Assessing Assessing Aid

15 January 1999 | Briefings

For low income countries aid remains an essential source of external finance, however, bilateral donors have been cutting back their aid budgets and rethinking the importance and value of foreign assistance. In response to this trend the World Bank has produced Assessing Aid, What Works, What Doesn’t and Why, which examines how aid can be used more effectively. The report finds that aid does work in countries with good economic environments. The basic message is that aid cannot induce unwilling governments to implement reforms therefore policy based lending should be focused on those countries that can demonstrate a commitment to the reform process via the policies they implement. Selecting countries according to their commitment to reform will help to ensure aid resources are used more effectively.

The report has been well received by donors and the media and it is likely to influence donors allocation of resources between countries. What follows is a summary of the report’s main findings and conclusions and a critique of these.

Aid works: in countries that have good economic policies and institutions.

The basic message is that aid works – it helps create growth and therefore poverty reduction – when it is directed to those countries that will make best use of it, i.e. those countries with good economic policies and institutional structures.

Previously, donors have tried to “buy” policy reforms, implying that they have provided most aid to countries with poor economic environments. According to the Bank’s analysis donors have been putting their money precisely where it will have least effect. Aid can be very useful to support reform efforts, i.e. where there is already government ownership of the reform process, but it cannot induce reform.

“When countries 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 usually failed.” [p2]

Selectivity: give more aid to reforming countries where it can be most effectively used.

Because policy based aid (structural adjustment lending) can only nurture (not induce) policy reform, it should be given only to those countries which can demonstrate a strong track record of reform or a willingness to do so.

“Clearly, poor countries with good policies should receive more financing than equally poor countries with weak economic management.” [p4]

Those countries which are not committed to reform should receive technical assistance and project finance which are useful means for disseminating ideas and communicating best practice techniques and stimulating informed policy debate amongst civil society.

Poverty reduction: selectivity can achieve more poverty reduction

Donor agencies have a mixed record of targeting finance to low income reformers. Yet the poorest countries, which account for a large proportion of the worlds poor, are amongst those that are reforming the most and have the best policy environments. Therefore if aid is targeted more selectively to these countries it will be more effective at reducing poverty.

Fungibility: aid effectively supports the entire public sector therefore the general functioning of the economy is important.

Aid is fungible, i.e. a donor may give aid to a particular sector but it may not lead to increased spending in that sector if the government reduces its own spending to it and spends the money saved elsewhere. Therefore, because aid effectively goes to the whole economy, the overall quality of government spending and its ability to use money efficiently is important.

“Government commitment to particular sectors is more important than targeting aid.” [p69]

A good economic environment: good policies and good institutions are essential.

A good economic environment requires good policies and good institutions: these are defined as an open trade regime, low inflation, low budget deficit, protected property rights and efficient public bureaucracies.

Aid is effective in a good economic environment because it can help to crowd in private investment by increasing the confidence of private investors and supporting delivery of important public services: it is a catalyst for growth. On the other hand, in a poor economic environment aid can help governments to maintain poor economic policies and put off reforms which discourages private investment.

A new strategy for aid: good policies and institutions matter more than money.

The rationale for giving aid has been based on the belief that countries have savings gaps (domestic savings are not sufficient to fulfil investment needs) and foreign exchange gaps (foreign exchange earnings are not sufficient to purchase necessary imports). The World Banks analysis suggests that developing countries need finance but they are held back more by policy and institutions gaps. Poor countries have been held back not by a financing gap but by an “institutions gap” and a “policy gap”. [p33]

Therefore the Bank proposes a two pronged development strategy:

  • help governments to adopt market oriented, growth enhancing policies;
  • ensure governments are able to provide services that cannot be well and equitably supplied by the private sector (e.g. infrastructure services and education).

For countries with good policy environments:

  • direct budget support should be given to those governments which allocate aid well and provide “effective, quality services. Non-project finance saves project preparation and implementation costs for donors, and enables better aid co-ordination and less monitoring and reporting by recipients;”
  • in countries where capacity for delivering services (effectiveness of spending) is poor project aid should be targeted at strengthening budget procedures and service delivery.

For countries with poor policy environments and poor institutions:

  • donors should reduce funding and increase support for policy dialogue and institution building.

Strengthening institutions: the quality of spending is as important as the quantity spent.

The quality of government spending is as important as the quantity and allocation of government spending. Donor projects have tended to be insulated, and have operated in isolation, from the rest of the economy, whilst increasing capital stock they have not often contributed to improving government capacity to provide efficient services. For countries lacking effective budgetary procedures or effective service delivery, projects are more important for transferring knowledge. Aid money might be fungible but knowledge learned through projects is not.

” The most critical contribution of projects is not to increase funding to particular sectors, but rather to help improve service delivery by strengthening sectoral and local institutions.” [p3]

Projects should be regarded as vehicles for transferring skills, knowledge and technology, i.e. aid should combine financial support with help to create local knowledge so that governments can improve the quality and effectiveness of public services. The emphasis is on creating and disseminating local knowledge rather than transferring knowledge. Therefore there should be more support for innovative projects. The only rationale for a project is that things should be done differently than they are now.

“This generation and dissemination of knowledge is one of the biggest contributions that development assistance can make” [p92]

“Development projects can be a testing ground for ideas or concepts that are new to a country, demonstrating to the government and citizens what does (or does not) work” [p97]

Stimulating reform in poor environments: encouraging policy dialogue.

In countries with bad policies and institutions aid can be used to support reform minded groups and individuals locally and nationally. Aid can be help reformers to test and refine their ideas by putting them into practice. Aid can also help to promote policy reform over the long-term by disseminating development ideas, training the next generation of leaders, and stimulating policy debate amongst civil society.

“It is possible to assist development in countries with weak institutions and policies, but the focus needs to be on supporting reformers rather than disbursing money.” [p116]

“There are few hard research results to show that disseminating good ideas, sending students abroad, or stimulating policy debates in civil society leads to developing country policy reform and better performance. But case studies suggest that these reforms are often important. Moreover, such activities are inexpensive.” [p57]

Participation: an active civil society improves public services.

Top down, technical approaches to project design and service delivery have not worked in areas critical for development. A participatory approach often results in huge improvements.

Civil liberties peoples ability to speak out matter. The probability of a project failing is 50% higher in countries with fewer civil liberties.

Involving civil society requires longer project preparation but this conflicts with donors practices, which are often motivated by “moving money”; skewing aid in favour of large projects and centralised agencies to reduce administration and time costs. Aid agencies need to redesign their incentive and reward systems and operational practices.

Donor co-ordination: co-ordination aids effectiveness.

Large numbers of donors leads to a proliferation of projects and poor co-ordination among them, which hampers aids effectiveness and increases the burden for recipients of co-ordinating, monitoring and reporting on project implementation. Cases of successful assistance are characterised by strong partnership amongst donors and a focus on large, long-term transformations rather than individual projects.

“Results oriented donors should worry less about planting their flags on particular projects and more about how communities, governments, and donors working together can improve services.” [p118]

Criticisms of the report

The Assessing Aid report makes some useful comments. In particular, focusing aid resources on helping governments to improve their administrative procedures and capacity to deliver effective services will be useful. This will be greatly helped if aid donors are able to improve their procedures, especially by co-ordinating their support and addressing their incentive structures. However, some of the assumptions implicitly made in the report are doubtful; the analysis is limited; and aspects of the “new strategy” appear to conflict with other aims such as ownership or simply repackage elements of the current process.

Ownership verses selectivity

The report acknowledges that “donor financing with strong conditionality but without strong domestic leadership and political support has generally failed to produce lasting change”. However, given the tight policy formulation identified in the report, ownership and selectivity can only be complementary when recipient governments programme reflects what the World Bank considers to be good policy. The implication is that governments wishing to take an alternative economic approach must expect to forego policy based lending. Alternatively, if finance needs are pressing they may feel compelled to design programmes which reflect the World Banks view point even if they are not committed to it.

This raises concerns about how the selectivity approach can be reconciled with the Banks aim to move towards a “partnership approach” to developing and funding programmes. Whereby, the objective is to encourage governments to develop their own programmes in consultation with civil society, and for donors to co-ordinate their aid to support the governments reform agenda.

Selectivity: another form of conditionality?

As the report acknowledges, often conditionality attached to loans has not been adhered to. However, rather than offering a new approach, the selectivity approach simply repackages conditionality; forcing governments to implement conditionality (i.e. adhere to a specific policy formula) “up-front”, that is, before they receive any money.

Adjustment loans are provided in order to assist the adjustment process, if governments are required to adjust before they receive the loans then adjustment may become much harder to achieve and costlier in terms of the impact on civil society.

Can donors be selective?

Political considerations and colonial relationships have tended to skew aid, particularly bilateral aid, in favour of certain countries. Whilst this may not be the most efficient way of allocating aid resources it is unlikely that donors can change their allocation habits that easily, for example, there are institutional costs attached to shifting aid between countries which might be disproportionately large for small donors. Donors may perceive that they are locked into relationships with countries because of historical relationships, especially if these countries are particularly poor.

A mechanism which locks in donors and recipients is defensive lending, i.e. when a donor carries on providing aid to a recipient so that they can repay previous loans made to them. The IMF and World Bank have been particularly engaged in this activity. Without substantial debt relief it is likely that the pressure to provide new aid for defensive lending will continue whether a country has a good economic environment or not.

It is also likely that donors have a tendency to be overly optimistic about governments commitment to there form process. For example, the Bank identified Kenya as a worthy recipient of its Higher Impact Adjustment Lending, even though it considered Kenya to be a “poor complier”, on the grounds that the government had signalled that it was now committed to reform. However, the government did not comply with its HIAL programme and it broke down. In another case, the IMF has repeatedly halted loans to Zimbabwe, but after renewed government commitment to the reform programme, the IMF has once again resumed lending.

Poor economic methodology

The background study by Dollar and Burnside, on which Assessing Aid is based, has been criticised for its poor methodology. At a meeting with the World Bank, one academic commented that it was not clear that the study was based on any relevant econometric work. Criticisms of the study are:

  • Dollar and Burnside analyse the impact of aid on growth since the early 1970s even though economic reform was not a specific aim of donors until the 1980s. By lumping all types of aid together it assumes all aid affects growth to the same extent, i.e. the relationship between aid and growth is stable whatever its purpose. However, intuitively it can be anticipated that emergency aid, for example, will have a different impact on growth to structural adjustment support.
  • The analysis of the effects of economic policies on growth is too narrow and incomplete, and only succeeds in explaining a third of growth performance. Given that Dollar and Burnsides model does not capture all the determinants of growth, if aid is correlated with any omitted variables then the equation will be subject to specification error and the coefficients biased, and thus they should not to be used for inference.
  • The study fails to take account of any possible lags which may exist in the aid-growth relationship. It cannot however be assumed that the impact of aid on growth is constant either across countries or across time, thus it is likely that the aid coefficient in the regression is unstable.
  • Dollar and Burnside also fail to consider other factors such as the pressure to use aid to repay loans and whether options such as debt forgiveness might be equally as important for improving aid effectiveness.
  • Data for developing countries tends to be particularly poor, raising concerns about the robustness of the conclusions derived on the basis of such a data set.

Good policy environment

Whilst it can generally be agreed that aid will be more effective in an economic environment in which policies are coherent and consistently implemented and supported by sound institutions, the Bank defines “good economic policies” very narrowly without examining other policy formulations. Thus although the Bank claims to be objective, it has subjectively chosen only to test those policies which reflect its own view point.

“good policy is not something subjectively decided by the World Bank; lessons emerge from the experiences of developing countries. Good management is, objectively, what has increased growth and reduced poverty in those countries.” [p54]

Not only is a “good” policy environment too narrowly defined, the consensus on what defines “good” policies is subject to change. What may have been regarded as a good policy yesterday may not be today. For example, in the 1970s the Bank supported statist-type policies, in the 1980s it promoted free-market based reforms, and now Joseph Stiglitz, World Bank Chief Economist, has acknowledged problems with the current “Washington Consensus” and is propounding a “Post-Washington Consensus.”

It may also be the case that aid worked in a “good policy” environment because in these cases governments agreed with the policies advocated by the Bank and Fund and so they were committed to the reform process. A governments willingness to sustain the reform effort, may be as important as the policy mix, but this is not investigated.

Inequality and poverty reduction

Inequality is not built into the policy matrix and there is no analysis of whether certain patterns of growth are more important for reducing poverty than others. The paper implicitly assumes that growth will have a positive impact on poverty or at least will not worsen it; absent is any analysis of which “good policies” lead to poverty reduction.

From good to bad policies

The report does not consider what would happen to a country in receipt of aid if its policy environment deteriorated. No mechanisms are outlined for determining at what stage deterioration would be considered permanent rather than temporary, and how donors should revise their aid allocation decisions. As mentioned above, donors often have great difficulty in reducing aid allocations to countries and will put off such decisions until the donor recipient relationship has completely broken down, especially if countries have previously been considered “good reformers”.

Fungibility of aid

Assessing Aid assumes that all aid is fungible, therefore there is no benefit in allocating aid to particular sectors etc. However, this is not necessarily so, in some cases selecting sectors for additional support can be effective. Therefore, the effectiveness of aid in particular countries needs to be assessed on a case by case basis.

Moreover, whilst it is correct to suggest that donor funds will be allocated to marginal projects, i.e. those which the government cannot finance from its own resources, it cannot be assumed that these projects do not have very significant economic benefits. Government spending in the poorest countries is very seriously constrained, particularly because a large proportion is used to pay interest charges. In these cases, there are likely to be many worthy projects that the government would like to finance but cannot.

Innovative projects

The report suggests that donors should fund innovative projects as a means of gaining and transferring knowledge: failed projects can be as useful as successful projects in this respect, i.e. they identify what not to do. However, if projects are financed with loans rather than grants this raises the question of how the loan scan be repaid if the projects do not earn money or whether they should be repaid. The IMF and World Bank have no mechanism for forgiving loans for failed projects.

Deficit before grant

Governments are often not allowed to use donor funds productively. The IMF does not include grant money in its assessment of governments revenue, therefore, spending is very tightly constrained. The implication is that governments are often unable to use aid money for productive purposes and they are required to retain it in the central bank to accumulate as reserves. If aid is to be used more productively thenthis practice must change.

Incentives and donor co-ordination

Donor practices and procedures have contributed considerably to aid’s ineffectiveness. The report makes some useful recommendations for improving donors aid allocation and evaluation procedures and internal incentive structures; donor reform will be a crucial

Incentives and donor co-ordination

Donor practices and procedures have contributed considerably to aid’s ineffectiveness.

The report makes some useful recommendations for improving donors aid allocation and evaluation procedures and internal incentive structures; donor reform will be a crucial element in improving aids effectiveness. However, institutional inertia will make such a task extremely difficult, as the Banks experience with its internal reform process demonstrates. Conflicting incentives can make donor co-ordination unmanageable. Reciprocal conditionality, i.e. conditionality which the donors will abide by, may help this process.

Angela Wood, Bretton Woods Project, January 1999.