W O R L D    B A N K    P O L I C Y    A N D    R E S E A R C H

The money—and ideas—of aid

Money matters in a good policy and institutional environment, and aid can be the midwife of good policies and institutions. But in a bad policy and institutional environment, ideas are far more appropriate than money.

Foreign aid has at times been a spectacular success. Botswana and the Republic of Korea in the 1960s, Indonesia in the 1970s, Bolivia and Ghana in the late 1980s, and Uganda and Vietnam in the 1990s are all examples of countries that have gone from crisis to rapid development. Foreign aid played a significant role in each transformation, contributing ideas about development policy, training for public policymakers, and finance to support reform and an expansion of public services. Foreign aid has also transformed entire sectors. The agricultural innovations, investments, and policies that created the green revolution-improving the lives of millions of poor people around the world-were financed, supported, and disseminated through alliances of bilateral and multilateral donors. Internationally funded and coordinated programs have dramatically reduced such diseases as river blindness and vastly expanded immunization against key childhood diseases. Hundreds of millions of people have had their lives touched, if not transformed, by access to schools, clean water, sanitation, electric power, health clinics, roads, and irrigation-all financed by foreign aid.

On the flip side, foreign aid has also been, at times, an unmitigated failure. While the former Zaire's Mobuto Sese Seko was reportedly amassing one of the world's largest personal fortunes (invested, naturally, outside his own country), decades of large-scale foreign assistance left not a trace of progress. Zaire (now the Democratic Republic of Congo) is just one of several examples where a steady flow of aid ignored, if not encouraged, incompetence, corruption, and misguided policies. Consider Tanzania, where donors poured a colossal $2 billion into building roads over 20 years. Did the road network improve? No. For lack of maintenance, roads often deteriorated faster than they were built.

Sadly, experience has long since undermined the rosy optimism of aid-financed, government-led, accumulationist strategies for development. Suppose that development aid only financed investment and investment really played the crucial role projected by early models. In that case aid to Zambia should have financed rapid growth that would have pushed per capita income above $20,000, while in reality per capita income stagnated at around $600 (figure 1).

Foreign aid in different times and different places has thus been highly effective, totally ineffective, and everything in between. The checkered history of assistance has already led to improvements in foreign aid, and there is scope for further reform. The pressing question: How can development assistance be most effective at reducing global poverty?

The answer is needed urgently. While there has been more progress in poverty reduction in the past 50 years than in any comparable period in human history, poverty remains a dire global problem. It is ironic-and tragic-that just as economic reform has created the best environment in decades for effective assistance, donors have cut aid back sharply. In 1997 OECD donors gave the smallest share of their GNPs in aid since comparable statistics began in the 1950s-less than one-quarter of 1 percent. It would take roughly a 50 percent increase even to restore aid to its 1991 share.

What works—what doesn't

New empirical evidence on the money and ideas of aid provides insights into the puzzle of what is effective aid and what is ineffective aid:

• Financial aid works in a good policy environment. Financial assistance leads to faster growth, poverty reduction, and gains in social indicators in developing countries with sound economic management. And the effect is large: with sound country management, 1 percent of GDP in assistance translates into a 1 percent decline in poverty and a similar decline in infant mortality. In a weak environment, however, money has much less impact. A $10 billion increase in aid would lift 25 million people a year out of poverty-but only if it favors countries with sound economic management. By contrast, an across-the-board increase of $10 billion would lift only 7 million people out of their hand-to-mouth existence.
• Improvements in economic institutions and policies in the developing world are the key to a quantum leap in poverty reduction. True, there have been sharp improvements in governance and policies in the past decade, but further reform of the same magnitude would lift another 60 million people a year out of poverty (box 1). When societies desire reform, foreign aid can provide critical support-in ideas, training, and finance. Efforts to "buy" policy improvements in countries where there is no movement for reform, by contrast, have typically failed.
• Effective aid complements private investment. In countries with sound economic management, foreign aid does not replace private initiative. Indeed, aid acts as a magnet and "crowds in" private investment by a ratio of almost $2 to every $1 of aid. In countries committed to reform, aid increases the confidence of the private sector and supports important public services. In highly distorted environments, aid "crowds out" private investment, which helps explain the small impact of aid in such cases.
• The value of development projects is to strengthen institutions and policies so that services can be effectively delivered. Aid brings a package of knowledge and finance. Most aid is delivered as investment projects in particular sectors such as roads, water supply, or education. Project finance, however, often does not increase spending in a sector any more than an untied grant would have-that is, aid finance is typically fungible. Thus choosing such laudable sectors as primary health or education cannot ensure that money is well used. Aid is financing the entire public sector, and the overall quality of policies and institutions is the key to securing a large return from this finance. These findings highlight that the most critical contribution of projects is not to increase funding for particular sectors, but to help improve service delivery by strengthening sectoral and local institutions. The knowledge creation supported by aid leads to improvements in particular sectors, whereas the finance part of aid expands public services in general.
• An active civil society improves public services. One good idea that many projects have supported in recent years is a participatory approach to service delivery, often resulting in huge improvements. The best aid projects support initiatives that change the way the public sector does business. The top-down, technocratic approach to project design and service delivery has not worked in areas critical for development -rural water supply, primary education, natural resource management, and many more.
• Aid can nurture reform in even the most distorted environments-but it requires patience and a focus on ideas, not money. In some of the poorest countries of the world, the government is not providing effective policies or services, which is why government-to-government transfers have yielded poor results. Still, there are often champions of local or sectoral reform, and aid at times has been effective supporting these initiatives. This work is staff-intensive and results in little disbursement of funds. Successful assistance here aims to help reformers develop and test their ideas.

Box 1: Defining sound management: good policies and institutions

Sound management consists of the institutions and policies that will lead to rapid development and poverty reduction in a particular country. Developing countries learn about good and bad policies from their own experience and each other's experience. Sound management is difficult but not impossible to measure using a number of proxy indicators. The index of economic policy used in box figure 1 combines three factors shown in empirical studies to affect developing countries' growth: inflation, the budget surplus, and trade openness. A country with poor policies would be one with high inflation, large fiscal imbalances, and a closed trade regime (Nicaragua in the 1980s, for instance). An example of good economic policy would be Uganda in the mid-1990s. The measure of institutional quality involves an assessment of the strength of the rule of law, the quality of public bureaucracy, and the pervasiveness of corruption. As the crisis of 1997-98 has shown, Indonesia is a country with poor institutional quality. Botswana, with its high-quality institutions, is a different story. As donors make more of an effort to support good management, they likely will want to broaden the measure beyond the macroeconomic and institutional features here. For example, efforts to improve education and health are critical for successful development. And government support to agricultural research and extension and to community solidarity efforts contributed importantly to East Asia's success. The general point is that the definition of "good management" emerges from the actual experiences of developing countries.

Five policy reforms

Making aid more effective in reducing poverty requires five policy reforms.

1. Financial assistance must be targeted more effectively to low-income countries with sound economic management. In a good policy environment financial assistance is a catalyst for faster growth, more rapid gains in social indicators, and higher private investment. In a poor policy environment, however, aid has much less impact. Clearly, poor countries with good policies should receive more financing than equally poor countries with weak economic management. Up until the early 1990s, however, finance has gone in equal amounts to well-managed countries and to poorly managed ones. Furthermore, much of aid continues to go to middle-income countries that do not need it. It is possible to make aid more effectively targeted to poor countries and to better management simultaneously.
2. Policy-based aid should be provided to nurture policy reform in credible reformers. Experience shows that donor financing with strong conditionality but without strong domestic leadership and political support has generally failed to produce lasting change. Continued flows to governments that pay only lip service to reform have been a major problem. Policy-based financing should go only to countries with a strong track record or where there is a demonstrable basis for optimism (to support, for example, the concrete actions of domestically initiated reform efforts or a government newly chosen on a reform platform). New governments in postconflict situations are often good candidates for support. In countries with poor policies and no credible reform movement, assistance should assume the more modest and patient role of disseminating ideas, transmitting experiences of other countries, training future policymakers and leaders, and stimulating capacity for informed policy debate within civil society. These measures are relatively inexpensive and do not conflict with the proposal that the bulk of finance should go to countries with sound economic management.
3. The mix of aid activities should be tailored to country and sector conditions. Even where institutions and policies are weak, donors have tried to find something useful to finance. Surely it must be a good thing to finance primary health care or basic education? The evidence, however, is that aid is often fungible, so that what you see is not what you get. In circumstances where similar projects would have been undertaken anyway, donor money for particular projects and sectors does not necessarily "stick"-it simply expands the government's budget. Thus even rigorous project selection or reallocation of donor finance to laudable activities cannot guarantee the effectiveness of aid in a distorted environment. To measure the effect of their finance, donors must look at overall allocations and, even more important, at the efficacy of public spending.

   The allocation of expenditures alone does not guarantee success, for the quality of public spending is as important as its quantity. In countries with sound economic management (of both macroeconomic policy and delivery of public services), more aid can be in the form of budget support, which would simplify administration and reduce overhead. In countries with basically sound policies but weak capacity for delivering services, project aid should be a catalyst for improving the efficacy of public expenditures. Countries without good policies, efficient public services, or properly allocated expenditures will benefit little from financing, and aid should focus on improvements in all three areas.

4. Projects need to focus on creating and transmitting knowledge and capacity. The key role of development projects should be to support institutional and policy changes that improve public service delivery. Even where money may not stick, the local knowledge and institutional capacity created by the catalyst of aid projects can. Where projects are innovative, it is crucial to have objective and rigorous evaluation of outcomes and dissemination of new information. Knowledge about what works in service provision -and what does not-is one of the most important outputs of development assistance. In many cases innovative approaches to service delivery will involve greater participation by local communities and decentralization of decisionmaking.
5. Aid agencies need to find alternative approaches to helping highly distorted countries, since traditional methods have failed in these cases. Communities and governments are heterogeneous, and even in the most difficult environments there will be pockets of reform. Donors need to be patient and flexible and look for windows of opportunity to nurture these reform efforts. Typically, ideas will be more useful than large-scale finance. Donors' ability to work in these environments has been hampered by an "approval and disbursement culture" that does not value small-scale, staff-intensive activities. In the past agencies have too often focused on how much money they disburse and on narrow physical implementation measures of the "success" of the projects that they finance. It turns out that neither measure tells much about the effectiveness of assistance. The evaluation of development aid should focus instead on the extent to which financial resources have contributed to sound policy environments. It should focus on the extent to which agencies have used their resources to stimulate the policy reforms and institutional changes that lead to better outcomes. These are not easy questions to answer, but independent reviews of development agencies-with participation of developing country policymakers and project beneficiaries-can help establish whether agencies are doing a good job.

   Developing countries are to a large extent masters of their own fate. Domestic economic management matters more than foreign financial aid. Economies that lag are held back more by policy and institutional gaps than by a financing gap. Aid as money has a large impact only once countries have made substantial progress with reform of policies and institutions. Poor countries with good policies should get more aid than ones with mediocre policies-but in fact they get less (figure 2).

Called for are independent reviews of development agencies with strong input from developing countries and focusing on two questions: Has the bulk of financing gone to sound institutional and policy environments? And have agencies contributed to policy reform and institutional change? Evaluating the right things should feed back into the management and incentives within agencies. With better management and evaluation, development agencies should become:

• More selective-putting more money into economies with sound management.
• More knowledge-based-using resources to support new approaches to service delivery, expanding knowledge about what works, and disseminating this information as a core business.
• Better coordinated-results-oriented agencies should worry less about planting their flags on particular projects and more about how communities, governments, and donors working together can improve services.
• More self-critical-agencies should continually ask themselves: Why do we do what we do? And what is the impact?

With a better understanding of development and aid effectiveness and with the end of Cold War strategic pressures, there is reason to be optimistic that reform of aid agencies will succeed.

Drawn from World Bank, Assessing Aid: What Works, What Doesn't, and Why (New York, Oxford University Press, 1998).