Aid, the Public Sector and the Market in Less Developed Countries

The article in this JOURNAL by Mosley et al. (I987) concludes its analysis of aid effectiveness as demonstrating '... inability of development aid over more than twenty years to provide a net increment to overall growth in the Third World' (p. 636). This note purports to show that this conclusion, which is ostensibly 'distressing' as the authors claim, is not warranted. The section of the study which is the basis of the conclusion cited and the subject of this note is a cross-section regression analysis, full details of which are shown in Mosley et al. Table 3. The note is concerned with the relation between the dependent variable specified as the growth rate of real GDP per cent per annum (hereafter GDP) and aid specified as a percentage of GDP (hereafter Aid).' The Aid results, with only two significant positive coefficients in one subsample, certainly give no basis for claiming aid-effectiveness in terms of GDP growth. The intention in this note is to interpret the results which are assumed to be valid. This is made possible by the fact that the authors have helpfully provided the full input set of mean values derived from the raw data. From this it is possible to calculate the GDP/Aid relationship for each country for the second and third period. The procedure for deriving these relationships is as follows. For each country the mean values of the time series of GDP and Aid are extracted from the data for I 960-70 and for I 970-80. Using I 960-70 as the base, the signs of the change in the levels of the GDP and Aid mean values between periods are recorded. This is repeated for I980-4 with I970-80 mean values as the base. The relationship between the two signs is an indication of the effect of Aid on GDP in the period in which the change in levels is effected. Thus in the case of an increase in Aid being combined with a decrease in GDP the latter must have fallen relatively to the former and aid-effectiveness is negative; Mutatis mutandis, for the other three sign combinations, thus generating the appropriate GDP/Aid relationship for each country in each period. A frequency distribution of the four sign-combinations, relating to the full sample, is presented in Table I. Since the signs of the changes in the levels of GDP and Aid are derived by making explicit the arithmetic relationship implicit in the data they form a valid base from which to draw conclusions without requiring any test of