In their widely cited paper, Alberto Alesina and David Dollar find that “the direction of foreign aid is as dictated by political and strategic considerations as by the needs and the policy performance of the recipients” (Alesina and Dollar 2000). Colonial past and political alliances, they add, are major determinants of foreign aid. To see the extent to which this holds, one can compare French and British aid to each other’s colonies. If colonial past matters, French colonies would always receive more aid from France than they do from England, and vice versa, regardless of poverty levels among aid recipients.

Before doing that, however, it is important to consider that former French and British colonies differ in their population size and poverty rates ¹. Figure 1 shows that there are far more poor people in former British colonies than in former French colonies. The average former British colony has approximately 10 million more poor people than that of France’s. That is, on average, the number of poor people in former British colonies is three times that in former French colonies.

Assuming that poverty levels are the principal determinant of aid allocation, one would expect that former British colonies receive more aid per capita than those of France. However, that is not always the case. Figure 2 shows net per capita bilateral aid flows from France to the two groups of countries. On average, France disburses $6 per capita to its former colonies annually and only $0.3 per capita to former British colonies. In other words, France gives out 19 times more bilateral aid per capita to its former colonies.

On the other hand, the U.K disburses an average of $0.9 per capita annually to former British colonies and only one third of that to France’s (See Figure 3). To its credit, the U.K’s former colonies have roughly three times the number of poor people. Consequently, though this might be just a coincidence, one can argue that the U.K allocates its aid neutrally with respect to colonial past and in a matter consistent with poverty levels among aid recipients.

Combining the information of Figures 2 and 3 one can clearly see that France and the U.K give out substantially different levels of per capita aid to their corresponding former colonies (Figure 4). As mentioned earlier, the average former French colony receives $6 per capita in aid from France while the average former British colony receives $0.9 per capita from the U.K. On the other hand, France and the U.K. disburse the same level of bilateral aid to each other’s colonies, both giving out an average of $0.3 per capita to non-former-colonies (Figure 5).

But why is the U.K’s level of aid per capita lower? One might suspect that this is a result of the bigger population of former British colonies. However, that is not the case. France simply gives out more bilateral aid than the U.K does, Figure 6. On average, former French colonies collectively receive 40 million dollars more annually from France than former British colonies receive from the U.K.

To conclude, it is clear that France does not allocate its bilateral aid neutrally. Rather, it heavily favors its former colonies despite their relatively low poverty levels compared to those of the U.K. For fairness, however, France does give out more bilateral aid than the United Kingdom. The latter, nonetheless, seems to allocate its bilateral aid flows in a manner consistent with recipients’ poverty levels.

References

Alesina, A., & Dollar, D. (2000). Who gives foreign aid to whom and why? Journal of economic growth, 5(1), 33-63.

Data and Code:

Download the data and code.

There are two ways to group countries by their GNI per capita. The one followed by the post Do Poor People and The Poorest Countries Receive The Most Aid tracks a group of countries based on their current per capita GNI. For instance, countries which currently fit the criteria for the middle income category are tracked over time, even though some of these may have been richer or poorer through the time period of analysis. Obviously this is limited in that, going back over time, these countries may have actually been low, not middle, income countries.

Alternatively, one can set GNI per capita brackets for various groups and allow countries to fall within these groups based on those income brackets. So, throughout the time period of analysis, a country can start as a low income, move to lower middle income, and end up as an upper middle income. This is the approach followed here.

Figures 1′, 2′, and 3′ are a reproduction of Figures 1, 2, and 3 in the post Do poor people and the poorest countries receive the most aid, respectively. These figures produce much stronger results against the claim that ODA doesn’t get allocated towards the poorest countries. Unfortunately, figures 4 and 5 are hard to reproduce using this method because complications arising from the availability of poverty data.

Data Sources:

All the data are from WDI.

Does foreign aid work? This question can be broken down into two parts: one is whether aid reaches the poorest; the other is, granted that aid reaches the poor or poor nations, does it help reduce poverty? This post addresses the first question by carefully looking at whether, or not, low income nations with the largest proportion of poor people receive most foreign aid as measured by ODA (official development assistance). Aid critics often argue that most ODA does not go to poor countries, much less the poor. As Angus Deaton argues in the Great Escape for example, “In direct contradiction to any supposed poverty mandate, much ODA does not even go to low-income countries, let alone to those countries where the poor live.”

Many countries receive aid, and several for decades. Since 1960, the highest average poverty rates have been observed in the Democratic Republic of Congo, Burundi, and Liberia, where an average of 87.7%, 84.0%, and 83.8% of the people in these nations, respectively, live in poverty. On the other hand, in terms of the total amount of ODA received during the same period, India, Egypt, and Bangladesh ranked first, with average net ODA receipts of 4, 3, and 2 billion dollars, respectively. As a percentage of recipients’ gross national product, GNI, the top ODA recipients were the Marshall Islands, Palau, and the Federated States of Micronesia with 37%, 38%, and 41% of their GNI, respectively. Clearly, these examples illustrate that individual countries’ ODA receipts and poverty rankings can be reported in various ways. One way to avoid such ambiguity is to group countries into income categories and compare poverty and aid levels across country groupings.

The remainder of this analysis does just that. Developing countries are grouped into the World Bank’s three income categories: Low Income Countries, Lower Middle Income Countries, and Upper Middle Income Countries. The grouping is based on countries’ GNI per capita in 2014. Data on poverty and aid between 1980 and 2012 are aggregated across these categories in three-year intervals.

Figure 1 illustrates the levels of ODA received across the three income categories in the periods between 1980 and 2012. It shows that the poorest countries, low income countries, received less aid than lower middle income countries almost every year. In 2005, low income countries actually received, as a group, the least amount of aid. On average, lower middle income countries received 45% more ODA than low income countries.
At first glance, Figure 1 suggests that Deaton is right. However, these numbers may be misleading. Rather than comparing the total amount of ODA received, it is important to account for the relative populations and incomes of the three country groups.

Figure 2 shows total ODA received by country grouping on a per capita basis. The results are now the reverse of those shown in Figure 1. Low income countries actually received the highest ODA per person, lower middle income countries received the second highest level of ODA per person, and so on. In contrast with Figure 1, on a per capita basis, low income countries received twice as much ODA than lower middle income countries, and four times more than upper middle income countries. Of course, this does not necessarily mean that donors allocate most of their aid to the poorest countries. High ODA per capita can be a consequence of a low population size in the category of low income countries. Nonetheless, Deaton is wrong from this perspective.
Accounting for nations’ income level leads to results identical with those above. Figure 3 shows that low income countries received the highest share of ODA out of their total GNI between 1980 and 2012, with an average share of 8%. Lower middle income countries received only 2% of their GNI in ODA. Once again, this does not necessarily mean that donors allocate the majority of their aid to the poorest countries. Instead, it may be the case that, holding ODA constant, low income countries have extremely low incomes or few alternative sources of income in the formal economy so that the ratio of ODA to their GNI is bound to be high.
While considering income and population levels is important, the most appropriate method to compare across country category is to control for the number of poor people across country income categories. The logic behind this is straightforward. To alleviate poverty, the number of poor people in a country provides a more relevant measure than its overall population. Figure 4 shows the ratio of ODA to the number of poor individuals, or headcount, at a $1.25 PPP poverty line.

Figure 4 indicates that the poor in low income countries consistently received the highest share of aid prior to 2001. More importantly, ODA per poor person to low income countries has been growing rapidly since 1998, and it has doubled since then. Yet, ODA per poor person to upper middle income countries has grown even faster. Between 2001 and 2004, the share of ODA per poor individual going to upper middle income countries climbed sharply from $40 to $140 and dropped to nearly $80 in 2010 before rising again. Two factors contribute to this jump in ODA per poor person. First, this era coincides with a surge in aid to upper and middle income countries. For instance, Figure 1 shows that ODA to upper middle income countries more than doubled between 2002 and 2005. Second, there was a sharp fall in poverty in upper and middle income countries during this time period. Poverty in upper middle income countries fell from 20% to 11% of the population (Figure 5).
So what does this mean? These figure indicate that it is safe to conclude that aid is most likely distributed such that the poorest receive the most. The claim that aid is allocated regardless of poverty levels doesn’t seem to hold here. Of course, one can argue that the method of grouping countries based on their GNI per capita in 2014 is misleading since some of the countries have historically moved across each group. For instant, China was a low income country until 2001, lower middle income country until 2009, and upper middle income since then. In this appendix countries are aggregated in three income groups based on their historical GNI per capita, allowing nations to move in and out of their groupings as their GNI per capita level changes. The results are much stronger in refuting the claim that ODA does not go the lower income countries.

References

Deaton, Angus (2013-09-22). The Great Escape: Health, Wealth, and the Origins of Inequality (p. 278). Princeton University Press. Kindle Edition.

Data Sources:

All the data are from WDI
Download the data used for this post.