Foreign Aid And The Business Cycle

In this paper, we document some key business cycle properties of foreign aid flows to developing countries. We identify two striking empirical regularities. First, aid flows are highly volatile over time -- on average, two to three times as volatile as the recipient's output. Second, for most African countries, net aid inflows are strongly positively correlated with their domestic output. Outside of Africa, we find a similar, if somewhat less pronounced, pattern of aid procyclicality.To see why these empirical regularities are important, recall that output fluctuations in developing countries are much stronger than in industrialized economies. Indeed, we document that the gross domestic product of an aid recipient is on average six times as volatile as that of a donor. For developing countries, though, customary ways to smooth out the impact of output fluctuations on domestic consumption are likely to be very onerous. For instance, moral hazard and repudiation risk imply that heavily indebted nations are often denied new loans (or are asked to repay old ones) precisely when their economies suffer adverse shocks -- see, e.g., Atkeson-1991). At the same time, foreign aid is a sizeable source of income to recipients, especially in Africa, where it averages 12.5% of gross domestic product and constitutes the main source of foreign capital. In such an environment, foreign aid flows have the potential to play a key role in smoothing out developing countries' output fluctuations. Our results imply that, all in all, aid does not play that role.Admittedly, it might be argued that, except for emergency relief, the chief purpose of foreign aid is not to act as an insurance device but, instead, to fuel economic development, in which case it is not clear a priori whether one should expect aid flows to be procyclical or countercyclical. It is well known, however, that output fluctuations affect growth negatively -- see, e.g., Hamilton (1989) and Ramey&Ramey (1995). Hence, even if aid were meant solely to help foster growth, serious concerns should nonetheless arise from the fact that aid disbursement patterns contribute to the volatility of developing countries' disposable income.Our findings are robust. Our data set comprises various yearly aid and output series for sixty-three recipient and eighteen donor countries between 1969 and 1995. We find few differences between the cyclical behavior of multilateral as opposed to bilateral aid disbursements, even though multilateral aid flows are relatively more volatile than their bilateral counterparts. Likewise, aid commitments fluctuate more than actual net disbursements, but both commitments and disbursements are procyclical. We also pay special attention to Africa, because it is the region where aid is largest relative to recipient GDP and aid procyclicality is most striking. We show that, regardless of the domestic output measure used, net aid receipts are procyclical for at least two-thirds of the thirty-eight countries in our African subsample and are countercyclical for, at most, two of them. Key components of African aid, such as grants or technical assistance, are as strongly procyclical as total aid flows. Finally, we can find no evidence that, among African countries, the procyclicality of aid might be a function of the recipient's former colonial power, choice of exchange rate regime or some other criterion.We complete the paper by analyzing the cyclical properties of aid flows from the donors' perspective. For almost all donor countries, we find that total aid disbursements are strongly positively correlated with the donor's output. In a clear majority of cases, however, those same donors' bilateral aid flows to the sample countries are not positively correlated with the donor's output. A corollary is that the procyclicality of aid inflows experienced by aid recipients is not the mere result of the conjunction of (i) positive comovements between North-South business cycles [Kouparitsas (1998); Agenor&Prasad (1999)] and (ii) a positive correlation between donors' aid policies and their business cycles.