The Effectiveness of Aid to Development. Focus on the Aid-Growth literature.
par François Defourny
Facultés N-D de la Paix de Namur - Université Catholique de Louvain - Master in International and Development Economics 2005
Our initial objective was to identify some conclusions about the macroeconomic effectiveness of aid to development. Since important implications are at stake, it has given birth to an abundant literature. Our first statement has been that the overwhelming majority these studies focus on the impact of foreign aid on recipient country's economic growth. After four decades of various publications, the discussion seems, at first sight, to be rather inconclusive. A couple of articles have already tried to clarify the debate but they are not really compatible. So, we decided to design our own framework of the literature.
Actually the whole aid-growth literature can be divided in two major periods separated by a double turning-point in the second half of the nineties. The first part has been dominated by inconclusive debate but it ended with the pessimistic agreement around the «micro-macro paradox». At the present time, this early literature appears not to be very reliable as it is characterised by particularly weak theoretical basis, poor econometric methodology and fragile databases. Then, some kind of methodological (Boone, 1996) and ideological (Burnside and Dollar, 1997) turning points took place in the late nineties. These two influential articles really boosted the production of new studies. Especially the innovative statement that aid works only if sound economic policy, has given rise to numerous reactions. The related principle of selectivity has extensively been discussed and criticized. Nowadays, though there remain a few pessimistic resistances, the large majority of recent scientific articles express the idea that foreign aid does globally work in enhancing economic growth.
Nevertheless, intense discussions remain about the extent of this efficiency as well as about the factors that could improve it.
More fundamentally, there are also some disagreements about the opportuneness to restrict the essential of the discussion on development effectiveness and economic growth. First, there exist many kind of aid to development. Only a limited part of the entire assistance targets investment that could have an impact on growth in the short run. Using the aid-growth relationship as a macroeconomic evaluation tool does give little indications about the efficiency of the majority of aid flows. Secondly, although there is an undeniable link between growth and poverty reduction, the two concepts are not synonymous. Unfortunately, economists are often inclined to confuse both notions or to accept the underlying assumption that growth is automatically beneficial for the poor. This has not been verified everywhere. Therefore, we may ask the question of the value of such an improved growth if it does not help the poor to catch up.
Throughout all of this work, we have tried to relate many aspects of the aid-growth literature. Actually, the contradictions we find between similar articles are mainly due to some striking shortcomings of this empirical approach based on cross-country growth regressions. First of all, the results appear to be extremely sensible changes in model specification. Concretely, the estimation of growth regressions is troubled by many technical econometric problems such as endogeneity, error correlation, parameter heterogeneity or influential observations. Furthermore, non-economic and structural factors may have constraining implication on aid effectiveness. Even if all these limitations do not prevent to recognise its global utility, aid is definitely not equally efficient everywhere. This may be due to either aid's characteristics or recipient country's features.
In consequence, it is particularly difficult to derive any robust policy recommendations. In light of all these technical difficulties, it should be wise to abandon this cross-country approach in favour of individual in-depth studies of countries that specify a growth model appropriate to the observed economy. In this context, we may also wonder whether it is really useful to look for global estimations of the effectiveness of foreign aid as a whole. We personally think of cross-country growth regressions to be quite inappropriate to assess aid effectiveness. Despite some improvements, the criticisms addressed to the early literature
remain mostly valid for recent studies. Doubtful theoretical background, unreliable data and technical difficulties are still topical.
The temptation is large to look for one single answer to the question of whether aid works or not. Whatever the conclusion, this approach is often misleading and it could have severe detrimental effects. Indeed, few influential aid-growth regressions have given colossal political implications despite their poor explaining power and their weak robustness. Since development process is not just economic growth, future investigations should concentrate on the impact of foreign aid on other issues such as poverty or health indicators. Provided the important contextual disparities of recipient countries, preference should also be given to individual case analysis. In any case, there remains much work to do and forthcoming studies should be very careful in their conclusions as they are very likely to provoke contradicting reactions.
9Impact, le film from Onalukusu Luambo on Vimeo.