[Back to index] [To Volume12#3 -- full graphics]

From Africa Recovery, Vol.12#3 (December 1998), page 1

World Bank calls for better aid targeting

Major new study evaluates the effectiveness of aid

By Jacqueline Irving

Aid will work effectively only if it is targeted to countries with "sound" policies and institutions, the World Bank states in a major report published in November. According to Assessing Aid -- What Works, What Doesn't, and Why, poverty reduction efforts would be more successful if donors directed their assistance to the right countries, for aid is "wasted" when it is disbursed to countries with "unsound" policies and institutions. "An across-the-board $10 bn increase in aid would lift seven million people out of poverty, while a targeted increase could lift 25 million out of poverty," the report claims.

At a time when development aid has hit the lowest level in its 50-year history, some have interpreted the report as a possible justification for aid cuts. But the Bank's Chief Economist Joseph Stiglitz argues that "the report encourages donors to contribute more aid where it will have the most impact." By stirring debate, the report helps draw needed attention to acknowledged inefficiencies of aid distribution and the shrinking size of the overall aid "pie."

Between 1991 and 1997, the report notes, official development assistance (ODA) from the 21 biggest donors fell from an average of 0.33 per cent to just 0.22 per cent of their gross national product (GNP). Of all these members of the Organization for Economic Cooperation and Development (OECD), only four reached or exceeded the UN target of 0.7 per cent of GNP last year. And for Africa, the world's most aid-dependent region, ODA from OECD countries fell to $15.4 bn in 1996 from $20.7 bn in 1992. (However, a few donors, including the UK and Sweden, recently announced plans to increase their aid to Africa in 1999.)

Assessing Aid's publication is well timed since it coincides with "rather acrimonious debate about the effectiveness of aid amidst donor fatigue," notes Mr. Kwesi Botchwey, former finance minister of Ghana and currently Director, Africa Research and Programmes of the Harvard Institute for International Development. He commended the report for its role in "structuring the debate and anchoring it against some general observations that are rooted in empirical study rather than just opinions and prejudices."

Indeed, the report frequently cites findings and evidence from multi-country research, World Bank-financed investment projects and aid effectiveness case studies. One of its findings is that "the simple relationship between aid and per capita growth in developing countries is weak, if it exists all." Some countries that have received plenty of aid have grown slowly (such as Zambia) while other countries also receiving substantial assistance have grown rapidly (such as Botswana). Adding to the complexity, there are even countries where significant aid disbursements have been followed by periods of slower or even negative economic growth.

The report argues that the picture becomes clearer when aid recipient countries are divided into two groups according to the "soundness" of their economic policies. Aid can be an important stimulus to economic growth and lead to a faster decline in poverty in those countries with "sound economic management." This is defined as free-market policies that aim for low inflation, low budget deficits, openness to trade, and so on, complemented by strong institutions that discourage corruption and protect property rights. In countries with "unsound economic management" -- where production incentives are weak and institutions undermined by corruption provide inadequate protection of property rights -- the authors cite research findings showing that "whatever the amount of aid growth was miniscule, or even negative."

According to the report, aid flows to countries pursuing "sound" economic policies have the bonus effect of boosting private-sector confidence, thereby "crowding in" further capital inflows. The authors put specific figures to this claim: aid finance equivalent to 1 per cent of gross domestic product (GDP) attracts an additional 1.9 per cent of GDP in private investment. But in "unsound" policy countries, aid has a "crowding out" effect on private investment.

Assessing Aid "makes the important point that good policies have a major bearing on poverty reduction," Mr. Kevin Watkins, Senior Policy Advisor of Oxfam, a major UK-based international non-governmental organization, commented to Africa Recovery. But in "making a case for cost-efficient aid, what is the policy conclusion? That you should put all aid in high performing countries? It's an incredibly 19th century, utilitarian perspective. The cost of Oxfam operating in a country like Mozambique is incredibly high as compared with a country like Bangladesh. Is the inference that we should withdraw resources from [Mozambique], ignore the poor there and put those resources in Bangladesh?"

By addressing the issue of how aid as a scarce resource should be allocated most effectively, the report "points us in the direction of a very interesting question," Mr. Watkins acknowledged. "But they are defining 'effectively' purely in terms of rate of return. So they haven't really helped us answer the question."

Inefficient aid

The report states that aid money was disbursed "in equal amounts to well managed countries and to poorly managed ones," up until the last few years. But many poverty-stricken, "good" policy countries, such as Uganda and Ethiopia, have received a smaller per capita share of aid than have countries pursuing less "sound" policies.

Aid allocations by bilateral and multilateral donors in the 1970-93 period were driven largely by political or strategic considerations -- and so were not necessarily directed to where they would yield the highest returns, the report explains. Further inefficiencies in aid allocation are due to less densely populated countries generally getting more aid per capita than do countries with higher population levels. This discriminates against larger countries, contributing to the overall weakness of the relationship between a country's aid intake and income, the report says.

But the authors believe there is now broad scope for a more focused approach to aid disbursement. They appear to discount any strong link between mass poverty and "unsound" policies, pointing to the many countries that have mass poverty but are also conforming to the global trend of economic liberalization in recent years. Setting up four sub-categories to better assess the relationship between aid money and poverty reduction, the authors say aid has "high impact" in 32 countries with "poverty rates of greater than 50 per cent" but pursuing "above average" policies. This is the "good policy-high poverty" sub-category.

Aid is less effective when given to countries in the "poor policy-high poverty" and "good policy-low poverty" sub-categories. In the latter case, aid works, but the returns are not as high as they would be under the "good policy-high poverty" scenario because "there is less work for [aid] to do." Unsurprisingly, the report says that aid is least effective when it is given to countries in the "poor policy-low poverty" sub-category.

In fact, 75 per cent of the world's poor live in high impact countries, said Mr. David Dollar, the report's lead author, speaking at a Washington, DC, press conference held in November to launch the report. "One of our messages is that there have been a number of successful reforms in Africa during the 1990s, so we really have a very good environment for effective aid," Mr. Dollar added, specifically citing Ethiopia, Uganda and Ghana as being in the "high impact" group.

However, notes Mr. Watkins of Oxfam, "policy environments change quickly. If aid is a long-term business, you can't behave like a capital speculator, moving aid in and out of the biggest markets."

While generally stating its arguments coherently, Assessing Aid sometimes appears inconsistent in its conclusions. For example, in one analysis of the conditions for most effective aid, drawing on data covering 1970-93, the report concludes, "When low-income countries such as Mali, put good policies in place, they perform better, but they still do not match the results for the middle-income countries." Stating that there are several reasons for this outcome, the report speculates that "these countries may have other characteristics that hurt growth -- being landlocked, for instance. But it is also possible that their ability to save and invest is hampered by poverty itself, even when good policies are in place."

But elsewhere in the report, drawing on other data, the authors deem aid targeted to the world's poorest, "good" policy countries more effective in reducing poverty than aid to less-poor countries pursuing "good" policies. However, the latter conclusion is not accompanied by an explanation of how the "other characteristics" that could limit growth in the world's poorest countries are overcome. So the strength of the assumed linkage between growth and poverty reduction varies across different sections of the report.

Money and ideas

Even if a developing country's policies are not "sound," such a country could still benefit significantly from donor assistance, the report emphasizes. Assistance should take a form other than financial aid allocations, however. In these cases, aid "requires patience and a focus on ideas, not money." Indeed, a key theme of Assessing Aid is that improving aid effectiveness requires a redefinition of the role of aid -- as a combination of money and ideas, with the appropriate mix tailored to suit a country's specific situation and problems.

More specifically, the authors recommend that assistance to countries with "unsound" economic management 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." The logic behind this prescription is that, "even where money may not stick," aid projects that focus on transfers of knowledge and advice on institutional capacity improvements can have more lasting benefits.

In trying to explain why aid money continues to go to poor countries pursuing "bad" policies, Assessing Aid points to donors' perception of aid as an incentive for policy reform. But the report's research shows that the relationship between aid and the quality of a country's policies is not a "simple" one and "aid cannot take the lead in promoting reform if there is little local movement in that direction."

"The study highlights the absolute importance of national ownership in the origination of a reform programme and a strong body of national consensus behind it," commented Mr. Botchwey. "If there's no sufficient consensus for reform in the country, no amount of conditionality can force the country to comply if a loan is given [so] the likelihood of a breach is very strong."

Not only does Assessing Aid deem conditional aid ineffective where reforms lack "serious domestic support," but it also points to situations where conditional lending might have been "counterproductive." For example, the report cites the case of Zambia to show that aid can actually enable governments to delay reforms. While Zambia's policies "were poor and getting poorer throughout 1970-93," its aid receipts increased, reaching 11 per cent of gross domestic product by the early 1990s. In Ghana, on the other hand, "bad" policies attracted little aid, while policy changes after 1983 were followed by strong aid flows.

The report's authors consider the possibility of donors correctly anticipating shifts in recipient countries' policies (from "bad" to "good") and timing aid disbursements to coincide with these policy shifts to spur reform. But they conclude that donors have not been successful in predicting these "turning points," since the data show that there is "little relationship between changes in aid and policy reform."

But appropriate timing of aid allocation can be a key means of maximizing aid's effectiveness, notes Assessing Aid. Indicators of progress on reform (such as a newly elected government) can provide cues for appropriate timing of aid disbursements. Aid disbursed too late is akin to a missed opportunity. Aid disbursed too early could delay what might have otherwise been a concrete economic policy reform programme.

Target countries, not sectors

Another common donor practice that generally does not work, according to Assessing Aid, is the disbursement of aid for specific sectors and/or projects. Because aid money typically is fungible -- meaning that a recipient government can use the money as it chooses -- much of the targeted money never reaches the sectors or projects for which it is intended, the report points out. In this part of the analysis, the report says neither corruption nor other administrative interference explains this characteristic of aid flows. Rather, it is due to the economic principle that a person or country will "rationally" allocate (or reallocate) scarce resources. And even when successfully channeled, aid's impact is limited unless it produces more systemic, economy-wide change, the report says.

The authors find that channeling sector or project-specific aid to many African countries generally is easier because donors tend to have stronger control of projects in this region. But this tendency -- which is due to the fact that the bulk of capital expenditure is donor-financed (up to 87 per cent for projects in Uganda) and "the lack of strong budgetary control by either [the recipient] government or society over public investment"- raises other problems. Cited by the report as a "major weakness of aid in Africa," the resultant pattern of donor-driven priorities often impedes the ability of recipient governments to set their own priorities and undermines their commitment to projects that need continuing finance.

While Assessing Aid devotes considerable attention to certain donor practices that do not work, the vital question of how donors could ensure that a larger portion of aid money actually reaches its destination remains largely unanswered. The report notes that the administrative costs associated with a "typical" World Bank project are about $1 mn. And it is widely known that a significant amount of aid money is not actually spent in recipient countries. Much of it goes into donor agency overheads, expatriate consultancy fees and procurement from donor country suppliers. Calling for a review of the ways in which aid is structured and provided, UN Secretary-General Kofi Annan recommended last April that donors aim to ensure that "at least 50 per cent of their aid to Africa is spent in Africa."

The report does suggest that donor agencies themselves need some reform. "Loan quantity had been given primacy over loan quality," concluded World Bank and other reviews of project loan effectiveness. The report also refers to a 1992 World Bank study that noted the rise of "approval cultures" in aid agencies, which used disbursement volume to gauge success. However, Assessing Aid gives few details of the changes now under way in the multilateral donor agencies, other than stating that "sector-wide approaches are replacing individual projects."

In declaring that "when donor projects fail, it is often because of weak [recipient country] institutions and public organizations,"

Assessing Aid seems to put the lion's share of the blame for failed projects and the onus for policy change on recipient countries. The report even seems to endow the "soundness" of a country's institutions and policies with the ability to determine how well coordinated donor efforts will be in that country. "Well-managed countries force coordination on donors, but in the weak environments donors often run amok. It is hard to explain this, except that different donors like to 'plant their flags' on something (anything) tangible," the report states. Non-governmental organizations and African governments have often blamed weak aid coordination on competition among donors, as well as the frequent shifts in donor priorities.

Although it clearly states its preference for aid targeted at specific countries rather than specific projects or sectors, Assessing Aid hardly evaluates how aid has been used when it reaches "good" policy countries.

"It is important to now do a more country-specific analysis of particular types of aid, how the aid was used and the conditions that came with the money so we can form a more rigorous view of why aid of a certain quantum would have a certain effect in one country and not another," Mr. Botchwey told Africa Recovery.

According to Mr. Dollar, the Bank is preparing a subsequent study that will look more closely at how aid is working in eight African countries.

In spite of the inefficient patterns of aid allocation in the past, the authors of Assessing Aid are optimistic. The post-Cold War period has reduced competition among donors to cultivate strategic allies through aid allocations, they say. At the same time, Mr. Stiglitz notes in the report's foreword that "with the end of the Cold War, there is a group that is questioning [aid's] very existence in a world of integrated capital markets." The report -- issued by one of the world's foremost providers of development assistance -- insists that aid remains important. It expects the global trend toward economic liberalization to increase the number of countries in the "sound economic management" grouping, with positive implications for more efficient aid use.

------------------------------------------

BOX 1:

Some World Bank findings and observations on aid

-- There is little or no relationship between aid and per capita growth in developing countries.

-- There is "surprisingly little" or no relationship at all between the amount of aid that countries receive and the quality of their policies.

-- Being a former colony of a major donor is more valuable in attracting bilateral aid than having good management.

-- In countries with good management, aid spurs growth, reduces poverty and helps attract private investment.

-- There is a "long legacy of failed adjustment lending" where there was no strong domestic constituency for reform. "It is worth thinking about how to create reformers and popular movements for reform, but it will not be easy."

-- Aid largely finances government consumption, which has no positive effect on growth. This helps explain why aid is not fueling growth in many developing countries.

-- Well-managed countries "force coordination on donors, but in the weak environments donors often run amok."

-- Donors should worry less about planting their flags on particular projects and more about how communities, governments and donors are working together to improve services.

-- Studies commissioned by multilateral development banks have concluded that loan quantity has been given priority over loan quality.

[Back to index] [To Volume12#3 -- full graphics]

Material from this article may be freely reproduced, with attribution to "Africa Recovery, United Nations".
We would appreciate a copy of the reproduction.

Africa Recovery
Room S-931
United Nations
New York, NY 10017 USA

Tel: (212) 963-6857
Fax: (212) 963-4556
Email: africa_recovery@un.org

Website: www.africarecovery.org
Contact us by email: africa_recovery@un.org