Assessing the impact of aid on growth: systematic new approach ”counts chickens when they hatch”

Foreign aid neither can nor should be the main driver of growth in developing countries. That is the conclusion of research by Dr Michael Clemens and colleagues, which takes the three most influential studies of the relationship between aid and growth and subjects them to systematic scrutiny.

Their study, published in the June 2012 issue of the Economic Journal, takes account of the key fact that different kinds of aid projects can affect growth at different times and to different degrees. It finds that when aid rises by 1% of a recipient country’s GDP, economic growth typically rises by between 0.1 and 0.2 percentage points within the following five to ten years. This positive but modest effect tends to decrease at very high levels of aid receipts.

National budgets for overseas development assistance are under severe pressure from austerity measures across Europe, North America and Japan. To understand the effects of a potential global contraction in foreign aid, it is important to understand the typical effects of aid on recipient countries.

Will an aid contraction lead to slower growth in the developing world? Or to put it differently, does aid achieve one of its central goals, that is, causing recipients’ economies to grow faster than they would have without aid?

Remarkably, it has been difficult for economists to tell. The most influential recent research papers arrive at different conclusions about the effect of aid on growth.

Some studies find that countries that receive foreign aid typically grow faster than countries that don’t. Some find no such robust relationship in the same data. And some find that the relationship is conditional – countries that receive aid grow faster than those that don’t, but only when the recipients pursue sound economic policies, such as keeping inflation low.

Understanding this mess is the goal of this new study. The authors propose that the reason that studies reach different conclusions about how aid affects growth is related to the assumptions they make about the timing of the effects.

Borrowing a phrase from Aesop’s fable ‘The Milkmaid and Her Pail’, they seek to ‘count chickens when they hatch’ – in other words, to test whether growth effects happen at the time they’re expected to happen.

The authors propose and test two reasons why the most influential studies have reached incompatible conclusions about whether aid causes growth.

· First, prior research has primarily tested whether aid receipts in the aggregate are associated with growth in the same time period. But different kinds of aid projects can affect growth at different times. For example, funding for a new road might affect economic activity in short order; funding for a vaccination campaign might only affect growth decades later; and humanitarian assistance may never affect growth.

· Second, it is difficult to disentangle mere correlation from causal relationships. For example, poor growth performance – or its causes – might bring in aid. So if aid inflows are associated with poor growth, it is critical to know which came first. Did poor growth arrive and bring in aid – or did aid fail to boost growth?

The researchers begin with the exact same data and analysis that are used in the three most influential research papers on aid and growth. They then alter them in a series of steps:

· First, they hold constant any confounding factors that might bring both poor growth and poor aid in the same country over time, or the short-term effects of growth on aid.

· Second, they consider only those portions of aid that might be reasonably expected to cause growth within a few years. For example, aid for a new road or seaport is included, but aid for primary education or emergency humanitarian assistance is not.

· Third, revealing that some of the most-cited studies do not use all the data available even when they were written, they expand the analysis to all possible countries and years.

After these adjustments, the incompatible results in the previous studies disappear. The analysis yields one finding, which is that when aid rises by 1% of a recipient’s GDP, the recipient’s economic growth typically rises by 0.1-0.2 percentage points within the following five to ten years.

Because the study systematically rules out alternative explanations for this relationship, the finding is very likely to represent the effect of aid on subsequent growth.

The effect is seen across all aid recipients collectively on average and not necessarily in any given recipient; indeed, it varies greatly from recipient to recipient. Because this positive effect is modest and tends to decrease at very high levels of aid receipts, the results do not suggest that aid can or should be the main driver of growth.

‘Counting Chickens When They Hatch: Timing and the Effects of Aid on Growth’ by Michael Clemens, Steven Radelet, Rikhil Bhavani and Samuel Bazzi is published in the June 2012 issue of the Economic Journal.