 05/112, Washington, DC. these countries, the priority for domestic reform must Bergin, Paul, Robert Feenstra, and Gordon Hanson. be to address governance and corruption issues asso- 2006. "Outsourcing and Volatility." University of ciated with distributing the revenues from resources. California. This must be supported by the global community Betcherman, Gordon, Karina Olivas, and Amit Dar. through increased efforts to discipline the activities of 2004. "Impacts of Active Labor Market Pro- firms and governments in countries demanding these grams: New Evidence from Evaluations with resources. Particular Attention to Developing and Transition 36. Countries with smaller agricultural and larger Economies." Social Protection Discussion Paper urban sectors, with strong political institutions, with Series, World Bank, Washington, DC. January. higher per capita incomes, and that are more open to Bhagwati, Jagdish, Arvind Panagariya, and T. N. trade tend to have higher collection efficiencies for Srinivasan. 2004. "The Muddles over Outsourc- VAT (Aizenman and Jinjarak 2006). ing." Journal of Economic Perspectives 18: 93­114. Bhorat, Haroon, and Paul Lundall. 2004. Employ- ment and Labour Market Effects of Globaliza- References tion: Selected Issues for Policy Management. Aggarwal, M. 1995. "International Trade, Labor Stan- Employment Strategy Papers No. 2004/3, ILO, dards and Labor Market Conditions: An Evalua- Geneva. tion of the Linkages." U.S. International Trade Blinder, Alan. 2006. "Offshoring: The Next Industrial Commission, Office of Economics Working Paper Revolution?" Foreign Affairs 85(2): 113­28. No. 95-06-C, Washington, DC. Borjas, George. 2003. "The Labor Demand Curve Is Aizenman, Joshua, and Yothin Jinjarak. 2006. "Glob- Downward Sloping: Reexamining the Impact of alization and Developing Countries--A Shrinking Immigration on the Labor Market." Quarterly Tax Base?" NBER Working Paper No. 11933, Journal of Economics 118(4): 1335­74. National Bureau of Economic Research, Cam- Borjas, George, Richard Freeman, and Lawrence Katz. bridge, MA. 1997. "How Much Do Immigration and Trade Af- Amiti, Mary, and Shang-Jin Wei. 2004. "Fear of Out- fect Labor Market Outcomes?" Brookings Papers sourcing: Is It Justified?" NBER Working Paper No. on Economic Activity 1: 1­90, Washington, DC. 10808, National Bureau of Economic Research, Bravo-Ortega, Claudio, and Daniel Lederman. 2006. Cambridge, MA. "The Effect of Chinese and Indian Patenting 136 N E W P R E S S U R E S I N L A B O R M A R K E T S Activity on Latin America's Patent Counts: Esti- de Ferranti, David, Guillermo Perry, Indermit Gill, and mating the Effects of Cumulative Learning and Luis Servén. 2000. Securing Our Future in a Contemporaneous Patent Tournaments." Back- Global Economy. Washington, DC: Latin Ameri- ground paper for the Office of the Chief Econo- can and Caribbean Studies, World Bank. mist for Latin America and the Caribbean Re- de Ferranti, David, Guillermo Perry, Daniel Lederman, gional Study: "Latin America and the Caribbean and William Maloney. 2002. From Natural Re- Respond to the Growth of China and India." sources to the Knowledge Economy: Trade and World Bank, Washington, DC. Job Quality. Washington, DC: Latin American Brenton, Paul. 2006. "Adjusting to Globalisation: Pol- and Caribbean Studies, World Bank. icy Responses in Europe and the US." In Global- de Mello, Luiz, and Mombert Hoppe. 2005. "Educa- isation and the Labour Market: Trade, Technol- tion Attainment in Brazil: The Experience of ogy and Less-Skilled Workers in Europe and the FUNDEF." OECD Economics Department Work- United States, ed. Robert Anderton, Paul Bren- ing Papers No. 424, OECD, Paris. ton, and John Whalley. London and New York: Dehejia, Vivuk, and Yiagadeesen Samy. 2002. "Trade Routledge. and Labour Standards--Theory, New Empirical Brofenbrenner, Kate, and Stephanie Luce. 2004. "The Evidence and Policy Implications." CESifo Work- Changing Nature of Corporate Global Restruc- ing Paper No. 830, December. turing: The Impact of Production Shifts on Jobs in Edwards, L. 1999. "Trade Liberalization, Structural the US, China and Around the Globe." Washing- Change and Occupational Employment in South ton, DC: US-China Economic and Security Re- Africa." Paper presented at the Trade and Indus- view Commission. trial Policy Secretariat (TIPS) Annual Forum, Brown, D., A. Deardorff, and R. Stern. 2002. "The Effects Muldersdrift, September 19­22. of Multinational Production on Wages and Working Eifert, Ben, Alan Gelb, and Vijaya Ramachandran. Conditions in Developing Countries." In Challenges 2005. "The Business Environment and Compara- to Globalization, ed. Robert Baldwin and L. Alan tive Advantage in Africa: Evidence from Invest- Winters. Chicago: University of Chicago Press. ment Climate Data." Paper presented at the Card, David. 2005. "Is the New Immigration Really So 17th Annual World Bank Conference on Develop- Bad?" NBER Working Paper No. 11547, National ment Economics (ABCDE), January 27, Dakar, Bureau of Economic Research, Cambridge, MA. Senegal. Cline, William. 2006. "Exports of Manufactures and Fajnzylber, Pablo, and William F. Maloney. 2005. Economic Growth: The Fallacy of Composition "Labor Demand and Trade Reform in Latin Revisited." Institute for International Economics, America." Journal of International Economics Washington, DC. 66: 423­46. Cortes, Patricia. 2005. "The Effect of Low-skilled Feenstra, Robert, and Gordon Hanson. 2003. "Global Immigration on US Prices: Evidence from CPI Production Sharing and Rising Inequality: A Data." Massachusetts Institute of Technology, Survey of Trade and Wages." In Handbook of Cambridge, MA. International Trade, ed. E. K. Choi and J. Harrigan. Cravino, Javier, Daniel Lederman, and Larcelo Blackwell Publishing. Olarreaga. 2006. "Substitution between Foreign Forrester Research. 2002. "3.3 Million US Services Capital in China, India and the Rest of the World: Jobs to Go Offshore." Cambridge, MA. Much Ado about Nothing?" Background paper for Fox, L., G. Betcherman, V. Chandra, B. Eifert, and A. the Office of the Chief Economist for Latin America Van Adams. 2004. "Realizing the Potential of the and the Caribbean Regional Study, "Latin America Labor Force in Africa: Barriers and Opportuni- and the Caribbean Respond to the Growth of China ties." World Bank, Washington, DC. and India," World Bank, Washington, DC. Freeman, Richard. 2005. "The Great Doubling: Labor Criscuolo, Chiara, Jonathan Haskel, and Matthew in the New Global Economy." Harvard Univer- Slaughter. 2004. "Why Are Some Firms More sity, Cambridge, MA. Innovative? Knowledge Inputs, Knowledge ______. 2006. "Labor Market Imbalance: Shortages, Stocks and the Role of Global Engagement." or Surpluses, or Fish Stories?" Paper presented University College London. at Boston Federal Reserve Economic Confer- Dar, Amit, and P. Zafiris Tzannatos. 1999. "Active ence, "Global Imbalances--As Giants Evolve," Labor Market Programs: A Review of the Chatham, MA. Evidence from Evaluations." Social Protection Friedman, Thomas L. 2005. The World Is Flat: A Brief Discussion Paper No. 9901, World Bank, History of the Twenty-First Century. New York: Washington, DC. January. Farrar, Straus and Giroux. 137 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Geroski, P. A. 1990. "Innovation, Technological Joskow, P. L. 1987. "Contract Duration and Opportunity, and Market Structure." Oxford Relationship-Specific Investments: Empirical Economic Papers 42(3): 586­602. Evidence from Coal Markets." American Eco- Ghose, Ajit K. 2003. Jobs and Incomes in a Globaliz- nomic Review 77: 168­85. ing World. Geneva: International Labour Office. Kanter, James. 2006. "EU Fund to Ease Globalization Grossman, G., and E. Rossi-Hansberg. 2006. "The Rise Pain." Reprinted from The International Herald of Offshoring: It's Not Wine for Cloth Anymore." Tribune, March 2, 2006, available at Yale Global Princeton University, Princeton, New Jersey. Online (http://yaleglobal.edu). Hall, Robert, and Charles Jones. 1999. "Why Do Some Katz, Larry F., and David H. Autor. 1999. "Changes in Countries Produce So Much More Output per the Wage Structure and Earnings Inequality." In Worker Than Others?" Quarterly Journal of Handbook of Labor Economics, Vol. 3A, ed. Economics 114: 83­116. Orley C Ashenfelter and David Card. Amsterdam: Hasan, Rana, Devashish Mitra, and K. V. Ramaswamy. Elsevier. 2003. "Trade Reforms, Labor Regulations, and Kirkegaard, Jacob Funk. 2005. "Outsourcing and Labor-Demand Elasticities: Empirical Evidence Offshoring: Pushing the European Model Over from India." East-West Center Working Papers the Hill, Rather than Off the Cliff!" IIE Working No. 59. Honolulu, Hawaii. Paper Series No. WP05-1, Institute for Interna- Heckman, James, and Carmen Pagés. 2000. "The Cost tional Economics, Washington, DC. of Job Security Regulation: Evidence from Latin Kletzer, Lori. 2001. "A Prescription to Relieve Worker American Labor Markets." NBER Working Paper Anxiety." Institute for International Economics No. 7773, National Bureau of Economic Re- Policy Brief No. 01-2, Institute for International search, Cambridge, MA. Economics, Washington, DC. Helpman, Elhanan. 2004. The Mystery of Economic Krishna, Pravin, Devashish Mitra, and Sajjid Chinoy. Growth. Cambridge, MA: Harvard University 2001. `Trade Liberalization and Labor Demand Press. Elasticities: Evidence from Turkey." Journal of Hoekman, Bernard, and Beata S. Javorcik. 2006. International Economics 55: 391­409. "Lessons from Empirical Research on Interna- Krugman, Paul. 1995. "Technology, Trade and Factor tional Technology Diffusion through Trade Prices." NBER Working Paper No. 5355, and Foreign Direct Investment." In Global Inte- National Bureau of Economic Research, Cam- gration and Technology Transfer, ed. Hoek- bridge, MA. man and Javorcik. Washington, DC: World Bank. Kuhn, P. J., ed. 2002. "Losing Work, Moving On." W. Hoekman, Bernard, and L. Alan Winters. 2005. E. Upjohn Institute for Employment Research, "Trade and Employment: Stylized Facts and Kalamazoo, MI. Research Findings." Mimeo. Lardy, Nicholas. 2004. "Do China's Abusive Labor ILO (International Labour Office). 1998. Labor and Practices Encourage Outsourcing and Drive Social Issues Relating to Export Processing Down American Wages?" Senate Democratic Zones. Geneva: ILO. Policy Committee Hearing, March 29. Available ______. 2004. A Fair Globalization: Creating Oppor- at www.iie.org. tunities for All: Report of the World Commission Leamer, E. 2006. "A Flat World, A Level Playing Field, on the Social Dimension of Globalization. a Small World After All, or None of the Above?" Geneva: ILO. Forthcoming in Journal of Economic Literature. ______. 2006. Competing for Global Talent. Geneva: Lederman, Daniel, Marcelo Olarreaga, and Isidro ILO. Soloaga. 2006. "The Growth of China and India in Iversen, Torben, and Thomas Cusack. 2000. "The World Markets: Opportunity or Threat for Latin Causes of Welfare State Expansion." World American Exporters?" Background paper for the Politics 52: 313­49. Office of the Chief Economist for Latin America Jaumotte, Florence, and Nigel Pain. 2005. "An and the Caribbean Regional Study, "Latin Amer- Overview of Public Policies to Support Innova- ica and the Caribbean Respond to the Growth of tion." OECD Economics Department Working China and India." World Bank, Washington, DC. Papers No. 456, OECD, Paris. Majid, Normaan. 2004. "What Is the Effect of Trade Jensen, J. Bradford, and Lori Kletzer. 2005. "Tradable Openness on Wages?" Employment Strategy Services: Understanding the Scope and Impact of Papers No. 2004/18, Employment Analysis Unit, Services Outsourcing." IIE Working Paper Series ILO, Geneva. No. WP05-9, Institute for International Econom- Martin, J. 2000. "What Works among Active Labour ics, Washington, DC. Market Policies: Evidence from OECD Countries' 138 N E W P R E S S U R E S I N L A B O R M A R K E T S Experiences." OECD Economic Studies, No. 30. Economics Working Papers tepica-193, Univer- OECD, Paris. sity of Toronto, Ontario. Moran, Theodore. 2004. "Trade Agreements and Rama, Martin. 2003. Globalization and Workers in Labor Standards." The Brookings Institution Pol- Developing Countries. World Bank Policy Re- icy Brief No. 133, available at www.brookings. search Working Paper No. 2958, Washington, edu. DC. January. Nolting, Louvan, and Murray Feshbach. 1980. "R and Raynauld, André, and Jean-Pierre Vidal. 1998. Labor D Employment in the USSR." Science 207 (Feb. 1): Standards and International Competitiveness: A 493­503. Comparative Analysis of Developing and Indus- Nunn, Nathan. 2005. "Relationship Specificity, In- trialized Countries. Cheltenham, U.K. and complete Contracts and the Pattern of Trade." Northampton, MA: Edward Elgar. University of British Columbia. Forthcoming in Reinecke, G., and R. Torres. 2001. Studies in the Social Quarterly Journal of Economics. Dimension of Globalization: Chile. Geneva: In- OECD. 1996. Trade, Employment and Labour Stan- ternational Labour Office. dards: A Study of Core Workers' Rights and In- Robbins, Donald. 1997. "Trade and Wages in Colom- ternational Trade. Paris: OECD. bia." Estudios de Economia 24(1): 47­83. ______. 2000a. International Trade and Core Labour Rodrik, Dani. 1996. "Labor Standards in International Standards. Paris: OECD. Trade: Do They Matter and What Do We Do ______. 2000b. Policy Competition and Foreign Direct about Them?" In Emerging Agenda for Global Investment: A Study of Competition among Gov- Trade: High Stakes for Developing Countries, ernments to Attract FDI. Paris: OECD Develop- Overseas Development Council Essay No. 20, ed. ment Centre Studies. R. Z. Lawrence, D. Rodrik, and J. Whalley. ______. 2004. OECD Employment Outlook. Paris: Washington, DC: John Hopkins University Press. OECD. ______. 1997. Has Globalization Gone Too Far? ______. 2005a. "The Impact of Structural Policies on Washington, DC: Institute for International Trade-Related Adjustment and the Shift to Ser- Economics. vices." Economics Department Working Papers Sapir, André. 2005. "Globalization and the Reform of No. 427, OECD, Paris. European Social Models." Background document ______. 2005b. Trade and Structural Adjustment: Em- for the presentation at ECOFIN Informal Meet- bracing Globalization. Paris: OECD. ing in Manchester, U.K., September 9. ______. 2005c. "The Share of Employment Potentially Scheve, Kenneth, and Matthew Slaughter. 2002. Affected by Offshoring--An Empirical Investiga- "Economic Insecurity and the Globalization of tion." Working Party on the Information Econ- Production." NBER Working Paper No. 9339, omy, DSTI/ICCP/IE(2005)8/FINAL. National Bureau of Economic Research, ______. 2005d. OECD Employment Outlook. Paris: Cambridge, MA. OECD. Slaughter, Matthew J. 2001. "International Trade and Ottaviano, Gianmarco, and Giovanni Peri. 2005. "Re- Labor-Demand Elasticities." Journal of Interna- thinking the Gains from Immigration: Theory and tional Economics 54: 27­56. Evidence from the US." NBER Working Paper Spencer, Barbara, and Larry Qui. 2001. "Keiretsu and No. 11672, National Bureau of Economic Relationship-Specific Investment: A Barrier to Research, Cambridge, MA. Trade?" International Economic Review 42: Oxfam. 2004. "Stitched Up: How Rich-Country 871­901. Protectionism in Textiles and Clothing Stern, Robert M. 2003. "Labour Standards and Trade Trade Prevents Poverty Alleviation." Briefing Agreements." RSIE Discussion Paper No. 496, Paper No. 60, Oxfam, U.K. Available at available at www.spp.umich.edu. http://www.oxfam.org.uk/what_we_do/issues/ Stiglitz, J. 2000. "Democratic Development as the trade/downloads/bp60_textiles.pdf. Fruits of Labor." Keynote Address, Industrial Perry, Guillermo, and Marcelo Olarreaga. 2006. Relations Research Associations, Boston, January. "Trade Liberalization, Inequality and Poverty Trefler, Daniel. 2005. "Offshoring: Threats and Op- Reduction in Latin America." Paper presented at portunities." Paper presented at the Brookings the Annual World Bank Conference on Develop- Trade Forum, May, Washington, DC. ment Economics (ABCDE) 2006. UNCTAD (United Nations Conference of Trade and Puga, Diego, and Daniel Trefler. 2005. "Wake Up and Development). 2004. India's Outward FDI: A Smell the Ginseng: The Rise of Incremental Inno- Giant Awakening? UNCTAD/DITE/IIAB/2004/1. vation in Low-Wage Countries." Department of New York and Geneva: United Nations. 139 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 ______. 2005. World Investment Report: Transna- ______. 2005. "Workers and Labor Markets." In tional Corporations and the Internationalization World Development Report 2005. Washington, of R&D. New York and Geneva: United Nations. DC: World Bank. Wood, Adrian. 1997. "Openness and Wage Inequality in ______. 2006. World Development Report: Equity Developing Countries: The Latin American Chal- and Development. Washington, DC: World lenge to East Asian Conventional Wisdom." World Bank. Bank Economic Review 11(1): 33­58. Yusuf, Shahid, Kaoru Nabeshima, and Dwight Perkins. World Bank. 2002. Globalization, Growth, and 2006. "China and India Reshape Global Indus- Poverty: Building an Inclusive World Economy. trial Geography." In Dancing with Giants: China, Washington, DC: World Bank. India, and the Global Economy, ed. L. Alan ______. 2003. Lifelong Learning in the Global Econ- Winters and Shahid Yusuf. Washington, DC and omy: Challenges for Developing Countries. Singapore: World Bank and the Institute of Policy Washington, DC: World Bank. Studies. 140 5 Managing the Environmental Risks to Growth The gains from growth and globalization to depress growth rates below the low-growth could be undermined by their environmental scenario presented here. It is more likely that side effects. Because increases in production decades will pass before the most severe effects magnify cross-border pollution, while im- of climate change begin to be felt. Even so, the provements in technology make it possible to collective response of today's global leaders is expand or intensify the exploitation of scarce almost certain to have far-reaching implica- global resources, decisions at the national tions for the welfare of future generations. level are having a growing impact on other Technological progress and rising demand countries. International institutions will thus have increased efforts to harvest fish from the be required to play a larger role in a wide open seas, degrading ocean environments and spectrum of issues--all involving global driving some valuable species to near- public goods--where exclusive reliance on the extinction. Longstanding efforts to limit decisions of individual governments or the marine catches to sustainable levels have private market can lead to adverse outcomes. met with only a few successes. Why? Because Such goods include maintaining global secu- institutional weaknesses, technical difficul- rity, keeping the trading system open and ties, and fishing subsidies impede sustainable nondiscriminatory, and ensuring global finan- management. cial stability. As developing countries enlarge The growing interaction of national their role on the global stage, their integration economies through trade and movements as full partners in multilateral solutions to of people, while broadly beneficial, has global problems will be essential. increased the risk of spreading contagious Mitigating climate change, containing diseases. HIV/AIDS (human immunodefi- infectious diseases, and preserving marine ciency virus/acquired immune deficiency syn- fisheries are three additional global public drome) is one example. The severe acute res- goods that demonstrate the need for--and piratory syndrome (SARS) is another. The benefits of--international policy cooperation. most prominent current threat is the avian Rising industrial output means increasing influenza virus. concentrations of greenhouse gases in the at- These examples of the side effects of mosphere, which will have detrimental effects globalization--one long-term, one medium- on future productivity and--more generally-- term, and one immediate--pose risks to the on human welfare around the globe. Even in progressive expansion of the global economy, the next decade or two, scientists underscore and to developing countries in particular. Some the (unlikely) possibility that global warming of the more catastrophic climate-change could cause natural disruptions severe enough scenarios, if they materialize, could undermine 141 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 the development prospects of whole countries though every country will be affected, there is and even regions through their effects on agri- little systematic relationship between the size culture, water, and ecosystems. Similarly, failure of most countries' efforts to reduce carbon to contain an epidemic could bring global com- emissions and the damage these countries merce to a sudden halt, isolate some popula- experience from climate change. tions, and impose huge losses on affected devel- Ensuring that developing countries reap the oping countries. Unrestrained marine fishing, benefits from global public goods is particu- while less potentially calamitous than climate larly difficult. Developing countries typically change or a flu pandemic, could permanently account for a small share of international exhaust a critical global food source and destroy transactions, so they often lack the clout to irreplaceable deep-sea habitats and biodiversity. ensure that decisions made in international Effective multilateral collaboration is fora adequately reflect their interests. Many needed to ensure that economic growth and developing countries lack the financial and poverty reduction will proceed without causing technical resources to participate effectively in irreparable harm to future generations. Devel- international negotiations on many issues. For oping countries are central to the management example, the simultaneous negotiations on a of these risks. Although these countries are rel- variety of critical environmental issues forces atively small contributors to global warming governments with inadequate resources to today, the projections in chapter 2 imply that limit their participation (Esty and Ivanova they will soon enough become large contribu- 2002). Developing countries also lack the tors to global warming. And if no action is resources required to effectively address many taken, the standard of living that they could common problems. For example, malaria kills otherwise expect may well be put at risk. millions in developing countries, but research Given the limited supply of medical facilities on pertinent vaccines is limited, although and nursing care in the developing world, a flu some efforts are now under way. pandemic could have horrific consequences. While the three cases spotlighted here differ In many developing countries, people depend in the agreement on the extent of risks, there on fish for an important share of their diet, is a sufficient scientific consensus to move and the poor would suffer if the price of fish, forward on all of them. The needs and the as well as substitutes, were to skyrocket as methods to protect against the spread of supplies dwindled. (selected) contagious diseases are well known, The degree of international coordination although the efficacy of particular strategies required varies greatly from issue to issue, (quarantine, stockpiling of available vaccina- depending on the nature of the issue and the tions) in limiting the spread of avian flu is in geographical spread of its causes and effects dispute. The overexploitation of marine fish (table 5.1). The need for international coordi- stocks is well understood, although disagree- nation falls with the degree to which an indi- ment remains on the amount of resources to vidual country can benefit from its own efforts commit, the limits on fishing to impose, and to provide the good (or mitigate the evil), how to allocate access to fisheries. There is an and rises with the number of countries in- international consensus that human activity is volved (Barrett 2004). For example, the U.S.- contributing to climate change, but the precise Canadian agreement on reducing acid rain implications of different levels of greenhouse was facilitated in part because only two coun- gas concentrations for climate change remain tries had to agree and because each country uncertain. While disagreements over the facts gained an important benefit from its own of each case have affected efforts at interna- efforts to reduce pollution. By contrast--as tional cooperation, they have not been the shown below--negotiations over climate major impediment to progress. In reviewing the change are intractable in part because even state of knowledge in each area, this chapter 142 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H Table 5.1 Progress in providing many global public goods is limited Examples of global public goods Good Role of developing countries Progress of international efforts Global commons Climate change Limited current contributors, but major Current mitigation efforts insufficient to future source, of carbon emissions; stabilize global temperature potentially disastrous impact on many countries Biodiversity and ecosystems Main reservoir of many species Rate of species extinction rising; tropical forest cover declining Water resources Over 600 million people face acute Little international effort beyond freshwater shortage increasing awareness; 2­3 billion people may face severe freshwater shortage by 2020 Fisheries Many countries dependent on ocean 75 percent of commercial fish stocks fisheries for exports and domestic exploited at or above sustainable levels consumption Human issues Infectious diseases Developing countries could suffer severe Flu pandemic avoided (for now); limited losses in a global flu pandemic; already progress in containing malaria, measles, suffer millions of deaths from tropical AIDS in developing countries diseases Peacekeeping Millions killed in civil wars and Some interventions successful (Kosovo in intercountry conflicts the Republic of Serbia); others less so (Sierra Leone) Poverty 1 billion people living on less than Asia expected to see continuing decline in $1 a day people living in extreme poverty; Africa likely to see rise Regulatory framework Trade Developing countries account for Trade rules effective, but limited progress 27 percent of global merchandise exports; on removing trade barriers critical to goods and services exports represent developing countries 33 percent of developing countries' gross domestic product Financial architecture Total fiscal costs of systemic crises in Crisis interventions have mixed success; developing countries since 1975 exceeds little change in global rules that would $1 trillion dampen volatility Sources: Rischard 2002; World Bank, World Development Indicators; Honohan and Laeven 2005. discusses some of the key economic issues that quarter-century the global economy will con- constrain or support effective action to protect tinue to be at risk of sharp downturns from the environment and sustain growth. disruptions caused by contagious diseases. For example, a human flu pandemic similar to the 1918 Spanish flu could reduce global GDP The immediate risk of epidemics by 3 percent over a one-year period, with the G lobalization has increased the volume and more severe effects (in percentage terms) felt in speed of cross-border transactions, thus developing countries (World Bank 2006). increasing the potential for the transmission of The potential for a devastating global out- contagious diseases. The international trans- break of contagious disease underscores the mission of disease is nothing new.1 But air importance of international cooperation. Indi- travel and international contacts have greatly viduals benefit directly from access to vacci- accelerated its potential speed. Over the next nation, and individual countries benefit from 143 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 controlling disease within their borders--both countries--whether they had reported SARS benefits reduce the role that international cases or not--designated special treatment institutions must play. Nevertheless, other coun- centers ("SARS hospitals") and put in place tries do have a critical interest in containing dis- quarantine procedures. In some of the places ease, as containment reduces the probability of most severely hit, the measures were very further transmission. Measures by individual strict. Hong Kong (China) imposed restric- countriestocontaininfectiousdiseasemaybein- tions on peoples' movement between city dis- sufficient from a global perspective, in part be- tricts, and Singapore used TV surveillance and causeindividualcountriesmaylacktheresources radio bracelets to monitor and control the to take all measures that the international com- movements of persons who had come in con- munity might consider prudent. And the supply tact with SARS patients and of patients dis- of informational goods--for example, research charged from SARS hospitals. The World on vaccines and knowledge of treatment and Health Organization (WHO) collected and quarantine procedures--is likely to be impaired disseminated up-to-date information on the in the absence of effective international coopera- development of the disease and how to re- tion. This section discusses these issues in the spond, and coordinated scientific efforts to context of the recent SARS epidemic and the control and identify the virus causing the potential for an avian flu pandemic. sickness. Given the crucial role of air traffic in the spread of SARS between countries and SARS was a case study in virus continents, WHO issued the first emergency proliferation and containment travel advisory in its history. Travel bans were Five months after initial reports from East imposed for major affected areas in April Asia (in February 2003) of an atypical respi- 2003 (Bell and Lewis 2004). ratory disease, more than 8,000 cases had The combined efforts by local, national, been reported in close to 30 countries.2 The and international authorities to contain the disease, labeled SARS, was highly contagious threatening pandemic were successful: newly and life threatening: almost 10 percent of re- reported cases, which increased rapidly in ported cases ended with the patient dying. March and April of 2003, peaked in early The global response to the rapidly spread- May and thereafter declined rapidly (fig- ing disease was swift and determined. Many ure 5.1). Although no cure has yet been found Figure 5.1 The SARS epidemic was contained in a matter of months Number of cases 9,000 8,000 7,000 6,000 5,000 Cumulative Cumulative cases recoveries 4,000 Active cases 3,000 2,000 Cumulative deaths 1,000 0 0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105 Days into outbreak Source: http://www.adrianboyko.com/SARS (WHO data). 144 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H for the disease, it was successfully contained outbreak have been discussed widely at both by late 2004, and no new cases have been the national and international level (WHO reported since 2005. The containment of 2006; CDC 2004; Osterholm 2005; Sturm- SARS was facilitated by the fact that the dis- Ramirez 2006). A large number of actions to ease is less transmissible than some influenzas address a potential avian flu pandemic can be experienced in the past. The SARS experience envisaged: highlights the role of the technological devel- opments and rapid communications that pro- · Reducing the incidence of avian flu in mote globalization both in speeding the birds would reduce the probability of spread of contagious diseases and in providing human infection. Effective systems to the tools to combat them. monitor flocks is required, coupled with compensation for damage if birds have Has the risk of avian flu been contained? to be slaughtered, and punishment for The virus causing avian flu (H5N1) affects failure to report, as otherwise breeders primarily birds, although cases of human in- are likely to conceal incidences of the fection, recorded since 1997, have increased disease in their flock. There also is a need significantly since 2003, while remaining low. to regulate bird breeding and marketing An outbreak of avian flu through human-to- methods that facilitate the occurrence or human transmission could have catastrophic spread of the disease. implications for welfare, particularly in devel- · Developing a vaccine that is certain to be oping countries where public health systems effective is impossible, given present are weak, and could result in a sharp, short- knowledge, because the form of the future term interruption in global growth. mutation of the virus is unknown. There is hope among researchers that it may be Possible pandemic. The avian flu is a subject possible to develop flu vaccines that will of great concern principally because more than be effective against whole classes of flu half of all infected persons have died from the viruses, including future mutations. Many disease, and flu viruses have the potential to experts argue that the likely success rate mutate into a form that is easily transmitted from the use of existing vaccines is suffi- between humans (the "Spanish flu" pandemic ciently high to make stockpiling vaccines of 1918­19 killed up to 50 million people). The a key ingredient in a comprehensive re- rapid expansion of the poultry population, and sponse strategy. in particular the close proximity between · An effective surveillance system will be humans and animals in East Asia, has increased essential to detect and report cases--even the likelihood that such a mutation may occur.3 suspected cases--before they have a And if it does, the greatly increased speed and chance to spread. The SARS episode scope of human travel would facilitate a rapid showed how important early detection spread of the disease worldwide. WHO can be for an effective containment projections reckon that the mutation of the strategy. avian flu virus permitting human-to-human · Steps to treat victims and contain the transmission would, under best-case scenarios, disease may range from administering entail the spread of the disease among humans appropriate medicines to implementing across all continents (WHO 2005). Other quarantine procedures for contagious estimates that assume a more virulent virus patients. It is uncertain whether current involve much higher numbers of deaths. antiviral drugs would be effective against a future mutation of the avian flu virus. Prevention and countermeasures. Alternative Even so, many experts argue that exist- measures of preventing or responding to an ing antiviral drugs are sufficiently likely 145 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 to be effective to justify stockpiling them independent efforts because they benefited as part of any response strategy. directly, thus reducing the burden on interna- · Should both prevention and containment tional cooperation. Contrast this (so far) in the early stage of the disease fail, trade success story with the failure to eradicate and travel restrictions, quarantine pro- other endemic diseases.5 The burden of infec- cedures, the transformation of existing tious diseases is greatest for developing coun- buildings into emergency hospitals, and tries, which often lack the resources to effec- general efforts to deal with the multitude tively distribute vaccines and treatments. of disruptions accompanying any such Would these diseases still be so prevalent if catastrophe would be required. industrial countries continued to be vulnerable to them? Avian flu threat receding. At this writing there are encouraging reports that the avian influenza The medium-term risks is indeed in retreat. New cases are rarely reported, and the countries where the most to marine fisheries M human infections have occurred (Vietnam, arine fishing, on the high seas and Thailand, and China) report that cases within many nations' 200-nautical-mile observed in both poultry and humans have exclusive economic zone (EEZ), is reaching declined steeply.4 Officials of WHO and the its limits. Increased demand and technologi- World Organisation for Animal Health (OIE) cal improvements have led to increasing pres- credit countries' aggressive countermeasures sures on marine fish, as well as on the fragile for the apparent success in getting the disease ecosystems in which they live.6 Excluding under control. Global communications and data from China, the accuracy of which has cooperation clarified the risks, publicized been questioned, production has declined advanced methods to contain it, and induced since about 1990 (FAO 2004 and figure 5.2). valuable international cooperation at various The acceleration of global growth envisioned levels. Virtual unanimity among national in chapter 2 is likely to increase pressures on administrations on the need to act rapidly marine fisheries over the medium term. With- enabled WHO and OIE to implement out efforts at conservation, the global econ- appropriate supranational measures without omy is likely to confront dwindling supplies delay. Nevertheless, these efforts have failed of commercially exploited marine fish, cou- to prevent the disease from becoming fully pled with rising demand for fish with grow- endemic in several countries, and reducing the ing incomes. scope of the disease remains a high priority to limit the risk of another flu pandemic. Marine fish are under increasing pressure A significant number of the world's most valu- Global cooperation may able fish stocks have been depleted through prevent contagion overfishing, habitat degradation, pollution, The success of international efforts to contain or other causes (Bolton 2005). Fully 75 per- infectious disease is in part rooted in the cent of the world's marine fish stocks are being nature of the problem. The threat of a global exploited either at or above their maximum pandemic is immediate, well understood, and sustainable level (FAO 2004). While reduc- potentially catastrophic for the industrial tions in fish stocks from fishing are not new, countries that have the resources to act. In they have accelerated over the past few addition, while international cooperation has decades owing to technological advances played a critical role in reacting to SARS and that have enabled large-scale commercial avian flu, individuals and individual govern- fishing fleets to increase their exploitation of ments have been willing to make major, traditional waters and expand to new areas 146 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H Figure 5.2 Total marine fish catch has leveled off Total catch, millions of tons 100 90 China, including Hong Kong and Macao 80 70 60 50 40 World, excluding China 30 20 10 0 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 Source: FAO 2004. in the Indian Ocean and the seas around Straddling Fish Stocks and Highly Migratory Antarctica.7 And improvements in fishing gear Fish Stocks, called the UN Fish Stock Agree- and onboard storage technology have made ment (effective in 2001), established basic new species commercially viable (Kura and standards for fisheries management for highly others 2005). Natural phenomena, for exam- migratory species, such as tuna, and for so- ple the impact of El Niño on Chilean and called straddling stocks--species that range Peruvian fisheries and of warmer water in the between EEZs and the high seas North Atlantic on North Sea cod, have con- (Lodge 2005). But high-seas bottom-dwelling tributed to the reduction in fish stocks species are in a jurisdictional vacuum. Few of (Schmidt 2002). Climate change (discussed the regional fisheries management organiza- below) will increase the acidity of the ocean as tions are mandated to manage bottom fishing increasing amounts of carbon dioxide dissolve on the high seas (Gianni 2004), and most of in sea water, with potentially serious implica- these fisheries should be considered unregu- tions for ocean environments and the sustain- lated (FAO 2004). Even when a mandate ability of some fish species (U.K. Government exists, the regional fisheries management or- 2006). And some deep-sea fish are particularly ganizations established for this purpose often vulnerable to overfishing owing to their long suffer from inadequate resources or insuffi- lives and few offspring (Shotton 2006). Sub- cient political support, and face several sidies, estimated globally at between $12 and important obstacles: $20 billion a year, have also contributed to overexploitation of fish resources (Milazzo · Data on catch volume and area, the num- 1998; APEC 2000; WWF 2001).8 ber and size of fish, the number of juveniles that develop to maturity, interactions with Managing high-seas fisheries is not easy other species, and the impact of environ- The UN Convention on the Law of the Seas mental factors are often inadequate to (effective in 1994) helped define property define sustainable catches (Kura and rights by enabling coastal states to establish others 2005). Overall, estimates of fish EEZs of up to 200 miles.9 The UN Agreement stocks may be off by as much as 30 percent for the Conservation and Management of (Berrill 1997), and single-species stock 147 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 assessments have failed to predict rapid of the increased trade in highly migratory stock declines in a number of cases (Pauly species since the late 1970s (Webster 2006).11 and others 2002). About 250 million people in developing coun- · It can be difficult to set limits on fish catch tries depend directly on the fishing sector that impose the right incentives. The (including inland fishing) for food and allocation of licenses, the most common income, and fish provide nearly 20 percent of system for controlling fish effort animal protein consumed by people in devel- (Cunningham and Greboval 2001), can oping countries (World Bank 2004). be thwarted by expanding the capacity of Many developing countries lack the re- individual boats or improving technology. sources to police their own coastal waters and Limits on the total catch (at which point a economic zone, much less establish effective re- fishery is closed) can result in the harvest gional institutions to manage nearby marine being caught more rapidly and using fisheries. The expansion of distant-water fish- more resources than would be the case if ing fleets to new, mostly unregulated areas has quotas were allocated to individual fish- spawned conflicts with traditional fishers in the ers (difficult for regional fisheries man- poorer developing countries, while largely un- agement organizations, though it is done controlled fishing has led to the depletion in areas controlled by single countries), of valuable fishery resources in the southern and may encourage more dangerous fish- seas--for example, the sharp decline in orange ing, such as during inclement weather roughy and in Patagonian toothfish (Chilean (Kura and others 2005).10 sea bass), as worldwide demand for them · Monitoring and enforcing limits on fish- increased. While some developing countries ing can be problematic, because fishers earn significant foreign exchange by selling understandably are reluctant to provide fishing rights in their waters (several West information that can affect their compet- African nations without significant industrial itiveness (Shotton 2006). Some fishers fleets have done so), this practice has inten- report data to national authorities under sified competition for small-scale fishers confidentiality agreements that prohibit (Kura and others 2005).12 Fish resources in the release to international authorities. shallow waters off the west coast of Africa may · Fishers can attempt to evade enforce- have declined by half from 1985­90 because of ment of conservation measures by regis- distant-water fleets. tering under a flag of convenience (with countries that exercise little control over International cooperation aims to ensure their ships); at least 2,800 large fishing sustainable fishing vessels either have a flag of convenience There is a global consensus, expressed at the or no registry at all (WWF 2001). World Summit on Sustainable Development in Johannesburg in 2002, that depleted fish Developing countries are stocks should be restored to levels that can particularly vulnerable produce their maximum sustainable yield Developing countries, important participants by 2015. Progress has been considerable in in large-scale commercial fishing, confront setting the institutional framework for con- particular weaknesses in managing fishery serving the ocean's fish, both through defining resources. In 2001, 6 of the top 10 marine ownership rights (setting an EEZ of 200 miles) fishing nations were from the developing and in setting up multilateral institutions world, with China and Peru (numbers 1 and 2) (regional fisheries management organiza- alone accounting for more than a quarter of tions). Effective management plans have been total marine capture in metric tons. Develop- implemented for a few highly migratory fish ing countries also account for the vast majority stocks, for example the North Atlantic 148 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H swordfish and Atlantic bigeye tuna (Webster consequences for growth. Moreover, the ap- 2006). Still, most commercially exploited proach adopted to reducing carbon emissions marine fish species face increasing pressures. could entail costs to growth, particularly if po- Regional management is clearly inadequate in litical constraints prevent the adoption of effi- many fisheries. And further uncontrolled ex- cient policies. Developing countries, at the cen- ploitation could lead to the irretrievable loss ter of this issue, are likely to suffer the worst of valuable sources of the world's food. A fur- consequences of climate change and have the ther strengthening of domestic and multilat- least ability to adapt. They also are the largest eral institutions, particularly the regional fish- future source of additions to carbon emissions, eries management organizations, is a high and thus will have an important role in negoti- priority for international action. ations to limit emissions. Global temperatures are rising The long-term risk The burning of fossil fuels produces gases that of climate change trap incoming solar radiation, leading to a C limate change induced by carbon emis- rise in global average surface temperature.13 sions already has had significant impacts Measurements show that the average world on the global environment, and continuing temperature has increased since the start of emissions at current levels are likely to have the Industrial Revolution (figure 5.3). Models severe implications for human welfare over the of the determinants of temperature change long term. The threat of climate change is inex- that take into account the addition of green- tricably linked with the scenario for global house gases (GHGs) into the atmosphere from growth over the next 25 years, because there human activities (second panel of figure 5.4) is a risk that climate change could accelerate, provide much more accurate explanations of entailing greater-than-expected near-term historical trends in temperature than models Figure 5.3 Temperatures have increased rapidly since the Industrial Revolution Departure in temperature ( C) from the 1961­90 average, Northern Hemisphere Data from thermometers 0.5 0.5 0 0 0.5 0.5 Data from tree rings, corals, ice cores, and historical records 1.0 1.0 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 Year Source: Intergovernmental Panel on Climate Change (IPCC) 2001. 149 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Figure 5.4 Temperatures and greenhouse gas emissions have risen Natural forcing only Natural anthropogenic forcing Temperature anomaly in C Temperature anomaly in C 1.0 1.0 1.0 1.0 Observations Observations 0.5 0.5 0.5 0.5 0 0 0 0 0.5 0.5 0.5 0.5 Model results Model results 1.0 1.0 1.0 1.0 1850 1900 1950 2000 1850 1900 1950 2000 Year Year Source: IPCC 2001. that ignore this addition (first panel of fig- Temperatures will continue to rise ure 5.4). There is general agreement that The extent of climate change will depend on human activity has contributed to the rise in future GHG emissions (which will be deter- GHG concentrations and climate change mined largely by growth, technological devel- since the start of the Industrial Revolution. opments, and policies that determine incentives Climate change, while generally viewed for carbon efficiency) and on the ultimate effect as a long-term problem, has already had of those emissions on climate. The Intergovern- significant effects. Ice coverage has declined at mental Panel on Climate Change (IPCC) has the two poles (box 5.1), mountain glaciers are developed scenarios that relate forecasts of retreating worldwide, ocean temperatures output, population, and technological develop- are rising, the sea level is rising, the perma- ments to future CO2 (the most important frost is thawing, growing seasons in mid- to GHG) concentrations in the atmosphere, and high-latitude areas are lengthening, and the thus to climate change. While the scope for ranges of some animal and plant species are limiting future GHG concentrations and the moving toward the poles and higher altitudes associated climate change remains great, past (IPCC 2001). Controversy remains about the and current GHG emissions will continue to precise quantitative impact of anthropogenic influence the global climate for some time. GHG emissions on the climate. Nevertheless Even if emissions peak in the 21st century and there is widespread concern that a continua- then decline below current levels, global surface tion of the rapid economic growth experi- temperature will continue to rise for centuries, enced since the beginning of the Industrial and sea levels will rise for several millennia Revolution (and as a second industrial revo- (figure 5.5). lution unfolds in China and other major The IPCC scenarios cover a wide range of rapidly growing developing countries), sup- growth paths, with the four principal scenarios ported by the continuing exploitation of fossil ranging from 1 percent to 3 percent growth fuels, will induce significant changes in global in per capita income during 2000­30. and regional climates.14 Although these scenarios were developed in the 150 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H Box 5.1 The vanishing polar ice O ne effect of the rise in global temperatures has summers in the Arctic could become near ice-free by been the drastic reduction in large bodies of ice the end of the century. in the Arctic and in Antarctica, which appears to Consecutive satellite images also have revealed the have accelerated recently. Average temperatures in collapse of the Larsen B ice shelf on the Antarctic the Arctic region are rising twice as fast as elsewhere Peninsula during the 2002 Antarctic summer, fulfill- in the world. Arctic ice is thinning, melting, and ing predictions made by British Antarctic Survey rupturing. The largest single block of ice in the (BAS) scientists. The collapse of the 3,250 km2 ice Arctic, the Ward Hunt Ice Shelf, had been around shelf is part of the ongoing developments in a region for 3,000 years before it started cracking in 2000. of Antarctica that has experienced unprecedented Within two years it had split all the way through warming over the last 50 years. and is now breaking further into smaller pieces. Continued melting of polar ice could induce The polar ice cap as a whole is shrinking. significant rises in sea levels, with potentially Images from National Aeronautics and Space catastrophic implications for many coastal areas, Administration (NASA) satellites show that the area and raise the possibility of interrupting the Gulf of permanent ice cover is contracting at a rate of Stream, which could drastically reduce European 9 percent each decade. If this trend continues, temperatures. The summer arctic ice field is shrinking The Larsen B ice shelf collapsed North Pole March 5, 2002 March 5, 2002 February 17, 2002 February 17, 2002 January 31, 2002 January 31, 2002 Summer Arctic Sea Summer Arctic Sea ice boundary 1979 ice boundary, 1979 late 1990s, a recent review finds that they are would imply considerable potential for reining roughly consistent with projections undertaken in carbon emissions over the medium term, since then (Van Vuuren and O'Neill 2006). The given strong international efforts to slow cli- scenario outlined in chapter 2 envisions global mate change. While achieving reductions in per capita growth of 2.2 percent. Thus, the path emissions (as envisioned in figure 5.5) by im- of carbon emissions implicit in this scenario is proving efficiency holds considerable promise, roughly similar to that envisioned in many of the the near-term prospects for reducing carbon IPCC scenarios. As discussed in chapter 2, this emissions through alternative energy sources are scenario assumes a steady improvement in the limited (box 5.2). The world is not yet on a path technical efficiency of energy use but no major toward the emissions reductions that will be policy initiatives that would raise the price of essential even to stabilize global temperatures fossil fuels. Thus, the forecasts in this book (at significantly higher levels than at present). 151 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Figure 5.5 Greenhouse gas emissions have long-term effects Magnitude of response Time to reach equilibrium CO2 emissions peak Sea-level rise due to ice melting: 0­100 years several millennia Sea-level rise due to thermal expansion: centuries to millennia Temperature stabilization: a few centuries CO2 stabilization: 100­300 years CO2 emissions Today 100 years 1,000 years Source: IPCC 2001. Box 5.2 Can efficiency and renewables be the answer? S ignificant reductions in GHG emissions can be have the opportunity now to adopt more efficient achieved through improvements in energy choices for infrastructure and technology that could efficiency and increased use of renewable energy drastically reduce GHG emissions for decades to come. sources. Nevertheless, the share of renewables in Energy efficiency is often the most cost-effective energy use is not expected to increase much within and low-risk approach to reducing the need for energy, the forecast period, and the demand for hydrocarbons and can also generate significant environmental is set to rise by more than 50 percent (IEA 2004, benefits. Considerable potential exists for adopting 2005). This underlines the need for policies to en- more efficient technologies in transport, industry, courage energy savings and improve the profitability buildings, and power generation. of alternative energy sources. Developing countries have much greater potential · In transport, new materials, compact engines, and than industrial countries for reducing emissions, advanced fuel systems can lead to lighter and more in great part because they are moving toward the fuel-efficient vehicles, while hybrid vehicles can technological frontier in existing industry and provide substantial fuel savings. If all technical infrastructure. For example, China could use some means were implemented, the International Energy 20 percent less coal if its plants were as efficient as Agency estimates that a 40 percent improvement the average plant in Japan, and the potential for in fuel economy of gasoline engines is achievable adopting proven energy savings in cement and pulp in the coming decades. The prospects for hydrogen and paper is significant. Moreover, rapidly growing and fuel cell vehicles are less promising over the developing economies can invest directly in energy- forecast period because they require significant efficient technologies, thereby leapfrogging earlier, cost reductions, performance improvements, and inferior processes. For example, expansion of development of fuel cell vehicle markets and low-power, white-light-emitting diodes that run on hydrogen infrastructure. batteries charged by solar panels could enable the · Many new buildings could be 70 percent more rural poor in some countries to bypass the need for energy efficient than the existing stock through centralized electrical grids. Developing countries the use of new technologies in windows, 152 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H Box 5.2 (continued) insulation, furnaces, air conditioners, appli- to accelerated expansion is competition with other ances, lighting, and standby power. demands for biomass, particularly for use as food. · In industry there is a large potential to improve · Wind generation is expected to rise, buoyed by the efficiency of motors, boilers, pumps, and sharp declines in the cost from economies of scale heating and cooling systems. In addition, large with the use of larger turbines. Nevertheless, wind amounts of energy can be saved through new still has problems of intermittency, low reliability, processes in individual sectors, such as direct problems connecting to the grid, and (more casting in iron and steel, and biofeedstocks in recently) difficulties in siting land-based turbines. the production of petrochemicals. · Generation from geothermal sources is concen- · In the electric power sector, switching from coal trated in a few countries. While geothermal is a to natural gas would reduce emissions, both very competitive and reliable source of power, because natural gas emits only about half as and its potential is enormous, it is a site-specific much CO2 as coal per kilowatt hour and be- resource that can only be accessed in certain cause the latest combined-cycle gas plants attain parts of the world. efficiencies of 60 percent, compared to 46­49 per- · Solar power is expected to account for less than cent for the best available coal-fired plants. 0.5 percent of total power supplies by 2030, as its Nuclear energy offers emission-free technology investment and generating costs are the highest but faces high capital costs, problems of waste of all commercially deployed renewable energy storage, risks of accident, public opposition, and sources, although the range of costs varies widely possible proliferation of nuclear weapons. depending on the amount of sunshine available. There will also be some rise in solar thermal Renewable energy sources. Renewable energy power, whose generation costs are typically now accounts for 14 percent of world energy de- double those of conventional energy sources. mand, and while the authors of this chapter antici- · Tide and wave generation is still in its infancy. pate that its use may rise by 50 percent by 2030, its Projects need to be large-scale if they are to share of total energy is not expected to change withstand offshore conditions, and these are very greatly unless vigorous policies encourage switching costly and carry high risks. Site-specific environ- from nonrenewable energy sources. mental effects also need careful assessment. · Renewable electricity generation is dominated by Biofuels may provide a significant alternative fuel hydropower, which accounts for 16 percent of option for transportation over the forecast period, with global electricity production. Hydropower is the ethanol from sugarcane (from Brazil, for example) of- cheapest source of power in many areas. There is fering the best chance of commercial viability. Other considerable potential for expansion, particularly feedstocks, such as corn, have much higher costs owing in the form of small hydro plants, although to lower yields and are unlikely to be financially viable concerns over undesirable environmental and without government support. If ethanol can be pro- social impact have been important barriers. duced from cellulose using biomass as the fuel for the · Biomass generation can be highly economic, conversions process, net GHG emissions from well to particularly for co-firing other hydrocarbon- wheel basis (that is, through the complete chain of fuel based plants. New technologies are expected production and use) could be reduced to zero, accord- to reduce costs further, but the largest barrier ing to the International Energy Agency. Climate change could have a catastrophic temperatures, the precise climate changes in- impact on some countries volved, and the links between climate change The effects of climate change on human wel- and human activity. Available calculations fare are uncertain, depending as they do on indicate that the aggregate global economic the magnitude and timing of increased impact of a small rise in temperatures would 153 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 be significant, but not enormous. Tol (2002) benefit from a modest rise in temperatures, finds that the impact of a rise in the global while mid-latitude countries, many of them mean surface air temperature of 1 degree Cel- high-income, are likely to face small net effects sius (the temperature rise anticipated over the from climate change through this century (the first half of the 21st century) could range any- rise in the sea level may inundate some coastal where from an annual increase of world GDP areas and increasing severity of hurricanes and by 2.3 percent to a decrease of 2.7 percent, de- cyclones could increase coastal damages, pending on assumptions made about the value while agricultural yields in other areas could of nonmarket goods and services. Examples improve). Over the long term, and in the ab- include how to value human lives lost and sence of successful mitigation efforts, climate gained, and the damage to ecosystems and change is likely to be disastrous for all coun- biodiversity--a quarter of the world's known tries. Examples of the possible damage include animals or plants, or more than a million the following: species, are likely to die out because of the forecast warming over the next 50 years · A rise of 1 degree Celsius could lead to (Grubb 2006b).15 A more recent analysis esti- an 80 percent loss of coral reefs; further mates that failing to address climate change increases in extreme precipitation caus- could reduce welfare by an amount equal to a ing drought and landslides; a 20 to 5­20 percent fall in per capita consumption 35 million ton loss in cereal production (box 5.3). and an approximately 10 percent decline These estimates also do not capture low- in yields of various African crops (such probability risks that could imply severe as barley and rice). consequences for the global economy over a · A rise of 2 degrees Celsius could lead to relatively short timeframe. For example, if the large-scale displacement of people in the Gulf Stream stalls as melting ice introduces Mahgreb as rainfall declines by at least more fresh water into the northern Atlantic, 40 percent; the total loss of summer European temperatures could plummet. And Arctic sea ice; the likely extinction of the there is potential for a rapid, almost self- polar bear and walrus; millions more perpetuating acceleration of climate change if people at risk to malaria, particularly in the large methane deposits in arctic tundra are Africa and Asia; and a 50 percent loss of released as climate change proceeds. the Chinese boreal forest. Estimates of the aggregate economic impact · A rise of 3 degrees Celsius could lead to of climate change mask extreme variations in massive changes in habitats, such as the costs and benefits for different countries. The collapse of the Amazon rainforest and brunt of the damage from climate change will the Great Lakes wetland systems; the in- be felt by low-latitude developing countries, undation of the Ganges delta region, un- with the extent of harm critically dependent dermining the agricultural system that on how much temperatures increase (so that feeds a quarter of a billion people; the damages are likely to rise as time goes on). spread of desert-like conditions in Africa Many developing countries are more vulnera- as the Kalahari dunes become mobile; ble to climate change because they are already additional millions of people at risk of warmer than developed countries and suffer hunger; two to three hundred million from high rainfall variability, they are heavily more exposed to malaria; hundreds of dependent on agriculture (the sector most millions more exposed to dengue; several vulnerable to climate change), poor public ser- tens of millions displaced from coastal vices increase the potential welfare loss, and areas because of rising sea levels; and low incomes impede adaptation (U.K. Gov- billions more subject to increased water ernment 2006). Countries near the poles could stress (Warren 2006). 154 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H Box 5.3 Stern Review:The Economics of Climate Change T he government of the United Kingdom recently or through establishing tradable quotas, although issued a report on the economics of climate regulation may also be used where market-based change prepared by Sir Nicholas Stern at Treasury. mechanisms are ineffective. However, setting an ap- The report underlines the very serious global risks propriate price for carbon may not lower emissions posed by climate change, and the urgency of steps to sufficiently, due to uncertainty on future pricing reduce carbon emissions. The principal finding is policies, barriers to technology development in key that the benefits of strong, early action on climate sectors related to climate change, and external bene- change exceed the costs involved, reflecting two fits to technology development (for example, inspir- insights. ing ideas for new technologies) that are not captured First, the continued growth of carbon emissions by investors. Thus, the public sector also should at current rates raises the risk of serious, irre- promote low-carbon and high-efficiency technolo- versible damage to global welfare. Absent changes gies through increasing support for research and de- in policies, carbon emissions could rise by the mid- velopment, demonstration projects, and early-stage dle of this century to a level that would eventually commercialization investments in some sectors. Gov- commit the world to a rise in average temperatures ernments also should focus on removing barriers to of more than 5 degrees Celsius above preindustrial behavioral change--such as transaction costs, organi- revolution levels, equal to the amount of global zational inertia, and a lack of reliable information-- warming that occurred between the last ice age and through regulation (for example, minimum stan- today. The total cost of the climate change resulting dards for buildings and appliances), labeling, and from "business as usual" emissions over the next sharing best practices, and financing the upfront two centuries is estimated to equal a minimum re- costs of efficiency improvements. duction in global per capita consumption of 5 per- Adaptation also will be essential to limiting the cent. Taking into account the nonmarket impacts (on negative impact of inevitable climate change. While the environment and human health) of these emis- individuals will undertake adaptation in reaction to sions, the potential for feedbacks that would amplify market or environmental changes, governments can climate change, and an increase in the weight ac- provide policy guidelines as well as economic and corded to the poorer regions, and "business as usual" institutional support. climate change results in a reduction of about 20 per- The Stern report emphasizes the importance of in- cent in global per capita consumption. ternational collective action to respond to climate Second, atmospheric greenhouse gas concentra- change. Cooperation should cover all aspects of tions could be stabilized at levels that greatly reduce emissions reductions policies. It is necessary to create the risk of climate change damages, at relatively low a broadly similar carbon price signal around the cost. Carbon emissions can be cut by reducing the de- world, and to promote carbon finance to accelerate mand for emissions-intensive goods and services, in- action in developing countries. An equitable distribu- creasing energy efficiency, switching to low-carbon tion of effort that takes into account income, historic technologies, and reducing non­fossil fuel emissions responsibility, and per capita emissions would have (from deforestation and in agriculture). A series of industrial countries undertaking emissions reductions model-generated estimates of the annual cost of cut- of 60 to 80 percent (from 1990 levels) by 2050. ting emissions to a level consistent with stabilizing at- mospheric greenhouse gas concentrations at 550 parts per million average 1 percent of global GDP by 2050.ª aAnything higher than 550 parts per million would Reducing emissions efficiently requires pricing substantially increase the risk of harmful impacts on global carbon to reflect fully the risks of climate change. welfare while reducing the expected costs of mitigation by This can be done through setting a tax on emissions comparatively little. 155 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 These measurements generally consider the polluters to seek the least costly method of welfare impacts of a particular level of global reducing the risk of climate change are average temperatures, and thus may exclude some significant risks. Climate change may · Setting a uniform price for the emission increase the uncertainty surrounding, and of GHG (a uniform global carbon tax, variability of, weather, which could increase for example), and an equivalent subsidy costs. For example, a higher mean sea level for measures reducing atmospheric GHG could make storm surges more devastating; concentrations and other conditions that and higher average temperatures may have a can cause climate change.17 smaller impact on agricultural productivity · Setting a global emissions target and es- than increases in long, hot, dry spells (Weyant tablishing a market for emission permits. 2000). The speed of climate change is also Industries and nations with high (or low) important, as many species may have trouble abatement costs would buy (sell) such adapting to rapid increases in temperatures. permits, up to the point where abatement costs were equalized across industries and What can be done to reduce national economies, resulting in a uni- GHG emissions? form price of emissions permits.18 In the absence of intervention, global CO2 emissions could reach between two and four Location is important in determining costs times current levels by 2100, resulting in because the marginal cost of combating cli- much greater GHG concentrations than envi- mate change differs widely between countries sioned in most models of climate change and economic sectors. For example, the cost (Grubb 2006b). Thankfully, there is a wide of a 100-million-ton reduction in carbon emis- variety of possible methods to reduce GHG sions by 2010 was estimated to be less than emissions, for example improving energy $5 per ton of carbon (in 1985 dollars) for the efficiency and relying more on renewable United States, about $40 for the European energy sources (see box 5.2), switching to Union, and almost $400 for Japan (Ellerman, fuels with lower GHG emissions (from coal Jacoby, and Decaux 1998). Costs tend to be to natural gas, for example), capturing and even lower for developing countries. The time storing carbon emissions, sequestering car- span over which emissions are required to fall bon through reforestation, changing lifestyles also affects the cost of abatement: longer time to reduce demand for energy, reducing spans reduce costs because existing plants and growth in output, and geoengineering (to equipment need not be retired before the end change the reflectivity of the atmosphere, of their useful life, while shorter time spans oceans, and land).16 Some measures, such improve the credibility of compliance targets. as reducing subsidies that support high levels The future path of global growth may well be of energy intensity, may have low or even affected by efforts at mitigation, depending on negative costs, while others involve very their severity and the attention paid to ensur- expensive regulatory intervention. ing that mitigation is achieved at least cost. The costs of measures to achieve a given In addition to reducing carbon emissions, level of emissions reductions, and thus the efforts to adapt to climate change will also be consistency of mitigation efforts with an accel- required. Even if the world succeeds in eration of global growth, will depend critically markedly reducing carbon emissions in the on technological developments and the poli- near future, the GHGs already in the atmo- cies adopted. Technological developments are sphere imply increases in global temperatures uncertain, although the dangers posed by and rises in sea level for many years to come climate change encourage attention to subsi- (see figure 5.5). The welfare impact of climate dizing research. Policies that would induce change on developing countries is likely to 156 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H be all the more devastating because develop- between the locations where GHG emissions ing countries bear the brunt of the anticipated originate and where major damages induced by damages and have less ability than industrial climate change are likely to occur, most coun- countries to adapt, and because the welfare tries gain very little direct benefit from their impacts of income declines are greatest for own mitigation efforts. Individual countries the poor. thus face a strong incentive not to make efforts Given the critical nature of this issue for to reduce emissions, and to minimize their own developing countries, the World Bank Group commitments to international efforts. The is rapidly expanding its activities to achieve problem is exacerbated because likely damages a low-carbon economy. The Bank currently are distributed unevenly around the globe. manages nine funds devoted to developing the While the very existence of some island states carbon market, with a total investment of may be threatened by rising ocean levels, coun- $2 billion. The Global Environment Facility is tries with large Arctic areas may actually bene- the largest source of multilateral grant financ- fit from (modest) climate change.19 ing for low-carbon technologies, with a total Developing countries can be major players investment of $1 billion. The Bank is on track in global efforts to reduce global climate to meet its 2004 commitment to a 20 percent change--they certainly will be greatly affected average annual growth in new renewable by success or failure. As discussed, developing energy and energy-efficiency commitments countries are likely to bear the worst costs of between fiscal year 2005 and fiscal year 2009 climate change. At the same time, they bear (Sierra 2006). little responsibility for the current stock of GHGs in the atmosphere and are understand- Agreeing on policy is difficult ably loath to impede their own growth to Model-based analyses suggest that the global resolve a problem that is largely the creation net benefits of a coordinated international of industrial countries. Still, future increases in policy regime to reduce GHG emissions far GHG emissions will occur mainly in develop- exceed the benefits of individual countries act- ing countries (figure 5.6), so that any policy ing on their own (Nordhaus and Yang 1996). strategy that excludes these major future emit- However, as there is no systematic relationship ters is unlikely to be effective. Moreover, Figure 5.6 Carbon emissions from developing countries are set to rise Annual carbon emissions, billions of tons 18 18 16 16 14 Major 14 emitters 12 12 World 10 Annex 1 10 8 8 6 6 4 4 Industrial 2 2 countries 0 0 1990 2000 2010 2020 2030 2040 2050 Source: OECD Green Model simulations. Note: Annex 1 includes the industrial countries plus some countries in Eastern Europe and the former Soviet Union. Major emitters include Annex 1 countries plus large developing countries: China, India, Indonesia, Mexico, and South Africa. 157 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 agreements with limited geographic coverage Mechanism). The provisions for carbon may lead to the migration of key polluting in- trading and the Clean Development dustries to nonparticipating countries, thereby Mechanism have provided useful practical undermining success.20 experience on how to manage such mech- Other aspects of climate change impede anisms, which are likely to be part of future international agreement. With a few highly hy- agreements on climate change. pothetical exceptions, the most severe impacts The Kyoto Protocol represents a major of climate change are not expected for several attempt by the international community to decades, which raises uncertainty, leads to dis- come to grips with climate change, and by agreement on how costs should be distributed signaling future policy actions to reduce GHG over time and across generations, and discour- emissions, it may encourage investors to adopt ages action by political leaders concerned more efficient technologies. But it has been with short time horizons (the next election). subject to many criticisms. The sharp cuts in And as elaborated above, considerable uncer- emissions required of some participating tainty remains over the costs of mitigation and countries restrained some countries from sign- the precise impact of different levels of GHG ing, particularly as no constraints were im- concentrations on human welfare. posed on other countries where emissions will be growing fastest in the foreseeable future. What has been done to reduce The transaction costs involved in the Clean GHG emissions? Development Mechanism and the Joint Imple- Despite the difficulties involved in reaching mentation Framework make it difficult for international agreements and the incentives countries to meet their obligations at the lowest for free riding, some progress has been made global cost. It is too early to judge compliance in reducing GHG emissions, both through (emission curtailment obligations are legally international agreements and by individual binding only for the 2008­2012 period). But countries and regions. emissions from transition economies are well below their Kyoto targets owing to the major Kyoto Protocol. The Kyoto Protocol, which decline in economic activity after 1990, while came into force in February 2005, committed emissions from most industrial country signa- most industrial countries and some of the tories exceed their targets.21 The penalties for transition economies (together referred to as the noncompliance are not likely to change be- "Annex B countries") to targets that implied havior. Countries that fail to meet their targets reductions by 2008­12 of some 5 percent of the during 2008­12 must make up for this short- GHG emissions recorded in these countries fall in the subsequent commitment period, in 1990. Countries may either reduce actual plus a 30 percent penalty. A country liable for GHG emissions or enhance the amount of the penalty could fail to ratify the extension, carbon captured in "carbon sinks" (by or insist on raising its emissions limit as a con- sequestering GHG from the atmosphere), for dition of participation. Unlike the World example, through reforestation programs. Trade Organization (WTO) agreement, other The protocol also allows countries to achieve countries are not provided with the means of their emission-reduction obligations together, enforcing compliance (Aldy, Barrett, and to buy emission rights from other Annex B Stavins 2003). countries whose emissions are below the limits, and to receive emission reduction credits for Country and local efforts. Individual coun- sponsoring GHG mitigation or sequestration tries, as well as some local governments, projects in other Annex B countries (Joint have taken steps to limit carbon emissions. Implementation Framework) or in devel- While these have not yet had a major impact oping countries (the Clean Development on the size of total emissions, they do help to 158 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H encourage similar initiatives, provide some momentum to efforts to limit climate change, Figure 5.7 Global trading in carbon emissions has mushroomed and provide useful information on the feasibility of different approaches. Finland, Millions of tons of CO2 equivalent the Netherlands, Norway, and Sweden 800 adopted a carbon tax in the 1990s, and the 700 United Kingdom has imposed a "climate 600 change tax" on electricity generated by using 500 fossil fuels since 2001. 400 In the United States, the California legisla- 300 ture recently passed a law that would cut 200 carbon gas emissions 25 percent by 2020; 100 Oregon has mandated cuts of 75 percent by 2050; 279 cities have signed a commitment to 0 2001 2002 2003 2004 2005 comply with the Kyoto targets; northeast states have set up the Regional Greenhouse Gas Ini- Total emissions traded tiative to control emissions; and 22 states have Emissions traded to comply with Kyoto Protocol Other emissions transactions adopted so-called renewable portfolio stan- dards to encourage renewable energy sources Source: Capoor and Ambrosi 2006. (Rabe 2006). Of course, a host of policies af- fect emissions, including many not designed to contain climate change. For example, high taxes on gasoline can help to reduce gasoline opportunity to develop projects and validate consumption and thus reduce emissions. their reduction credits under the Kyoto Protocol will soon start to close. Carbon trading. The Kyoto Protocol and regional initiatives have created a carbon The way forward market that trades reductions in GHG Because the establishment of institutions to emissions, supported by efforts from the address climate change requires considerable government of the Netherlands and the World lead time, it would be desirable to start Bank (notably through the Prototype Carbon building such institutions immediately. And Fund, which began operations in April 2000). because lack of agreement on international The overall market rose from about 13 million cost distribution remains an important imped- tons of CO2-equivalent in 2001 to 704 million iment to policy implementation, reaching a tons in 2005 (figure 5.7), when its value compromise on this question is a high priority. totaled $11 billion. The value of the market Progress in international negotiations con- continues to rise--it was $7.5 billion in the cerning optimal climate policies would first quarter of 2006 alone. The vast majority strengthen private incentives for energy of transactions are aimed at complying with efficiency and government incentives for the Kyoto Protocol, and the market is appropriate policies. Progress would be en- dominated by the European Union's Emissions couraged by an agreement that the results of Trading Scheme. Developing countries such efforts will be recognized in any future accounted for almost half of global revisions of contractual obligations. transactions in 2005 through the Clean Proposals for a new agreement to succeed Development Mechanism. Given the huge the Kyoto Protocol should be evaluated uncertainties about the post-2012 climate according to several criteria (Aldy, Barrett, policy regime, the volume of project-based and Stavins 2003). The emissions targets transactions may decline, as the window of should reduce climate change to an acceptable 159 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 level. Participation should be as broad as pos- Conclusions and policy sible. The policies should be efficient, either recommendations by maximizing the net benefits to society vian flu, the depletion of marine fisheries, compared with alternatives, or at a minimum, A and climate change are very different issues by representing the least costly means of representing different threats to global welfare. achieving an agreed-upon goal. The obliga- But in some respects, they are similar. All could tions and results of the policies should be involve substantial economic and human costs. viewed as equitable, both across countries In all cases, the risks posed have been intensi- and, given the long-term issues surrounding fied by globalization and the related accelera- climate change, across generations. (A conflict tion in growth and technological progress. And currently exists between different notions of in all cases, the necessary solutions will require equity: the industrial countries are most re- a high degree of international policy coordina- sponsible for climate change and have the tion. No one country can, by itself, stem the rise greatest ability to pay, while developing coun- of GHGs sufficiently to avoid a continued in- tries are likely to be most affected.) Policies crease in global temperatures and potentially should be flexible enough to take account of catastrophic effects. Similarly, cooperation by new information; this is critical given the time all countries is required to contain a potential scale involved and the potential impact of flu pandemic that could result in millions of fa- technological developments on emissions, talities. And ensuring the sustainability of ma- mitigation efforts, and countries' ability to rine fisheries requires cooperation on sustain- adapt. Finally, the design of the rules and the able management and observance by many institutions established must effectively ad- countries of agreed-on fishing limits. dress the substantial difficulties involved in monitoring performance and ensuring compli- ance with treaty provisions. Institutional effectiveness varies Obviously there are important trade-offs from case to case among these goals. Targets that achieve large The effectiveness of the current institutional reductions in GHG emissions may not attract frameworks for addressing these issues varies. sufficient participation, and flexible arrange- International efforts to contain the short- ments may not reflect sufficient commitment term threats of the SARS epidemic and avian to environmental targets. And the criteria for flu virus have been swift and effective, al- judging some targets are subjective. Individu- though avian flu remains endemic in several als and governments may have different views countries and thus a continuing threat. The of what level of climate change is acceptable, generally adequate legal framework governing or how much future generations should pay the management of marine fish stocks is often for mitigation. So, designing an optimal agree- rendered ineffective by inadequate enforce- ment to limit climate change is ultimately a ment and inappropriate incentive systems. political, rather than a technical, exercise. The international institutions required to Negotiations for a successor to the Kyoto confront the longer-term threat posed by cli- Protocol have already started within the frame- mate change have been generally ineffective. work of the United Nations Framework Con- The Kyoto Protocol represents an initial effort vention on Climate Change. Failure to seize to limit GHG emissions, and it has provided this opportunity to come up with an effective valuable experience in the implementation of treaty curtailing the risk of climate change may controls. However, it lacks the participation seriously endanger not only the benefits of of major current and future GHG emitters, achieving the Millennium Development Goals, enforcement of its provisions is problematic, but also the welfare of entire future generations and, in its present form, it is neither an in industrial and developing countries alike. effective nor an efficient response to the 160 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H climate change problem. There is at present no benefits from prevention efforts, so interna- international institution able to coordinate an tional efforts to contain (some) infectious dis- effective response to climate change. eases have been relatively effective. The threat Achieving strong international coordina- to the sustainability of marine fisheries occu- tion to address threats to global welfare is eas- pies an intermediate position: there is little iest where there is a general consensus on the disagreement over the dangers of overexploita- nature of the problem and what to do, where tion of marine fish stocks, while the extent to the threat is immediate, where individuals and which individual government efforts to manage countries have strong private incentives to ad- fisheries generate private benefits varies de- dress the problem in ways that have external pending on the species involved--and particu- benefits, and where the number of countries larly on whether fish tend to migrate to the high that must be involved in negotiations is seas or between exclusive economic zones. limited (table 5.2). The greatest difficulties in achieving effec- Some policy priorities are clear tive international cooperation are presented Climate change. Understanding how the by climate change. Although scientific under- lack of effective international institutions standing of the relationship between GHG impedes an effective response to climate emissions and global warming is sufficient to change focuses attention on policy priorities. justify action, the implications for welfare of Discussions are already under way under the both problems and solutions are difficult to aegis of the UN Framework Convention on forecast. The most severe damages from cli- Climate Change to replace the Kyoto mate change will likely take several decades to Protocol, which expires in 2012, with a more occur, leading to disagreements on the appro- comprehensive and ambitious agreement. priate discount rate to apply to welfare calcu- Meanwhile, it may be useful for the global lations and the equitable division of costs community to start putting in place the among generations. No one country gains pertinent institutions, such as a global system much relief from the threat of climate change for trading emission permits, as well as through its own efforts to control emissions. improved means of monitoring emissions And an effective response requires gaining (particularly in developing countries), which agreement from all major polluters. It is no will allow a rapid implementation of effective surprise that international institutions have policies once these are agreed upon. made little headway, despite the potentially Negotiations over the next agreement must catastrophic costs of failure. By contrast, there take into account the position of developing is general agreement on the short-term threat countries. Since industrial countries are the posed by a flu pandemic, and individuals and major source of the current stock of GHGs individual countries gain substantial private in the atmosphere, there is a compelling Table 5.2 Uncertainty and incentives affect international institutions Degree of Time scale scientific Benefit of country's Number of countries Effectiveness of Threat of threat consensus own mitigation efforts involved in solution institutions Flu pandemic Short-term High High Many High Marine fisheries Medium-term High High/moderate Many or limited Moderate (depends on species) Climate change Medium- to Moderate Limited Many Low long-term Source: Authors. 161 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 argument that they should assume the lion's broader application against groups of viruses. share of the costs. Nevertheless, future growth Because uncertainty concerning use, the large in emissions will occur mainly in developing sunk costs involved, and lack of effective countries. Industrial countries taking on a demand from many potential consumers in larger burden can be reconciled with achieving developing countries limit private investment universal participation through a system of in vaccines, this is an urgent area for public appropriate transfers, for example through investment by the industrial countries. the allocation of emission permits. Individual governments should focus on While international agreement is critical to the distribution of vaccines, arrangements for limiting GHG emissions, individual countries quarantine, financial incentives for reporting need not delay action. A large number of mea- disease, and sanctions for failure to report sures could be adopted to limit atmospheric within their own jurisdictions. Industrial GHG concentrations while simultaneously countries might consider it in their interest to raising current welfare. For example, eliminat- subsidize such activities in developing coun- ing subsidies on fossil fuels could reduce the en- tries, which may lack financial resources ergy intensity of production and thus unneces- adequate to the task. And international sarily high GHG emissions. These efforts could discussions could be useful to provide for ap- also have a substantial role in improving health propriate burden-sharing among the countries by reducing local pollution. For industrial able to assume a portion of such costs. countries, the health benefits from reduced pol- lution may offset a large share of mitigation Marine fisheries. Strengthening the system of costs (see Burtraw and others 2003; Proost and regional fisheries management organizations, Regemorter 2003; Aunan and others 2004; and establishing them where none exist, may McKinley and others 2005). Choosing energy- merit further contributions by industrial efficient technology for the tens of trillions of countries. The UN General Assembly fund to dollars in global infrastructure investment will aid developing countries in implementing the have irreversible impacts on GHG emission Fish Stocks Agreement appears to have gotten paths throughout the century (Grubb 2006a). off to a slow start, with some developing The focus on international coordination is countries calling for increased contributions essential, given the nature of the problem. But and others noting that the fund is underused international negotiations typically proceed at owing to a lack of knowledge by many the slow pace required to achieve consensus, potential recipients (Fiji UN Mission 2006). A while there is an urgent need for action now reduction in fish subsidies, a redirection of to slow the accumulation of GHGs. Further subsidies toward assisting with exit from the delays in addressing climate change would industry, and the financing of general support increase the costs of future, necessary mitiga- to fishing through taxation of the fishing tion efforts and greatly increase the risks of industry would help limit overcapacity and severe damage to global welfare. The scientific overfishing. Limits on the exploitation of consensus is sufficient to demonstrate that the environment at the bottom of the sea prudence lies on the side of addressing climate are sensible until more knowledge has change. Achieving policy consensus is more accumulated on how fishing and other difficult, but it is now urgent. activities affect that environment. Sustainability would be enhanced by imple- Avian flu. Research remains a priority in menting an ecosystem-based approach to combating future pandemics, particularly fisheries management, which focuses on efforts to speed the development of vaccines in sustainable exploitation while safeguarding response to the next mutation of the flu virus, the ecosystem's structure, function, and or--even better--to develop vaccines with productivity. The uncertainties involved in 162 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H determining sustainable levels of fish stocks 2. Subsequent investigations established that the argue for allowing a safe margin of error when disease originated in Guangdong province in China in setting management regimes and catch limits. late 2002, but news of the early cases of the sickness had not been made public by the Chinese government. The need for international 3. In China alone over the last 40 years the human population has increased by two-thirds, while the poul- cooperation will grow try population has expanded more than 10-fold. Similar The scenario presented in chapter 2 envisions increases in both human and animal populations have some acceleration in global growth and trade occurred in other Asian countries (Osterholm 2005). over the next quarter century. A deepening Additional concerns arise from the presence of the of globalization will lead to faster poverty re- H5N1 virus in migratory birds (without showing clinical duction and a general improvement in global symptoms), which can lead to the transmission of the welfare. But continuing globalization will disease between continents, and the virus's ability to adapt to other species (including various mammals). also increase the risks that countries face. 4. The disease outbreaks have been found to be This chapter has highlighted the short-term strongly linked to the cold season, and the behavior of risks from infectious disease, the medium- the disease over the coming months will be critical. term risk of depletion of marine fish re- 5. The one successful case of global eradication was sources, and the medium- to long-term risk smallpox, which succeeded in part because of the nature posed by climate change. Of these, only of the disease: no nonhuman host, potential for effective climate change appears capable of seriously diagnosis and surveillance, ability to interrupt person- to-person transmission, and vaccination (Barrett 2004). derailing global growth over the next quarter 6. Bottom trawling, where the trawling rig is century. The strength of global institutions dragged along the sea floor, can damage vulnerable designed to meet these problems will have ecosystems on the sea bottom. Studies in Australia important implications for the likelihood of indicate that the sea floor ecosystem had not recovered achieving this growth path. from bottom trawling 15 years after an area is closed Many other problems will, to differing to fishing (FAO 2004). The use of explosives has dam- degrees, require a global solution--among aged coral reefs, and poisons have killed nontarget species (Whole Systems 2006). The FAO estimates that them preserving biodiversity, achieving an marine fish discards (fish caught other than the target intellectual property regime that encourages fish and thrown away) total about 10 million metric innovation while limiting excessive monopoly tons per year (Kura and others 2005), although dis- rents, and reducing the transmission of cards declined since the early 1990s (FAO 2004). macroeconomic instability. The countries in- 7. For example, improved ships and freezer facili- volved in each case, the importance of the ties enable ships to stay at sea for long periods. Sonar, risks and benefits, and the scope for interna- satellite navigation systems, depth sensors, and air sur- veillance, combined with detailed maps of the ocean tional action will vary considerably. But the floor, help locate fish and improve the accuracy of net interrelated phenomena of growth, technolog- casting (Parsell 2002). ical progress, and globalization will intensify 8. All subsidies do not threaten the sustainability the need to find cooperative international of fish stocks, and few studies have attempted to link solutions to all of these problems, while the value of subsidies quantitatively to their effect on diminishing the ability of any single country fish stocks (FAO 2000). For example, subsidies to arti- to resolve critical issues on its own. sanal fishing may not raise catch levels enough to en- danger sustainability, and some subsidies already are designed to facilitate exit from the industry. Quantita- tive modeling is extremely difficult owing to the lack of adequate data on subsidies and the multiple causes Notes of changes in fisheries stocks (Tallontire 2004). The 1. The Antonine Plague, either smallpox or measles, impact of subsidies will also depend on the effective- is estimated to have killed 5 million people in the ness of management of fish stocks. second century A.D., and major episodes of bubonic 9. The Convention also provides that the freedom plague occurred in the 6th and 14th centuries, the to fish on the high seas is subject to the general duty latter killing a quarter of Europe's population. to cooperate in conservation and management and to 163 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 maintain or restore populations so as to obtain the of sunlight from the Earth through the use of a giant maximum yield. space mirror (Hall 2005). While considered of little 10. Examples of other approaches, rarely used in practical relevance only a decade ago, geoengineering is regional fisheries management organizations, are a gaining more serious consideration today (Broad 2006). total ban on fishing for several years to allow replen- 17. The existence of various GHGs requires that ishment of fish stocks, restrictions on the capture of they be taxed in proportion to their contributions to females or immature fish (to allow them reproduce), climate change. Similarly, subsidies for alternative and closure of the fishery during spawning season. activities that reduce climate change potential (such as 11. Note that these data include China, which reforestation) should be proportional to their effect. many believe has overstated fish captures (FAO 2004). While easy to formulate, the implementation of this 12. When industrial fishing fleets move close to principle is not a trivial task. shore they can damage the sea-bottom habitat, damage 18. In a world without uncertainty, setting a price local fish nets, and drastically reduce fish species on for emissions is equivalent to setting a quota. If the cost which local craft fishers depend. of reducing emissions is uncertain, the welfare effects 13. Climate change refers to the incremental effect of either setting prices or quantities of emission may of anthropogenic GHG emissions on the average global differ (Weitzman 1974). surface temperature and related changes in weather pat- 19. The present value of benefits from a coordinated terns. The natural greenhouse effect, caused by the solution may be negative for some countries (including pre­Industrial Revolution contents of GHGs in the at- the United States), thus further undermining incentives mosphere, is estimated to raise average global surface for participation (Nordhaus and Yang 1996). While side temperature by some 32 degrees Celsius from what it payments might be envisioned to encourage par- would be without natural radiative forcing, allowing ticipation by those who suffer from a coordinated solu- human life to exist. So far anthropogenic emissions tion, the fairness involved in paying the world's largest of CO2, the most important GHG, have raised atmos- GHG emitter to restrain emissions is problematic. pheric concentration of CO2 from 280 parts per million 20. While the problem of carbon leakage is gener- (ppm) at the start of the Industrial Revolution to ally recognized, there remains disagreement concerning 380 ppm, coinciding with an increase of average global its quantitative importance: alternative model simula- surface temperature by about 0.6 degrees Celsius. tions come to different results as to the amount of 14. Given existing stocks of fossil fuels relative carbon leakage likely to occur in response to a given to current and projected economic growth, the past policy for a given subregion (Burniaux and Oliveira close link between output and fossil-fuel use could con- Martins 2000). tinue for a sufficiently long period to lead to a multiple 21. Emissions from the European Union (EU15) increase in the atmospheric GHG concentrations that countries are estimated at 0.8 percent below 1990 prevailed before the Industrial Revolution. The carbon levels, compared with a target of 8 percent. Japan's contents of estimated fossil-fuel reserves are approxi- emissions are estimated to be 7.4 percent, and Canada's mately five times the current atmospheric carbon 29 percent, above 1990 levels, compared to a target of content (in the form of CO2) and more than 600 times 6 percent (UNFCCC 2006). current annual anthropogenic carbon emissions. 15. These estimates are clearly subject to multiple uncertainties (Grubb 2006b). In addition to the usual caveats for forecasts of growth and population, the References economic and welfare impact of climate change is dif- Aldy, Joseph E., Scott Barrett, and Robert N. Stavins. ficult to measure. Many market effects are not included 2003. "Thirteen Plus One: A Comparison of owing to lack of data. The valuation of nonmarket Global Climate Policy Architectures." Available effects is problematic and can raise ethical issues (life is at www.nicholas.duke.edu/solutions/documents/. typically valued in terms of the willingness to pay to APEC (Asia-Pacific Economic Cooperation). 2000. preserve it, thus leading to high valuations in rich "Study into the Nature and Extent of Subsidies countries and low valuations in poor countries). Tran- in the Fisheries Sector of APEC Member sition costs are typically not included, and the sensitiv- Economies." Available at www.APEC.org. ity of existing systems to transitional rises in tempera- Aunan, K., J. Fang, H. Vennemo, K. Oye, and H. M. ture is largely unknown. Seip. 2004. "Co-Benefits of Climate Policy: 16. These may include, among other ideas, wind Lessons Learned from a Study in Shanxi, China." scrubbers to filter carbon dioxide from the air, "fertil- Energy Policy 32 (4): 567­81. ization" of oceans with iron to encourage growth of Barrett, Scott. 2004. "Critical Factors for the Success- plankton, petrification of carbon dioxide, and deflection ful Provision of Transnational Public Goods." 164 M A N A G I N G T H E E N V I R O N M E N T A L R I S K S T O G R O W T H Background Study for the International Task ______. 2004. "The State of World Fisheries and Aqua- Force on Global Public Goods. Available at culture." FAO, Rome. www.gpgtaskforce.org. Fiji UN Mission. 2006. "Report on the Review Con- Bell, Clive, and Maureen Lewis. 2004. "The Economic ference on the Agreement for the Implementa- Implications of Epidemics Old and New." World tion of the Provisions of the United Nations Economics 5 (4): 137­74. Convention on the Law of the Sea of 10 De- Berrill, M. 1997. The Plundered Seas: Can the World's cember 1982 relating to the Conservation and Fish be Saved? Vancouver, BC: Greystone Books. Management of Straddling Fish Stocks and Bolton, D. A. 2005. "Managing Living Marine Resources Highly Migratory Fish Stocks." United Nations, Multilaterally: Some Threshold Questions." In New York. May. Deep Sea 2003: Conference on the Governance and Gianni, Matthew. 2004. "High Seas Bottom Trawl Management of Deep-Sea Fisheries, ed. Ross Shot- Fisheries and Their Impacts on the Biodiversity ton. Rome: Food and Agriculture Organization. of Vulnerable Deep-Sea Ecosystems: Options for Broad, William J. 2006. "How to Cool a Planet International Action." World Conservation (Maybe)." New York Times, June 27, F1. Union, Gland, Switzerland. Burniaux, Jean-Marc, and Joaquim Oliveira Martins. Grubb, Michael. 2006a. "Framing the Economics of 2000. "Carbon Emission Leakages: A General Climate Change: An International Perspective." Equilibrium View." OECD Economics Department Submission to the Stern Review on the Economics Working Paper No. 242, Organisation for Eco- of Climate Change, HM Treasury and Cabinet nomic Co-operation and Development, Paris. Office, London. Burtraw, D., A. Krupnick, K. Palmer, A. Paul, M. ______. 2006b. "Climate Change Impacts, Energy and Toman, and C. Bloyd. 2003. "Ancillary Benefits Development." Paper presented at the Annual of Reduced Air Pollution in the U.S. from Mod- World Bank Conference on Development erate Greenhouse Gas Mitigation Policies in the Economics, Tokyo, May 30. Electricity Sector." Journal of Environmental Hall, Wayne. 2005. "Climate Change and Geoengi- Economics and Management 45: 650­73. neering." Available at www.globalresearch.ca. Capoor, Karan, and Philippe Ambrosi. 2006. "State Honohan, Patrick, and Luc Laeven. 2005. Systemic and Trends of the Carbon Market." Unpublished Financial Crises: Containment and Resolution. paper, World Bank, Washington, DC. Cambridge, U.K.: Cambridge University Press. CDC (Centers for Disease Control). 2004. "Interim Rec- IEA. 2004. World Energy Outlook. Paris: Interna- ommendationsforInfectionControlinHealth-Care tional Energy Agency. Facilities Caring for Patients with Known or Sus- ______. 2005. World Energy Outlook. Paris: Interna- pected Avian Influenza." Available at www.cdc.gov/ tional Energy Agency. flu/avian/professional/infect-control.htm. IPCC (Intergovernmental Panel on Climate Change). Cunningham, Steve, and Dominique Greboval. 2001. 2001. Third Assessment Report--Climate Change Managing Fishing Capacity: A Review of Policy 2001. Geneva. Available at www.ipcc.ch. and Technical Issues. FAO Fisheries Technical Kura, Yumiko, Carmen Revenge, Eriko Hoshino, and Paper No. 409, Food and Agriculture Organiza- Greg Mock. 2005. "Fishing for Answers: Making tion, Rome. Sense of the Global Fish Crisis." World Resources Ellerman, A. Denny, Henry D. Jacoby, and Annelene Institute Report, Washington, DC. Decaux. 1998. "The Effects on Developing Coun- Lodge, Michael W. 2005. "Improving International tries of the Kyoto Protocol and Carbon Dioxide Governance in the Deep Sea." In Deep Sea 2003: Emissions Trading." World Bank Policy Research Conference on the Governance and Management Working Paper No. No. 2019, Washington, DC. of Deep-Sea Fisheries, ed. Ross Shotton. Rome: Esty, Daniel H., and Maria H. Ivanova. 2002. Food and Agriculture Organization. "Revitalizing Global Environmental Governance: Lomborg, Bjorn, ed. 2004. Global Crises, Global A Function Driven Approach." In Global Envi- Solutions. Cambridge, U.K.: Cambridge Univer- ronmental Governance: Options and Opportuni- sity Press. ties, ed. Daniel H. Esty and Maria H. Ivanova. McKinley, G., M. Zuk, M. Hojer, M. Avalos, I. New Haven, CT: Yale School of Forestry and Gonzalez, R. Iniestra, I. Laguna, M. Martinez, Environmental Studies. P. Osnaya, L. M. Reynales, R. Valdes, and J. FAO (Food and Agriculture Organization). 2000. Martinez. 2005. "Quantification of Local and "Report of the Expert Consultation on Eco- Global Benefits from Air Pollution Control in nomic Incentives and Responsible Fisheries." Mexico City." Environmental Science and Tech- FAO, Rome. nology 39: 1954­61. 165 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Milazzo, Matteo. 1998. "Subsidies in World Fisheries: Fisheries Products." Food and Agriculture Orga- A Reexamination." World Bank Technical Paper nization, Rome. No. 406, Fisheries Series, Washington, DC. Tol, R. S. J. 2002. "New Estimates of the Damage Nordhaus, William D., and Zili Yang. 1996. "A Costs of Climate Change, Part I: Benchmark Esti- Regional Dynamic General-Equilibrium Model of mates." Environmental and Resource Economics Alternative Climate-Change Strategies." American 21 (1): 47­73. Economic Review 86 (4): 741­65. U.K. Government. 2006. "The Stern Review: The OECD (Organisation for Economic Co-operation and Economics of Climate Change." Available at Development). 2005. Why Fish Piracy Persists: www.hm-treasury.gov.uk/independent_reviews. The Economics of Illegal, Unreported and Unreg- UNFCCC (United Nations Framework Convention on ulated Fishing. Paris: OECD. Climate Change). 2006. "National Inventory Osterholm, Michael T. 2005. "Preparing for the Next Submissions 2006" (data on emissions by coun- Pandemic." New England Journal of Medicine try). Available at unfccc.int/national_reports/ 352 (18): 1839­42. Available at www.nejm.org. annex_i_ghg_inventories/national_inventories_ Parsell, D. L. 2002. "High-Tech Fishing is Emptying submissions/items/3734.php. Deep Seas, Scientists Warn." National Geo- Van Vuuren, Detlef P., and Brian C. O'Neill. 2006. graphic News, February 26. Available at news. "The Consistency of IPCC's SRES Scenarios to nationalgeographic.com/news/2002/02/0225_ 1990­2000 Trends and Recent Projections." 0226_fishcrisis.html. Climatic Change 75 (1­2): 9­46. Pauly, Daniel, Villy Christensesn, Sylvie Guenette, Warren, Rachel. 2006. "Impacts of Global Climate Tony J. Pitcher, U. Rashid Sumaila, Carl J. Change at Different Annual Mean Global Walters, R. Watson, and Dirk Zeller. 2002. Temperature Increases." In Avoiding Dangerous "Toward Sustainability in World Fisheries." Climate Change, ed. J. Schellnhuber, W. Cramer, Nature 418 (August 8): 689­95. N. Nakicenovic, G. Yohe, and T. M. L. Wigley. Proost, S., and D. V. Regemorter. 2003. "Interaction New York: Cambridge University Press. between Local Air Pollution and Global Warming Webster, D. G. 2006. "Leveraging Competitive Advan- and Its Policy Implications for Belgium." Interna- tages: The Role of Developing Countries in Inter- tional Journal of Global Environmental Issues national Fisheries Management." Available at 3 (3): 266­86. biology.usc.edu/PDF/WebsterDevcos.pdf. Rabe,Barry.2006."RacetotheTop:TheExpandingRole Weitzman, Martin. 1974. "Prices vs. Quantities." of U.S. State Renewable Portfolio Standards." Pew Review of Economic Studies 61 (4): 477­91. Center on Global Climate Change, Arlington, VA. Weyant, John P. 2000."An Introduction to the Rischard, Jean-François. 2002. High Noon: Twenty Economics of Climate Change Policy." Pew Global Problems, Twenty Years to Solve Them. Center on Global Climate Change, Arlington, VA. New York: Basic Books. WHO (World Health Organization). 2005. Avian In- Schmidt, Carl-Christian. 2002. "Fish Crisis: A Problem fluenza: Assessing the Pandemic Threat. http:// of Scale." OECD Observer 233 (August). www.who.org. Sender, Elena. 2006. "Les retards du plan grippe ------. 2006. "Avian Influenza, Including Influenza A aviare" (Bird flu plan delayed). Science et Avenir (H5N1), in Humans: WHO Interim Infection (March): 57­59. Control Guideline for Health Care Facilities Shotton, Ross. 2006. Deep Sea 2003: Conference on Alert." Available at www.who.int/csr/disease/ the Governance and Management of Deep-Sea avian_influenza/guidelines/infectioncontrol1/en. Fisheries. FAO Fisheries Proceedings 3/2. Rome: Whole Systems. 2006. "Ocean Fish Production." Food and Agriculture Organization. Availableatwww.whole-systems.org/fisheries. html. Sierra, Katherine. 2006. "An Investment Framework World Bank. 2004. "Saving Fish and Fishers: Toward for Clean Energy and Development." Presenta- Sustainable and Equitable Governance of the Global tion at the Dialogue on Climate Change, Clean Fishing Sector." Agriculture and Rural Development Energy, and Sustainable Development, October Department, World Bank, Washington, DC. 3­4. Monterrey, Mexico. ______. 2006. Global Development Finance. Washing- Sturm-Ramirez, Katharine. 2006. Unpublished briefing ton, DC: World Bank. memorandum, St. Jude's Children's Hospital, WWF (World Wildlife Fund). 2001. "Hard Facts, Memphis, TN. Hidden Problems: A Review of Current Data on Tallontire, Anne. 2004. "Trade Issues Background Fishing Subsidies." World Wildlife Fund Techni- Paper: The Impact of Subsidies on Trade in cal Paper. Washington, DC. 166 Appendix Regional Economic Prospects Table A.1 East Asia and the Pacific forecast summary Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 GDP at market prices (2000 US$)b 8.3 8.8 9.0 9.0 9.2 8.7 8.1 GDP per capita (units in US$) 6.9 7.8 8.1 8.1 8.3 7.8 7.2 PPP GDPc 8.9 9.2 9.2 9.3 8.8 8.2 Private consumption 7.3 6.1 7.3 7.7 6.0 6.5 7.3 Public consumption 7.3 5.3 6.3 5.9 4.0 6.3 6.4 Fixed investment 9.7 17.0 11.8 9.6 8.9 11.7 9.0 Exports, GNFSd 12.4 17.8 22.4 17.7 16.1 12.7 11.9 Imports, GNFSd 12.0 17.0 19.4 12.5 13.1 14.0 12.8 Net exports, contribution to growth 1.2 5.3 7.0 9.6 11.4 11.3 11.2 Current account balance/GDP (%) 0.4 3.5 3.4 5.8 7.0 6.4 5.9 GDP deflator (median, LCU) 6.6 3.4 4.2 3.0 2.7 4.7 3.4 Fiscal balance/GDP (%) 0.8 2.5 1.8 1.3 1.1 1.1 1.0 Memo items: GDP East Asia, excluding China 5.9 5.5 6.1 5.4 5.4 5.7 5.9 China 9.5 10.0 10.1 10.2 10.4 9.6 8.7 Indonesia 3.3 4.9 5.1 5.6 5.5 6.2 6.5 Thailand 3.6 7.0 6.2 4.5 4.5 4.6 5.0 Source: World Bank. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. c. GDP is measured at PPP exchange rates. d. Exports and imports of goods and nonfactor services. 167 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.2 East Asia and the Pacific country forecasts Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Cambodia GDP at market prices (2000 US$)b 7.0 10.0 13.4 8.9 6.5 7.0 Current account balance/GDP (%) 3.6 3.9 6.4 13.7 10.7 7.7 China GDP at market prices (2000 US$)b 9.5 10.0 10.1 10.2 10.4 9.6 8.7 Current account balance/GDP (%) 1.5 2.8 3.6 7.1 8.5 7.5 7.0 Fiji GDP at market prices (2000 US$)b 2.1 3.0 5.3 0.7 3.1 2.2 2.5 Current account balance/GDP (%) 3.1 8.3 17.0 17.1 10.3 6.1 1.9 Indonesia GDP at market prices (2000 US$)b 3.3 4.9 5.1 5.6 5.5 6.2 6.5 Current account balance/GDP (%) 0.4 3.5 0.6 0.3 0.8 0.2 0.5 Lao PDR GDP at market prices (2000 US$)b 6.1 6.4 7.0 7.3 6.6 6.9 Current account balance/GDP (%) 8.2 14.3 19.9 14.6 24.9 23.6 Malaysia GDP at market prices (2000 US$)b 5.4 7.2 5.2 5.5 5.5 5.5 Current account balance/GDP (%) 12.9 12.9 15.6 14.8 14.8 15.2 Papua New Guinea GDP at market prices (2000 US$)b 3.9 2.7 2.9 3.0 3.8 4.0 4.0 Current account balance/GDP (%) 2.3 11.2 2.2 13.5 8.0 7.0 4.9 Philippines GDP at market prices (2000 US$)b 3.1 3.6 6.2 5.0 5.5 5.7 6.0 Current account balance/GDP (%) 0.2 4.4 1.9 2.5 2.6 1.8 1.5 Samoa GDP at market prices (2000 US$)b 2.7 1.0 3.1 3.0 3.0 3.5 3.5 Current account balance/GDP (%) 8.4 5.0 4.5 7.3 0.1 0.2 0.0 Thailand GDP at market prices (2000 US$)b 3.6 7.0 6.2 4.5 4.5 4.6 5.0 Current account balance/GDP (%) 1.2 5.5 2.9 1.5 0.2 2.2 2.5 Vietnam GDP at market prices (2000 US$)b 7.0 7.3 7.8 8.4 8.0 7.5 7.5 Current account balance/GDP (%) 4.9 2.0 0.4 2.4 0.1 1.7 Source: World Bank. Note: Growth and current account figures presented here are World Bank projections and may differ from targets contained in other Bank documents. Kiribati, Dem. Rep. of Korea, N. Mariana Islands, Marshall Islands, the Federated States of Micronesia, Mongolia, Myanmar, Palau, American Samoa, Solomon Islands, Timor-Leste, and Tonga are not forecast owing to data limitations. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. 168 R E G I O N A L E C O N O M I C P R O S P E C T S Table A.3 Europe and Central Asia forecast summary Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 GDP at market prices (2000 US$)b 0.2 5.9 7.2 6.0 6.4 5.7 5.5 GDP per capita (units in US$) 0.4 5.9 7.2 6.0 6.3 5.6 5.5 PPP GDPc 0.4 6.2 7.4 5.9 6.5 5.8 5.6 Private consumption 1.2 6.0 8.1 7.9 7.8 6.3 5.8 Public consumption 0.5 2.9 2.3 2.9 3.1 3.0 2.9 Fixed investment 4.6 10.4 12.7 11.7 10.9 8.9 7.9 Exports, GNFSd 3.8 12.7 13.4 7.3 10.0 9.5 9.9 Imports, GNFSd 2.8 15.7 17.7 10.5 12.8 10.6 10.2 Net exports, contribution to growth 0.5 2.2 0.5 1.1 2.5 3.2 3.5 Current account balance/GDP (%) 1.0 0.3 0.9 0.8 0.6 1.4 GDP deflator (median, LCU) 104.7 4.3 6.2 4.0 6.0 5.1 5.0 Fiscal balance/GDP (%) 2.6 0.6 1.4 1.9 2.2 1.5 Memo items: GDP Transition countries 2.6 4.8 6.7 5.5 5.8 5.2 5.2 Central and Eastern Europe 2.2 4.3 5.5 4.6 5.7 5.3 5.3 Commonwealth of Independent States 3.7 7.7 8.0 6.7 7.3 6.4 6.0 Poland 4.5 3.8 5.3 3.4 5.4 5.1 5.2 Russia 3.4 7.3 7.2 6.4 6.8 6.0 5.5 Turkey 3.5 5.8 8.9 7.4 6.0 5.0 5.0 Source: World Bank. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. c. GDP is measured at PPP exchange rates. d. Exports and imports of goods and nonfactor services. Table A.4 Europe and Central Asia country forecasts Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Albania GDP at market prices (2000 US$)b 4.7 6.0 5.9 5.5 5.0 6.0 5.8 Current account balance/GDP (%) 5.6 8.1 5.5 7.8 8.1 7.1 6.5 Armenia GDP at market prices (2000 US$)b 2.6 13.9 10.5 14.0 9.5 8.5 7.5 Current account balance/GDP (%) 6.8 4.5 3.9 4.7 4.6 4.5 Azerbaijan GDP at market prices (2000 US$)b 5.1 11.2 10.2 26.2 22.7 25.7 19.9 Current account balance/GDP (%) 27.8 30.0 1.1 15.1 25.6 34.1 Belarus GDP at market prices (2000 US$)b 1.1 7.0 11.0 9.2 9.3 4.5 3.3 Current account balance/GDP (%) 2.2 5.2 1.5 0.2 3.2 3.9 Bulgaria GDP at market prices (2000 US$)b 0.9 4.5 5.7 5.6 5.6 5.6 5.6 Current account balance/GDP (%) 2.3 5.5 5.8 11.3 12.5 12.0 11.3 Croatia GDP at market prices (2000 US$)b 0.8 5.3 3.8 4.3 4.5 4.0 4.0 Current account balance/GDP (%) 7.2 5.4 6.6 6.7 5.1 5.0 Czech Republic GDP at market prices (2000 US$)b 1.5 3.2 4.2 6.1 6.8 6.0 6.3 Current account balance/GDP (%) 6.4 6.2 2.1 3.0 3.1 3.0 (continued) 169 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.4 (continued) Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Estonia GDP at market prices (2000 US$)b 0.0 6.7 7.8 9.8 9.2 8.0 6.8 Current account balance/GDP (%) 12.1 13.0 11.0 11.8 11.2 10.5 Georgia GDP at market prices (2000 US$)b 7.2 11.1 6.2 8.5 7.5 6.5 6.0 Current account balance/GDP (%) 7.2 8.3 8.4 9.9 11.5 11.0 Hungary GDP at market prices (2000 US$)b 2.1 3.4 5.2 4.1 3.8 2.5 3.2 Current account balance/GDP (%) 5.4 8.7 8.6 7.4 8.0 6.7 6.0 Kazakhstan GDP at market prices (2000 US$)b 2.5 9.3 9.6 9.4 9.0 9.0 8.9 Current account balance/GDP (%) 0.9 1.1 0.9 7.0 2.4 1.9 Kyrgyz Republic GDP at market prices (2000 US$)b 3.2 7.0 7.1 0.6 4.3 5.5 4.8 Current account balance/GDP (%) 5.2 3.4 8.3 11.0 9.8 7.7 Latvia GDP at market prices (2000 US$)b 1.6 7.2 8.5 10.2 9.8 7.5 6.0 Current account balance/GDP (%) 8.2 12.9 12.4 13.5 12.0 11.5 Lithuania GDP at market prices (2000 US$)b 2.8 9.7 7.0 7.5 7.0 6.5 6.0 Current account balance/GDP (%) 7.0 7.7 7.0 8.5 8.4 8.0 Macedonia, FYR GDP at market prices (2000 US$)b 0.3 2.8 4.1 4.0 4.0 4.0 4.5 Current account balance/GDP (%) 3.3 7.7 1.4 3.1 3.9 3.9 Moldova GDP at market prices (2000 US$)b 8.2 6.6 7.4 7.1 3.0 3.0 5.0 Current account balance/GDP (%) 7.1 2.0 9.8 21.2 17.6 9.8 Poland GDP at market prices (2000 US$)b 4.5 3.8 5.3 3.4 5.4 5.1 5.2 Current account balance/GDP (%) 3.5 2.1 4.2 1.4 1.5 1.9 2.4 Romania GDP at market prices (2000 US$)b 0.3 5.2 8.3 4.1 5.8 6.2 6.3 Current account balance/GDP (%) 6.7 5.8 8.2 8.7 11.4 12.9 13.6 Russian Federation GDP at market prices (2000 US$)b 3.4 7.3 7.2 6.4 6.8 6.0 5.5 Current account balance/GDP (%) 8.2 10.2 10.9 9.7 5.2 2.9 Slovak Republic GDP at market prices (2000 US$)b 1.9 4.5 5.4 6.1 6.7 7.1 5.7 Current account balance/GDP (%) 0.9 3.1 8.5 7.2 4.2 3.5 Turkey GDP at market prices (2000 US$)b 3.5 5.8 8.9 7.4 6.0 5.0 5.0 Current account balance/GDP (%) 1.1 3.4 5.2 6.4 8.0 7.5 6.4 Ukraine GDP at market prices (2000 US$)b 7.2 9.4 12.1 2.6 6.0 4.5 5.5 Current account balance/GDP (%) 5.8 10.5 3.1 1.0 3.4 4.1 Uzbekistan GDP at market prices (2000 US$)b 0.2 4.2 7.7 7.0 6.0 4.0 4.0 Current account balance/GDP (%) 8.7 9.9 14.3 17.0 17.2 14.8 Source: World Bank. Note: Growth and current account figures presented here are World Bank projections and may differ from targets contained in other Bank documents. Bosnia and Herzegovina, the Republic of Montenegro, the Republic of Serbia, Tajikistan, and Turkmenistan are not forecast owing to data limitations. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. 170 R E G I O N A L E C O N O M I C P R O S P E C T S Table A.5 Latin America and the Caribbean forecast summary Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 GDP at market prices (2000 US$)b 2.9 2.0 6.0 4.5 5.0 4.2 4.0 GDP per capita (units in US$) 1.4 0.5 4.5 3.1 3.7 2.8 2.7 PPP GDPc 3.5 2.1 5.6 4.3 4.9 4.1 4.0 Private consumption 2.5 2.5 5.3 4.7 5.0 3.8 3.6 Public consumption 1.5 5.9 0.9 3.4 3.3 1.9 1.1 Fixed investment 5.2 3.2 14.6 7.7 9.0 8.1 6.9 Exports, GNFSd 7.4 2.7 12.4 7.8 6.2 5.7 6.8 Imports, GNFSd 8.9 2.0 14.4 11.3 9.0 7.6 7.2 Net exports, contribution to growth 0.6 1.6 1.3 0.6 0.1 0.6 0.7 Current account balance/GDP (%) 2.8 0.5 1.0 1.6 1.9 1.4 1.0 GDP deflator (median, LCU) 10.1 7.2 8.3 9.1 7.3 6.1 6.1 Fiscal balance/GDP (%) 0.1 0.2 0.2 0.7 0.1 0.4 Memo items: GDP Latin Amer. & the Carib., excluding Argentina 2.8 1.0 5.5 3.8 4.6 3.9 4.0 Caribbean 3.1 3.2 2.5 6.5 7.6 5.0 4.8 Central America 3.1 1.6 4.3 3.2 4.5 3.6 3.6 Argentina 3.3 8.8 9.0 9.2 7.7 5.6 4.0 Brazil 2.6 0.5 4.9 2.3 3.5 3.4 3.8 Mexico 3.0 1.4 4.4 3.0 4.5 3.5 3.5 Source: World Bank. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. c. GDP is measured at PPP exchange rates. d. Exports and imports of goods and nonfactor services. 171 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.6 Latin America and the Caribbean country forecasts Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Antigua and Barbuda GDP at market prices (2000 US$)b 3.1 4.9 5.2 5.0 7.1 3.9 4.1 Current account balance/GDP (%) 5.3 10.1 18.7 15.9 20.4 18.4 16.7 Argentina GDP at market prices (2000 US$)b 3.3 8.8 9.0 9.2 7.7 5.6 4.0 Current account balance/GDP (%) 3.1 6.2 1.9 2.7 2.2 1.4 0.9 Belize GDP at market prices (2000 US$)b 4.8 9.4 4.6 3.1 2.6 2.6 3.3 Current account balance/GDP (%) 7.2 20.3 17.6 18.5 18.8 24.9 24.9 Bolivia GDP at market prices (2000 US$)b 3.2 2.8 3.6 4.1 3.1 3.1 3.2 Current account balance/GDP (%) 6.1 0.8 3.5 5.2 5.3 4.0 3.9 Brazil GDP at market prices (2000 US$)b 2.6 0.5 4.9 2.3 3.5 3.4 3.8 Current account balance/GDP (%) 2.1 0.8 2.0 1.9 1.4 1.1 0.8 Chile GDP at market prices (2000 US$)b 5.6 3.7 6.1 6.3 5.0 5.3 5.3 Current account balance/GDP (%) 2.8 1.5 1.7 0.6 3.5 2.7 2.0 Colombia GDP at market prices (2000 US$)b 2.3 4.1 4.8 5.1 4.7 4.2 4.0 Current account balance/GDP (%) 1.9 1.2 1.0 1.9 2.3 3.0 3.7 Costa Rica GDP at market prices (2000 US$)b 5.0 6.5 4.1 5.9 5.0 4.6 4.1 Current account balance/GDP (%) 3.6 5.3 4.6 4.9 5.7 4.1 4.4 Dominica GDP at market prices (2000 US$)b 1.8 0.0 3.6 2.4 3.0 3.0 3.0 Current account balance/GDP (%) 16.3 19.5 23.0 23.2 24.2 24.5 24.0 Dominican Republic GDP at market prices (2000 US$)b 5.9 0.4 2.0 9.3 8.5 5.5 5.0 Current account balance/GDP (%) 3.2 6.3 5.3 0.4 3.2 4.1 3.6 Ecuador GDP at market prices (2000 US$)b 1.3 2.7 7.9 4.7 3.5 3.0 3.0 Current account balance/GDP (%) 2.3 1.7 0.9 0.3 0.7 1.1 2.7 El Salvador GDP at market prices (2000 US$)b 4.2 1.8 1.5 2.8 3.2 3.1 3.1 Current account balance/GDP (%) 2.0 5.1 3.9 4.4 5.7 4.7 4.3 Guatemala GDP at market prices (2000 US$)b 3.7 2.1 2.7 3.2 4.1 4.0 4.0 Current account balance/GDP (%) 4.6 4.2 4.3 4.4 4.1 4.0 3.4 Guyana GDP at market prices (2000 US$)b 4.3 0.6 1.6 3.0 3.5 3.3 3.6 Current account balance/GDP (%) 19.9 6.3 8.9 19.9 26.1 22.3 15.4 Honduras GDP at market prices (2000 US$)b 3.0 3.5 4.6 4.2 4.5 4.5 4.0 Current account balance/GDP (%) 7.7 4.6 5.3 0.5 1.5 1.3 1.2 Haiti GDP at market prices (2000 US$)b 1.7 0.4 3.8 1.5 2.5 2.7 3.0 Current account balance/GDP (%) 1.6 0.4 0.4 0.7 1.2 1.4 1.5 Jamaica GDP at market prices (2000 US$)b 0.7 2.3 0.9 2.0 3.0 3.5 3.0 Current account balance/GDP (%) 2.7 9.4 5.8 8.8 10.4 8.4 5.0 172 R E G I O N A L E C O N O M I C P R O S P E C T S Table A.6 (continued) Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Mexico GDP at market prices (2000 US$)b 3.0 1.4 4.4 3.0 4.5 3.5 3.5 Current account balance/GDP (%) 3.7 1.4 1.0 0.6 0.1 0.2 0.4 Nicaragua GDP at market prices (2000 US$)b 3.4 2.3 5.1 4.0 3.7 4.2 4.6 Current account balance/GDP (%) 28.6 18.1 18.7 18.8 18.1 19.4 19.9 Panama GDP at market prices (2000 US$)b 4.1 4.3 7.6 6.4 6.3 5.7 5.5 Current account balance/GDP (%) 4.8 3.9 7.8 5.2 4.6 5.0 6.2 Peru GDP at market prices (2000 US$)b 3.7 4.0 4.8 6.7 6.6 5.5 5.0 Current account balance/GDP (%) 5.5 1.5 0.0 1.4 1.1 0.5 0.4 Paraguay GDP at market prices (2000 US$)b 1.7 2.6 4.1 3.0 3.2 3.0 3.1 Current account balance/GDP (%) 2.0 2.2 0.3 0.2 0.3 0.4 0.3 St. Kitts and Nevis GDP at market prices (2000 US$)b 4.1 2.1 6.4 4.9 3.7 4.0 4.1 Current account balance/GDP (%) 18.8 51.8 24.4 21.6 21.0 20.0 20.0 St. Lucia GDP at market prices (2000 US$)b 2.4 3.0 4.0 5.4 5.5 3.4 3.3 Current account balance/GDP (%) 11.3 18.6 13.0 25.2 15.3 10.0 10.0 St. Vincent and the Grenadines GDP at market prices (2000 US$)b 2.0 4.5 4.3 4.9 4.3 4.1 4.2 Current account balance/GDP (%) 19.0 15.5 19.4 23.6 24.3 25.0 25.8 Trinidad and Tobago GDP at market prices (2000 US$)b 2.9 13.2 6.5 7.0 12.0 6.2 6.5 Current account balance/GDP (%) 0.2 9.4 15.4 18.9 23.2 17.2 17.3 Uruguay GDP at market prices (2000 US$)b 2.7 2.5 12.3 6.6 5.5 4.4 3.8 Current account balance/GDP (%) 1.5 0.5 0.3 0.5 1.7 2.2 2.5 Venezuela, R. B. de GDP at market prices (2000 US$)b 1.1 7.7 17.9 9.3 8.5 6.0 5.5 Current account bal/GDP (%) 2.6 13.7 12.6 18.1 17.1 12.6 7.6 Source: World Bank. Note: Growth and current account figures presented here are World Bank projections and may differ from targets contained in other Bank documents. Barbados, Cuba, Grenada, and Suriname are not forecast owing to data limitations. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. 173 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.7 Middle East and North Africa forecast summary Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 GDP at market prices (2000 US$)b 3.8 4.4 4.8 4.4 4.9 4.9 4.8 GDP per capita (units in US$) 1.9 2.7 3.0 2.6 3.1 3.0 3.1 PPP GDPc 3.9 4.6 4.8 4.4 5.2 4.9 4.9 Private consumption 3.5 3.7 6.3 4.8 5.0 5.0 6.5 Public consumption 3.6 3.1 2.8 6.2 9.2 5.3 5.2 Fixed investment 3.2 5.9 10.0 5.4 10.1 9.5 3.7 Exports, GNFSd 3.8 3.8 6.2 4.8 6.6 4.7 5.2 Imports, GNFSd 0.9 3.8 12.9 7.2 12.5 8.7 7.2 Net exports, contribution to growth 4.1 0.0 1.9 2.6 4.4 5.7 6.4 Current account balance/GDP (%) 0.5 0.0 2.5 6.6 6.8 3.6 2.3 GDP deflator (median, LCU) 7.7 4.4 6.9 14.5 8.7 4.1 4.8 Fiscal balance/GDP (%) 4.3 0.9 2.4 1.2 0.4 0.1 0.1 Memo items: GDP MENA Geographic Regione 3.1 5.7 5.0 5.3 5.5 5.2 5.0 Resource poor-labor abundantf 4.8 4.0 4.8 4.0 5.0 5.1 5.3 Resource rich-labor abundantg 2.8 5.1 4.9 4.7 4.7 4.6 4.4 Resource rich-labor importingh 2.3 7.4 5.3 6.7 6.5 5.7 5.2 Algeria 1.8 6.8 5.2 5.3 3.0 4.5 4.3 Egypt, Arab Rep. of 4.4 3.1 4.2 4.9 5.8 5.6 5.8 Iran, Islamic Rep. of 2.9 5.0 5.1 4.4 5.8 5.0 4.7 Source: World Bank. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. c. GDP is measured at PPP exchange rates. d. Exports and imports of goods and nonfactor services. e. Geographic region includes high-income countries: Bahrain, Kuwait, and Saudi Arabia. f. Egypt, Jordan, Lebanon, Morocco, and Tunisia. g. Algeria, Iran, the Syrian Arab Republic, and the Republic of Yemen. h. Bahrain, Kuwait, Oman, and Saudi Arabia. 174 R E G I O N A L E C O N O M I C P R O S P E C T S Table A.8 Middle East and North Africa country forecasts Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Algeria GDP at market prices (2000 US$)b 1.8 6.8 5.2 5.3 3.0 4.5 4.3 Current account balance/GDP (%) 3.3 13.0 13.1 21.2 24.2 17.5 15.8 Egypt, Arab Rep. of GDP at market prices (2000 US$)b 4.4 3.1 4.2 4.9 5.8 5.6 5.8 Current account balance/GDP (%) 0.9 4.5 4.3 3.3 1.7 1.5 0.7 Iran, Islamic Rep. of GDP at market prices (2000 US$)b 2.9 5.0 5.1 4.4 5.8 5.0 4.7 Current account balance/GDP (%) 1.2 7.8 0.9 7.5 5.6 2.2 2.0 Jordan GDP at market prices (2000 US$)b 4.9 4.1 8.4 7.3 6.3 5.0 5.0 Current account balance/GDP (%) 4.3 11.6 0.2 18.2 21.6 20.3 16.2 Lebanon GDP at market prices (2000 US$)b 4.9 6.3 1.0 5.5 4.5 2.9 Current account balance/GDP (%) 27.5 23.7 21.7 21.5 23.1 23.5 Morocco GDP at market prices (2000 US$)b 1.6 5.5 4.2 1.7 7.0 3.5 4.5 Current account balance/GDP (%) 1.4 3.5 1.9 2.4 1.2 0.7 0.9 Oman GDP at market prices (2000 US$)b 4.0 1.3 3.1 4.8 6.5 5.5 5.0 Current account balance/GDP (%) 3.7 4.0 2.2 14.6 25.2 19.1 14.4 Syrian Arab Republic GDP at market prices (2000 US$)b 4.1 1.1 3.9 5.1 4.0 3.7 3.5 Current account balance/GDP (%) 1.0 3.4 1.1 4.0 2.5 4.9 6.7 Tunisia GDP at market prices (2000 US$)b 4.3 5.6 6.0 4.2 5.3 5.6 6.0 Current account balance/GDP (%) 4.3 2.9 1.7 1.1 1.2 1.4 1.2 Yemen, Republic of GDP at market prices (2000 US$)b 5.3 3.1 2.6 3.8 3.9 2.5 3.0 Current account balance/GDP (%) 4.3 1.4 2.0 5.0 4.9 8.4 11.5 Source: World Bank. Note: Growth and current account figures presented here are World Bank projections and may differ from targets contained in other Bank documents. Djibouti, Iraq, Libya, and the West Bank and Gaza are not forecast owing to data limitations. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. 175 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.9 South Asia forecast summary Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 GDP at market prices (2000 US$)b 5.0 7.8 8.0 8.1 8.2 7.5 7.0 GDP per capita (units in US$) 3.2 6.1 6.3 6.4 6.7 5.9 5.6 PPP GDPc 5.6 8.0 8.1 8.2 8.3 7.5 7.1 Private consumption 3.8 6.7 6.3 8.2 7.8 7.0 6.3 Public consumption 5.1 4.6 8.4 4.4 5.3 4.2 4.2 Fixed investment 5.8 11.5 8.2 10.9 12.6 12.1 10.3 Exports, GNFSd 9.4 11.5 12.9 19.0 22.3 15.5 13.8 Imports, GNFSd 10.2 11.3 21.9 19.6 23.6 16.9 13.3 Net exports, contribution to growth 2.4 0.4 1.1 1.3 1.7 2.2 2.2 Current account balance/GDP (%) 1.6 1.4 0.8 1.4 2.2 2.5 2.5 GDP deflator (median, LCU) 8.1 4.5 7.6 6.3 8.1 7.4 6.5 Fiscal balance/GDP (%) 7.6 7.8 7.2 7.1 7.1 6.7 6.1 Memo items: GDP South Asia, excluding India 3.9 5.1 6.1 6.9 6.5 6.6 6.4 Bangladesh 4.5 5.3 6.3 6.2 6.7 6.2 6.5 India 5.4 8.6 8.5 8.5 8.7 7.7 7.2 Pakistan 3.4 5.0 6.4 7.8 6.6 7.0 6.5 Source: World Bank. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. c. GDP is measured at PPP exchange rates. d. Exports and imports of goods and nonfactor services. Table A.10 South Asia country forecasts Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Bangladesh GDP at market prices (2000 US$)b 4.5 5.3 6.3 6.2 6.7 6.2 6.5 Current account balance/GDP (%) 0.4 0.3 0.4 0.9 0.9 0.4 0.6 India GDP at market prices (2000 US$)b 5.4 8.6 8.5 8.5 8.7 7.7 7.2 Current account balance/GDP (%) 1.2 1.1 0.8 1.3 2.2 2.5 2.4 Nepal GDP at market prices (2000 US$)b 4.4 3.1 3.8 2.7 1.9 3.7 4.5 Current account balance/GDP (%) 6.3 2.1 2.9 2.2 2.4 3.9 2.9 Pakistan GDP at market prices (2000 US$)b 3.4 5.0 6.4 7.8 6.6 7.0 6.5 Current account balance/GDP (%) 3.7 4.3 0.8 3.1 3.9 4.4 5.3 Sri Lanka GDP at market prices (2000 US$)b 4.7 6.0 5.4 6.0 7.0 6.5 6.0 Current account balance/GDP (%) 4.6 0.6 3.2 2.8 4.9 4.1 3.5 Source: World Bank. Note: Growth and current account figures presented here are World Bank projections and may differ from targets contained in other Bank documents. Afghanistan, Bhutan, and the Maldives are not forecast owing to data limitations. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. 176 R E G I O N A L E C O N O M I C P R O S P E C T S Table A.11 Sub-Saharan Africa forecast summary Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 GDP at market prices (2000 US$)b 2.3 4.2 5.2 5.5 5.3 5.3 5.4 GDP per capita (units in US$) 0.0 1.9 3.0 3.2 3.3 3.3 3.5 PPP GDPc 3.2 3.8 5.4 5.7 5.6 5.7 5.7 Private consumption 1.9 0.6 5.6 5.8 5.3 4.5 4.6 Public consumption 2.9 7.2 5.7 5.7 5.0 5.9 5.9 Fixed investment 3.8 7.7 13.6 9.0 13.6 8.7 8.7 Exports, GNFSd 4.3 7.5 6.0 6.8 5.7 7.2 7.1 Imports, GNFSd 4.3 7.3 9.5 9.2 10.3 7.7 7.7 Net exports, contribution to growth 0.7 1.7 3.0 3.9 5.6 5.9 6.3 Current account balance/GDP (%) 2.1 1.0 0.2 0.8 0.3 0.2 0.9 GDP deflator (median, LCU) 10.0 5.7 6.4 6.7 5.8 4.6 5.0 Fiscal balance/GDP (%) 4.4 2.6 2.4 1.3 1.0 1.2 0.9 Memo items: GDP Sub-Saharan Africa, excluding South Africa 2.6 5.0 5.7 5.9 5.8 6.2 6.1 Oil exporters 2.4 6.7 6.6 7.0 6.9 7.5 7.2 CFA countries 2.5 3.5 5.0 4.3 4.1 3.6 4.4 Kenya 1.7 3.0 4.9 5.8 4.9 5.1 4.9 Nigeria 2.2 10.7 6.5 6.2 4.8 5.1 5.4 South Africa 1.9 3.0 4.5 4.9 4.6 3.9 4.3 Source: World Bank. a. Growth rates over intervals are compound averages; growth contributions, ratios and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. c. GDP is measured at PPP exchange rates. d. Exports and imports of goods and nonfactor services. 177 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.12 Sub-Saharan Africa country forecasts Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Angola GDP at market prices (2000 US$)b 0.9 3.4 11.1 18.7 16.9 22.3 15.7 Current account balance/GDP (%) 6.0 5.0 3.5 8.7 11.9 15.3 13.1 Benin GDP at market prices (2000 US$)b 4.3 3.9 3.1 3.5 4.3 4.2 4.1 Current account balance/GDP (%) 6.8 9.8 7.9 7.3 7.4 7.4 7.4 Botswana GDP at market prices (2000 US$)b 4.4 6.7 4.9 4.0 5.2 4.3 4.1 Current account balance/GDP (%) 8.4 6.0 9.9 14.0 14.0 12.1 9.5 Burkina Faso GDP at market prices (2000 US$)b 3.2 8.0 4.6 7.1 6.5 4.9 5.2 Current account balance/GDP (%) 5.6 12.2 13.2 12.2 6.9 6.5 5.0 Burundi GDP at market prices (2000 US$)b 2.2 1.2 4.8 0.9 5.3 5.7 5.4 Current account balance/GDP (%) 3.4 4.8 8.1 10.5 15.6 14.3 13.7 Cameroon GDP at market prices (2000 US$)b 1.8 4.2 3.6 2.4 4.1 3.9 4.1 Current account balance/GDP (%) 3.6 6.3 3.1 2.0 0.5 0.2 0.2 Cape Verde GDP at market prices (2000 US$)b 5.6 5.0 4.4 5.9 5.8 5.9 5.6 Current account balance/GDP (%) 8.3 11.1 14.6 4.5 9.0 8.6 8.2 Central African Republic GDP at market prices (2000 US$)b 1.7 4.6 1.8 2.8 3.6 3.9 4.3 Current account balance/GDP (%) 4.3 2.2 4.5 2.8 3.1 2.9 3.0 Chad GDP at market prices (2000 US$)b 1.2 14.3 33.2 8.4 3.9 2.8 2.7 Current account balance/GDP (%) 5.5 43.9 3.8 4.1 8.7 7.0 4.8 Comoros GDP at market prices (2000 US$)b 1.8 2.1 0.2 4.2 1.3 2.1 2.7 Current account balance/GDP (%) 6.7 4.1 4.1 4.6 4.7 4.2 3.7 Congo, Rep. of GDP at market prices (2000 US$)b 1.3 0.8 3.6 7.7 6.8 1.1 6.5 Current account balance/GDP (%) 16.5 14.1 20.7 19.6 25.5 25.4 25.6 Côte d'Ivoire GDP at market prices (2000 US$)b 2.3 1.5 1.5 1.8 1.7 2.2 2.7 Current account balance/GDP (%) 4.0 2.0 1.6 0.1 1.7 2.5 2.3 Equatorial Guinea GDP at market prices (2000 US$)b 18.5 14.0 29.4 8.1 8.2 8.3 12.3 Current account balance/GDP (%) 33.0 147.6 23.8 13.4 7.0 8.4 8.6 Eritrea GDP at market prices (2000 US$)b 3.0 2.8 4.5 1.7 1.9 2.4 Current account balance/GDP (%) 11.0 5.9 0.6 1.1 1.7 1.9 Ethiopia GDP at market prices (2000 US$)b 3.8 3.9 12.3 8.7 5.8 5.6 5.5 Current account balance/GDP (%) 0.9 2.6 4.4 7.6 7.7 5.5 4.9 Gabon GDP at market prices (2000 US$)b 1.8 2.2 1.4 2.9 2.7 1.9 2.7 Current account balance/GDP (%) 5.6 9.5 10.9 15.9 21.3 19.7 17.2 Gambia, The GDP at market prices (2000 US$)b 3.0 6.9 5.1 5.0 4.4 3.8 3.6 Current account balance/GDP (%) 4.5 5.7 11.8 12.7 9.1 6.9 5.9 178 R E G I O N A L E C O N O M I C P R O S P E C T S Table A.12 (continued) Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Ghana GDP at market prices (2000 US$)b 3.8 5.2 5.8 5.4 5.6 5.7 5.8 Current account balance/GDP (%) 6.5 1.9 2.7 7.6 7.6 7.1 6.9 Guinea GDP at market prices (2000 US$)b 3.8 1.2 2.6 3.1 4.1 4.7 3.9 Current account balance/GDP (%) 5.7 2.9 5.2 2.9 4.0 3.2 3.1 Guinea-Bissau GDP at market prices (2000 US$)b 1.0 0.6 1.6 2.4 3.8 2.9 3.1 Current account balance/GDP (%) 24.0 10.9 3.1 7.1 5.2 7.8 7.0 Kenya GDP at market prices (2000 US$)b 1.7 3.0 4.9 5.8 4.9 5.1 4.9 Current account balance/GDP (%) 1.6 0.4 2.7 2.2 3.5 5.5 4.7 Lesotho GDP at market prices (2000 US$)b 3.0 3.3 2.7 1.3 1.7 1.8 2.1 Current account balance/GDP (%) 13.3 10.7 2.3 13.4 16.3 17.9 19.6 Madagascar GDP at market prices (2000 US$)b 2.4 9.8 5.2 4.6 4.9 5.3 5.5 Current account balance/GDP (%) 7.8 8.0 9.3 11.2 10.6 9.6 8.3 Malawi GDP at market prices (2000 US$)b 2.6 3.9 5.1 2.1 8.1 4.7 5.1 Current account balance/GDP (%) 8.5 7.9 9.7 8.1 4.7 6.3 5.8 Mali GDP at market prices (2000 US$)b 3.9 7.6 2.3 6.8 5.7 5.0 4.8 Current account balance/GDP (%) 5.7 13.0 6.1 7.9 6.6 5.8 5.4 Mauritania GDP at market prices (2000 US$)b 4.5 6.4 5.2 5.4 17.9 9.8 14.7 Current account balance/GDP (%) 0.6 9.4 19.2 40.0 4.7 1.5 3.7 Mauritius GDP at market prices (2000 US$)b 4.6 4.4 4.7 2.5 3.8 2.9 2.7 Current account balance/GDP (%) 1.6 1.7 1.6 3.9 4.8 6.2 6.4 Mozambique GDP at market prices (2000 US$)b 5.1 7.8 7.5 6.6 6.9 6.5 6.7 Current account balance/GDP (%) 17.2 14.1 8.6 10.8 12.3 13.9 13.7 Namibia GDP at market prices (2000 US$)b 3.4 3.5 6.0 3.3 3.5 3.9 4.1 Current account balance/GDP (%) 4.1 5.4 8.6 8.8 9.0 5.9 2.3 Niger GDP at market prices (2000 US$)b 1.5 3.8 0.6 7.1 4.1 4.0 4.0 Current account balance/GDP (%) 6.9 12.2 12.2 10.8 7.7 7.4 6.6 Nigeria GDP at market prices (2000 US$)b 2.2 10.7 6.5 6.2 4.8 5.1 5.4 Current account balance/GDP (%) 0.7 16.3 17.7 23.1 18.5 17.6 15.2 Rwanda GDP at market prices (2000 US$)b 0.4 0.9 4.0 6.5 5.1 6.1 5.7 Current account balance/GDP (%) 3.5 7.5 2.7 3.6 9.3 10.2 9.7 Senegal GDP at market prices (2000 US$)b 3.0 6.5 5.6 5.5 3.8 5.1 5.2 Current account balance/GDP (%) 6.0 7.6 7.1 9.9 9.9 8.9 8.0 Seychelles GDP at market prices (2000 US$)b 4.3 6.3 2.0 2.3 1.8 0.4 0.9 Current account balance/GDP (%) 7.4 2.3 4.2 13.1 4.1 3.9 3.7 (continued) 179 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 7 Table A.12 (continued) Annual percent change (unless otherwise indicated) Estimate Forecast 1991­2000a 2003 2004 2005 2006 2007 2008 Sierra Leone GDP at market prices (2000 US$)b 5.6 9.3 7.4 7.2 6.9 6.1 6.2 Current account balance/GDP (%) 9.0 7.1 4.3 8.4 6.9 6.0 5.6 South Africa GDP at market prices (2000 US$)b 1.9 3.0 4.5 4.9 4.6 3.9 4.3 Current account balance/GDP (%) 0.2 1.4 3.5 4.2 5.9 5.7 5.3 Sudan GDP at market prices (2000 US$)b 4.9 6.0 5.2 7.9 11.8 10.1 9.2 Current account balance/GDP (%) 6.8 5.4 3.4 11.0 5.1 3.7 3.5 Swaziland GDP at market prices (2000 US$)b 2.8 2.4 2.1 1.8 1.2 1.1 0.9 Current account balance/GDP (%) 2.6 1.7 1.4 1.9 2.5 3.1 4.0 Tanzania GDP at market prices (2000 US$)b 2.7 5.7 6.7 6.9 5.5 7.1 6.8 Current account balance/GDP (%) 12.5 0.6 3.0 4.7 7.3 7.2 7.6 Togo GDP at market prices (2000 US$)b 2.3 1.3 4.6 1.5 2.8 2.7 3.1 Current account balance/GDP (%) 8.5 9.9 7.6 11.0 9.0 7.2 7.0 Uganda GDP at market prices (2000 US$)b 6.2 6.5 5.5 6.3 5.1 5.7 5.8 Current account balance/GDP (%) 7.0 5.1 1.7 2.8 6.5 7.4 7.0 Zambia GDP at market prices (2000 US$)b 0.7 5.1 5.4 4.8 5.1 4.9 4.6 Current account balance/GDP (%) 10.5 8.1 10.3 7.8 6.1 6.9 7.3 Zimbabwe GDP at market prices (2000 US$)b 0.4 10.4 3.8 6.5 3.3 2.9 2.1 Current account balance/GDP (%) 7.5 6.1 19.4 20.6 7.6 8.7 9.4 Source: World Bank. Note: Growth and current account figures presented here are World Bank projections and may differ from targets contained in other Bank documents. The Democratic Republic of Congo, Liberia, Mayotte, São Tome and Principe, and Somalia are not forecast owing to data limitations. a. Growth rates over intervals are compound averages; growth contributions, ratios, and the GDP deflator are averages. b. GDP is measured in constant 2000 U.S. dollars. 180