Journal Issue

Aid Spurs Growth-in a Sound Policy Environment

International Monetary Fund. External Relations Dept.
Published Date:
January 1997
  • ShareShare
Show Summary Details

Foreign aid to developing countries has been criticized as wasteful and even counter productive. Careful examination of the recent experience with foreign aid shows, however, that it can be an effective investment when a recipient country’s economic policies are sound before aid is provided.

FOREIGN AID accelerates growth and poverty reduction in developing countries that pursue sound economic policies. It has had no measurable effect in countries with poor policies. These are the main findings of our study (see the end of this article). By “sound policies,” we mean measures that have been shown in a wide range of studies to promote growth: open trade regimes, fiscal discipline, and avoidance of high inflation.

These results imply that aid would be more effective overall if it were well targeted to poor countries with sound policies. We examined the allocation of official development assistance (ODA) during 1970–93. About two-thirds of ODA is bilateral. Since the allocation of bilateral assistance has been influenced by strategic and political factors, countries with poor policies and those with sound policies received this aid in equal amounts. In contrast, multilateral aid—including that provided by the World Bank’s affiliate, the International Development Association (IDA)—favored countries with sound policies in this period.

The last question we take up in our study is whether the amounts of aid that countries receive affect their policies. Conservative critics of aid argue that large amounts of such financing encourage poor policies. We found no systematic evidence of this. There are some well-known examples of countries that have received large amounts of money and persisted with poor policies for decades (as Zambia did until its recent reforms), but there are just as many countries that have received large amounts of aid and implemented sound policies (for example, Ghana) or that have received virtually no aid and persisted with poor policies (for example, Myanmar and Nigeria).

The academic literature on policy formulation generally finds that foreign aid plays, at most, a supporting role. Countries’ policies are determined largely by domestic social and political factors. Once a society has generated a thorough reform program, foreign aid can play an important supporting role through advice, training, and financial assistance.

If we want to get the maximum impact on growth and poverty reduction, then financial assistance should be directed to low-income countries that have successfully put sound policies into place. Countries with poor policies should be encouraged to reform, and providing them with advice, workshops, and training may have a high payoff. It is wasteful, however, to put significant financial resources into countries with poor policies.

Aid and growth

Other studies have looked at the relationship between aid and growth and have generally found no evidence that more aid leads to higher growth. What is different about our study is that we distinguish among countries on the basis of their economic policies. There have been a range of studies in the 1990s that have identified policies that promote long-run growth: open trade regimes, fiscal discipline, and avoidance of high inflation. We form an index of these three measures. For example, consistently good performers would be Botswana or Indonesia, and poor performers would be Tanzania or Zambia before 1993. Our analysis has been conducted both across countries and over time. Bolivia and Ghana are examples of countries that had very poor policies in the early 1980s and then quite sound policies during 1990–93.

For our study, we assembled data on 56 developing countries over 24 years (1970–93). Our results are particularly strong for the 41 low-income countries in the sample. (Middle-income countries such as Brazil or Mexico have received very small amounts of ODA, and it is hard to find any measurable effect of this aid on their growth.) When we focused on low-income countries with good economic policies, we found that those receiving large amounts of aid grew faster than those receiving small amounts of aid: 3.5 percent per capita growth per year compared with 2.0 percent (Chart 1).

Chart 1Per capita GDP growth in low-income countries with good policies

(growth per annum)

Source: Craig Burnside and David Dollar, 1997, Aid, Policies, and Growth, World Bank Working Paper No. 1777 (Washington).

Among developing countries, 2.0 percent per capita growth is a good performance. Sound policies are, in and of themselves, good for growth. But what is new in our findings is that the sound-policies, high-aid group grew at 3.5 percent per capita. If we look at the developing countries with poor policies, though, we find virtually no per capita growth for either those receiving small amounts of aid or those receiving large amounts. It is the fact that some countries have received large amounts of aid for decades and yet have shown no growth that has given aid a bad reputation. What has often been missed by observers, however, is that there are aid recipients with sound policies and that this group of countries has fared very well. Examples of countries in our sample that have received large amounts of aid and had good economic policies in the early 1990s are Bolivia, El Salvador, Ghana, Honduras, and Mali.

There are, of course, influences other than policies and aid on a country’s growth. In our study, we use a series of growth regressions in an effort to control for these other influences. We also adjust for the fact that some aid is a response to serious, but temporary difficulties in countries. If, for example, a country is given more aid because a typhoon has wiped out part of its rice crop, the data will seem to indicate a high level of aid associated with poor growth. There are, however, ways to adjust statistically for such temporary fluctuations in aid receipts. The finding that aid affects growth in a sound policy environment but not in a poor one holds true in all instances, once adjustments are made to take such temporary influences into account. Our statistical work has provided a rough estimate of the magnitude of the effect: 1 percent of GDP in the aid a country receives increases the growth rate of a good policy performer by 0.4 percentage points, thus implying a rate of return on aid of about 30 percent.

The central objective of ODA is not growth per se, but poverty reduction. We repeated this analysis with an important social indicator—infant mortality—and found an analogous result: in a sound policy environment, an increase of 1 percent of GDP in aid leads to a reduction in infant mortality of 0.9 percent. In a poor policy environment, aid has no effect on infant mortality.

These findings are very intuitive. On the one hand, in a highly distorted policy environment, there are weak incentives for, and low efficiency of, investment. There is no reason to think that pouring money into such an environment would lead to satisfactory outcomes. On the other hand, if poor countries stabilize their macroeconomic situations and liberalize their trade, they can create a sound environment for investment and growth. Our results show that these policies will produce improved results even without an increase in a country’s receipts of aid. However, growth and poverty reduction occur significantly faster in countries to which aid is increased. When a country such as Mali reforms, it is still constrained by its own poverty and by the reluctance of private investors. In such a situation, aid can be used to make investments in human resources and infrastructure that have high returns. Furthermore, increasing inflows of foreign aid, in and of themselves, increase the confidence of private investors in a country’s economy. Our preliminary follow-up results show that in a sound policy environment, aid attracts private investment, whereas in a poor policy environment, it displaces private investment.

Allocation of aid

If aid is to have the maximum impact on growth and poverty reduction, it should be given to poor countries with sound policies. In an effort to evaluate recent experience with aid, our study looked at how ODA was allocated during 1970–93, distinguishing between bilateral and multilateral aid. We found, consistent with other studies, that the provision of bilateral aid is influenced by political and strategic considerations. The influence of political considerations cancels any tendency for bilateral aid to be given to countries with sound policies.

For countries with a given income level, poor policy performers and good policy performers received equal amounts of bilateral aid (Chart 2). For the allocation of multilateral aid, political and strategic considerations did not exert a significant influence. Multilateral aid has favored sound policy environments (Chart 3). For a lower-middle-income country, aid received by the typical good performer was 50 percent higher than that received by the typical poor performer. However, at very low levels of income, the difference between amounts of aid provided to countries with sound policies and to those with poor policies has been minor.

Chart 2Allocation of bilateral aid, 1970–93

(share of aid in real GDP)

Source: Craig Burnside and David Dollar, 1997, Aid, Policies, and Growth, World Bank Working Paper No. 1777 (Washington).

Chart 3Allocation of multilateral aid, 1970–93

(share of aid in real GDP)

Source: Craig Burnside and David Dollar, 1997, Aid, Policies, and Growth, World Bank Working Paper No. 1777 (Washington).

Because the majority of ODA is bilateral, if we add bilateral and multilateral assistance and examine the allocation of total assistance, we find that there has been very little relationship between the distribution of aid and the quality of countries’ policies. In the Cold War period, much assistance was driven by strategic considerations. Although it may have achieved its donors’ strategic goals (a question beyond the scope of our study), we now know that the aid that went to countries with poor policies promoted neither growth nor poverty reduction. Thus, the end of the Cold War has created an opportunity to make ODA more effective overall by strengthening the link between the provision of assistance and the soundness of recipient countries’ policies.

Aid and policy reform

The final question we take up is whether the amounts of aid countries receive affect their policies. In testimony before the US Senate Foreign Relations Committee, Nicholas Eberstadt of the American Enterprise Institute asserted that “Enormous and steady flows of concessional external finance from developed countries have permitted Third World governments to pursue ‘development’ policies that have been wasteful, ill-conceived, unproductive—or even positively destructive.” This is a very strong statement, implying that aid actually makes people worse off by causing bad policies. We wanted to see if there was systematic evidence for this.

The first thing that we found is that there are clear examples that might lead one to believe the above statement. In Zambia, for example, the index of trade, fiscal, and monetary policies got almost continually worse (up through 1993) while aid flows rose higher and higher (Chart 4). This correlation does not, in and of itself, prove anything, but it would be consistent with the argument that aid permitted the maintenance of policies that were bad for the people of Zambia.

Chart 4Zambia: aid and policy

For each Zambia, however, there is a Ghana. Ghana received very little aid during the period it had poor policies, but donor support has been strong since it reformed (Chart 5). Case studies of Ghana generally find that foreign financing has helped consolidate a good reform program, although there have been some recent policy slippages.

Chart 5Ghana: aid and policy

Source: Craig Burnside and David Dollar, 1997, Aid, Policies, and Growth, World Bank Working Paper No. 1777 (Washington).

11.0 = average developing country policy.

The academic literature on policy reform in developing countries finds that the relationship between aid and reform is complex and not necessarily the same across countries. There are countries such as Vietnam that initiated serious reform after aid (in Vietnam’s case, Soviet aid) declined. But Myanmar and Nigeria are counterexamples of countries in which aid has virtually stopped but no serious reforms have ensued.

What our study finds is consistent with the academic literature: there is no systematic relationship between the amount of aid that countries get and policy reform (or the lack of it). It is not true, as a general proposition, that large amounts of aid lead to poor policies, nor does the absence of aid promote sound policies.

We conclude that policy reform is largely driven by domestic social and political forces. Multilateral organizations can provide useful policy advice and exposure to other countries’ experience. Once the quality of a developing country’s policies has reached a certain level, financial aid has a high return, and providing it should increase the probability that reforms will persist by strengthening the benefits the country will receive in terms of economic growth and poverty reduction.

Our study is not primarily about adjustment lending, but has implications for it. The impetus for reform has to come from countries themselves, and there is no point in doing adjustment lending in an unreceptive environment. That one-third of World Bank adjustment loans have failed suggests that it has not been sufficiently selective in providing them. Where countries have developed good reform programs, adjustment loans can be a useful vehicle for providing financing. For the money to earn a high return, the quality of the borrowing country’s policies must be good. When decisions are made about the provision and structuring of adjustment loans, more weight should be placed on demonstrated reform and less on promises that are to be fulfilled after loans have been provided.


Aid can be a powerful tool for promoting growth and reducing poverty. To do this effectively, it should be given to countries that are already helping themselves by putting growth-enhancing policies into place. In the Cold War period, donors—and bilateral donors in particular—did not do this effectively. The aid that went to countries with poor policies was wasted, although it could have helped growth and poverty reduction in countries with sound policy environments.

To have a sense of the magnitude of this inefficiency, we conducted a hypothetical reallocation of aid by eliminating the influence of donors’ strategic interests and doubling the weight placed on policy. This reallocation in effect took some aid away from the group of countries in our sample having poor policies and gave it to the group with sound policies. The estimated impact of this reallocation would be to increase the growth rate of the developing countries in our sample by one-third. Thus, there is the potential to do significant good by allocating aid more efficiently.

A final point worth noting is that there has been a worldwide trend toward economic liberalization in the 1990s. Hence, there are many low-income countries with sound policies today. Ethiopia, India, and Vietnam are three examples of countries that have very low per capita incomes and large populations, and that have made important strides in policy reform in recent years. The international climate for effective assistance is the best it has been in decades. Ironically, at the same time, industrial countries are cutting back on their assistance: in 1996, developed countries gave the lowest share of their combined GNPs in assistance (0.27 percent) since comparable statistics on aid first became available in 1950.

There are a number of reasons for the decline in aid. The end of the Cold War has reduced aid’s strategic role, and fiscal problems in most of the developed countries have also contributed to the cuts. But a third reason is a certain disenchantment with foreign aid fueled by notable failures. The notable failures—aid to Mobutu’s Zaire is a good example—are countries with poor policies getting large amounts of money. The logical response to these failures is to stop funding countries with poor policies—not to stop providing foreign aid altogether. It may be convenient for some donor countries with fiscal problems to think that aid does not work, but the evidence is clear that aid can significantly reduce poverty if it is managed wisely.

This article is based on the authors’ paper, Aid, Policies, and Growth, World Bank Working Paper No. 1777 (Washington, June 1997).

Privatization and Capital Market Development

Strategies to Promote Economic Growth

By Michael P. McLindon

Foreword by Ceslav Ciobanu, Minister, Ministry of Privatization, Moldova

This is the first book to discuss comprehensively the linkages between privatization and capital market development, and how they support and reinforce each other to promote economic growth. It examines four types of privatization, including mass privatization, that are critical to developing capital markets and promoting growth, and analyzes the experiences of 14 countries that have adopted these strategies. It is intended for policy makers in developing countries and in the international donor agencies, for academics and practitioners working in development, and for the internationally minded private sector.

Michael P. McLindon, Ph.D., CFA, has advised on privatization, capital markets and economic reform in more than 20 emerging market countries. He currently works with The Institute for Public-Private Partnerships in Washington, D.C.

Praeger Publishers

88 Post Road West, Box 5007

Westport, CT 06881-5007

Phone: 203-226-3571


Fax: 203-222-1502


ISBN: 0-275-95066-2 • 172 pages

Price: $55.00 (Cloth)


Craig Burnsidea Canadian national, is an Economist in the World Bank’s Development Research Group.

Davod Dollar, a US national, is a Research Manager in the World Bank’s Development Research Group.

Other Resources Citing This Publication