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
Actually, two major turning-points took place in the literature during the second half of the nineties. Though they are very different in their philosophy, together, they gave birth to a new generation of aid-growth empirical studies leading to both improved methodology and modifications in the conclusions.
Boone's (1996) pessimistic conclusions are quite in line with most of precedent articles and so it confirms Mosley's idea of the «micro-macro paradox». As we already mentioned, he finds that aid does not increase growth nor benefits the poor because it tends to finance consumption rather than investment. So, the main result of aid seems to be an increase of the size of governments. Nevertheless, this article is a kind of turning-point in the sense that it is probably the first one to initiate the movement of a methodological revolution. Compared to previous studies the new generation of articles brings several significant innovations.
First, they are now working with panel data covering a larger number of countries and years. As Hudson (2004, p1 87) explains, data are more reliable than ever: «The problems relating to data are declining gradually over time as both more and better quality data becomes available.» Secondly, authors are now taking account of innovations in matters of growth theory. Variables reflecting the quality of institutions and economic policies of the beneficiary countries are now included in the regression. Third, it has become the rule to address explicitly the possibility of endogeneity of aid21. Last but not least, the inclusion of regressors such as aid² and aid×policy enables a non-linear aid-growth relationship22. Quibria (2004, p9) confirms this major innovation: «Perhaps the principal contribution of the Boone study has been to stimulate further professional interests in a research area that was fast approaching rapid diminishing returns.»
21 The amount of aid allocated to a country is not independent from its initial level of GDP since aid appears to go primarily to poor countries.
22 Cfr. Arguments discussed earlier
In 1997, in this context of aid's legitimacy crisis and constantly decreasing international assistance flows, Burnside and Dollar wrote their «extraordinarily influential paper»23: Aid, Policies and Growth. Initially known as a World Bank working paper it was only published three years later in the American Economic Review. By means of a new data base of the World Bank, they want to revisit the pessimistic conclusions of Boone (1996) and other recent articles. In this sense, it gave a new impulse to the debate. The following statements confirm this new dynamism:
(Roodman 2004, p2): «The work of Burnside and Dollar has brought corroboration and challenges.»
Easterly (2003, p26): «The Burnside and Dollar (1997) paper gained prominence because it addresses the scepticism implied by Boone and by the lack of consensus from the early literature». Although, from a methodological point of view their innovation was quite minor, their conclusions brought new arguments to the discussion about aid effectiveness.»
Quibria (2004, p9): «The paper that has done the most to stimulate the interest in the topic is the one of Burnside and Dollar (1997).»
McGillivray (2005, p2): «The turning point in the literature is defined by two, very well-known studies. The first is Burnside and Dollar (1997) and the second is a Assessing Aid: What Works, What Doesn't and Why (World Bank, 1998). The latter report results presented in the former.»
Their major novelty is the inclusion in the equation of an interaction term between foreign aid and economic policies. This «aid×policy» interaction term enables to make a distinction between countries with adequate economic policies and those with unfavourable ones. Hence, the simple correlation of aid and growth appears to be close to zero24 but this is only true for country with mean policy level. Indeed, the impact of aid on growth appears to be significantly positive in good policy environments. Their conclusion is that aid effectiveness in the growth process is directly dependent on the quality of economic policies25. Whereas Boone's (1996) conclusions offered no opportunity to increase aid effectiveness, Burnside and Dollar (1997) suggest a more encouraging discourse since better selectivity in aid allocation should enhance its impact on economic growth.
Obviously, such a finding has led to important political consequences: «This work has clear implications for how to make foreign aid more effective... If they (donors) want to have a
23 See Easterly, Levine and Roodman (2003, p1)
24 In this sense Burnside and Dollar (2000) is in consistent with Boone (1996). 25 They also found that foreign aid suffers from diminishing marginal returns.
larger impact on growth and poverty reduction, then they should place greater weight on economic policies of recipient.» (Burnside and Dollar, 1997, p4) In the years that follow this publication many development agencies adapted their allocation behaviour to adopt the selectivity principles recommended by Burnside and Dollar. The Economist (June, 1999) also expressed this concept of selectivity that became so popular at that time: «Countless studies have failed to find a link between aid and faster economic growth.... Rich countries should be more ruthless about how they allocate their largess, whether earmarked or not. Emergency aid is one thing. But mainstream aid should be directed only to countries with sound economic management. In other words, aid could work if properly directed.» Such strong statements gave rise to abundant reactions. As we will see, the application of the selectivity principle as suggested by Burnside and Dollar (1997) was probably not a very opportune decision.
9Impact, le film from Onalukusu Luambo on Vimeo.