Blog
Aid Volatility & Aid Heterogeneity

by George Mavrotas

Aid volatility issues of growing importance for the attainment of the MDGs

Recent years have witnessed a growing interest in aid effectiveness issues and although recent empirical work seems to suggest that overall aid works in enhancing growth a number of issues still remain unresolved. An issue that has attracted particular attention quite recently is related to the volatility of aid inflows.

• The issue of volatility of aid inflows is now becoming particularly important in view of the ongoing discussion and debate on how to finance the Millennium Development Goals (MDGs). Aid volatility, and in particular the unpredictability of aid flows, is of crucial importance for the attainment of the MDGs.
• Aid volatility issues have also been raised recently within the context of the British proposal for an International Finance Facility (IFF), a mechanism to frontload aid flows so that MDGs can be met by 2015.
• Very recently, aid volatility issues have also been briefly discussed in connection with aid to difficult partnership countries and fragile states.
• Finally, there has been much recent debate on whether more aid can be spent effectively in developing countries, particularly in sub-Saharan Africa, in view of potential absorptive capacity constraints and diminishing returns to aid.

However, despite these pressing policy questions, there has been surprisingly little empirical work on the measurement of aid volatility. Using simple measures of aid volatility a few papers have found that aid volatility is bad for economic growth, ceteris paribus, and that aid is often among the most volatile sources of foreign exchange income.

More recently, it was found that aid flows are more volatile than domestic fiscal revenues and tend also to be pro-cyclical; another finding was that fiscal planners are highly uncertain of aid receipts, the information content of aid commitments being either very small or statistically insignificant. (1)​

Linking together aid volatility and aid heterogeneity is crucial

One drawback of much of the recent work on aid volatility (and actually of the aid effectiveness empirical literature as a whole) is the use of a single aggregate for aid. However, distinguishing among different aid modalities is crucial, because different types of aid are likely to have different degrees of volatility. Furthermore, different aid modalities have different conditionality. Certain types of aid such as emergency aid and, arguably, programme aid, should exhibit a high degree of volatility, since they are designed to deal with local economic and social crises. On the other hand, sector aid volatility, which is designed to promote investment in physical and human capital, is more likely to be detrimental to long-term economic and social development.

New evidence on aid volatility

The present article briefly discusses new evidence obtained on the volatility of aid by a very recent WIDER empirical study on the subject conducted by Fielding and Mavrotas (2005). The key purpose of the above study (which was prepared within the WIDER project 'Development Aid: A Fresh Look') was to determine the factors driving the cross-country variation in aid volatility. The approach employed differs from previous studies in a number of ways: 

• First, the important issue of aid heterogeneity was taken into account instead of treating aid as a single aggregate. In view of this, the WIDER study focused on two types of aid, namely sector-specific aid (i.e. 'project' aid) and non-sector allocable aid ('programme' aid) which, if taken together, make up more than 95 per cent of total aid volumes.
• Second, the key volatility concept that was employed was that of a shock to aid, so the aid series were conditioned on an information set of lagged macroeconomic variables.
• Third, the work explored the factors affecting aid volatility by using a modelling framework that includes the size of aid flows, per capita income, institutional quality and policy regime in aid recipients.
• Finally, the study examined aid volatility using data for 66 aid recipients spanning the period 1973- 2002 thus incorporating more recent data on aid volatility as compared to earlier work. 

Key findings of the WIDER aid volatility study

Sector aid volatility is positively correlated with programme aid volatility:

• a 1 per cent increase in aid as a proportion of GNI is associated with a fall in conditional sector aid volatility of around 0.16-0.17 per cent, and a 1 per cent increase in per capita income is associated with a rise in volatility of around 0.35-0.40 per cent; similar results were found regarding programme aid volatility, despite the fact that programme aid is rather more volatile than project assistance;
• the institutional quality of the aid recipient seems to reduce the volatility of sector aid;
• more open economies, which tend to be smaller and richer, ceteris paribus, experience more volatile sector aid flows;
• the quality of institutions and the degree of openness in aid recipients do not seem to be important factors concerning the volatility of programme assistance.

Some policy implications

The above empirical findings emphasize inter alia the importance for donors of speedy implementation of the Rome Declaration on Aid Harmonization, which will lead to substantial reductions in aid volatility.

More generally, it is imperative that donors explore new sources of financing accompanied by less volatility (for example, the IFF) so that the MDGs can be attained by 2015.

Last but not least, the WIDER study points to the importance of aid heterogeneity in explaining the volatility of aid inflows. This has crucial policy implications with respect to progress towards achievement of the MDGs, since the differing degrees of volatility would not be apparent if a single aggregate for aid were employed. This is particularly important for aid recipient governments who are attempting to manage aid volatility by some combination of adjustment to tax and spending plans, adjustment of foreign exchange reserves or domestic non-monetary financing. For these countries, improved forecasting of both short- term and medium-term aid is also crucial. 

(1) Buli0, A. and J. Hamann (2003), 'Aid Volatility: An Empirical Assessment', IMF Staff Papers, 50: 65-89. 

George Mavrotas is a Research Fellow & Project Director at UNU-WIDER currently co-directing two WIDER research projects on 'Development Aid: A Fresh Look' with Mark McGillivray and 'Financial Sector Development for Growth and Poverty Reduction', with Basudeb Guha-Khasnobis