Southern African Regional Poverty Network (SARPN) SARPN thematic photo
Regional themes > General Last update: 2020-11-27  

 Related documents

What would doubling aid do for macroeconomic management in Africa?

Mick Foster and Tony Killick

Overseas Development Institute (ODI)

April 2006

SARPN acknowledges the World Institute for Development Economic Research as the source of this document.
[Download complete version - 166Kb ~ 1 min (43 pages)]     [ Share with a friend  ]

Background and approach

Britain and other donor governments have promised to double aid to Africa by 2010. The prospect of such a ‘scaling-up’ sounds a blessing to Ministers of Finance struggling to balance their budgets but the reality is more complex. Remember that many African countries already received large aid inflows relative to the scale of their economies and budgets. A doubling of this could bring in extra foreign exchange, investible resources and government revenues on a scale typically equivalent to a fifth of national income, 100% of existing investment and a very large proportion of tax revenues. Whatever the benefits, a ‘macroeconomic shock’ on this scale could not fail to have large balance of payments and monetary and fiscal consequences, posing real challenges for economic management. There are also related questions about absorptive capacity. The purpose of this paper is to explore the consequences of aid scaling-up for macroeconomic management, to report on the results of recent research and to consider the policy implications.

  1. Coverage of study

    This short synthesis paper tries to distil some general lessons and findings on the macroeconomic consequences of scaling up aid flows. It draws on three main sources:-

    1. four country case studies commissioned by ODI (Mauritania, Mozambique, Sierra Leone, Tanzania);

    2. a recent IMF study of the macroeconomic consequences of scaling up aid that presented five country case studies based on a common analytical framework (Ethiopia, Ghana, Mozambique, Tanzania, Uganda) (IMF, 2005; Aiyar et al., 2005); and

    3. a review of the literature on commodity export booms, to explore whether it provides lessons relevant to the macroeconomics of increased aid.

    Two of the countries covered by ODI were also studied by the IMF, giving a total sample of seven countries. The ODI studies were commissioned before the IMF study was produced, and did not have a common methodology. In order to facilitate the drawing of conclusions, an attempt has been made to apply a broadly similar analytical framework to present some of the material from the two ODI cases not covered in the IMF study.

    The focus on Africa is appropriate, given that the region has by far the highest relative aid levels, and is expected to receive the steepest increase in future aid in order to reach the MDGs. According to projections by OECD/DAC, aid to Africa is expected to more than double between 2004 and 2010.1

  2. Data problems

    There are serious disparities in data on aid between different sources. The OECD/DAC publishes data on donor commitments and disbursements, based on donor self-reporting, and is the source for statistics published in the World Bank’s World Development Indicators. Independently, governments collect data on donor commitments and disbursements. These tend to differ from the OECD/DAC statistics and to be generally lower for a number of reasons:

    1. donors do not declare the full cost of technical assistance to government and may include quasi-administrative costs of managing the programme in aid statistics, which they may also decline to declare;

    2. not all aid goes to government, and donors do not necessarily report support channelled through NGOs. There are also risks of double counting for such support;

    3. aid given to a particular country may come through more than one route, and data on centrally managed programmes may not always be known by donors reporting locally;

    4. donors do not always report their disbursements, or may report them only to the recipient institution, which may not necessarily report them to the Ministry of Finance;

    5. only a proportion of aid notionally designated as going to the government passes through its budgetary processes. Concentrating on this part of aid may give an inaccurate impression of the overall trend; and

    6. donor data may refer to different financial years and different currencies, and may be reported as ‘disbursed’ at a different point in the processing of the transaction.

    In varying degrees, all our country studies struggled with the absence of reliable, let alone definitive, data on aid receipts and no doubt the same difficulties affected the IMF study. In some cases, these problems were compounded by weaknesses in country macroeconomic data, most notably in the case of Sierra Leone, whose statistical series were severely disrupted by civil war. The Mozambique paper accurately refers to aid data as being ‘seriously and famously not accurate’.

    The reliability of the following analysis, and conclusions reached, are inevitably qualified by these data problems. Indeed, one of the policy inferences – unfortunately an all too familiar one – is the urgent need for improved data if the management and effectiveness of aid are to be improved.

    In all honesty, it should also be admitted that we found it difficult to identify unambiguous and relevant country cases of large aid surges. This reinforces the need for caution when drawing general conclusions from the cases identified. In Mauritania, for example, what is treated as an aid surge could equally well be regarded as a fluctuation around a rather flat trend, and the value of the Sierra Leone study is limited by its special circumstances as a country emerging from a recent civil war.

  3. Approach and structure

    ODI did not seek to impose a uniform methodology on its country studies, although the terms of reference were common to all of them. It was believed that each case would have important particularities and that to impose a single framework would risk losing a good deal of richness in the analysis. Against this, the IMF (2005) study, which was more in the nature of desk research, did follow a common framework across its countries and the advantages of that were revealed by the comparability of their country results.

    In what follows, we seek to combine the strengths of both approaches, by applying the basic IMF methodology to all cases, so far as the data permit, while at the same time going beyond the IMF framework in a number of respects.

    One important area in which we depart from the Fund approach is by making more of a distinction between the short run (which is what the Fund paper is mainly concerned with) and the longer term. The point of this is that, while it is justifiable for short-term analysis to take institutions as given and to disregard the supply-side effects of aid inflows, for the longer term both factors need to be taken into account more.

    For the purposes of short-term analysis and in the spirit of the IMF paper, a key distinction is made between the absorption and spending of aid inflows. When aid is transferred to an economy, it is useful to distinguish between the transfer of foreign exchange and the spending that it helps to finance. The foreign exchange accrues in the first instance to the central bank reserves, while the recipient of the aid (usually the government, although it could be an NGO) is credited with the value in domestic currency, which is available to spend or save. If the aid is provided ‘in kind’, the spending of the aid is simultaneous with the use of the foreign exchange. Even in this case, the distinction is valuable, however, since the aid might pay for imports that would otherwise have been funded from other sources, increasing both the foreign exchange available for other purposes and the domestic bank balances available for spending. We follow the IMF in referring to the utilisation of the foreign exchange as ‘absorption’, and to the utilisation of the domestic counterpart as ‘spending’ of the aid (Box 1).

    Box 1: The distinction between absorption and spending of aid

    When aid is transferred to an economy the foreign exchange accrues in the first instance to central bank reserves, while the recipient government is credited with the value in domestic currency. We refer to the utilisation of the foreign exchange as ‘absorption’, and to the utilisation of the domestic counterpart as ‘spending’ of the aid. Aid is ‘absorbed’ when the balance of payments current account deficit (excluding aid) increases, either because more is imported or increased domestic demand causes producers to export less. Aid is ‘spent’ when the fiscal deficit (excluding aid) increases, either as a result of higher government expenditure, or lower domestic revenue.

    • Absorption is defined as the widening of the current account deficit (excluding aid) due to more aid. Absorption depends on both exchange rate policy and on policies that influence the demand for imports.

    • Spending is defined as the widening of the fiscal deficit (excluding aid) due to additional aid.
    The importance of this distinction is that aid only enables an economy to invest and consume more by financing an increase in imports. If the aid is simply spent on domestically produced goods and services, it does nothing to increase their supply. Unless there is spare capacity in the economy, the inevitable result is an increase in inflationary pressures.

    The importance of the distinction is that aid only enables an economy to invest and consume more than it otherwise would, by financing an increase in net imports of goods and services. Real resources are only transferred when the aid results in an increase in net imports. If the aid is simply spent on domestically produced goods and services, it does nothing to increase their supply. Unless there is spare capacity in the economy, extra local expenditure financed by aid will simply squeeze out existing customers for the goods and services bought with the aid. If the increased expenditure financed by aid does not result in increased net imports, it achieves nothing that could not be achieved through expansionary domestic fiscal and monetary policy.

    The qualification in the last paragraph concerning the existence of spare capacity is potentially important, however. All economies operate within their production possibility boundaries, perhaps especially so in Africa, and in that sense there is always spare capacity. However, the obstacles to eliminating this are often profound and long-term, and it is only exceptionally the case in African economies that much of the spare capacity is a result of demand deflation. In that sense, the assumption of full capacity has justification, at least in the short term. However, we should note as a special case the situation of economies bouncing back after civil conflict or some other major catastrophe, such as Sierra Leone. In such cases, it is indeed possible to achieve quite rapid increases in output by improved utilisation of production capabilities that had been dislocated by the catastrophe.

    When aid is spent on local goods and services, the normal route is that donors provide foreign exchange to the central bank, which credits the government account with the local currency equivalent. Absorbing the aid requires the central bank to sell the foreign exchange to finance imports. This may require some appreciation of the real exchange rate to persuade the market to buy the foreign exchange. The extent of the real appreciation will depend on the responsiveness of imports and exports to changes in the level of demand and in relative prices. The necessary adjustment may come through first round effects (government uses the aid to finance imports) or through indirect second round effects (those from whom government makes purchases increase their net imports). The necessary adjustment can come either through increased demand for imports, or through reduced supply of exports as firms respond to increased demand and better relative prices by diverting some of their output to the domestic market. In African-type economies, there are typically limited possibilities of switching export production to meet domestic demand, so the main adjustment is likely to be effected by increased imports.

    The adjustment in relative prices that may be required to enable the increased aid to be absorbed may raise fears of ‘Dutch Disease’, whereby the traded goods sectors of the economy lose competitiveness as a consequence of exchange-rate appreciation. If the traded goods sectors are regarded as particularly important because of their leading role in transferring technology, it is argued that the need for a real appreciation can damage the long-run growth potential of the economy. The benefits of the aid-funded expenditures would need to be offset against any negative effects.

    The remainder of this paper is structured as follows. Section 2 sets out some salient facts about the countries studied, particularly the recent histories of aid flows into them. Section 3 then briefly considers what might be an ideal policy response to an aid surge, drawing on lessons from past commodity booms and examining the factors which might determine an optimal policy response. The determination of an optimal response turns out to be more complex than might have been expected.

    Section 4 presents findings from the country studies, applying the absorptionspending distinction just described, and Section 5 takes this further by focusing on the important question of whether an aid surge is likely to generate Dutch Disease symptoms. Section 6 concludes and examines the implications of the study for the scaling-up of aid.

  1. See

Octoplus Information Solutions Top of page | Home | Contact SARPN | Disclaimer