Research Brief
Aid and Government Fiscal Behaviour

What Does the Evidence Say?

Donors are concerned about how their aid is used, especially how it affects public spending. For low-income countries that receive significant amounts of aid relative to GDP, most of the aid spent in the country is given to the government either directly, or by financing services that would otherwise be a demand on the budget. To illustrate, in 1997 aid accounted for almost a third of government spending on average in low-income or sub-Saharan African countries, and over 100% in some countries. On average during the 1990s aid amounted to over 100% of government spending in Burundi, Ethiopia and Sierra Leone; during 2001-07 it was over 100% in Sierra Leone and Zambia, 70% in Uganda, 50% in Ghana, and 20% in Kenya. It follows that aid should have a direct and significant impact on the level (relative to GDP), evolution (increases over time) and composition (the allocation of spending to different types of public goods and services) of government expenditure. Aid may also affect the level of tax revenue, either because it influences tax effort or because policy reforms associated with conditional lending affect tax rates or the tax base.

My WIDER working paper ‘Aid and Government Fiscal Behaviour: What Does the Evidence Say?’ reviews the recent evidence on the fiscal effects of aid, on government spending and tax effort, in recipient countries. Given severe data limitations in analysing the relationship between aid and spending, little more can be claimed other than that aid finances government spending. Spending may not increase by the full amount of aid, either because the aid is used to reduce domestic borrowing (often a requirement of multilateral agencies) or is not actually reflected in the budget. Considerable amounts of aid are disbursed through donor projects that do not appear as spending in the budgets; indeed, to the extent the government is aware of these projects it may reduce its own spending in these areas. Thus, the direct link between aid and government spending is weak, although there are indirect effects whereby spending increases in line with the allocation favoured by donors over time.

Country studies of fiscal response show that aid does contribute to increased expenditure in total and in the sectors favoured by donors, and that aid has no consistent effect on tax revenue (there is no robust evidence that aid, whether grants or loans, reduces tax effort). As governments in low-income countries have limited ability to alter tax revenue in the short term but can readily alter borrowing, the observed association between the change in aid and spending in any year is largely determined by changes in borrowing behaviour. There are legitimate concerns about how aid is used and whether it is fungible, but there is very little evidence that these present a challenge to aid effectiveness, at least in fiscal terms. The fiscal environment for aid has improved markedly in many recipients, presenting an opportunity in many countries to reduce the transactions costs by delivering aid in the form of budget support that enhances the fiscal effectiveness of aid.

Three generalizations are permitted by the evidence: (i) aid finances government spending; (ii) the extent to which aid is fungible is over-stated and even where it is fungible this does not appear to make the aid less effective; (iii) there is no systematic effect of aid on tax effort. Beyond these conclusions effects are country-specific.