This paper was prepared under the direction of Nicholas Stern, Chief Economist and Senior Vice President Development Economics (DEC); and Ian Goldin, Director of Development Policy. Halsey Rogers, a Senior Economist, was the lead author. Many other staff from across the Bank contributed. Sections 4 and 5 draw heavily on the experience of the Operations Evaluation Department; the interpretation of the material is the responsibility of DEC. The paper was presented at the United Nations International Conference Financing for Development held in Monterrey, Mexico, on March 18-22, 2002. It is available online at http://econ.worldbank.org/. Comments and questions may be sent to research@worldbank.org.

Foreign aid is increasingly a catalyst for change, and is helping to create conditions in which poor people are able to raise their incomes and to live longer, healthier, and more productive lives. The past 50 years have seen remarkable successes, as well as failures, in development assistance. Better policies in developing countries, together with improved allocation of aid since the end of the Cold War, imply that aid is more effective today at reducing poverty than ever before.

Yet much of the developing world saw little progress over the past several decades, and in some places, such as parts of Sub-Saharan Africa, living standards declined. Moreover, huge challenges remain, such as reversing the Acquired Immune Deficiency Syndrome (AIDS) and malaria epidemics, and finding ways to spur growth and empower poor people in countries with weak institutions, governance, and policies. An estimated 1.2 billion people subsist on under $1 per day, and the majority of the developing world’s population lives on less than $2 per day. (Dollar amounts in this paper are U.S. unless otherwise noted.) Moreover, the world’s population will increase by 2 billion in the next 30 years, with almost all of the growth coming in developing countries. How effective the development community is in helping poor societies respond to these challenges will depend on continued learning and on improvements in the allocation, design, and delivery of foreign aid.

This study, The Role and Effectiveness of Development Assistance: Lessons from World Bank Experience, takes a broad view of the relationship between development experience and official development assistance (ODA) over the past 50 years, with particular emphasis on the World Bank’s experience in recent decades. This report finds that progress in improving well-being has been rapid, if uneven, and that--notwithstanding some significant shortcomings and failures--ODA has often helped to underpin and support success and is becoming more effective in doing so.

The complexity of social and economic change means that the impact of aid cannot be separated easily from other factors. Developing countries themselves bear most of the burdens of development, and rightly claim credit when development succeeds. Assistance works best and can only be sustained when the recipients are strongly committed to development and in charge of the process. In addition, successful projects that draw on foreign assistance in their early stages may later become self-sustaining and serve as models for replication elsewhere, but without any foreign involvement at all. For these and other reasons, the positive impact of ODA can be very large. Nevertheless, identifying cause and effect and attributing outcomes to particular actions is often difficult; furthermore, any excessive attempt to claim credit for the successes of foreign aid can devalue the idea and practice of partnership.

To tackle these analytical problems, this study draws on a variety of types of evidence. We first survey the rapid but uneven progress over the past 50 years and the sources of development success (Sections 1 and 2). Looking more closely at the role of aid, we then considers cross-country statistical analysis (Section 3) and various types of evidence on the impact of the project, program, and sector lending of the World Bank (Section 4). We conclude with a discussion of global initiatives. Country examples and insights from country case studies are included throughout the text. Overall, the study presents persuasive evidence that foreign assistance is an increasingly effective tool for reducing poverty and building a more inclusive world. It also recognizes that there are failures as well as successes, and that we can and should learn from both.

In recent decades, development progress has taken place at unprecedented rates, not only in advanced economies but also in the poorer regions and countries of the world:

• Health. Over the past 40 years, life expectancy at birth in developing countries increased by 20 years. It is likely that the previous 20-year increase in longevity took millennia. The improvement resulted partly from higher incomes and better education, particularly of women and girls, but also in large measure from improved knowledge and understanding about the prevention and treatment of disease, and new programs to share this knowledge and put it into practice.
• Education. Over the past 30 years, illiteracy in the developing world has been cut nearly in half, from 47 percent to 25 percent of all adults. Steady expansion of school enrollments worldwide and increases in education quality made key contributions to this improvement, as did better infrastructure and nutrition.
• Income poverty. The number of people subsisting on less than $1 per day rose steadily for nearly two centuries, but in the past 20 years it has begun to fall. As a result of better and more market-oriented economic policies through much of the developing world--but most importantly in China and India--the number of poor people worldwide has fallen as much as 200 million, even as the world’s population has risen by about 1.6 billion since 1980.

• Since 1965, the per-capita gross domestic product (GDP) of the developing world as a whole has increased by an average of some 2.2 percent per year, more than doubling the income of the average developing-country resident. Again, this is a huge change by historical standards and substantially higher than growth rates achieved by the developed countries in the nineteenth century and most of the twentieth century.
• Since 1990, developing countries’ economies have on average grown faster in per capita terms than those of Organisation for Economic Co-operation and Development (OECD) countries (1.9 versus 1.6 percent). This is largely as a result of improved policies, including a stronger market orientation, which have created better environments for private-sector growth in many countries.

• Programs like Progresa in Mexico and the Bolsa Escola in Brazil have used financial incentives and parental involvement in school management to induce families to keep their children in school, substantially raising school enrollments among the poorest children. Progresa, for example, increased secondary enrollment rates by 8 percent for girls and 5 percent for boys in just four years.
• Targeted action by a global public-private partnership has eliminated the disease of riverblindness from much of West Africa, following the global community’s earlier success in eradicating smallpox from the planet.
• Bangladesh has cut infant mortality in half (from 140 to 71 per 1,000 live births in the past 30 years), reduced fertility sharply (from 7 births per woman in 1970 to 3.2 births in 1999), and achieved near-universal primary enrollment for girls in an environment where they historically faced high barriers. The country’s innovative NGOs played an especially important role in this progress, but outside assistance was also strongly supportive. While the growth rate of per-capita incomes has not been high in Bangladesh, it has increased in each decade since independence.
• Market-oriented reforms in Vietnam and China have helped catalyze domestic and foreign investment there, lifting tens of millions of people out of abject poverty in the former and hundreds of millions in the latter.

These are just a few examples. More systematic evidence presented below and in greater detail in the full paper supports the case that good policies and external support have been the key to progress.

Some regions and countries have grown very slowly or declined in recent decades.

• Most notably, Sub-Saharan Africa (SSA) as a region saw no increase in its per-capita incomes between 1965 and 1999, even with some improvement in the 1990s.
• Although Africa did make steady progress on health and education indicators over much of that period--despite the lack of income growth--the AIDS epidemic has sharply reversed progress on life expectancy in the region. For the region as a whole, it fell from 50 years in 1990 to 47 years in 1999, and several countries have suffered double-digit declines in life expectancy.
• Many of the transition economies of Eastern Europe and Central Asia suffered deep declines in living standards and sharp rises in poverty during the 1990s--although the recession associated with transition has now ended in these countries, and the countries of Central Europe and the Baltics have grown well since the mid-1990s.

The challenge is to extend to all regions and countries the progress that has already improved the well-being of so many people. To do so, the development community must learn from past failures, as well as understand the origins of the successes. Like aid recipients, who have often followed weak policies or allowed institutions to deteriorate, donors also have made mistakes that slowed development.

• Too often during the Cold War, aid allocations were driven by geopolitical aims rather than by poverty-reduction goals. Given the diversity of motives, it is not surprising that some of this aid failed to have the direct effect of spurring growth and reducing poverty.
• Initially, donors placed too much emphasis on the role of what were often isolated projects, neglecting the quality of the overall country environment for growth--a mistake that adjustment (or policy-based) lending was intended to overcome. Prescriptions for reform were too often formulaic, ignoring the central need for country specificity in the design, sequencing, and implementation of reforms.
• Adjustment lending had its own problems. Donors incorrectly believed that conditionality on loans could substitute for country ownership of reforms. Too often, governments receiving aid were not truly committed to reforms. Moreover, neither donors nor governments focused sufficiently on poverty in designing adjustment programs.
• In many countries, donors underestimated the importance of governance and institutional reforms and of social investments as a complement to macroeconomic and trade reforms.

From these cases and others, donors have drawn lessons, and are now showing greater selectivity in lending, shifting resources toward governance and institutions, emphasizing ownership, and making room for diverse responses to local needs. These new approaches and procedures have begun to pay off. However, it is clear that there is still much to learn: for example, on the question of how best to catalyze and support reforms and institution-building in countries with very weak policies, institutions, and governance.

The understanding of development has evolved. It is now widely accepted that poverty reduction efforts should address poverty in all its dimensions--not only lack of income, but also the lack of health and education, vulnerability to shocks, and the lack of control over their lives that poor people suffer. This understanding of poverty in some cases implies different approaches than in the past--for example, an increased focus on public service delivery to vulnerable groups, as well as greater attention to early disclosure of information that poor people can use. This multidimensionality of poverty is embodied in the Millennium Development Goals (MDGs) adopted by heads of state at a United Nations summit in 2000.

Knowledge about what works and what does not work has also improved. Experience has shown that neither the central planning approach followed by many countries in the 1950s and 1960s, nor the minimal-government free-market approach advocated by many people in the 1980s and early 1990s, will achieve these goals. Most effective approaches to development will be led by the private sector, but with effective government to provide the governance framework, facilitation or provision of physical infrastructure, human capital investments, and social cohesion necessary for growth and poverty reduction. Institutional development has too often been neglected in past policy discussions, but is now recognized to be essential to sustained poverty reduction.

While a number of key principles for effective development are clear, there is no single road to follow. Countries must devise their own strategies and approaches, appropriate for their own country circumstances and goals.

Experience and analysis show that countries reduce poverty fastest when they put in place two pillars of development:

• Create a good investment climate--one that encourages firms and farms, both small and large, to invest, create jobs, and increase productivity.
• Empower and invest in poor people--by giving them access to health, education, infrastructure, financial services, social protection, and mechanisms for participating in the decisions that shape their lives.

Understanding of economic growth and its causes has improved. We now understand that creating an investment climate that sustains growth requires progress in a number of areas: macroeconomic stability and trade openness; governance and institutions (including a good education system, an effective legal and judicial system, a professional bureaucracy, a strong and well-regulated financial sector, and vigorous competition); and adequate infrastructure.

This improved understanding does not mean that all growth challenges have been solved. One major issue is the appropriate sequencing and selection of policy and institutional reforms. No poor country has the capacity to move forward with equal vigor on all these fronts at once, so it will be important for the country, with external support, to focus on identifying and grappling with the main obstacles to growth. A second challenge is consistency: many countries manage to achieve growth spurts of several years, but find it very difficult to achieve the two or three decades of consistent growth necessary for sustained poverty reduction.

Poverty reduction depends heavily on sustained economic growth. On average across countries, income distribution does not worsen during periods of economic growth, so that on average the incomes of poor people rise at a similar rate to those of wealthier people. Countries that grew rapidly in the 1990s--such as China, India, Vietnam, and Uganda--managed to reduce the share of their people in absolute poverty by 5 to 8 percent per year.

While growth is essential, it is also important to recognize that these are just averages, and that there is a good deal of variation across countries, regions, and groups. Countries can accelerate their reduction of income poverty by acting to ensure that poor people have the tools necessary to participate in growth; in other words, they can promote pro-poor growth. Of special importance here are health and education, as they also have a direct impact by reducing poverty in key non-income dimensions. Other services that equip poor people to participate include infrastructure and financial services, which often fail to reach them. Furthermore, participation of poor people in the decisions that most affect their lives--for example, parental involvement in school management, or participation by beneficiaries in project design--is a key part of fostering the involvement of poor people in the growth process.

Policies and investments aimed directly at reducing non-income dimensions of poverty can be highly effective. Countries can accelerate health and education progress far beyond what would result simply from economic growth. Without targeted measures, economic growth only slowly improves health and education. On the other hand, progress on these dimensions can serve as a major stimulus to growth.

Globally, we see the powerful effects of policy in the dramatic reductions in infant mortality, which has fallen steadily at each level of income as a result of improvements in technology, knowledge, and policies and institutions. For example, the average country with $8,000 per capita income in 1950 (measured in 1995 dollars) would have had an infant mortality rate of 45 per 1,000 live births; by 1995, an average country at the same income level would have had an infant mortality rate of just 15 per 100 live births, a reduction of two-thirds.

The development progress that we have seen in many countries can be attributed primarily to actions by the countries themselves: specifically, improving the investment climate and investing in people.

Macroeconomic stability and openness have improved throughout the developing world over the past two decades. The median inflation rate of developing countries was cut in half between 1982 and 1997, from about 15 to 7 percent. Average tariff rates have also declined sharply in all developing regions. In South Asia, for example, the unweighted average tariff fell from about 65 percent in 1980-85 to about 30 percent in 1996-98; and in Latin America and the Caribbean, from about 30 percent to under 15 percent.

A group of countries that has integrated most quickly with the global economy--thanks to greater openness and improved investment climates--has grown very rapidly. This group that showed the strongest advances in trade integration, accounting for some 3 billion of the developing world’s 5 billion people, saw per-capita incomes increase by a remarkable 5 percent per year in the 1990s. Even with fast-growing China excluded, the average was 3.5 percent. By comparison, per-capita growth in the high-income economies was only 2 percent. Integration, it should be stressed, is not about applying a simplistic liberalization formula. Both China and India are part of this group because of their rapid trade growth and strong movement toward more open markets, even though their economies are not among the most open in the developing world in terms of import liberalization.

Complementary actions by rich countries are essential. One key area is in trade policy. Trade protection in the rich countries, particularly in agriculture and textiles, blocks the exports and undermines the competitiveness of many countries. Poor countries bear the brunt of this protection, as they face tariffs twice as high on average as those applied to the products of rich countries. Protection greatly reduces the growth rewards to reform in poor countries.

The economic decline in much of Sub-Saharan Africa stems in part from events beyond African countries’ control, including large and persistent declines in the prices of commodities that they export and high country risk perceptions that deter investment. But many African countries compounded the external problems by failing to promote good policies and institutions. Although policies have improved sharply in recent years, governance and institutions remain a major challenge in most of these countries.

Countries that have not grown rapidly--in Africa and elsewhere--have often failed to make progress on key features of the investment climate. For example, they may have achieved macroeconomic stability but not social stability; or they may have lowered trade barriers but not built the basic infrastructure necessary for international trade; some have done none of these things. As we have learned from experience, macroeconomic stabilization and liberalization are important steps, but they are not generally enough by themselves to promote lasting development.

Nevertheless, there are countries in which the rewards to reform have been far less than predicted. These countries have made many of the reforms advocated by external parties, but have either seen little growth response or have not seen growth translate into rapid poverty reduction. In some cases, geographic challenges have hindered growth; in other cases, a history of poor policies has led investors to doubt the permanence of recent improvements. In such cases, the development community should work to help sustain these reforms and increase their returns.

Resource transfer is an important part of development assistance, and the quantitative analysis in this section will focus largely on assessing its effects. But finance is only one of the instruments used to support development, and in some situations, it is not even the most useful one. Development assistance also includes analysis, advice, and capacity-building. Many of the country boxes in this paper highlight the importance of these nonlending tools, especially in situations--such as early in a reform era--when finance is not likely to contribute to poverty reduction. The mixture of instruments to be used in any particular country depends on the specifics of that country’s needs and capacity.

Really successful development assistance will have effects that reverberate far beyond the confines of the project itself, either because the ideas in the project are replicated elsewhere, or because the intervention has helped institutionalize new approaches. Levels of development assistance are small relative to other financial flows and to the scale of the challenge at hand. Development aid totaled about $54 billion in 2000; this was only one-third as much as foreign direct investment in developing countries ($167 billion), which itself was only a small fraction of total investment (nearly $1.5 trillion). This underscores the point that when aid makes a major difference in the fight against poverty, it does so through demonstration effects or improvements in institutions, not simply through resource transfer.

Furthermore, successful development strategies and actions generally depend on strong country ownership, as well as good partnership among donors; both make it difficult, and even counterproductive, for any external actor to claim full credit for a reform or project. For these reasons, this paper makes use of several types of evidence--cross-country statistical analysis, program and project evaluations, case studies, and analyses of global programs--in trying to assess the role and effectiveness of development assistance.

Because each country is unique, the role of aid can be understood best through careful analysis of individual countries. In the body of this paper, detailed country examples are included in each section. In the interests of brevity, this executive summary presents some snapshots of aid in action at the country level before proceeding to the cross-country statistical analysis.

Uganda, Mozambique, China, Vietnam, India, and Poland are all examples of countries where, within the past two decades, policy and institutional reforms have sparked rapid development. In all cases, the country and its government have been the prime movers for reform. Development assistance from many sources supported these transformations. The examples below highlight the ways in which the World Bank assisted in reform and development, while recognizing that countries themselves are responsible for their achievements and that the Bank is only one among many development partners.

• China. The past twenty years of reform in China have contributed more to poverty reduction than any other growth episode in history. Unquestionably, this process of growth and development was driven by China itself. The composition, sequencing, and timing of reforms were designed at home, and built on China’s existing strengths in such areas as literacy and basic health. At the same time, support from outside helped make reform happen and contributed to the structure of the reforms. In the early stages of market-oriented reform, the World Bank provided advice on laying the foundation for the private investment and productivity growth that has buoyed the country’s remarkable progress. The Bank provided the government with the first in-depth overall analysis of China’s economic problems, and it helped China engage with the outside world (again, on the country’s own schedule) through advice on liberalization, exchange-rate unification, and port modernization. The Bank’s rural development projects and analytical work complemented strongly these improvements in the overall business environment by targeting poverty where it was most prevalent--in the countryside. The Bank, as both a knowledge and lending institution, thereby made a significant supporting contribution to the massive reduction in rural poverty: from 34 percent of the rural population in 1985 to just 18 percent in 1998.
• India. Throughout the 1960s and 1970s India was weakly integrated into the international economy and relied heavily on planning and licensing; as a result, economic growth and poverty reduction were unimpressive. Growth accelerated in the 1980s but was based in large measure on unsustainable public spending and foreign borrowing. With the entry of a reformist government in 1991, the Bank provided support for trade and other reforms to stabilize and open up the economy. Over the past five years, the Bank has supported India’s decentralization process, working closely with state, local, and municipal governments committed to reform. Powerful demonstration effects are beginning to emerge.
• Mozambique. Mozambique is a recent example of successful post-conflict reform. The country emerged in 1992 from a long civil war, which--combined with a socialist experiment--had left the country one of the poorest in the world. Since then the World Bank has helped the government to design and implement exchange rate reform, trade liberalization, financial liberalization, and privatization. In this more stable and open environment, GDP has grown at an average rate of 8.4 percent--in part due to revitalized agricultural growth and to increasing exports, which had been stagnant for a decade. The private sector has responded: foreign direct investment (FDI) grew some 500 percent between 1992 and 2001. Today, the focus has shifted towards two areas of continued weakness: strengthening the social sectors (Mozambique is struggling against the AIDS epidemic), and reforming judicial and tax systems.
• Uganda. Uganda’s new government in the mid-1980s inherited a country that was devastated by years of conflict and economic mismanagement. Starting with advice, the Bank helped the government learn from the comparative experience of Ghana and other countries and helped it design and implement key measures on fiscal adjustment, exchange rate reform, and trade liberalization. Aid and the conditionality associated with Bank-supported adjustment lending helped generate policy reforms in the late 1980s and early 1990s, a period during which multilateral assistance from the Bank and other lenders was particularly important. Since that period, Uganda has achieved a remarkable recovery: it has increased private investment, reversed capital flight, increased external trade, and privatized commercial public enterprises. It has made great strides in primary education, with several million additional children attending school during the first year of a Bank adjustment loan. Uganda has also reversed income poverty sharply, from 56 percent in 1992-93 to 35 percent by the year 2000.
• Poland. Poland was the first country to emerge from the transition recession in 1992 and has since maintained an average GDP growth rate of 3.7 percent, the highest among transition economies. Positioned to gain access to the European Union (EU) in 2004, Poland has led the way in many reforms, often taking major risks. Outside assistance has also helped. From the early phases of Poland’s economic transformation, the Bank provided advisory and financial support. Activities included aiding macroeconomic reforms, supporting the creation of an institutional and regulatory framework, helping with the restructuring and privatization of industries, upgrading infrastructure with private-sector participation, and helping to restore Poland’s creditworthiness. Finally, the Bank also helped to improve public understanding of the government’s economic strategy.
• Vietnam. Vietnam has also moved strongly to reform its economy and reduce poverty over the past dozen years, beginning when it was still politically and economically estranged from major donors and therefore could not receive large-scale aid. The Bank began to provide advice to Vietnam in 1989, at a time when the country’s disastrous economic policies had produced a crisis of hyperinflation, falling economic activity, and mass exodus of economic migrants. Although it did not provide finance until 1993, the Bank advised the government on stabilizing the macroeconomy, opening to foreign trade and investment, and reforming property rights. As reforms took hold, the Bank later shifted its focus to infrastructure and primary education. The results have been remarkable: the income poverty rate was cut from 58 to 37 percent in just six years.

These six country provide recent examples of Bank-supported progress, but the past 50 years of development experience also provide examples--such as the Republic of Korea and Botswana--that illustrate how effective aid can be in supporting reform. Korea, for example, progressed from borrowing from the International Development Association (IDA), the Bank’s soft-loan facility for the poorest countries, to borrowing from the International Bank for Reconstruction and Development (IBRD), the lending arm for middle-income countries, and finally to borrowing solely in private markets. Korea is also an example of a country that, with donor support, built both pillars of development: it invested heavily in education and human development while also greatly improving the environment for growth and entrepreneurship.

It is important to note that impressive development results do not depend on reaching all goals simultaneously. Each of the countries listed here continues to face major development challenges, whether in governance or institution-building or capacity development. However, these examples show just how strong the returns can be to moving in the right direction, and how important it is to help countries that are committed to making this movement. Of course, the World Bank has also been involved in many countries, particularly in Sub-Saharan Africa, where results have been less impressive. Its role is to be a long-term partner, working consistently to support countries through challenging circumstances.

The statistical evidence shows that large-scale financial aid can generally be used effectively for poverty reduction, where reasonably good policies are in place. The World Bank and other donors have acted on these findings by tailoring support to local needs and circumstances. Thus the balance of support has moved toward providing large-scale aid to those that can use it well, and focusing on knowledge and capacity-building support in other countries.

By this criterion, donor financial assistance is targeted far more effectively at poverty reduction than it was a decade ago. At that time, Cold War geopolitics was still exercising a heavy influence on aid allocations, and too many recipient economies were poorly run, often suffering from excessive state intervention in economic activity and poor governance.

• In 1990, countries with worse policies and institutions--as assessed by Bank diagnostics--received $44 per capita in ODA from all sources (multilateral and bilateral), while those with better policies received less: only $39 per capita.
• By the late 1990s, the situation was reversed: better-policy countries received $28 per capita, or almost twice as much as the worse-policy countries ($16 per capita).

As a result, the poverty-reduction effectiveness per dollar of overall ODA has grown rapidly.

• In 1990, a one-time aid increase of $1 billion allocated across countries in proportion to existing ODA would have permanently lifted an estimated 105,000 people out of poverty; but by 1997-98, that number had improved to 284,000 people lifted out of poverty. (Poverty is defined here as living on less than $1 per person per day, adjusted for cross-country differences in living costs.) In other words, the estimated poverty-reduction productivity of ODA nearly tripled during the 1990s. And it must be emphasized that movements of people above the poverty line represents just one impact of the aid, which also helped increase income and other dimensions of development throughout the economy.

Yet despite the improvement in aid effectiveness, donor countries did not respond with an increase in aid flows; instead, aid levels fell.

• Aid flows dropped substantially over the 1990s in real terms, and by 2001 were 20 percent below the 1990 level in inflation-adjusted terms.
• With the growth in incomes in the rich countries over the 1990s, aid levels expressed as a share of donors’ GNP fell even more sharply--from 0.33 percent in 1990 to 0.22 percent in 2000.

IDA funds are relatively well focused on poor countries that have shown they can use aid effectively.

• As with ODA in general, too often in the past IDA allocations have been driven by considerations other than poverty reduction. And even where poverty has been the focus, the Bank has often been overly optimistic about the prospects for reform, thereby contributing to misallocation of aid.
• Nevertheless, even in 1990, much more IDA funding went to the good-policy countries ($4.7 per capita) than to the poor-policy countries ($2 per capita). By the late 1990s, targeting had improved still further: good-policy countries now receive $6.5 per capita, compared with $2.3 per capita in poor-policy countries.

As a result, an additional $1 billion allocated in the same way as existing IDA funds in 1997-98 would, by this measure, have lifted an additional 434,000 people permanently out of poverty. This means that IDA is not only 60 percent more efficient than it was in 1990, but also 50 percent more efficient than overall ODA. Again, the effect on numbers crossing the $1 per day poverty line is just one effect of additional IDA funding. IBRD (nonconcessional) lending also goes primarily to countries with good policies and institutions.

Well-targeted aid (such as IDA funding) helps to attract private investment, by improving the investment climate. As a result, it has very high overall returns.

• Much IDA lending is targeted at improving the investment climate, by supporting improved policies and building the governance, institutions, and infrastructure necessary for private-sector productivity.
• As a result, every dollar of IDA funding on average leads to an increase in gross investment of nearly two dollars. In good policy environments, aid also increases foreign direct investment substantially--by 60 cents for each dollar of aid.
• This "crowding-in" effect means that the returns to IDA funding far exceed the poverty-reduction returns. It is estimated that the overall return--in terms of the income growth spurred by IDA--may be as high as 40 percent.

Adjustment lending was developed as a necessary corrective, after distorted macroeconomic environments had been shown to undermine the effectiveness of project-level assistance. Early experience with adjustment lending was uneven and revealed problems: lending was not sufficiently selective, with too much money going to countries with poor reform and implementation records, and the design and implementation paid insufficient attention to social costs, institutions, and the dynamics of reform. Since then, adjustment lending has improved in targeting and design. It is now generally an effective way of providing strong reform support in countries with adequate policy and institutional frameworks. By the 1990s, adjustment lending was earning higher internal Bank evaluation ratings (as judged by the independent Operations Evaluation Department) than investment lending. And attention to the distributive impact of reforms has improved, although further learning is still necessary.

At the society-wide level, early evidence suggests that the PRSP process is increasing both ownership (through participatory discussion of reform alternatives) and poverty focus of development programs in low-income countries. Although it is too early to gauge outcomes on the ground, full PRSPs are so far associated with a 20-percent increase in spending in poverty-related areas. A similar shift in the balance of spending is associated with the Highly Indebted Poor Country (HIPC) debt relief programs developed since the mid-1990s and led by the World Bank and International Monetary Fund (IMF).

Bank operations have served in general to increase the economic productivity of borrowers. The minimum economic rate of return (ERR) expected from Bank-financed projects is 10 percent. Although there is considerable variation and there have been failures, actual results have been substantially better, reaching an average of 16 percent in the 1980s and rising to 25 percent in the 1990s. Bank operations also make a difference in the areas of health and education.

The Bank is the world’s largest external funder of education, having provided a cumulative $30 billion for education projects. It also is the world’s largest external funder of health programs, with new commitments of $1.3 billion a year for health, nutrition, and population projects. These projects had significant results in themselves and were often replicated outside of the original project area. Although it is not possible to quantify this overall effect, country experiences are indicative. In China, for example, a project in the late 1990s helped raise the proportion of households using iodized salt from 40 to 89 percent, reducing iodine deficiency in targeted age groups from 13 to 3 percent. Ultimately, this will result in average gains of 10 to 15 points in children’s IQ levels in affected communities. In Mali, a project in the early 1990s raised the share of children fully vaccinated from 0 to 24 percent. And beyond these direct effects, economic growth supported by the Bank often indirectly improves social indicators, although countries need to strengthen this link by prioritizing health and education for poor people.

Bank lending generally encourages good economic performance. Reviews of adjustment lending show that in the 1990s, it supported governments in maintaining their efforts in social areas and in poverty focus. Developing countries receiving Bank adjustment lending in 1990-97 maintained and even increased social expenditures, on average, more frequently than countries without such lending.

The outcomes of Bank lending have improved steadily. Project outcomes have improved sharply over the past decade. Despite the growing complexity and more demanding nature of its development agenda, the Bank has increased its share of projects rated as satisfactory from well below 60 percent in the late 1980s to above 80 percent today.

These improvements in outcomes come as a result of attempts to learn from experience and adapt to changing circumstances. In part due to increases in private capital flows and changing government priorities, as well as due to changing understanding of the Bank’s role, the Bank has shifted the sectoral composition of its lending. It has shifted resources away from direct infrastructure lending (which fell from three-quarters of Bank lending in the 1960s to one-third in the 1990s) and toward the social sectors (now one-fifth of the total). This shift has been made possible in part by the increased role of the private sector in some areas of infrastructure; in those areas, the Bank now focuses its efforts on facilitating the environment for improvements in infrastructure, including through regulatory reform.

The Bank has also increased both the quantity and quality of policy-based lending. As a result of greater selectivity, adjustment operations achieved their objectives in more than 80 percent of cases in the 1990s--up from 60 percent in the 1980s--and the share with a poverty focus rose sharply. The same shift towards social spending has occurred in the programs associated with debt relief for HIPC; indeed, this is a principal goal of these programs.

Other improvements include:

• Stronger quality management, by identifying key features of successful projects and applying quality controls at earlier stages of design and implementation.
• Sharper country focus. Bank assistance is provided through the framework of country assistance strategies. These are now developed in consultation with governments, and increasingly with NGOs, the private sector, and other development partners.
• Tighter social and environmental standards.
• Improved fiduciary performance, through country assessments of financial management and through procurement and fiduciary safeguards at the project level.

Development is a risky business. To develop ideas, the Bank has to take some risks, and some failures are inevitable. What is essential is to learn from those failures. The Bank has done so: for example, by launching a successful Higher Impact Adjustment Lending (HIAL) initiative in Africa in 1995 to improve design and country selection in line with lessons from evaluation; and by changing its approach to the power sector in Africa.

The Bank builds on success by scaling up initiatives that work, such as the high-return dairy cooperative program (Operation Flood) in India. This vast organization grew out of a single, small cooperative society. By 1996, 10 million farmer-members were supplying an average of 11,000 metric tons of milk per day through 55,000 village cooperative societies, and this program and associated policy changes increased the annual growth rate of Indian dairy production from 0.7 to 4.2 percent. In this case, external support from the Bank and others helped to encourage government policy changes that grew out of a homegrown initiative.

The Bank in the late 1990s strengthened its approach to managing by results, and also moved strongly to increase the sharing of knowledge. In both areas, there has been progress, but there remains considerable room for further improvement.

Global development challenges--such as the spread of infectious diseases, the challenge of building an international trade and financial architecture, loss of biodiversity, deforestation, and climate change--cannot be handled solely by individual countries and therefore require multilateral action. Such action is typically most effective when linked to country efforts. The Bank has contributed actively to these global programs, through financing, through advocacy, and through alignment with its country programs. Two examples illustrate the far-reaching potential and very large returns of such global initiatives.

• The West African Onchocerciasis Control Program, a collaborative effort of multilateral agencies, governments, NGOs, and the private sector. Since it was launched in 1974, the program has eliminated the scourge of riverblindness from 11 countries in West Africa. As a result, it has prevented an estimated 600,000 cases of blindness and added 5 million years of productive labor to the 11 countries’ economies.
• The Consultative Group for International Agricultural Research (CGIAR), a network of research centers that has created and promoted crop improvements in developing countries over the last 30 years, improving productivity and nutrition, and reducing rural poverty. These centers have produced more than 500 varieties of grain now planted in poor countries and have helped to increase average yields in target grains by 75 percent over three decades.

In recent years, the obvious need for such programs and their strong potential led the Bank to become involved in a growing number of global actions. This in turn led to a greater awareness of the need to maintain focus and tighten priorities. Since 2000, the Bank has moved to make its global programs more strategic by establishing a small number of priority areas for action. This focus has helped increase the resources available for high-priority programs, such as combating AIDS and preventing conflict.

Well-allocated foreign aid has been an effective means of supporting poor countries and poor people in their efforts to improve their lives; and with improved allocation and better design and delivery of assistance, aid is more effective today than ever before.

The international development community has recognized that assistance is most effective when recipient countries are the primary drivers of their own reforms and institutional development. Financial aid benefits most the countries that have shown they can use it well. Advice and capacity-building comprise an important complement, especially in the poorest countries and in those that do not have the capacity to absorb significant financial transfers. Advice is often more acceptable and credible when linked to resource transfer.

Since the end of the Cold War, there has been a strong shift in the direction of better allocation of aid, and better tailoring of assistance to country needs and circumstances. Improvement in allocation and the increase in the number of poor countries that are putting in place the governance, institutions, and policies to promote rapid, market-driven, pro-poor growth mean that aid is more effective today than ever before. Yet in the past 10 years, a period when many rich countries experienced remarkable growth, aid allocations have fallen. In inflation-adjusted terms, aid to poor countries was some 7 percent less in 2001 than in 1990. The past record of development progress, and the great efforts the developing countries are now making to develop themselves, provide the evidence that additional aid resources could be used effectively to reduce global poverty.

Research staff in the Development Economics Vice Presidency of the World Bank prepared this paper for the United Nations International Conference on Financing for Development held in Monterrey, Mexico, on March 18-22, 2002.¹ It has an ambitious purpose: to analyze in reasonably concise form the changing roles and effectiveness of development assistance during the past 50 years, with particular attention to the past two decades and to the experience of the World Bank. Such a paper cannot possibly be comprehensive. Rather, the authors attempt to describe broadly how the goals and forms of development assistance have changed over time. They conclude that as a result of these changes and, perhaps more important, of improvements in the policies, institutions, and governance of developing countries, aid is more effective at reducing poverty today than ever before.

The past 50 years have taught the international development community--developing countries, donors, and the international financial institutions--a great deal about development and poverty reduction. We now understand better what the goals of development should be and how to attain them. We have learned from both successes and mistakes, in particular about markets, governments, and institutions, and how they interact. This learning--a common and ongoing effort of the entire development community--has led to improved development performance overall and at the World Bank in particular.

At the same time, there is much that we still do not know. Perhaps most important, we do not understand fully how to help improve institutions and governance, especially in the poorest countries where the needs are greatest. And we are still learning how best to deal with pressing cross-border issues that threaten development, such as disease, environmental problems, and political instability.

In recent decades, development has progressed at historically unprecedented rates.²

For all the political, economic, and social disruptions of the second half of the twentieth century, it was also a period of unprecedented progress in living standards worldwide. Better technology, policies, and institutions not only spurred rapid growth in the advanced economies, they also made possible substantial improvements in the lives of poor people throughout much of the developing world. Progress has been uneven, to be sure. Yet key indicators of well-being testify to striking progress overall.

• Health. Over the past 40 years, life expectancy at birth in developing countries has increased by a remarkable 20 years--or about as much as had been achieved in all of human history before the 1960s. The improvement resulted partly from higher incomes and better education, particularly of women and girls, but also in large measure from improved knowledge and understanding about the prevention and treatment of disease, and new programs to share this knowledge and put it into practice.
• Education. Over the past 30 years, illiteracy in the developing world has been cut nearly in half, from 47 percent to 25 percent of all adults. Steady expansion of school enrollments worldwide and increases in education quality made key contributions to this improvement, as did improvements in infrastructure and nutrition.
• Income poverty. Absolute income poverty--defined as people subsisting on less than $1 per day--rose for much of the past two centuries, but over the past 20 years it has begun to fall. Thanks to better economic policies through much of the developing world--but most importantly in China and India--the number of poor people worldwide appears to have fallen (perhaps by up to 200 million), even as the world’s population rose by 1.6 billion.³ Similarly, a two-century decline in the estimated share of poor people in the global population has accelerated since 1950. In the 1990s alone, the share fell from an estimated 29 percent to 23 percent.⁴

There are various reasons why assessing the contributions of development assistance is inherently difficult. First, responsibility for development progress will always lie primarily with the developing country itself. No outside donor is able, or indeed has the right, to compel a country to follow policies likely to promote development. Nor would it have the local knowledge necessary to do so even if it had the influence. Country ownership of the reform program and development strategy is essential. That is, as the evidence in following sections will show, the country must be committed to and involved in shaping development strategies and projects.

This country ownership is one of several factors making it impossible to assess with any precision the effects of development assistance. If the country takes the lead in reform and institution-building, then by definition the external actor--whether a bilateral donor or a multilateral like the World Bank--plays only a supporting role and should not claim too much of the credit. As discussed in Section 3, other factors also make attribution difficult: the need for partnership across donors, as well as the tendency for the most effective projects to have benefits that extend far beyond the narrow confines of the project itself.

Moreover, levels of development assistance are small relative to other financial flows and to the scale of the challenge at hand. Development aid totaled about $54 billion in 2000; this was only one-third as much as foreign direct investment in developing countries ($167 billion), which itself was only a small fraction of total investment (nearly $1.5 trillion). Similarly, although the World Bank is the world’s largest external provider of assistance in the education sector, it typically provides less than $2 billion in direct assistance for education each year.⁵ By comparison, annual public spending on education in the developing world totals more than $250 billion. Given this discrepancy in scale, even if the World Bank were to greatly increase its lending in the sector, its effectiveness would have to come primarily through catalyzing institutional development and policy change in education, rather than through resource transfer alone.

This means that the direct effects of assistance (for example, in terms of income increases or reductions in mortality) will often be swamped by other factors. Successful development assistance necessarily will have effects that reverberate far beyond the confines of the project itself--either because the ideas are replicated elsewhere, or because the intervention helped to institutionalize new approaches. Yet when the positive effects of aid do spill beyond the bounds of the specific project or policy intervention, the effects become much harder to measure.

Despite these analytic limitations, there is strong evidence that outside assistance is often a powerful force for growth and poverty reduction, provided that the recipient country is committed to using outside resources well. This paper assembles several types of evidence supporting this conclusion. Each of these approaches has methodological problems, and none of them is above challenge. Nevertheless, all four provide evidence that assistance has made a difference, and together they provide a basis for optimism:

• Cross-country statistical analysis of the effects of aid, which shows that aid has accelerated growth when allocated correctly, and furthermore, that the allocation of aid by the World Bank and other donors is improving (Section 3).
• Project- and sector-level analysis of specific interventions, drawing primarily on World Bank experience, which shows high returns to development assistance--both in terms of income-focused cost-benefit analysis and in terms of other indicators of human development (Section 4).
• Country case studies, which show that external assistance has played an important role in supporting, cementing, and often helping to shape reform efforts that have delivered massive poverty reduction in recent years, in countries as diverse as Vietnam, Uganda, China, India, Poland, Mozambique, and Bangladesh (boxes in various sections).
• Evidence on global programs, which shows that global efforts--such as investment in agricultural research and programs to halt communicable diseases--often have been an important complement to country-specific efforts (Section 5).

These positive results have depended heavily on learning. The World Bank and other development agencies have adapted their approaches over time, in response not only to changed circumstances but also to lessons learned through experience, research, and evaluation. Some of these are lessons from successes, such as those listed above. But the Bank has also learned from failures. For example, it has learned that project lending in poor policy environments typically has much lower returns than in countries with good policies, and that loan conditionality does not alter the policies of recipient countries that are not already committed to change for their own reasons.

This paper focuses on lessons learned and the results that development assistance has already achieved. This is not to suggest there is nothing left to learn. To the contrary, there is much that the global development community and the Bank still do not know. For example, we know too little about how to help countries improve governance and how to support the creation of effective institutions. We also know too little about how to spark reform and growth in the poorest-performing, low-income countries--those that have been mired in a cycle of poor institutions and policies, economic stagnation, and often conflict. The development community recognizes that existing modes of aid and development cooperation have not worked well in these environments. The World Bank, like other development organizations, is now investigating fresh approaches to helping the desperately poor people in these countries.⁶

Despite the progress made in the past 50 years, an immense poverty challenge remains. Some 1.2 billion people still live on less than $1 per day, and the challenge will grow as the population of the developing world increases by another 2 billion in the next 30 years. To address a challenge of these dimensions, aid will need to have effects far beyond the value of the money alone. This means that aid must build the frameworks for private economic activity and social improvements--ensuring that its effects go far beyond any individual project--and it must contribute to greater capacity and greater knowledge. Continued learning is essential to these aims.

Ideas about policies and institutions, as well as the policies and institutions themselves, drive development. So understanding how those ideas have evolved, and how they can be changed, is crucial to understanding how we can contribute more effectively to development.

Development thinking has evolved continually over the past 50 years. In response to the lessons of experience and analysis, development practitioners have adapted their approaches to promoting development, and even the goals of development work. We have learned that strategies that seemed obvious to many at some point--for example, both the heavily statist and minimal-government free-market approaches--have had to be reconsidered and changed as part of a continuous learning process. This is one reason why a careful and measured look at experience is so important.

Early postwar aid was focused on reconstruction of the war-torn economies of Europe and Japan--a task that it contributed to with considerable success.⁷ For example, five of the World Bank’s first six loans went to countries of Western Europe, and the first four were explicitly for reconstruction. The poor countries of the world were not the first priority, and the focus was on raising production and income, rather than on broader notions of development.

With rapid progress in post-war reconstruction, development assistance began to focus instead on raising incomes in what came to be called the developing world. At first, the goal was largely confined to raising aggregate national incomes. With the recognition that population growth rates vary sharply, so that aggregate income did not necessarily give a clear picture of changes in living standards, attention turned to per capita incomes. Soon, however, with increased understanding of issues related to income distribution, simply raising average per capita incomes also was recognized as too limited a goal. By the 1960s and 1970s attention focused on the twin problems of growth and income distribution, and increasingly on the basic needs of poor people. Reduction of income poverty became a greater priority within the international financial institutions, as well as for governments.

In recent years, the goals of development have come to embrace the elimination of poverty in all its dimensions--income poverty, illiteracy, poor health, insecurity of income, and powerlessness. A consensus is emerging around the view that development means increasing the control that poor people have over their lives, through education, health, and greater participation, as well as income gains. This view comes not only from the testimony of poor people themselves, but also from advances in conceptual thinking about development.⁸ It is clear that the various dimensions of poverty are related, and income growth generally leads to strong progress against the non-income dimensions of poverty as well. It is also clear that direct action to improve these other dimensions can accelerate the reduction of both income and non-income poverty.

In summary, the objectives of development have evolved in response to changing circumstances and a deeper understanding of poverty. This evolution has made development assistance more complex and challenging. At the same time, it has focused the development community’s attention on our essential mission: making it possible for poor people to improve their lives.

1.2 Approaches to development: The roles of states, markets, and institutions ⁹

The development community’s understanding of the most effective way to achieve development objectives has also evolved over time, with the accumulation of evidence and experience. Approaches that appeared at the time to be both correct and obvious have been undermined by experience and closer analysis; in the same way, our current ideas will no doubt give way to others as experience accumulates and thinking evolves. This surely reminds us to beware of simplistic solutions or "silver bullets." Perhaps the most important question on which our understanding has deepened over the past decades is this: what are the respective roles of governments and markets in spurring development, and how do institutions fit into the picture? At the risk of oversimplification, we can identify at least three major phases in the evolution of our answers to these questions.

In practice, we recognize that there is a continuum of approaches, both in developed and developing countries, and that the phases described here do not match precisely the evolution of thinking in any particular region. Instead, they are intended to capture the broad shifts in the thinking of the development community and development practitioners. It is also the case that successful countries throughout this period have seen both state and market play positive roles. With those caveats, this broad-brush portrait can nevertheless provide a useful context for a discussion of development assistance, by suggesting where that assistance is most likely to be effective.

The 1950s and 1960s were a period of great confidence in government. Development practitioners and thinkers trusted government both for its intentions and for its ability to make economic progress happen, whether in the richer or poorer countries. Development thinking focused on market failures, which were especially prevalent in developing countries and seemed to provide a strong rationale for state intervention. The private sector was thought to be too uncoordinated, too poorly developed, and too focused on private interests to allow it to serve as the locomotive for growth. And in Africa, newly independent countries searched for a post-colonial model of development and a strengthened leadership role for the national state. In many countries around the world, the confidence in government was reflected in the heavy role of central planning and in the relatively closed (import substitution) trade policy.

This state-led approach had some initial development successes. Leading economies of Latin America, where state economic management did not completely crowd out the private sector, grew rapidly for decades under the import-substitution model. And even in some "tiger economies" of East Asia, industry managed to grow and become more productive behind high trade barriers, thanks to otherwise good economic management. Nevertheless, the costs of state economic control became clearer over time. State planners were not omniscient: they could not possibly acquire all the information needed to make decisions that reflected both efficiency considerations and people’s differing preferences.

Worse, governments revealed themselves to be collections of interests rather than disinterested and benevolent "social planners." Even had they been effective in their role as social planners, government officials would not have been able to create the entrepreneurial dynamism essential for sustained development and change. Behind protective barriers, firms in many countries (for example, India and Mexico, to name just two) became less efficient as they focused on obtaining government favors rather than improving productivity. Finally, fiscal and macro instability rose with the oil price shocks of the 1970s and early 1980s, contributing to the debt crisis and revealing the weaknesses in the statist model.

As a result of the disappointment with the state-led approach, the 1980s and early 1990s saw a strong reaction that stressed the primacy of markets in development. This reaction was a necessary corrective in many ways: it refocused attention on production efficiency and market signals, and it inspired the move to lower trade barriers as a means of spurring productivity. Macro stability and balanced fiscal accounts were seen as fundamental building blocks for development and became early priorities for reform. This period saw substantial improvements in both macro stability and openness through much of the developing world.

At the same time, this school of development thinking also failed to address key points. Once countries began to achieve macro stability and greater openness, it became clear that these were necessary but not sufficient for growth. In particular, the free-market view tended to neglect the institutional foundations of effective private markets. The importance of institutions was underscored by major shifts: the economic decline in the countries of the former Soviet Union; the continued growth in China, a country that moved forward with market-oriented reforms without excessive disruption of institutional foundations; and, later in the 1990s, the financial crisis in East Asia, to which institutional weaknesses contributed heavily. Furthermore, even as it performed a useful service by spotlighting government failure, the free-market reaction had minimized very real problems of market failure that are prevalent in the developing world. As a result, growth performance fell short of expectations in many parts of the developing world.

Recent years have seen a greater recognition in the policy debate of the complementarities between markets and governments. Clearly, experience shows that the private market economy must be the engine of growth; but it shows also that a vibrant private sector depends on well-functioning state institutions to build a good investment climate and deliver basic services competently.

This view of complementarities draws heavily on what we have learned in the past two to three decades in the more successful cases of income growth, such as East Asia and Chile. It also draws on learning from the transition process in the former Soviet Union, where a lack of institutional development combined with excessively optimistic expectations led to extremely disappointing development outcomes and demonstrated clearly the importance of a sound state in providing the environment for growth. The role of institutions has come through more strongly than it did in earlier views of development, and particularly than it did in the policy debate in the 1980s and early 1990s. As Section 2.3 shows, countries that have combined institutional improvements with market-oriented policy reforms and greater engagement with the world economy saw their per capita incomes grow in the 1990s at the very rapid pace of 5 percent per year.

This understanding of states, markets, and institutions has strong conceptual foundations, and it is based also in experience with what works in development. There is a strong and growing consensus about the value of such an approach. Nevertheless, there is much that the development community has yet to learn. The World Bank is committed to continued learning and to constant improvements in our overall approach to development. For example, while institutions have now taken a central role in the World Bank’s strategic approach to development, we are keenly aware that we know far too little about just how to build them.¹⁰

Our understanding of what development means has evolved. Poverty reduction efforts should address poverty in all its dimensions--lack of income, but also the lack of health and education, the vulnerability to shocks, and the lack of control over their lives that poor people suffer. This multidimensionality of poverty reduction is embodied in the Millennium Development Goals.

We have also learned more about the most effective broad model for development:

• Evidence from past successes and failures suggests strongly that neither the more statist approach of the 1950s and 1960s nor the more minimal-government free-market approach that dominated policy debate in the 1980s and early 1990s will achieve these goals.
• Effective approaches will be led by the private sector, but with effective government to provide the governance framework and to ensure the provision of physical infrastructure and human capital investments necessary for growth and poverty reduction. In fact, to set state and market against each other is to miss the central question: how can they best complement each other to promote growth and reduce poverty?
• A public-private development partnership is essential, especially in the area of health and education. Institutional development has too often been neglected in the development policy debate, but is now recognized to be essential to sustained poverty reduction.

Development depends on two pillars, which together support sustained growth and poverty reduction:

• Countries must build a good investment climate--an environment in which the private sector will invest and produce efficiently, in a way that generates jobs and productivity growth. The investment climate consists of all the factors that most influence private-sector decisions: macroeconomic stability and openness, governance and institutions, and infrastructure. The "private sector" should be understood not only (or even primarily) to include large firms and multinationals, but also farmers and small and medium enterprises (SMEs).
• Countries must empower and invest in poor people, so that they can participate in growth. We know that sustained growth is essential for poverty reduction, so the investment climate focus is itself a tool for poverty reduction. At the same time, governments need to target poverty more directly, notably by equipping poor people with the tools necessary to contribute to growth, such as education and health, and by giving them access to infrastructure and financial services. People are empowered when they are given the ability to shape their own lives, whether through greater capabilities or through participation in decisionmaking. Direct actions by government, international organizations, and NGOs are necessary to help this happen.

Many countries have built up these two pillars in recent decades and have seen the rewards: rapid growth and poverty reduction. Although too many other countries continue to fall behind economically, a majority of the developing world’s population lives in countries that have grown rapidly and are closing the gap with the rich countries. Even the countries that have stagnated economically have, for the most part, seen material improvements in social indicators, such as health and education measures.

Development assistance has helped accelerate this progress. It aims at helping countries build both pillars: improving the investment climate (through building the factors that contribute to investment and growth), and empowering people (through education, health, and social protection). Section 2 explicates the context for development assistance, by laying out in greater detail the channels through which assistance can affect growth--that is, the factors that give rise to growth and poverty reduction. It also summarizes the success and failure of countries at generating growth. Sections 3 through 5 below show that these efforts have had large rewards.

Approaches: Understanding sources of growth ¹¹

To increase incomes, countries need to improve one or both of two fronts:

• They can rely on factor accumulation: that is, they can accumulate more of the factors of production, such as physical capital (through investment) and human capital (through higher levels of education and health among the population).
• They can increase what economists call total factor productivity (TFP)--the efficiency with which these factors produce output. TFP represents both the quality of the policy and institutional environment, on the one hand, and knowledge and technology on the other.

In practice, virtually all sustained growth processes involve some combination of factor accumulation and increased productivity, with the respective shares of the two varying across countries and time periods. In any event, the real world usually defies a clean breakdown into these two factors: the question of how effectively inputs are used enters not just into productivity, but also into the accumulation of capital. Many development interventions are aimed at accelerating both factor accumulation and productivity growth.

Our understanding of the sources of growth of GDP and productivity is based on several types of research: cross-country regressions, growth decompositions, case studies, and microeconomic studies of the sources of productivity growth. All these approaches have their shortcomings. For example, regressions and growth decompositions are ways of assembling statistically aggregate indicators and trying to identify the most important explanatory factors. Yet while these methods reveal associations between statistical variables, they cannot prove that one factor causes another. Moreover, they place excessive emphasis on a simplistic aggregate production function model of accumulation and production. Much of the growth comes from reallocation of resources to more dynamic sectors; from different ways of organizing productive activity; and from intensification of entrepreneurship. None of these are illuminated by the aggregate production function approach.

The results from country case studies do not suffer from these weaknesses, but they may be too country-specific to provide lessons by themselves for other countries. Nevertheless, together these methods sketch out a good picture of what matters in spurring overall income growth. While no single variable stands up as a determinant of growth in all analyses of growth processes, certain factors appear to be particularly important; these are discussed below.

Broadly, it is clear that private-sector growth is essential to long-term development. While the state can have some success in financing or inducing factor accumulation over the medium term, it is far less successful at inducing productivity growth when it takes a lead role in the economy. Even when a state-dominated economy manages large quantities of investment, the quality of that investment is often low. Moreover, employment provides the main long-term route out of poverty, but the vast majority of the world’s poor workers are employed in the private sector, not the public sector. Moreover, few of these private-sector jobs are in large firms; most of the world’s poorest people work in small firms, most notably small farms. For these reasons, any development strategy must have at its core a focus on building a good climate for private-sector investment, productivity, and employment creation by all firms, small and large.

Given this background, which factors most affect long-term economic growth and poverty reduction? Although different analysts will have somewhat different lists, most would agree on many of the factors listed below. One way to organize these factors is to group them in terms of the major elements of the investment climate that are amenable to direct action by governments: macro factors such as macroeconomic stability and trade openness; governance and institutions, including basic service delivery in education; and infrastructure. Geography is an additional factor that cannot be directly influenced by policy but can have important effects on growth and development.

Unstable fiscal accounts and high inflation undermine the private sector by reducing predictability and discouraging investment.¹² The slowing of growth rates in much of the developing world after the 1960s and early 1970s stemmed in part from macroeconomic instability, which reduced confidence in government and in its ability to maintain a stable climate for productivity, jobs, and growth. While macro stability on its own is not enough to jumpstart growth and development, it is generally a prerequisite for sustained progress.

International trade and investment ¹³

There remains some debate about the degree of trade liberalization that is advisable for poorer developing countries--debate fueled by the observation that open domestic markets were not a precondition for rapid growth in the East Asian tigers, nor indeed historically in much of Europe and North America. Yet it is also clear that most recent periods of rapid growth have also been periods of increased engagement with the international economy. In the 1990s, for example, developing countries that had fast export growth also had, on average, 1 percent higher annual growth in GDP than those with slow export growth; in the 1980s, the difference was even greater.¹⁴ The domestic investment climate and the actions of the domestic private sector are the key drivers of development, but international engagement can play an important complementary role. Even in historically closed economies such as China, which was far from completely open during the early reform period, high growth was fueled in part by increased access to international capital and markets. Where local conditions for entrepreneurship and growth are good, exposure to international competition increases the productivity of domestic firms and lowers input costs for downstream users, while foreign direct investment brings with it new production and process technologies, organizational capacity-building, and marketing networks.

Many of the early analyses in the empirical growth literature found that the level of investment in physical capital plays a role in determining growth rates.¹⁵ But subsequent empirical research, together with the anecdotal experience of countries that achieved high investment rates but slow productivity growth (such as the Soviet Union), has shifted attention from quantity to quality and productivity of investment. Factor accumulation by itself leaves much to be explained in growth regressions, and the evidence is that total factor productivity growth--which captures quality improvement--explains a major share of most episodes of long-term economic growth.¹⁶ The lesson is that we need to focus on the investment climate: the constellation of underlying factors causing growth of productive investment.¹⁷

It is widely thought that an educated and healthy workforce also contributes to growth, with widespread basic education (primary and secondary) being especially important.¹⁸ Not all empirical analyses find such an effect. One reason that the effects of education have been hard to pin down empirically is that the quality of education matters as much as the quantity, and yet we are much better at measuring quantity than quality. Another reason is that, like physical capital, human capital may be relatively unproductive from a societal standpoint if a weak investment climate inhibits development.¹⁹ But while the case is not fully established empirically, strong basic education and reasonable levels of health have been a precursor to many development successes, and recent analyses suggest strongly that additional education does indeed spur development in the typical country. Finally, as discussed below, the evidence on the substantial returns to education at the microeconomic level is unambiguous: education lifts people out of poverty, raising their earnings by some 5 to 10 percent per year of schooling.

Stable and effective government institutions, respect for property rights, equal treatment under the law, the absence of bureaucratic harassment, a lack of corruption, and protection from organized crime all matter for growth. Investment and productivity depend on predictability, which in turn hinges on confidence that government will not act opportunistically or capriciously. The "soft infrastructure" of an effective legal and judicial system is critical, whether for achieving economic growth, empowerment of poor people, or security. Good governance also reduces the costs faced by producers. Sound supervision and regulation of financial institutions decreases the costs of capital to businesses and contributes greatly to macroeconomic stability. Indeed, it has been argued that institutional quality and governance are the underlying variables that drive all of the other growth-enhancing factors.²⁰

Vibrant competition is important in spurring productivity growth. Interventions that increase the extent of the market, including good domestic infrastructure, reduce costs and improve the selection of goods and services by widening the number of competitors. The dynamic effects of such changes, which work through the increased productivity growth they spur, may greatly exceed the one-time benefits to consumers of accessing goods from new suppliers at lower prices.²¹ Competitive pressures can come through various means. Openness to imports is often the most effective source of competition--and the one that requires least administrative capacity on the part of government--but other important sources have included well-functioning competition authorities and government-mandated "export tournaments."²² Ensuring that firms can both enter and exit markets is also an important element of competition.

Physical infrastructure also matters a great deal in terms of both investment climate and empowerment. Transportation and communications infrastructure connect markets and people, both domestically and internationally.²³ Better infrastructure thus diminishes the effects of geography, reducing costs to producers and improving the reach of government and private services for consumers. In Sri Lanka, for example, the arrival of telephone service in rural areas increased farmers’ share of the price crops sold in the capital city from 50-60 percent to 80-90 percent.²⁴ In Peru, households with access to modern infrastructure (water, sanitation, electricity, and telecommunications) had income growth about 45 percent higher than households without these services.²⁵ Availability of affordable and reliable energy supplies is another important spur to production, while good water and sewage infrastructure improves health and environmental outcomes. How government can best provide infrastructure, or facilitate its provision by the private sector, is a key question for development assistance, as are the questions of the prioritization and sequencing of infrastructure investment.

Geography seriously impedes growth for some countries, making development much more difficult.²⁶ If a country is landlocked, mountainous, and surrounded by poor neighbors, or if its population centers are in remote areas, it may encounter additional difficulties in developing domestic markets of efficient size, engaging in international trade, or acquiring technology from abroad. In such cases, it will be especially important to build effective infrastructure links, to improve transportation and communications both domestically and internationally. Regional integration and trade-creating customs unions may also be important in overcoming geographic barriers. Ecological fragility is another geography-linked barrier to development: ecological stresses may most directly affect poor people, and these stresses too require specific policy and institutional responses.

Although our understanding of the importance of these factors has grown over time, to put them together in a way that yields sustained growth remains a daunting challenge. One major challenge for governments is to decide where to focus their efforts as they strive to make the conditions for growth as favorable as possible. Strategies have to be determined in each country context, but it is clear that administrative capacities of low-income governments are typically so limited that an assessment of where they should be focused is essential: the government simply cannot push ahead effectively on all fronts at once. At the same time, sequencing is necessary. It is by now widely recognized that the East Asian financial crisis of 1997-98, which exacted a substantial cost in poverty and lost output, stemmed in no small measure from financial and capital-market liberalization that proceeded before the appropriate regulatory safeguards were in place. But while some sequencing problems are easier to identify, finding the best sequencing of steps in a particular country context is a great challenge, and it remains an area where our knowledge needs to expand further.

Another major challenge is sustaining growth. Rapid growth episodes of a few years or a decade are not uncommon. For example, countries that successfully emerge from civil war often experience relatively rapid economic rebounds for several years.²⁷ What has been much less common is sustained rapid growth over a period of decades, which is what is necessary to eliminate absolute poverty.²⁸

The need for consistency underlines the importance of attaining sustained productivity growth. Only a portion of growth is driven by increases in physical and human-capital intensity of production, which can be difficult to sustain over long periods. Countries also need rapid growth in productivity.

The World Bank and its partners--United Nations agencies, bilateral donors, and governments--have committed to a common set of poverty-reduction results, including the Millennium Development Goals agreed to by governments at major conferences in the 1990s. The goals are accompanied by numerical targets expressed in terms of changes between 1990 and 2015: reducing the share of people living in poverty worldwide by half; reducing infant and child mortality by two-thirds; reducing maternal mortality by three-quarters and improving access to reproductive health services; and halting the increase in incidence of communicable diseases (AIDS, malaria, TB) and reducing malnutrition. What do we know about the factors underlying the reduction of poverty, in both its income and social dimensions?²⁹

Economic growth is essential for sustained progress on poverty reduction. Countries that have reduced income poverty the most are those that have grown the fastest, and poverty has grown fastest in countries that have stagnated economically. Between 1992 and 1998, for example, the share of the population in poverty fell an average of 5 to 8 percent annually in fast-growing Uganda, India, Vietnam, and China.³⁰ In Nigeria, by contrast, per capita consumption fell 16 percent between 1992 and 1996, and the poverty share increased by half, from 43 percent to 66 percent of the population.³¹

Some have expressed fear that growth alone cannot be relied upon to result in significant poverty reduction in developing countries. Evidence shows that income distribution has not changed on average in periods of growth in the typical country, and that therefore overall growth has meant that the incomes of the poor have increased (again, on average) proportionately. In China, inequality did increase with reform; but the increase in inequality was an inevitable feature of the improvement in the incentive structure, which led to growth and poverty reduction. Indeed, growth was so strong that poverty fell sharply despite worsening income distribution. In other cases--Uganda, for example--income distribution improved at least modestly with reforms and growth.³²

Nevertheless, it is the case that countries with better income distribution see growth translate into faster poverty reduction. At the same rate of GDP growth, a country with highly equal distribution (that is, one with a Gini coefficient of 0.30) will see poverty fall twice as fast as a highly unequal country (Gini of 0.60).³³ In addition, the evidence suggests that greater inequality of important assets, such as land and education, may retard society-wide growth.

It is also the case that groups of poor people will experience reform and growth differently. A large increase in the income of one group may be offset by a smaller increase or even decline in the income of another group.³⁴ This underlines the importance of ensuring that there is adequate social protection in place as a complement to structural adjustment measures. Social protection helps build broader support for action, and it enables individuals to take risks involved in entrepreneurship. And social protection is not just an instrument for achieving growth: it also targets poverty directly, by reducing the income vulnerability that poor people identify as one of the defining elements of a life in poverty.³⁵

Once we recognize that poverty is about more than income, we see that there are other determinants of poverty reduction beyond growth. Social indicators--health and education--improved far faster in the twentieth century than we would have expected, given the rate of income growth. Most countries have made major progress in increasing educational attainment and health outcomes by targeting these goals directly, and by applying new knowledge and technologies, rather than just waiting for the effects of income growth to improve these indicators. At every level of income, infant mortality fell sharply during the twentieth century. For example, a typical country with per-capita income of $8,000 in 1950 (measured in 1995 dollars) would have had, on average, an infant mortality rate of 45 per 1,000 live births. By 1970, a country at the same real income level would typically have had an infant mortality rate of only 30 per 1,000; by 1995, only 15 per 1,000.³⁶ Similar reductions occurred all along the income spectrum, including in the poorest countries.

The improvements in social indicators are remarkable by historical standards. As noted in the preface, life expectancy in developing countries increased by 20 years over a period of only 40 years, as it shot from the mid-40s to the mid-60s. By comparison, it probably took millennia to improve life expectancy from the mid-20s to the mid-40s. Literacy improvements have also been remarkable: whereas in 1970 nearly two out of every four adults were illiterate, now it is only one out of every four.

These advances in education and health have greatly improved the welfare of individuals and families. Not only are education and health valuable in themselves, but they also increase income-earning capacity. Where macroeconomic analyses of the growth effects of education have been somewhat ambiguous, the microeconomic evidence of the returns to education is overwhelming and robust. Each additional year of education increases the average individual worker’s wages by at least 5 to 10 percent.³⁷

These trends make it clear that public policy matters. Government has a role not only in ensuring delivery of good basic services in health and education, but also in ensuring that technology and knowledge spread widely through the economy. The dramatic improvement in life expectancy at a given income level is attributable to environmental changes and is the result of public health actions. The control of diarrheal diseases, including the development of oral rehydration therapy to reduce child mortality, is one example; the education of women was an important component of these efforts. Smallpox eradication, made possible through a combination of advances in public health research and effective program management, is another example of a successful twentieth century public health effort.

Bangladesh (see Box 2.1) provides an excellent example of a country that has expanded educational access and improved neonatal and reproductive health dramatically and unexpectedly in a short time. Growth has not been rapid, although it has increased in each decade since independence, and Bangladesh has suffered from weak governance. However, because Bangladesh has taken targeted steps to achieve health and educational goals and has provided a framework for dynamic NGOs to fight poverty, the country has achieved much in a short time. As the box explains, the World Bank and other donors provided important analytical and financial support.

Similarly, Mexico has achieved substantial improvements in the educational attainment and health of its poor citizens through its Progresa program, which it launched on its own, without donor support (see Box 2.2). The program has been so successful that it is being emulated in a number of other countries, with Bank advisory and financial support. A third major example, discussed in Section 5, is the Onchocerciasis Control Program, which has virtually eliminated riverblindness in 11 countries of West Africa.

Within the overall picture of global development progress, the experiences of various developing regions and countries have differed sharply. Some countries have grown very rapidly, others hardly at

all in recent decades. Figure 2.1 provides an illustration of this, graphing median per-capita incomes from 1970 to 2000 for two groups of developing countries--ten of the fastest- and ten of the slowest-growing countries.³⁸ The divergence over that period has been extraordinary: while the median "high group" country has nearly quadrupled the average income of its people over 30 years, the slow-growing countries now find their people some 40 percent poorer on average than they were at the start of the period. Comparisons of individual countries make the same point: The Republic of Korea had a lower nominal GDP per capita than Ghana in 1960 ($155 versus $180). However, because of Korea’s growth, its average nominal income is now 35 times that of Ghana--even though Ghana is not among the world’s slowest-growing economies.

The excellent health achievements also hide regional and country divergences. Like income growth, improvements in health status and life expectancy have not been equally distributed. The health status and life expectancy of the poorest nations lag behind the rest of the world, and within countries, the health of the poor is worse than that of the rest of the population. Poverty is the most important underlying cause of preventable death, disease, and disability; and there is growing recognition that poor health, malnutrition, and large family size are key determinants of poverty.³⁹

The overall picture for the developing world of course lies in between these two groups, but in strongly positive territory. Since 1965, the per-capita GDP of the developing world as a whole has increased by an average of 2.2 percent per year, more than doubling the income of the average developing-country resident. Again, this is a huge change by historical standards.⁴⁰ And since 1990, developing countries have on average grown faster in per capita terms than the richer countries. Given the good policies that developing countries have been putting in place, we should expect this catch-up to continue.

The development success of countries of East Asia, starting with the "tiger economies" of Japan, Korea, Taiwan province of China, Hong Kong (China), and Singapore, is widely known and does not need to be recapitulated in detail here.⁴¹ But there are developing countries outside of East Asia, such as Botswana, that also have a consistent record of growth and poverty reduction over several decades. And other countries that are now high-income, such as Portugal and Ireland, achieved this status by growing more rapidly than other OECD countries over the past several decades.

A substantial number of countries, with a very large share of the developing world’s population, have grown very rapidly and reduced poverty in the last two decades. One group of 24 countries that sharply increased the trade intensity of their economies since the 1970s has seen per capita incomes increase by an average of 5 percent per year in the 1990s. These countries are home to 3 billion of the developing world’s 5 billion people, and such rapid growth in relatively large countries is responsible for the global progress in reducing absolute poverty. (For comparison, the rich countries grew at just 2 percent per year.) Even with fast-growing China excluded, per capita incomes in the remaining countries--such as Mexico, Thailand, Hungary, and Indonesia--grew an average of 3.5 percent per year. By historical standards, this is a very rapid growth pace, fast enough to double incomes in just two decades.⁴²

Policy reforms and institution-building have driven much of this growth. Greater integration with international markets was one factor, but by no means the only one, or even necessarily the most important. In most cases, growth came only with improvements in the broader investment climate: macro stability as well as greater openness, but also better governance, institutions, and infrastructure. (Box 2.3 discusses the World Bank supportive role in reforms that spurred rapid growth in India, and the role of external assistance in that country’s growth and development.)

Indeed, the 1980s and 1990s were periods of major policy reforms throughout the developing world. We can see this clearly, for example, in the areas of macro stability and openness. The median inflation rate of developing countries was cut in half between 1982 and 1997, from about 15 to 7 percent. In low-income countries, the improvements were even greater. More important, at the end of the 1990s only 5 percent of developing countries had inflation above 10 percent. Average tariff rates have also declined sharply in all developing regions. In South Asia, for example, the unweighted average fell from about 65 percent in 1980-85 to 30 percent in 1996-98; in Latin American and the Caribbean, from 30 percent to under 15 percent. Tariff averages are an imperfect measure of openness, but there have been even greater declines in the black-market exchange-rate premium, which is an indicator of macroeconomic instability as well as of the restrictiveness of the trade regime.⁴³

Another way of gauging development success is to look at the numbers of graduates from the World Bank’s soft- and hard-loan windows (IDA lending at concessional terms and IBRD lending at nonconcessional terms, respectively). We are not suggesting here or elsewhere that the Bank alone was responsible for these improvements; on the contrary, the Bank has, at best, supported countries’ own efforts and those of others. Twenty-three countries have seen their incomes increase to the point that they could graduate from IDA and remain IDA graduates, borrowing only from IBRD at nonconcessional terms. (A few others graduated from IDA but later had to return, because their economic performance had worsened and their creditworthiness for IBRD lending became limited.) In addition, 26 countries have graduated from IBRD. This includes Korea, which had started as a low-income IDA borrower but graduated from IBRD lending in 1994.⁴⁴ About ten IBRD-eligible countries have developed to the extent that they have gained broad, though volatile, access to international capital markets. Most of these countries have stopped borrowing from the Bank for extended periods in which they have had good access to commercial financing, but continue to value Bank support during periods in which they are buffeted by external shocks. They also continue to draw on the Bank for advisory and other non-financial services.

Not all countries have been this successful. Indeed, the number of countries that have grown slowly or stagnated over the past two decades is larger than the group of rapid integrators listed above, although the total population of these poor-performing countries is less than that of the more successful countries.⁴⁵ Most notably, Sub-Saharan Africa as a region saw no increase in its per-capita incomes between 1965 and 1999, even with improved performance in the 1990s. And although Africa did make steady progress on health and education indicators over much of that period, despite the lack of income growth, the progress in increasing life expectancy has been reversed by the AIDS epidemic in many countries, as well as by malaria and tuberculosis.⁴⁶

The economic decline in much of Sub-Saharan Africa stems in part from events beyond African countries’ control, including large and persistent declines in the prices of their export commodities. For example, non-oil exporters in Africa (other than South Africa) lost a cumulative 120 percent of their GDP between 1970 and 1997 as a result of changes in import and export prices. In addition,

Africa has faced geographic challenges and protected export markets, as well as many investors’ reluctance to change their risk perceptions of the continent.

However, it is not only external factors that explain Africa’s weak performance. Numerous shortcomings in domestic governance, weaknesses in institutions, and misdirected policies and investments have also contributed. Too many African countries have suffered from an investment climate that has not proven conducive to growth and productivity.⁴⁷

In many of the lagging countries, there are now encouraging signs of progress. Over the past decade, and especially in the past five years, Sub-Saharan Africa has achieved very significant improvements in macroeconomic stability and governance, including widespread democratization. For example, the median inflation rate has fallen from 12 percent in 1990 to 4 percent in 2000. In 1990, 15 countries had negative growth rates, and only 17 countries had growth rates greater than 3 percent. By 2000, only 8 countries had negative growth rates, while 22 had growth rates greater than 3 percent. Perhaps most remarkable among these improvements is the peaceful transition from apartheid to democracy in South Africa and that country’s transition from hyperinflation to macro stability.

The New Partnership for African Development (NEPAD), proposed by several governments in the region, reflects a welcome recognition of the importance of domestic responsibility. It calls for a partnership in which domestic reforms will be supported by increased external assistance, including debt relief and in reductions in protectionist trade policies in rich countries. The Bank, along with many others, welcomed this partnership and is committed to working with African leaders and other partners in ensuring that the next decade does indeed see a broad-based turnaround in Africa’s disappointing development record.

In Eastern Europe and the former Soviet Union, the experience of transition from a centrally planned to a market economy has varied greatly across countries.⁴⁸ Every country went through a "transition recession" before output began to recover, but the depth and length of those recessions differed. Countries in Eastern Europe--helped by prior experience of markets and nationhood, and buoyed by the prospect of accession to the European Union--more quickly followed policies to downsize unprofitable enterprises and encourage rapid entry of new firms. This policy package led to deep recessions, but these economies began to recover after several years of decline, and output exceeded its pre-transition levels by 2000. Today, the countries of Central Europe and the Baltics have made considerable progress with market-oriented reforms, to the point where they are poised for early accession to the European Union. (See Box 2.4 for a discussion of the transformation in a top-performing transition country, Poland.)

In contrast, the countries of the former Soviet Union experienced a prolonged recession, much deeper than the Great Depression. The severity of the recession stemmed from various factors. The dissolution of the Soviet Union, where industrialization had been based on cheap energy and subsidized transport, disrupted production and trade relationships. CIS countries also lacked any recent memory of market economies, and in a number of cases, they suffered from war and civil strife. With the exception of the Baltics, where output recovered rapidly despite the deep initial recession, vested private interests captured the state in most CIS countries early in the transition. The CIS also failed to stop fiscal and financial leakage to unprofitable enterprises and to discipline managers by developing and strengthening institutions of corporate governance. The result was large-scale corruption that undermined confidence in reforms. These countries also failed to put in place policies that would encourage new firms that can create wealth and contribute to economic growth.

Such measures include combating crime, both organized and otherwise; dealing with corruption; implementing legal and judicial reform to ensure secure property rights, streamlined business licensing and registration requirements; and building a tax system that encourages compliance rather than growth of a shadow economy. (Box 2.5 discusses governance challenges more generally, and the role of the World Bank in helping countries to address them over the past several years.)

Given the dangers described here, some of the early external policy advice appears, with hindsight, to have been simplistic: it concentrated too heavily on improving policies, and too little on building institutions. Furthermore, early expectations concerning the time it would take to complete the transition process turn out to have been highly optimistic. As a result of all these problems, it was not until after 1998 that the CIS countries as a group began to witness a recovery. (Recovery had

started in many of the individual countries earlier.) As recovery has begun, however, these countries have grown more rapidly than Eastern Europe during the period 1999 through 2001.

We have learned much about the overall sources of growth and poverty reduction.

• Experience and analysis show that countries reduce poverty fastest when they put in place two pillars of development:

• create a good investment climate--one that encourages firms, both small and large, to invest, create jobs, and increase productivity; and
• empower and invest in poor people--by giving them access to health, education, social protection, and mechanisms for participating in the decisions that shape their lives.

• Understanding of economic growth and its causes has improved greatly. We now understand that sustained growth depends on broad progress in a number of areas: macroeconomic stability and trade openness; governance and institutions, including a good education system, effective legal institutions, and professional bureaucracy; vigorous competition; and adequate infrastructure, especially in countries that are landlocked or face other geographical barriers.
• Poverty reduction depends heavily on sustained economic growth. On average, income distribution does not worsen during periods of economic growth, so the incomes of poor people rise at the same rate as those of wealthier people. Countries that grew rapidly in the 1990s--such as China, India, Vietnam, and Uganda--managed to reduce the share of their people in absolute poverty by 5 to 8 percent per year.
• But while growth is essential, countries can accelerate reduction of income poverty by acting to ensure that poor people have the tools necessary to contribute to growth, such as health and education.
• Policies and investments aimed directly at reducing non-income dimensions of poverty can be highly effective. Countries can accelerate health and education progress far beyond what would result simply from economic growth.
• Globally, we see this in the dramatic reductions in infant mortality, which has fallen steadily at each level of income as a result of improved technology, knowledge, and policies and institutions. The average country with $8,000 per capita income in 1950 would have had an infant mortality rate of 45 per 1,000 live births; by 1995, an average country at the same income level would have had an infant mortality rate of just 15 per 1,000 live births (a reduction of two-thirds).

Countries have acted on this knowledge by improving policies and institutions, often with very positive growth results.

• The development progress seen in many countries should be attributed primarily to actions by those countries themselves: specifically, improving the investment climate and investing in people.
• For example, macro stability and openness have improved throughout the developing world over the past two decades. The median inflation rate of developing countries was cut in half between 1982 and 1997, from about 15 to 7 percent. Average tariff rates have also declined sharply in all developing regions. In South Asia, for example, the unweighted average fell from about 65 percent in 1980-85 to 30 percent in 1996-98; in Latin American and the Caribbean, from 30 percent to under 15 percent.
• A group of countries that has integrated most quickly with the global economy--thanks to greater openness and improved investment climates--has grown very rapidly. This group of "rapid globalizers," which accounts for some 3 billion of the developing world’s 5 billion people, saw per capita incomes increase by an astounding 5 percent per year in the 1990s. Even with fast-growing China excluded, the average was 3.5 percent.

However, development has been far from universal, and the roots of slow growth and decline can typically be traced to a bad climate for investment and productivity.

• The economic decline in much of Sub-Saharan Africa stems in part from a bad environment: large and persistent declines in the prices of exported commodities, high tariffs on exports, as well as adverse economic geography. But many African countries have compounded these and other historical disadvantages by failing to adopt good policies and institutions. Although policies have improved in recent years, governance and institutions remain a major problem in most SSA countries.
• Countries that have not grown rapidly--in Africa and elsewhere--have often failed to make progress on key features of the investment climate. For example, they may have achieved macro stability but not social stability; or they may have lowered trade barriers but not built the basic infrastructure necessary for international trade.

Sections 1 and 2 have shown that rapid growth and development are possible, and that the development community has learned much about the policies and institutions needed for growth and poverty reduction--even if important gaps in our knowledge remain. Sections 3 through 5 focus on how best to promote improvements in policies and institutions from outside. Have external actors, particularly the World Bank, been successful in helping to foster development? What broad lessons emerge about ways to enhance the effectiveness of aid?

Before analyzing the effects of development assistance, it is worthwhile to summarize the ways in which external agents can contribute to development. Development assistance encompasses both financial and non-financial instruments used in support of a country’s own growth and poverty-reduction efforts. Resource transfer is an important part of development assistance, and the quantitative analysis in this section will focus largely on the effects of lending operations. But finance is only one of the instruments used to support development, and in some situations, it is not even the most useful one. Development assistance also includes analysis, advice, and capacity-building, either packaged with lending or on their own. Many of the country boxes in this paper highlight the importance of these nonlending tools, especially in environments--such as early in a reform era--when finance is not likely to contribute to poverty reduction. The mixture of instruments to be used in any particular country depends on the specifics of that country’s needs and capacity.

As discussed in the Executive Summary, any attempt to assess the effectiveness of development Assistance--and in particular of a single institution, such as the World Bank--confronts inherent analytical difficulties. The complexity of social and economic change means that the impact of aid cannot be separated easily from other factors. Because successful development requires country ownership, outside agents are unlikely to be the prime movers of development. Indeed, where they are central to the process, the sustainability of reform efforts may be in doubt. Even where external actors are clearly important, they will be most effective working in partnership with other external and internal partners, which makes it difficult to give credit or blame to any single institution. Finally, the need to "scale up"--to ensure that assistance is replicated or institutionalized, so that it has effects far beyond the boundaries of the specific intervention--means that even identifying the outcomes of a particular intervention is difficult.

To address these difficulties, we draw on as many types of evidence as possible. This section begins with the big-picture question of how to allocate and deliver large-scale financial assistance, drawing heavily on evidence from cross-country regressions. This type of analysis allows us to identify statistical correlations between aid flows and policy, on the one hand, and growth and poverty reduction on the other. Country cases (in boxes throughout this paper) and project evaluations (in Sections 4 and 5) complement the cross-country analysis by showing that results on the ground are consistent with the macro evidence.

At the economy-wide level, the evidence suggests strongly that aid promotes development, if the aid is provided under the right circumstances and with the right design. ODA in general has historically delivered substantial poverty reduction, and the poverty-reducing impact of ODA has increased in the past decade due to improved design and allocation. Empirical studies have shown that:

• Well-targeted aid increases investment. Each dollar of assistance provided through the Bank’s concessional lending arm, the International Development Association (IDA), leads to nearly two dollars of additional private investment, including 60 cents of additional foreign direct investment. Thus aid draws in private investment, rather than crowding it out.
• Well-targeted aid has high overall economic payoffs. Because aid creates new economic possibilities, improving the investment climate and increasing investment, its economy-wide returns are far greater than even the direct poverty-reduction returns--some 40 percent in the case of IDA.
• Aid has become far better targeted in recent years. With the end of the Cold War, donors became less interested in using aid to achieve geopolitical goals and more interested in using aid for poverty reduction. Large-scale financial assistance is being increasingly allocated to countries that have reasonably good policies and institutions--that is, the countries that can best use aid for poverty reduction. With this shift, the poverty-reduction effectiveness of official development assistance (ODA) tripled during the 1990s.

When all aid is lumped together, some analyses have found no clear relationship between aid and growth or poverty reduction.⁴⁹ But not all aid is aimed directly at poverty reduction, nor has aid always been provided in ways that will maximize growth. Moreover, because aid is often provided to help countries cope with external shocks, even if aid is reasonably well-designed and allocated--and thus effective in helping the poor--the positive impact of such aid may be obscured by the magnitude of the shocks.

Disaster relief, for example, is not aimed directly at long-term poverty reduction, and thus it is no surprise that such aid is not correlated with that result.⁵⁰ However, it does achieve its goal of helping to avert famine or assisting countries to recover from natural disasters. Similarly, large amounts of aid were directed at supporting the transition in Eastern Europe and Central Asia (ECA), for both political and economic reasons. There, the mandate in the early 1990s was explicitly to help transform these countries into market economies, rather than to focus directly on reducing poverty. More recently, as Box 3.1 explains, the World Bank has played an important role in the international effort to

promote peace and stability in the Balkans--a mission that is essential to poverty reduction, but not identical. Much aid during the Cold War was provided for military-oriented reasons, rather than developmental ones. In all of these cases, the lending--whether it is for transition or disaster relief or post-conflict stabilization--can help to build the foundations for poverty reduction, but the effects may appear only after a long lag. The end of the Cold War and the progress in transition have made possible a more direct targeting of aid to poverty reduction.

Beyond noting that aid with different goals has different results, research has also found that it is important to distinguish among the different environments in recipient countries. Aid to countries with relatively good policies and institutions has very different effects than aid to countries with worse policies and institutions. (Section 3.1 will address how best to assist the latter group of countries, given that large-scale aid is typically not effective in these environments.)

Finally, the returns to aid will often be difficult to gauge because of the circumstances in which it is provided. Some developmentally successful countries, such as Korea in 1998, have received large aid inflows during periods of economic setbacks--creating a misleading statistical correlation between aid and poor economic outcomes. Natural disasters will produce a similar correlation. In both cases, analysis of the proper counterfactual would often show that, relative to what would have happened otherwise, aid has contributed to better outcomes.

Research shows that ODA reduces poverty and spurs growth substantially when it goes to developing countries that have reasonably good policies and institutions.⁵¹

Experience and analysis have shown that it is generally not enough for donors simply to design projects well. In poor overall policy and institutional environments, the payoffs to flows of large-scale financial aid are on average very small. But where the environment is at least moderately good, then aid typically has large payoffs, as the remainder of this section will explain.⁵²

Why would the overall environment matter so much in determining the effectiveness of ODA? The first reason is very straightforward: policies and institutions affect project quality. For example, a major reason for the dramatic decline in measured World Bank project outcomes in the 1970s and 1980s was the deterioration in policy quality and governance in many borrowing countries. No matter how well designed, a project can easily be undermined by high levels of macroeconomic volatility or of government corruption.

The second reason is more subtle. Even if a project does seem to succeed--based on narrow measures of economic returns and attainment of project objectives--the actual marginal contribution of aid funneled through that project may be small or even negative. This is because government resources are often largely fungible: money can be moved relatively easily from one intended use to another.⁵³ Thus if donors choose to finance a primary education project, displacing local money that would have been used for education, that local money could then be shifted to less productive purposes, such as military spending. In a country with poor public expenditure management, the displaced money could even be diverted to the personal uses of corrupt officials. In this case, the indirect but very real effect of aid could be to promote corruption.

During the 1990s, overall aid allocation shifted in favor of countries with good policies. Aid is always driven by a variety of factors--poverty reduction goals among them, to be sure, but also geopolitical considerations, former colonial ties, human rights concerns, and other factors. During the Cold War, factors unrelated to poverty were especially important in driving aid allocations. This can be illustrated by both example and quantitative evidence:

• As an example, consider the case of Zaire under President Mobutu, which was clearly no one’s ideal of a state dedicated to widespread poverty reduction. Policies were abysmal, and government officials diverted much aid money for their personal uses. Nevertheless, the international community allocated large amounts of aid to Zaire for reasons that were largely geopolitical. With its huge size and strategic location, Zaire was seen as a buffer against the spread of communism in southern and central Africa.
• Looking across countries, it is clear that the ability of the recipient country to use aid well was not a significant factor in total aid allocations as recently as 1990. Research using the World Bank’s 5-point policy index from Country Policy and Institutional Assessments (CPIA) has shown that in 1990 a large difference (1 point on a 5-point scale) in the CPIA corresponded to only a very small (and statistically insignificant) increase in per capita aid flows (about 8 percent). Such a result is in keeping with the view that aid allocations up until this time were strongly influenced by factors other than the donor’s desire to reduce poverty.
• Similarly, if we divide countries into two groups according to policy quality, we find that in 1990, countries with good policies received an average of $39 per capita in overall aid, while those with bad policies received more, at $44 per capita. Even in the case of a single country, aid levels often failed to rise as policies improved, or fall as policies deteriorated. In Zambia, for example, aid levels rose as policy deteriorated in the 1970s and 1980s.⁵⁴
• Because the World Bank was less constrained than bilateral donors by geopolitical objectives, IDA was already relatively well-targeted in terms of poverty reduction in 1990, and has since improved further. The same holds for the Bank’s nonconcessional lending through IBRD, the allocation of which is strongly influenced by the Bank’s assessment of countries’ policies and institutions.

Since the end of the Cold War, and increasingly through the 1990s, donor countries and other aid agencies have reallocated assistance in ways that strengthened poverty reduction.

• By the late 1990s, good-policy countries received $28 per capita, or almost twice as much as the poor-policy countries ($16.4 per capita).
• How much has this shift increased the efficiency of poverty reduction? In 1990, another $1 billion allocated proportionately to ODA would have lifted an estimated 105,000 people permanently out of poverty. But with the improved 1997-98 allocations, an additional one-time expenditure of $1 billion in ODA would have lifted an estimated 284,000 people permanently out of poverty. According to these estimates, then, the poverty-reduction productivity of ODA nearly tripled during the 1990s.⁵⁵

It was not only waning geopolitical concerns, but also a more extensive knowledge base and greater transparency that have contributed to this shift. Research into aid effectiveness undermined both polar positions about aid: that aid was always effective or that aid was never effective. This knowledge helped to focus attention on the key question of whether the right countries were receiving aid. The continuing challenge is to keep development aid focused on poverty reduction and to improve its effectiveness.

Aid flows dropped substantially over the 1990s in real terms, and by 2000 stood at about 7 percent less than the 1990 level. With the growth in incomes in the rich countries over that period, aid levels expressed as a share of donors’ GNP fell even more sharply--from 0.33 percent in 1990 to 0.22 percent in 2000. By the end of this period, aid accounted for less than 1 percent of donor countries’ total government expenditure.⁵⁶

Like other types of ODA, IDA was affected by pressures to lend to countries for reasons not directly connected with poverty reduction, particularly during the Cold War. Nevertheless, the evidence suggests that IDA lending was less affected than overall ODA. We can see this in the statistics on the relationship between poverty, policy, and IDA. Even in 1990, much more IDA went to the good-policy countries ($4.7 per capita) than to the poor-policy countries ($2 per capita). And IDA was already more than four times as responsive to policy improvements than ODA was (in that a 1-point improvement in policy on the Bank’s rating scale led to a 36 percent increase in aid per capita).

By the late 1990s, IDA allocation had improved substantially even from this good base. Good policy countries now received $6.5 per capita, compared with $2.3 per capita in poor-policy countries. Or, to express it in terms of marginal efficiency of poverty reduction, an additional $1 billion allocated like average IDA would lift 434,000 people out of poverty. This means that IDA is about 60 percent more efficient than it was in 1990, and 50 percent more efficient than overall ODA.⁵⁷

High IDA repayment rates are also consistent with the view that the aid is going to countries that typically are able to make good use of it.

• IDA repayment rates are very high. As Table 3.1 shows, only about 5 percent of expected IDA repayments were in nonaccrual status in 2000 and 2001; uncollected payments were concentrated in a handful of states, many of them mired in extended conflict.⁵⁸ Given the evidence that IDA is well targeted at countries that are serious about growth and poverty reduction, this should not be so surprising. Countries that put in place decent policies and institutions generally earn enough of a return on aid that they can afford to make repayments.⁵⁹

• A substantial share (39 percent) of IDA reflows come either from countries that have graduated entirely from IDA (11 percent) or from "blend" countries that receive both nonconcessional IBRD lending and IDA lending (28 percent).
• Debt relief for the HIPC countries does not substantially change this picture, for two reasons. First, HIPC represents a recognition that a country’s debt service payments are not sustainable in a manner consistent with development; however, those unsustainable debts are not primarily IDA. Second, total estimated HIPC debt relief represents only a small share of the total outstanding IDA portfolio--about 13 percent of total credits outstanding.

Nonconcessional lending (IBRD) also goes primarily to countries with good ratings for policies and institutions. It does not typically go to countries with the highest ratings, for reasons noted above: these countries tend to have good, if volatile, access to capital markets and may not borrow from IBRD during periods of good access. Over the last 5 years (FY96-01), about 40 percent of all IBRD-eligible countries received good ratings (not including countries with the highest ratings), but these borrowers accounted for more than two-thirds of all new lending commitments.

Well-designed aid crowds in private investment, both domestic and foreign. Aid is not simply a transfer payment for the consumption of poor people, but an investment in improved policies and institutions. The best aid finances the costs of change, rather than the costs of not changing.

As a result, the investment climate improves when aid is well designed and well allocated. Research shows that IDA flows on average lead to increases in gross investment nearly twice as large as the value of the aid itself. In good policy environments, in both low- and middle-income countries, aid also increases foreign direct investment substantially--by 60 cents for each dollar of aid.

The "crowding-in" effect on investment means that the returns to aid are far higher than the poverty-reduction effects alone would suggest. Indeed, IDA has an estimated average rate of return as high as 41 percent.

China is an excellent example of a country in which aid investments in support of a country-owned development strategy helped to improve the investment climate and encourage rapid private-sector-led growth and poverty reduction (Box 3.2). But the crowding-in effects of aid extend also to smaller countries whose reforms are less well known to investors. In Mauritania, for example, aid-supported reform of telecommunications has prompted a surge of investments and economic activity (Box 3.3).

During the commodity price boom of the 1970s, many poor developing countries responded by increasing domestic expenditures. These expenditures were amplified through increases in external borrowing, including from multilateral, bilateral, and commercial sources. Both the borrowing countries and the lenders believed incorrectly that high commodity prices and export earnings would be long-lived.

When an oil-price shock and global recession hit in the late 1970s and early 1980s, the borrowing countries suddenly faced a debt service crunch. Commodity prices turned sharply against the non-oil commodity exporters, making it difficult to pay for both imports and debt service. Official lending increased to help cushion the effects of the shock, and to substitute for finance from commercial sources, which soon evaporated. Official lending and grants then continued as support for adjustment of policies to the new environment.

But in many of the highly indebted countries, the expected improvements in policy performance did not materialize, whether because of insufficient commitment by borrowers or because the design of the adjustment had not paid enough attention to social concerns. (Section 3.2 discusses the problems of spurring growth when borrowing countries are not sufficiently committed to reform.) In other cases, reforms did not lead to the expected supply and growth response. As a result, the GDP average growth rate between 1980 and 1987 of the 33 countries that were characterized in the mid-1990s as the most severely indebted low-income countries was just 1.9 percent--which translated into an income decline in per capita terms. The cumulative effect of the shock and economic decline was that a debt burden that had been very reasonable quickly became unsustainable. Between 1982 and 1992, the debt-to-export ratio of these 33 countries rose from 266 to 620 percent.⁶⁰

The international community has attempted to address the problem through a variety of debt-reduction mechanisms since the late 1980s. In 1996, it went a step further, creating the Highly Indebted Poor Countries (HIPC) debt relief initiative. Box 3.4 describes briefly the design and achievements of this initiative.

It is important to note that although official lending contributed to debt accumulation in these countries, and debt reductions through HIPC have generally been substantial, problem lending was only a small portion of total IDA lending. Most of the population of the low-income world is in countries that have been able to keep their debts at sustainable levels; as a result, the estimated total IDA debt forgiveness under HIPC amounts to about 13 percent of the portfolio.

While aid effectiveness requires that large-scale financial assistance be allocated to poor countries that have demonstrated the capacity to use aid well, the international community cannot simply abandon people who live in countries that lack the policies, institutions, and governance necessary for sustained growth and for effective use of aid. Poor people in these countries are among the poorest in the world and face the greatest hurdles in improving their lives. Experience suggests that current technical assistance for capacity-building efforts, as well as the promise of greater financial assistance if policies, institutions, and governance improve, are often insufficient to enable these countries to initiate and sustain reform. A World Bank task force is now exploring new approaches for helping people and fostering change in these countries, which are sometimes known as Low Income Countries under Stress (LICUS).

Of the two or three dozen countries with the poorest institutions and policies, only a few have made major improvements in the environments for growth and poverty reduction over the past decade, in contrast to the broad improvements in policies in other developing countries. Ethiopia, Mozambique, and Uganda are unusual among LICUS countries in having achieved significant progress.⁶¹ (See Box 3.5 for a discussion of the role of Bank assistance in Mozambique.) Other LICUS countries have seen little development progress, and the performance of the Bank lending portfolio in this group has been poor: projects have failed at double the rate for other countries.

Approaches that work in the typical low-income country may not be appropriate in the LICUS, as they typically lack the basis for county leadership of reform, and traditional lending conditionality has not worked well in inducing and supporting reform. Given the weaknesses in governance and central institutions, there may at times be a role for the development community to become more directly involved in the provision of basic health and education services.

These countries vary widely in their problems and opportunities. As is the case for the better-performing countries, no single strategy will be appropriate for all of the LICUS; each country has its specific challenges and must look to its unique solutions. Nevertheless, it is useful to distinguish approaches in LICUS countries from those that will work in countries with better policies, institutions, and governance.

• Large-scale financial transfers are unlikely to work well, because the absorptive capacity in these environments is quite limited.
• Instead, donors should focus on knowledge transfer and capacity-building to facilitate change. Given the constraints on government capacity, such efforts should concentrate on a limited reform agenda that is both sensible in economic terms (that is, mindful of sequencing issues) and feasible from a sociopolitical standpoint. Only when they develop greater capacity will these countries generally be able to make good use of large-scale aid.

• In the LICUS, there will often be a case for using aid to improve basic health and education services. To be effective, however, funding should probably be directed through channels other than the central government. The LICUS task force suggests wholesale-retail structures in which a donor-monitored wholesaling organization contracts with multiple channels of retail provision, such as the private sector, NGOs, and local governments.

Laying the foundations for growth typically requires considerable structural reform of the economy. To support this process, the Bretton Woods institutions have provided structural adjustment loans for the past two decades. Adjustment lending is intended to help member countries design and implement reform programs that contribute to the creation of economic conditions that are conducive to economic growth. Adjustment lending grew out of the recognition that, as discussed in Section 3.1, the overall policy and institutional environment heavily influences the success rate of projects; and that with the oil shocks of the 1970s, the overall environment had become far less favorable to development.⁶² With experience, the Bank has learned to better take into account the potential distributive impacts of some aspects of structural adjustment programs. The Bank now works with borrowers to strengthen social protection mechanisms in order to ensure that poor people do not bear an unfair share of the short-term adjustment costs.

In assessing the effectiveness of adjustment lending in spurring growth and poverty reduction, it is important to look separately at the impact of adjustment on growth and poverty reduction and the link between adjustment lending and the design and implementation of adjustment programs.

Some observers have questioned the existence of a systematic link in developing countries between significant economic reforms--or structural adjustment--and economic growth and poverty reduction.⁶³ However, as noted above, there is strong evidence of a clear and robust link between improving economic policies and growth, and between growth and poverty reduction.⁶⁴ First, economic growth in developing countries on average increases the income of poor people proportionally: that is, as noted earlier, a 1-percent increase in per-capita GDP tends on average to lead to a 1-percent increase in the incomes of the poorest quintile of the population. (It is important to stress that this is only an average, and that specific countries, regions, or groups may be left behind; adjustment design and implementation need to anticipate and respond to this variation in outcomes.)

Second, cross-country evidence shows that successfully implemented adjustment policies have tended to increase growth rates on average with little effect on income distribution. Policies such as trade openness, low inflation, moderate size of government, and strong rule of law have generally benefited the incomes of poor people as much as those of anyone else. Indeed, stabilization of high inflation has tended to benefit the poor more than others. Furthermore, since adjustment programs are often launched in times of economic crisis, their effect on growth and poverty reduction tends to be underestimated in cross-country analyses. Research simulating the effect of not adjusting as a counterfactual suggests that adjustment experiences were both better for economic growth and more equitable than not adjusting.⁶⁵

Generally, these findings support the view that structural adjustment can be a powerful instrument for growth and poverty reduction, but it is equally clear that in specific country experiences the distributional impacts can be quite different from the average. This suggests the need for carefully analyzing the poverty impact of adjustment policies under local conditions, to ensure that we can support an adjustment with positive effects on poor people.

Adjustment lending helps countries to undertake reforms and smooth the transition costs associated with adjusting to economic shocks--but only if critical reforms are actually carried out. The record of World Bank adjustment lending is mixed in this regard;⁶⁶ it includes both a number of success stories and a number of cases where adjustment lending was not followed by reform implementation.⁶⁷

Bank reviews and studies concluded from the experience of the 1980s that country commitment is one of the most critical conditions for ensuring the success of reforms supported by adjustment lending.⁶⁸ Evidence has shown that policy change is driven by the country’s own initiative, capacity, and political readiness, rather than by foreign assistance and associated loan conditionality.⁶⁹ Heavy reliance on conditionality is ineffective for several reasons: it can be difficult to monitor whether a government has in fact fulfilled the conditions, particularly when external shocks muddy the picture; governments may revert to old practices as soon as the money has been disbursed; and when assessments are subjective, donors may have an incentive to emphasize progress in order to keep programs moving. Without country ownership, adjustment lending has not only failed to support reforms, but has probably contributed to their delay. For example, case studies of Côte d’Ivoire, Kenya, Tanzania, the Democratic Republic of Congo, and Nigeria all concluded that the availability of aid money in the 1980s postponed much-needed reforms.⁷⁰

In practice, country commitment has often proved difficult to assess. For example, a government may be seriously committed to a reform program but subsequently find it impossible to implement key measures, sometimes for reasons not fully under the government’s control. In other cases, the government may be interested primarily in the funds, not in the reforms on which the funding is conditional. For this reason, the government’s track record--measured by the quality of the policies and institutions it has already put in place--is often a good indicator of its commitment to reform. Research has shown that in good policy and institutional environments, well-designed adjustment lending succeeds in accelerating, broadening, and deepening reforms and in enhancing their impact on growth and poverty reduction.⁷¹

Early adjustment operations also sometimes suffered from insufficient attention to the social dimensions of adjustment. As noted above, the evidence suggests that on balance poor people benefit from adjustment and reform as much as other groups.⁷² Nevertheless, even successful, market-friendly reforms have benefits and costs that are distributed unevenly, especially in the near term, and design and implementation of adjustment operations needs to take this factor into account.

The Bank has increasingly focused adjustment lending on borrowers with a satisfactory track record. At the same time, the Bank also supports reforms in countries where improvements in the policy environment suggest the potential for a turnaround. Judging from CPIA ratings, the strength of the link between IDA disbursements and country policy performance tripled between FY90 and FY97-98. The increasing performance-based selectivity in World Bank (IDA and IBRD) lending decisions also shows up in the share of adjustment loans to countries with above-average performance on a broad range of policies, again according to CPIA ratings. In FY95-00 that share amounted to 72 percent.

Increased country selectivity--due to learning and the post-Cold War decline in geopolitically motivated lending--led to dramatic increases in the development effectiveness of adjustment lending, as evaluated by the independent Operations Evaluation Department (OED).⁷³ OED’s outcome ratings typically measure to what extent an operation has achieved its objectives and thus provide a good indicator for the strength of the link between adjustment lending and the implementation of the program being supported by the loans. The share of operations rated satisfactory in outcome increased from 68 percent in FY90-94 to 86 percent in FY99-00.⁷⁴ In terms of lending volumes, the

increase was even more pronounced, from 73 percent in FY90-94 to 97 percent in FY99-00. OED ratings of long-term sustainability (an assessment of resilience to risk) and of institutional development (the contribution of operations to capacity building) also increased over the same period. The share of satisfactory ratings of sustainability increased from 62 to 81 percent, and that of satisfactory ratings of institutional development increased from 44 to 71 percent (both percentages refer to the share of lending in dollar terms).

As noted above, experience with adjustment lending has been mixed, although the quality of such lending has improved over time. Adjustment lending was most successful when there was strong government involvement in the design of the program, the country had relatively good institutions, and the program extended beyond the confines of a specific Bank operation.

A good example of the latter condition is Mexico’s rural sector. In the late 1980s and early 1990s, the Bank was involved in two agricultural adjustment loans, which supported privatization and price liberalization. However, some of the most important aspects of the government's adjustment program were not financially supported by Bank lending. Rather, they came about as a result of the dialogue and analysis that were initiated with adjustment loans (such as the land reform of the ejido sector, as well as the trade and price liberalization of their major crops, maize and beans).⁷⁵

This Mexico example is not unique. Other developing countries such as Vietnam (see Box 3.6) have also had strong country ownership of their reforms, and have charted their own course while making good use of international assistance (both knowledge-based and financial). But other countries have borrowed without fully accepting the conditions placed on the loans, and both reform design and implementation have suffered as a result.

The lesson that strong borrower ownership is essential for economy-wide reform, together with the need for improved coordination among donors and greater attention to poverty goals, led the World Bank, the International Monetary Fund, and their partners to adopt a new framework for development assistance organized around the recipient’s preparation of a Poverty Reduction Strategy Paper (PRSP).

In December 1999, the Boards of the IMF and the World Bank approved a new approach to the challenge of reducing poverty in low-income (IDA) countries, based on country-owned poverty reduction strategies that would serve as a framework for development assistance. As of December 2001, 41 countries had prepared Interim-PRSPs (I-PRSPs), and 10 countries had completed full PRSPs. The underlying goals of the PRSP process (reflecting the principles of the Comprehensive Development Framework) are to ensure broad-based country ownership of poverty reduction strategies; develop strategies that take a comprehensive, long-term perspective; focus on results that matter for the poor; and to build stronger partnerships between low-income countries and the international donor community.

After two years of implementation of the PRSP initiative, the IMF, the Bank, the PRSP teams and partners collected their experience in order to learn how to implement PRSPs most effectively. The PRSP approach, with its emphasis on country ownership and flexibility, challenges the multilaterals to change their way of working. The current review of the PRSP process reveals that there seems to be widespread agreement (among national PRSP teams, donors, and civil society organizations) on five key achievements:

• There is a strong sense of commitment among most governments of their poverty reduction strategies.
• The participatory processes have created a more open dialogue within governments and with at least some parts of civil society than had previously existed.
• Issues related to poverty reduction, including issues of governance and corruption, have taken a more prominent place in national policy making and expenditures (see bullet below).
• The need to base a poverty strategy on a comprehensive assessment of the poverty in each country has provided a substantial impetus to the poverty information and knowledge base.
• The donor community as a whole has embraced the principles of the PRSP approach, providing technical and financial support to countries in PRSP preparation and indicating their intention to align their assistance programs to support PRSPs.⁷⁶

Recent IMF staff analysis has compared budget allocations from 1999 (pre-PRSP) with the projected expenditure levels for 2001–02 for 32 low-income countries with active PRSP programs. The results clearly indicate that full PRSPs are associated with an increased emphasis on "pro-poor spending."⁷⁷ For the first group of seven countries with full PRSPs, the increase in "poverty-reducing" spending is about 20 percent larger from 1999 to 2000-01 than for the other PRSP countries analyzed. Overall increases in such spending for the full PRSP countries, relative to 1999, has been substantial, on the order of 30 percent of the budget as a share of GDP (that is, poverty-reducing spending is projected to rise from 7.7 to 10.2 percent of GDP between 1999 (pre-PRSP) and 2001-02). While these results are encouraging, this is obviously a narrow measure of success. It is important to bear in mind that these budgetary reallocations will translate into improved poverty outcomes only if there are corresponding improvements in budget execution, service delivery, and access for the poor, and in the investment climate. It is in large measure the investment climate that will determine whether there are job opportunities to match the growth in human capital.

Participants in the PRSP review also identified areas for improvement. They noted that the task of developing a strategy that deals comprehensively with macroeconomic policy, structural and sectoral reforms, governance issues, social inclusion, and the medium-term public expenditure program is extremely complex. Targets set by countries are often overly ambitious. Policy changes and programs to reach them should be better prioritized, so as to ensure realism and to facilitate effective implementation. Setting clearer priorities will require better understanding of the impact of policy actions on the poor, improved public expenditure management systems, and a greater investment in monitoring poverty outcomes. Another challenge concerns the human development sectors, such as health and education; a recent review suggested that these sectors (including their budgetary implications) were not fully integrated with the rest of the PRSP.

Finally, the interaction among the principles of country ownership, partnerships, and results orientation also gives rise to an issue that bears special mention: the potential tension between the principle of country ownership and the need for donors to be accountable for the effective use of their resources. Inevitably, there will be some divergence in donor and country views about the right policies and about capacities for implementation. Donors are being called on to align their assistance programs to support in a coordinated manner some, if not all, aspects of countries’ PRSPs, within the institutional and capacity constraints that each donor faces. Donors--including the World Bank--as part of their contribution need to apply the lessons on pro-poor growth and to support countries to undertake the necessary reforms. This is likely to be a permanent part of development assistance, and addressing it is part of the job of donors. With the PRSP process, this is beginning to happen.

• Both statistical evidence and country-level case studies demonstrate that large-scale financial aid can generally be used effectively for poverty reduction, where reasonably good policies are in place.
• By this criterion, donor financial assistance is targeted far more effectively at poverty reduction than it was a decade ago, when Cold War geopolitics still heavily influenced aid allocations.
• In 1990, countries with bad policies received $44 per capita, while those with good policies (as defined by the World Bank’s policy and institutional performance rating system) received less: only $39 per capita. By the late 1990s, the situation was reversed: good-policy countries received $28 per capita, or almost twice as much as the poor-policy countries ($16 per capita).
• As a result, the poverty-reduction effectiveness of each dollar of official development assistance (ODA) has grown rapidly. In 1990, another $1 billion allocated proportionately to ODA would have lifted an estimated 105,000 people permanently out of poverty; by 1997-98, that had number improved to 284,000 people lifted out of poverty. In other words, the estimated poverty-reduction productivity of ODA nearly tripled during the 1990s.

World Bank assistance is, on average, well targeted at reducing poverty.

• IDA is well allocated to places where it will reduce poverty. Even in 1990, much more IDA went to the good-policy countries ($4.7 per capita) than to the poor-policy countries ($2 per capita). By the late 1990s, targeting had improved further, so that good-policy countries now received three times as much aid as poor-policy countries.
• As a result, an additional $1 billion allocated like average IDA in 1997-98 would lift 434,000 people out of poverty. This means that IDA is not only 60 percent more efficient than it was in 1990, but also 50 percent more efficient than overall ODA.

Well-targeted aid (such as IDA) crowds in private investment, by improving the investment climate. As a result, it promotes broader economic growth and has high average overall returns.

• The effects of aid in the right environments extend well beyond poverty reduction.
• Every dollar of IDA flows on average leads to an increase in gross investment of nearly two dollars. In good policy environments, aid also increases foreign direct investment substantially--by 60 centsfor each dollar of aid.
• This "crowding-in" effect means that the returns to IDA far exceed the poverty-reduction returns, and may be as high as 40 percent, according to one careful study.

Adjustment (or policy-based) lending in the right environments--countries with reasonable policies and strong reform momentum--is effective in accelerating poverty reduction.

• Adjustment lending emerged in the early 1980s as a necessary corrective, in a period when distorted macro environments had undermined the effectiveness of project-level assistance.
• Adjustment lending suffered from early problems with weak commitment and implementation by countries, however, as well as insufficient attention to social costs of adjustment. Because each country situation is unique, borrowers need to have sufficient commitment and flexibility to be able to design and sequence reforms appropriate for their circumstances.
• Since its early years, adjustment lending has improved in targeting and design. It is now an effective way of providing strong reform support in countries with adequate policy and institutional frameworks. By the 1990s, it was earning higher evaluation ratings than investment lending.

One lesson from experience is that reform does not usually succeed without strong local ownership and a broad-based approach, which includes a consideration of institutions, governance, and stakeholder participation--a lesson that has provided the impetus for the Poverty Reduction Strategy Paper (PRSP) process.

• At the society-wide level, early evidence suggests that the PRSP process is increasing ownership and poverty focus of development programs in low-income countries.
• Although it is too early to gauge outcomes on the ground, full PRSPs are so far associated with a 20 percent increase in pro-poor spending.

The evolving understanding of development, which we have summarized in Section 1, is reflected in the changing nature of World Bank activities. Growth has been high on the Bank’s agenda throughout the period, and poverty reduction has been an explicit goal since the 1970s. Nevertheless, focal points for the Bank’s activities have evolved over the past four decades. In the 1960s, Bank activities included promoting food security; in the 1970s, addressing the energy crisis by encouraging new supplies and conservation; in the 1980s, tackling the debt crisis and promoting open trade policies; and in the 1990s, focusing more on poverty reduction and social development. Over this period the Bank has realigned its work at the sectoral and project level to focus on the quality of the policy environment and to focus on the country as the unit of account. Results have improved over the past decade as the Bank has incorporated its lessons of experience and altered its ways of doing business.

The Bank has substantially improved outcomes on the ground in the past decade. Measurement of these outcomes is reflected in various indicators, such as economic and financial rates of return of Bank-financed projects, growth rates of GDP or sectoral output in countries and sectors supported by adjustment programs, and improvements in social indicators. More systematic tracking of the Bank’s results on the ground is being focused on the Millennium Development Goals (MDGs). CASs increasingly set out objectives and measure results on the basis of MDG indicators.

Since 1972 the Bank has systematically collected data on project outcomes for every project it has financed and compared these outcomes with the expected results at appraisal. The result is a wealth of data showing results and trends over time by type of project, sector, and country. The older and most frequently used indicators have been the economic and financial rates of return of the projects.

The economic rate of return (ERR) is calculated for many projects, though not all: adjustment loans and projects in a number of sectors, such as education, health, and social protection, are not covered. Where they are available, the minimum ex ante economic rate of return (ERR) required of Bank-financed projects is 10 percent. Actual ex post results are considerably better. Figure 4.1 shows the average ERR in the 1990s, by sector, for disbursements in sectors for which sufficient data are available to provide a reliable analysis. The projects represented in the figure cover 47 percent of disbursements for all projects closed during 1996-2001, and 82 percent of disbursements in the sectors included in the analysis. On average, the return on these disbursements was nearly 25 percent, up from 16 percent in the 1980s.

These rates of return demonstrate the positive contribution of Bank-financed projects to increased output and productivity of borrowing countries. Bank research has shown (since about 1990) that individual project outcomes and ERRs are strongly associated with the quality of a country’s policy and institutional environment. Evaluation results confirm that the introduction of performance-based lending allocations has contributed to the increase in average ERRs for Bank projects.

The Bank’s private-sector arms--the International Finance Corporation (IFC) and the Multilateral Investment Guarantee Agency (MIGA)--also measure their economic impacts systematically. IFC’s Operations Evaluation Group (OEG) found that 63 percent of 102 real-sector projects begun in 1993-95 and evaluated near the end of that decade had ERRs in excess of 10 percent. In 65 percent of real sector projects, the ERR exceeded the financial rate of return, meaning that the benefits for customers, employees, suppliers and taxpayers were greater than the financial benefits to owners and financiers. Analogously, MIGA’s Operations Evaluation Unit (OEU) reported that as of 2000 it had issued 473 guarantees for just over $7 billion, which had facilitated investments of $36 billion; this means that each $1 of guarantee helped to generate more than $5 in investment in developing countries. While some of this investment might have occurred anyway, in a survey of MIGA clients, 73 percent reported that the guarantee was "critical" or "absolutely critical" for their firms to have made the covered investment.⁷⁸

Two caveats are warranted here. Taken alone, these ERRs should be used with considerable caution.

• As noted earlier, if government funds are fungible, then the marginal effects of the aid could be far lower than suggested by these calculations. The same is true if private-sector firms would have taken up the project in the absence of Bank action. For these reasons, high ERRs in a sector do not necessarily mean that investment in that sector should be increased.
• On the other hand, if a project is successful at demonstrating new approaches or making permanent improvements in institutions, the ERR calculations will be based on underestimates of the project’s benefits.

This is why it is essential to complement project-level evaluations with broader evaluations of progress at the sector or country level, as well as cross-country statistical analysis of the type presented in Section 3. Those earlier results on the effectiveness of IDA lending provide some confidence that these highly positive ERR calculations may at least be of the right magnitude.

The World Bank has long had projects aimed at improving health and education outcomes in developing countries. The Millennium Development Goals and the CDF/PRSP process (discussed in Sections 2 and 3) have helped raise the profile of these goals still further, so that they are now incorporated more systematically into the preparation of Bank country assistance strategies. A growing number of CASs reflect the objectives and monitor progress against them, and increased lending is going to human development sectors.

Having provided $30 billion for education projects, the Bank is the world’s largest external funder of education; it also is the world’s largest external funder of health programs, with new commitments of $1.3 billion a year for health, nutrition, and population projects. These social sector projects are achieving large, measurable improvements in human development indicators. (See Box 4.1 for examples from the health sector, and Box 4.2 for an example of AIDS prevention in Brazil.)

OED reviews of the health sector portfolio have found that the Bank has made important contributions to strengthening health, nutrition, and population policies and services worldwide.⁷⁹ Through its support to policy reform, technical assistance, and financing, the Bank has helped expand geographical access to basic health services, sponsored valuable training for service providers, and contributed advice and other important inputs to government basic health services. The Bank has also used its lending and nonlending services to promote dialogue and policy change on a variety of key issues, including family planning, health financing, and nutrition strategies. Clients appreciate the Bank's broad strategic perspective on the sector, and the Bank has taken a growing role in donor coordination. After an initial focus on government health services, the Bank is increasingly focusing on other key issues, such as private and NGO service delivery, insurance and regulation. In recent years, the Bank also has placed greater emphasis on client ownership and beneficiary assessments in project design and supervision.

At the same time, OED raised a number of shortcomings in the Bank approach to health, such as a slowness to incorporate good institutional analysis into reform design and to focus on improving quality of service. In addition, with some notable exceptions, the Bank has not placed sufficient emphasis on addressing determinants of health that lie outside the medical care system, including behavior changes and cross-sectoral interventions. Promoting health reform required strategic and flexible approaches to support the development of the intellectual consensus and broad-based coalitions necessary for change, and the Bank is in the process of adapting its instruments to place more emphasis on learning and knowledge transfer.

As Bank lending in support of structural and sector adjustment programs increased after 1980, its results were regularly assessed in several reports. The Bank’s internal reviews of the experience of the first decade concluded that, on average, countries that received adjustment lending had moderately higher growth of GDP, imports, and exports than countries that did not. Similarly, evaluations of IFC projects in the 1990s found that 60 percent contributed to overall economic growth in the country.⁸⁰ For MIGA projects, nearly half had a substantial or better macroeconomic impact, effects directly related to the size of the project relative the size of the country’s economy.⁸¹

At the same time, the Bank reviews also found that: (1) failure to address the social costs of adjustment accounted for some policy reversals; (2) supply response in low-income countries (especially in Sub-Saharan Africa) was slow, and more time was required for institutional reforms; (3) there was a need for more specific and monitorable conditionality, particularly for tranche release; and (4) a more selective approach to adjustment lending and closer alignment between the Bank’s and countries’ expectations were needed.

As highlighted in reports by the Bank and NGOs, although poor people might gain on average from adjustment over the medium term, at least some poor people suffer during the adjustment period; this requires increasing attention to the allocation of public expenditures and cushioning income declines for vulnerable groups.⁸² A broader time perspective on the outcomes of adjustment lending, and of the major improvements in the 1990s over the previous decade, is contained in the Bank’s recent Adjustment Lending Retrospective: Final Report (World Bank 2001b), and Adjustment from Within: Lessons from the Structural Adjustment Participatory Review Initiative (World Bank 2001a), as well as in three OED reports.⁸³ These reports document the mixed experiences of the earlier generation of adjustment operations and the improvements experienced in the 1990s (noted earlier). And they document the increased impact of successful adjustment operations on growth and, particularly, poverty reduction. For example, the share of explicitly poverty-focused adjustment operations increased from 47 percent in 1995 to 75 percent in 1999. In parallel, there has been a growing focus on social objectives: the share of conditions in adjustment loans directly supporting social sector reforms increased from 3 percent in the 1980s to 18 percent in the last three years. Finally, cross-country evidence shows that developing countries receiving adjustment lending in FY90-97 maintained and in some cases increased social expenditures, on average, more frequently than countries without such lending. An example of successful adjustment lending with positive impact on both growth and poverty alleviation is shown in Box 4.3.

The outcomes of Bank lending are approaching their highest levels in two decades. OED’s outcome indicators are performance-related, measuring project contributions to productivity or amelioration of social dimensions of poverty, as discussed above. (The Annex of this report provides more detail about the evaluation system.) Figure 4.2 shows outcome data for nearly 30 years, a period in which standards for good performance have increased, making it more difficult for a project to earn a satisfactory rating. The figure shows a decline in performance from 1977 to 1989, followed by an improvement since then.

The Bank has achieved improvements in performance over time by continually evaluating, learning, and adapting. Projects are evaluated by the independent Operations Evaluation Department, which reports directly to the Bank’s governing board (rather than through the President). In addition, there are systematic analyses of development approaches and project quality carried out by the Operations

Policy and Country Services unit (OPCS), the Quality Assurance Group, the Bank’s research department, and the thematic networks, all of which help to identify and learn from successes and failures. Through such learning, the Bank has made several major changes in approach that have contributed to improved outcomes over time.

As we explain below, changes in the Bank’s approach include: (1) shifts in lending towards human development, social protection, and governance; (2) the introduction of adjustment lending in the 1980s to eliminate major policy distortions; (3) moving beyond projects to adopt the country as

the main "unit of account," with Bank’s CASs and countries’ PRSPs becoming the central framework for ensuring alignment with countries and with the Bank and other external partners; (4) systematic managerial attention to operational quality; (5) tighter social and environmental standards; and (6) improved fiduciary performance. As shown below, these operational emphases have been driven by lessons of experience, new development challenges, and the growing realization that policies and institutions are keys to the effectiveness of development assistance.

The Bank has responded to changes in its environment by reallocating its lending. Successful development, plus the growth of international capital markets, meant that emerging markets now had greater access to private finance for revenue-earning development projects, such as transportation and energy. Rather than compete with the private sector, the Bank shifted its efforts--in countries with access to alternative sources of financing--away from financing physical investments to helping countries reform sector policies and establish good regulatory frameworks. There has thus been a sharp decline in the Bank’s financing of infrastructure, from about three-fourths of Bank lending in the 1960s to about one-third in the 1990s (Figure 4.3). Within the reduced lending for infrastructure, there has been a trend toward more projects--with smaller loan amounts--aimed at helping governments to establish regulatory frameworks, address environmental issues, pilot new approaches to public sector infrastructure delivery, and design mechanisms for private participation.⁸⁴

While the share of lending to the financial sector remained largely constant, the content of the Bank’s lending moved from financial intermediary lending through national development banks to supporting financial sector reform and bank restructuring.

In contrast, Bank lending for health, education, and social protection has risen steadily and now constitutes nearly one-fifth of total lending. During the 1990s, the Bank also increased lending for governance and public sector reform, private sector development, and environmental protection. As a result of these shifts, many of which are from the project to program level, Bank projects have become more relevant; but they have also become increasingly complex and demanding for both the Bank and borrower.

The IFC has complemented these shifts in recent years by aiming to crowd in private investment in sectors where private participation holds the most promise. It does this by targeting financial markets, infrastructure, and projects in the social sectors, and "frontier" regions or sectors within countries, as well as high-risk or low-income countries in recent years. In 2001, just under 70 percent of IFC approvals were in targeted sectors and over 40 percent in high-risk or low income-countries.⁸⁵

Section 3.2 has discussed in detail the evolution of adjustment (or policy-based) lending as a broad instrument of aid, but it is worth reiterating these messages briefly as an example of how the Bank has adapted its approach to improve performance. This type of lending was widely used by the Bank during the early years after World War II, but thereafter the focus shifted to investment lending. Program lending was resurrected in 1980, with the purpose of helping countries adjust their balance of payments after the 1979 oil shock. The focus soon shifted to helping countries remove obstacles to growth as they attempted to overcome the debt crisis of the 1980s. That focus was complemented in the 1990s with an explicit emphasis on poverty reduction and on avoiding adverse social impacts of adjustment. In recent years, as many countries have succeeded in removing major macroeconomic distortions and reducing excessive state intervention, adjustment lending has supported more complex institutional reforms.

The rapid expansion of adjustment lending during the 1980s, and its mixed outcomes during this period, lent great impetus to the learning process. Several OED and Bank reviews analyzed the economic and social impact of adjustment and fed back the lessons learnt into the design of adjustment operations whose results improved greatly. Adjustment operations achieved satisfactory or better outcomes about 60 percent of the time in the 1980s, and above 80 percent in the 1990s--following design improvements and increased country selectivity based on lessons of experience.⁸⁶

Along with this increased focus on the policy environment, Bank strategy moved beyond projects to adopt the country as the privileged "unit of account." Although the country had long been at the core of Bank programming, a qualitative change took place in 1990 with the introduction of the CAS. The CAS is a Bank-owned document that sets forth the Bank’s assistance program based on the country’s own vision and associated development program, the Bank's diagnosis of the country's policies, institutions, and private sector strength, and the Bank's comparative advantage, given the support provided by development partners. The CAS also sets out benchmarks for monitoring and evaluating the Bank's and the country's performance. From 1994, CASs were prepared in consultation with the government and, more recently, participation has been broadened to include NGOs, civil society, the private sector, development partners, and other stakeholders. In 1998, CASs for most countries began to be publicly disclosed. As countries increasingly formulate their visions and programs in PRSPs or other country-owned documents, CASs are expected to acquire more the character of a business plan for the Bank.

Evaluation (including self-evaluation) and borrower feedback have reinforced the country focus by assessing the impact of Bank assistance at the country level. A 1999 Client Feedback Survey found that borrower authorities gave the Bank high marks on: (1) incorporating country realities in project design, (2) technical competence, and (3) ensuring that project benefits justify the cost to the country. Key areas identified for improvement were: (1) building capacity at the community level, (2) strengthening local training and research organizations, (3) helping to strengthen the private sector, and (4) helping to strengthen and maintain sound macro and trade policies.

Experience has demonstrated that for projects to be most successful they should focus on results, respond to client needs, be cost-effective and innovative, and build on the lessons of experience. Evaluations carried out by OED have demonstrated the importance of client commitment to projects, and of client and beneficiary participation in the design and implementation of the project for achieving satisfactory outcomes. (See Box 4.4 for an example of a successful project that incorporated these elements.)

The Bank’s "safeguard policies" cover its regulations and standards governing the treatment of Environmental Assessments, Natural Habitats, Involuntary Resettlement, Pest Management, Indigenous Peoples, Forestry, Projects on International Waterways, Cultural Property, Safety of Dams, and Projects in Disputed Areas. Management attention to the implementation of these policies has increased in recent years, both as part of the overall quality agenda and as a response to concerns raised by NGOs. For example, a Quality Assurance and Compliance Unit now monitors Bank compliance with its social and environmental safeguards, producing implementation reports for management in real time. On the IFC side, two-thirds of evaluated projects met high standards for environmental sustainability, including many of those whose profitability for their owners fell short.⁸⁷

Inevitably, more rigorous implementation of these policies and compliance monitoring has been seen by some as raising the cost of doing business with the Bank and increasing risk aversion among staff and borrowers, especially with respect to the oversight of the independent Inspection Panel. There is anecdotal evidence that some managers are discouraging their staff from tackling operations involving safeguard policies, and that the safeguards have reduced the Bank’s involvement in some activities, such as sustainable forestry management and construction of major dams. But on balance the safeguard policies have improved the social and environmental quality of Bank operations and have come to be viewed as good-practice benchmarks within the development community.

In recent years, the Bank has tripled the resources allocated to ensuring that its funds are used for the intended purposes. The overall quality of fiduciary performance in Bank-financed projects has improved steadily, as the Bank has introduced formal assessments of the financial management and procurement arrangements of the entities that implement individual projects. With the growing importance of nontraditional projects such as adjustment operations and debt relief, it has become clear that fiduciary assurance and achievement of development objectives depend on satisfactory performance in the management of public finances more generally. The Bank therefore has introduced country assessments of financial management and procurement, to provide it with a solid underpinning for its assistance program in each country. Results of the assessments inform the design of assistance programs, including support for measures to address weaknesses revealed by the reviews. At the same time, the Bank is working with international standard-setting organizations to develop a global set of standards for good quality public financial management.

The Quality Assurance Group (QAG) evaluates the quality at entry of Bank operations and the quality of Bank supervision. OED evaluates the outcomes of Bank operations and the Bank’s own performance--after projects have been completed and all loan funds have been disbursed. QAG’s indicators of quality at entry have improved, from 69 percent satisfactory or higher in 1996 to 82 percent in 2001, while OED’s have increased from 62 to 79 percent. Since quality at entry is a leading indicator, the performance of completed projects is likely to continue to improve in the next few years. Similarly, an evaluation of IFC projects also concluded that where IFC had done its job consistently well, 90 percent of projects were successful developmentally (compared to 61 percent overall).⁸⁸

A theme running through this paper has been the necessity of learning from both failure and success. This section provides some additional examples of how the Bank has done so.

Development is a risky business and some failures are inevitable. While performance shortfalls have many causes, it is important for the Bank to learn from them and feed that learning back into subsequent operations. Evaluation evidence confirms that when project performance has been weak, the Bank has modified its approach. For example, Bank lending for power development in Sub-Saharan Africa has been less successful than in other regions. By the early 1990s, evaluation had disclosed that power projects were not meeting their goals, especially with respect to financial sustainability. A new sector policy, introduced in 1993 and in force today, identified guiding principles for Bank support in the energy sector. The results in Africa have been impressive: between 1990 and 1994, power outages dropped from 50 to 18 hours per customer (per year), losses fell from 20 to 17 percent and the number of employees per thousand customers decreased from 10 to 7.

A similar learning from costly lessons of experience has been evident in adjustment lending. In line with evaluation lessons, the Higher Impact Adjustment Lending (HIAL) initiative introduced in the Africa Region in 1995 aimed at enhancing adjustment lending through improved country selection and better design. During FY96-98, HIAL lending to Sub-Saharan Africa (SSA) exceeded US$2 billion through 21 operations in 17 countries. An evaluation of the impacts of the program found that the HIAL group outperformed non-HIAL SSA and all non-SSA IDA countries in terms of fiscal adjustment, exchange and interest rate policy, and structural reforms. The group also achieved better results in terms of macro variables such as inflation, current account, foreign exchange reserves, growth, and debt sustainability. Fiscal adjustment was associated with both inflation reduction and growth in the HIAL group. Also, the HIAL share of poverty-focused operations was higher relative to comparators. By enhancing the flexibility of resource flows in HIAL through varied tranching arrangements, governments gained increased freedom in the timing of reforms and greater ownership of programs. A HIAL operation had about half of the number of conditions of the SSA average during 1980-93. This is consistent with HIAL's objectives of improving the effectiveness of adjustment lending by providing greater flexibility in disbursement through fewer but more specific conditions.

Sometimes, building on success means helping to spread successes across countries and regions through partnership with other development actors. The "Green Revolution," which began in South Asia in the 1970s and spread to Africa and Latin America, has led to impressive gains in production of basic food crops across the developing world, as shown in Table 4.1. Between 1970 and 1997 yields of cereals in developing countries rose more than 75 percent, coarse grains 73 percent, root crops 24 percent and pulses nearly 11 percent.⁸⁹ The Bank supported this sweeping change through its lending for irrigation, rural infrastructure, and agriculture, and by mobilizing support with other donors through the Consultative Group for International Agricultural Research (described in Section 5.2 below).