Research Brief
Aid Volatility Across Development Sectors

A key pledge of the Paris Deceleration of 2005 was that aid flows would be made more predictable. This is a key goal as aid shortfalls can cause a government to disproportionately cut their investments, while sudden spikes in aid can lead to a dramatic boost in government consumption. The majority of the work done on the issue of aid volatility focuses on the effect these reactions to variation in total levels of aid has had on key macro variables such as growth and government expenditure. In the WIDER Working Paper 'Consequences of Aid Volatility for Macroeconomic Management and Aid Effectiveness' John Hudson gives a more detailed analysis of the issue by focusing on the impact of the aid volatility of different aid sectors, and on the effect such volatility has on specific targets such as healthcare and education.

Why disaggregate aid sectors and aid targets?

Aid volatility is often studied at the aggregate level. Researchers look at how total amounts of aid vary over time and use this data to draw conclusions about aid volatility. However, Hudson suggests that analyses of volatility that only look at total amounts cannot provide us with a full picture of the phenomenon. For instance, we should not assume that aid volatility will have the same effects across different sectors. It may be the case that volatility in one sector has only a small negative effect on the achievement of goals, while the negative effect is much larger in other sectors. Similarly, one should question whether volatility in aid aimed at social goals such as health is likely to have as big an effect on economic growth as volatility in aid for infrastructure. If only the volatility of total aid is looked at, these questions can't be answered with any degree of accuracy. Furthermore, Hudson suggests, researchers only looking at the totality of aid may underestimate aid volatility. For example if health aid and education aid are both highly volatile but also negatively correlated then this volatility will not show up in a study looking at volatility in the overall amount of aid.

It is also important to look beyond the macro variables, such as growth and government expenditure, as much aid does not even attempt to have a direct effect at this level. Aid used to build hospitals aims to improve healthcare, building a road between two cities promotes trade and movement between those two cities, and aid aimed at promoting education does just that. When judging the impact of aid volatility, we should be, Hudson argues, judging the impact that it has on the actual goals of the specific aid in question, together with any negative side effects.

Aid sectors and aid targets

Hudson begins by attempting to discover what role each of the component parts of aid play in terms of overall aid volatility. His results back his theory that focusing only on total aid fails to address important parts of aid volatility. For instance in 2004 aid volatilities in individual sectors compensated for each other as some sectors experienced increases and others reductions in aid, this therefore reduced the volatility of overall aid. In 2009, however, the opposite occurred, with sectoral volatilities in the same direction, which was thus reflected in overall aid volatility.

When looked at individually, the health education and social sectors tend to show less volatility. Debt aid , aid for industry, programme assistance and government aid are the most volatile sectors. In particular if debt aid is removed from the equation then overall aid volatility is much reduced. The sectors which are most individually volatile differ slightly from those which contribute most to overall volatility.

Trends in aid volatility

Hudson next looks at how donors adjust their aid distribution in the face of aid volatility. He identifies that positive/negative aid volatility is not simply short-lived and followed by a return to normal aid levels, but that there is a tendency for donors to compensate with negative/positive volatility in the following period. Hudson's results also show that there are some spillover effects on donor behavior; volatility in one aid sector does sometimes lead to volatility in another. In general these spillover effects are negative; a period of positive aid volatility in one sector is often compensated for by a period of negative volatility in other sectors. However Hudson points out that when it comes to government aid, programme assistance, and humanitarian aid there can be a positive spillover which causes aid levels to increase in other sectors.

Aid volatility and aid goals

In the final section of his paper Hudson turns to look at the impact that the different aid sectors, and aid volatility have on the outcomes of aid. While Hudson acknowledges that his data does not span as much time as would be desirable, his results do provide statistically significant evidence that aid volatility does have an impact on whether aid can achieve specific goals. For example both infrastructure and industry aid are important for achieving communications goals, and aid aimed at building social infrastructure is important for improving targets linked to education and healthcare. Understanding that volatility in individual aid sectors has an effect on the likelihood of achieving the goals of those sectors, provides further evidence that this kind of disaggregation is important when looking at aid volatility.

Hudson concludes that his research shows the importance of considering different aid sectors when discussing aid volatility. Looking at volatility at only the aggregate level hides the facts that the volatility of individual aid sectors is often greater than overall aid volatility, and that volatility has different effects in different aid sectors. Hudson finishes by suggesting that as more data becomes available, more aspects of aid volatility will be able to be studied. He highlights the question of how aid volatility differs between different donors as one particularly important area for future research.