Aid and Global Poverty
Abstract and Keywords
This article examines the relationship between foreign aid and poverty in developing countries, with the goal of determining whether donor governments are motivated and actively set out to reduce poverty in developing countries through the provision of aid but with the impact of aid on poverty reduction. It begins with an overview of the aid and poverty record based on global data from the 1980s onward, with particular emphasis on Official Development Assistance (ODA). It then considers the analytics of aid and poverty before reviewing the relevant literature, including studies that address the impact of aid on growth and growth elasticity of poverty. The article argues that aid has had a marginally positive impact on poverty reduction in developing countries, and that poverty would be slightly higher without it.
Aid, for the purpose of this chapter, is defined as resource transfers from governments of rich countries to relatively poor countries that purportedly are intended to assist the latter achieve higher living standards. Controversies surround aid. The linkage between aid and poverty in developing countries is especially controversial and has been for decades. There were lofty expectations regarding the poverty-reducing impacts of foreign aid in the early 1960s, when activists called on rich countries to transfer 1 percent of their combined incomes to developing countries to help fight poverty. It seemed to many that foreign aid was a golden wand: wave it enough times and it would rid the world of poverty. To others it might not necessarily rid the world of poverty, but it was the single most important thing rich countries could do to pull many millions of people in poor countries out of poverty. Other observers were skeptical or downright hostile toward aid. Many of these people saw aid as no more than an attempt by rich countries to exploit their poorer counterparts in the pursuit of their own political, commercial, and other self-interests. Some even suggested it might do more harm than good by entrenching the positions of inept or corrupt leaders, constraining the operation of private enterprise, and promoting dependency through the displacement of domestic savings. Rich countries did, to some extent, respond to the calls of the pro-aid activists. Few managed to provide aid to the level of 1 percent of their income, but they have certainly allocated substantial amounts of public money to poor countries. The most widely used measure of aid is Official Development Assistance (ODA).1 ODA levels have climbed considerably since the 1960s. More than $4.3 trillion in ODA has been provided to more than 150 developing countries over the period 1960 to 2011 (OECD 2012). Global ODA increased from $36 million in 1960 to $148 billion in 2010, the highest level ever recorded (OECD 2012).2 Yet despite these sums of money, some simple empirical facts remain: poverty levels have not fallen appreciably in the last 30 years, and billions of people still live in poverty. The poverty (p. 688) statistics are indeed quite sobering. In 2008, just under 1.3 billion people were living in developing countries in extreme income poverty on less than $1.25 per day. There were just under 2.5 billion living on less than $2.00 per day in 2008, also the highest level ever recorded (World Bank 2012). Present day critics of aid often see these numbers as proof that the skeptics were correct, with some insisting that aid made a bad situation worse and that it is part of the problem and not the solution to global poverty. Yet aid still has its supporters. Some argue that the problem is that donors have simply not provided enough aid, that it has reduced poverty in some countries but not others, and that it can work a lot better if it was more poverty focused.3 A more nuanced broadly pro-aid case is that poverty would have been much worse in the absence of aid; that we need to be cognizant about what aid can realistically achieve given that, since the 1970s aid has accounted for the equivalent of just 1.5 percent of developing country GDP; and that even if aid can reduce poverty, it is only one of many drivers of poverty.4 The international community’s stance on the Millennium Development Goals, which include the goal of halving the proportion of people living in extreme income poverty by 2015, is consistent with this recognition in that it views aid as one of a number of means by which this goal might be achieved. Whatever basic position might have the most validity, it remains the case that there is widespread disappointment that aid has not done better in terms of its poverty-reducing impact and that links between aid and poverty reduction remain highly controversial.
This chapter re-examines the relationship between aid and poverty in developing countries, attempting to cut through the controversy surrounding the topic. Its concern is not whether donor governments are motivated and actively set out to reduce poverty in developing countries through the provision of aid but with the impact of aid on poverty reduction. It does so mainly by looking at aid and poverty descriptive statistics—the aid and poverty record—and surveying relevant empirical studies. The chapter concludes that an objective and balanced examination of the literature suggests that income poverty would in all probability be higher in the absence of aid, although by how much remains a matter of much speculation. That said, it also argues that aid can play an important but at best marginal role relative to other drivers in promoting poverty reduction in developing countries, irrespective of how high or efficient aid flows might become. The chapter consists of a further four sections.
The second section examines the aid and poverty record, by looking mainly at global data from the 1980s onward. A key question considered in this section is whether the disappointment over the poverty-reducing impact of aid is, in its proper context, disappointment with the efforts of donors to reduce poverty as opposed to disappointment with the impacts of the aid that has actually been provided. It also uses these data to highlight some of the complexities in looking at empirical relationships between aid and poverty, in particular the problem of the “without aid” counterfactual. Unless we can be reasonably confident of what poverty would have been in the absence of aid, that is, knowing the counterfactual, one can say little of its impact on poverty.
The third section looks at the analytics of aid and poverty, highlighting the various ways that one can analyze the impact of the former on the latter. It makes the fundamental (p. 689) point that only a small proportion of aid has been provided directly to poor people, as if it were a form of welfare support. This has profound implications for how the poverty impact of aid is investigated, and how the results of this investigation are interpreted.
The fourth section provides the literature review. It commences by reviewing studies that look directly, in a strictly empirical sense, at the relationship between aid and poverty. The review then looks at studies that do not address poverty directly but instead look at the impact of aid on variables that might drive poverty reduction.
The fifth section concludes, suggesting that in all probability aid has had a marginally positive impact on poverty reduction in developing countries, implying that poverty would be slightly higher in its absence.
The Aid and Global Poverty Record
Prior to progressing any further, the ways that aid and poverty are conceptualized and measured in this chapter need to be specified. Aid can be defined and measured in different ways, but as is evident from the discussion in the first section, we focus on ODA. ODA has a specific definition. ODA represents flows that are provided by OECD country governments to countries classified as developing and which are concessional, either being in the form of grants or soft loans. ODA need not go directly from one government to another but can be provided through a multilateral agency (such as the World Bank or the many UN agencies) or a nongovernment agency. This definition, and the reporting consistent with it, is that of the OECD’s Development Assistance Committee, the membership of which is made up of 23 OECD countries and the European Union.5 In order for a flow to be classified as ODA it is supposed to have the promotion of development and welfare in developing countries as its main objective. Many observers of ODA would question with validity whether any particular ODA flow, be it to a country, region, project or program, actually has these outcomes as the main objective. In practice this requirement excludes military assistance as aid or other flows that are clearly nondevelopmental.
Our focus on ODA means that we take no account of so-called South-South development cooperation, including aid from China. Chinese aid has become an important part of the international aid scene, but an absence of reliable information on these flows prevents its inclusion in the analysis of this chapter. Focusing on ODA also means that we take no account of aid during the Marshall Plan. Much of the aid provided during this period would not be classified as ODA, and, even if it was, data availability remains an issue. While the Marshall Plan is judged to have been very successful in rehabilitating countries from the damage caused by World War II, and in doing so undoubtedly lifted large numbers of people out of poverty, we leave it to historical studies to consider its impacts on poverty. These points noted, it remains the case that ODA is by far the largest form of what might be broadly described as development aid, and for this reason our focus covers the vast majority of the aid landscape. (p. 690)
Poverty, like aid, can also be defined in many ways. Contemporary thinking rightly considers poverty to be multidimensional, involving more than just shortfalls in income. We focus on income poverty, measured using the well-known poverty headcount measure. The chapter does not focus on other poverty dimensions, such as those relating to health and education. Aid certainly has the potential to make an impact on poverty in these dimensions.6 This chapter does not focus on poverty in nonincome dimensions for two reasons. First, there remains ambiguity over poverty lines for these dimensions and, in turn, the extent of poverty within them. Second, there is little if any preexisting literature on the impact of aid on poverty in dimensions other than income. This is not to say that studies have not looked at the impact of aid on nonincome quality of life achievements. There is, for example, a body of literature examining the impact of aid on health and education achievements.7 But this literature does not look at impacts by population subgroup within countries, especially those at the bottom end of the relevant distributions. There is also a well-established literature looking at the relationships between aid and public sector expenditures on health and education.8 It would be difficult to draw inferences from the findings of this literature, for reasons that are explained below.
Let us now consider some simple empirics of the aid and income poverty relationship. Figure 30.1 looks at trends in aid and poverty since the earliest year for which comprehensive poverty data are available. The measures of poverty under consideration are the widely used $PPP1.25 and $PPP2 poverty headcounts, which show the number of people living on incomes below these respective poverty lines. The former is treated as a measure of extreme income poverty.9 There are obvious messages about the relationship between aid and poverty from Figure 30.1. Aid volumes trend upward for the period shown, despite stagnating from 1993 to 1996. Extreme income poverty trends downward for the whole period, as does the number of people living on less than $PPP2 per day from 1996. Putting aside the fact that there are many other drivers of poverty reduction (a point considered later), one might be tempted to conclude that aid (p. 691) is associated with declines in global poverty. It could be equally argued that Figure 30.1 is consistent with the disappointment about the poverty-reducing impact of aid, since the number of people living on less than $PPP2 per day is only slightly lower in 2008 than in 1981, and the declines in extreme income poverty are modest. Perhaps all one can conclude is that the data shown in Figure 30.1 do not support the view that aid is part of the problem with poverty being lower in the absence of this inflow. Measuring aid in constant price per capita terms, per head of recipient country population, does not provide terribly different insights, as Figure 30.2 suggests.10
The aid and poverty record in the two regions where poverty has been highest, South Asia and sub-Saharan Africa, is shown in Figures 30.3 to 30.6. The relationship between aid and poverty in South Asia is not especially clear, and identifying a relationship of any type is difficult. Aid to this region might at best be associated with declines in extreme income poverty from 2002. But at the same time, increased aid levels from the (p. 692) late 1990s are associated with increased numbers of people living on less than $PPP2 per day. These trends relate to aid in absolute numbers, not per head of recipient population. If per capita aid to South Asia is considered, lower aid numbers are associated with higher levels of poverty. This is shown in Figure 30.4.
The aid and poverty record in sub-Saharan Africa lends greatest support to the harshest critics of aid, those who argue that aid is part of the problem and not the solution. The picture emerging from Figures 30.5 and 30.6 is far from encouraging. Income poverty declined between 2005 and 2008. This is obviously good news. Yet for most of the years under consideration, as aid increased, so too did poverty. This also applies to the rise in per capita aid from 1999, shown in Figure 30.6. Perhaps the most aid-sympathetic interpretation is that the big push in aid to the world’s poorest region from the early 2000s led to a slowing and eventual reversal of the upward trend in income poverty. (p. 693)
Arguably the most telling information coming from the simple empirics just presented is that relating to per capita aid, shown in Figures 30.2, 30.4, and 30.6,. Whether aid may rise, fall, or remain the same in absolute terms does not matter so much from a general development-effectiveness perspective. What matters from this perspective are levels relative to the number of people at the receiving end in the recipient country. The number of people at the receiving end can be defined in many ways. It could simply be treated as the total population of the receiving country or the number of people living below and close to the chosen poverty line. Irrespective of what is the most appropriate measure of aid, it is clear that per capita aid, both globally and to South Asia and sub-Saharan Africa, has fallen for much of the period under question. It is lower in 2008 than in 1981 for both South Asia and sub-Saharan Africa. This might suggest that the criticisms of aid are best leveled not at the incremental impact of the aid actually provided but at the efforts of donor governments in seeking to fight poverty in developing countries.
A more incisive way of looking at the aid-poverty relationship, one that gets us closer to drawing conclusions about the impact of aid on poverty, is to examine scatter plots of these two variables. Scatter plots are shown in Figures 30.7 and 30.8, which focus on the poverty headcount based on the $PPP1.25 poverty line.11Figure 30.7 combines regional aid and poverty data, while Figure 30.8 combines data for all individual countries for which requisite data are available.12 In each case aid is measured in per capita terms. A line of best fit is included in these charts, which most closely resembles the relationship evident in the data. What is clear from the data is that higher aid levels are associated with lower poverty levels.
Supporters of aid might like this association, but it does not mean that aid has been effective in reducing poverty or that poverty would have been higher in aid’s absence. There are many determinants of poverty reduction, of which aid might (but might not) be one. The relationships shown in these figures might be driven by other variables and have nothing to do with aid. As such, the associations shown in these (p. 694) figures could well be spurious. A further complicating factor is that donors might respond to higher levels of poverty with more aid, with poverty leading to more aid rather than more aid leading to less poverty. If this were the case, we might observe a positive relationship between aid and poverty, irrespective of the impact of the former on the latter. The higher-aid-higher-poverty associations shown in a number of Figures 30.1 to 30.8 could actually be due to donors responding to higher poverty levels with more aid. In order to confidently answer the questions of whether poverty is higher, unchanged, or lower owing to aid we need to know what levels of poverty would have prevailed in the absence of aid. Put differently, we need to be able to speculate with accuracy the “no aid” poverty counterfactual. We address this in the remainder of this chapter.
(p. 695) The Analytics of Aid and Poverty: How Can Aid Impact Poverty?
Analyzing the impact of aid on poverty would be relatively straightforward if it was given directly to poor people, as if it were a form of welfare support, either by donor governments or agents acting on their behalf. What would matter is the amount provided and how the poor used it. The amount of aid would have to be sufficient to lift the recipients to, or above, an appropriate poverty line, and it would need to be used in a way that kept them on or above this line when the aid money ceased. Yet aid is not by and large given directly to poor people, nor do donor government agencies necessarily target poor people directly.13 Among the reasons for this is that they are often difficult to reach and attempting to do so comes with high risk, especially when donor country governments and taxpayers demand observable results and value for money (McGillivray et al. 2013).
Aid is instead provided mostly to recipient country governments, civil society, or even to the private sector in recipient countries by donor governments either bilaterally or through multilateral organizations in various forms with a range of specific purposes, some of which will have little direct relationship with poverty reduction. Donors can also provide aid directly, typically using private sector contracting organizations. The specific purposes for which aid is provided do not relate to high-level foreign policy or other objectives donors might pursue in the provision of aid per se but rather to areas or sectors to which it is directed in recipient countries. Among those identified by the OECD’s Development Assistance Committee (DAC), and under which OECD donor nations report, include health, education, water, sanitation, public sector administration, human rights, conflict prevention, transport, communications, banking and financial services, trade policy, mining, tourism, agriculture, and forestry (OECD 2012). While it is not uncommon for donors to identify poverty reduction in developing countries as an overall objective motivating the provision of aid, it is not among the specific purposes of aid under which donor nations report as an examination of OECD (2012) makes evident.
This has clear implications for how the poverty impact of aid is investigated, and how the results of this investigation are interpreted. One approach would be to speculate on the drivers of income poverty reduction and ask which forms of aid are likely to address these drivers. It is often claimed that in the context of income poverty, among the reasons why many poor people are poor is a lack of education, low levels of health, and a lack of access to basic services. If this is so, then it would seem reasonable to look for empirical associations between the levels of aid to these sectors and income poverty headcounts through the application of regression analysis, say by using cross country data and after controlling for other drivers of poverty. For example, analysis could be (p. 696) undertaken of the association between education aid and income poverty. Regression analysis could provide an overall association between education aid and income poverty, or an estimate of the total impact of the former on the latter, in that there is no attempt to establish the various causal channels through which such an association might arise. It could be, for instance, be that education aid augments domestic expenditure on primary schooling and thereby allows poor people to obtain more highly paid employment, or that it enables those involved in the delivery of pro-poor services to do so more efficiently, or a combination of these and other impacts.
While such an approach might seem intuitively appealing, it is not without its difficulties. Other forms of aid might have some impact on income poverty. This might not only lead to an incomplete picture of the overall impact of aid on poverty but also deem the results of the empirical analysis biased and misleading owing to omitted variable bias, a well-known problem with regression analysis. There are also good reasons for expecting that education aid, or for that matter aid for health, water, and sanitation, might not have any impact on income poverty or that this impact varies so much across countries that looking for an empirical association from cross-country data is a fruitless exercise. It is reasonably well established that health and education expenditures, along with spending on the provision of water and sanitation services, can have pro-rich biases, and, as a consequence, do not reach those living in income poverty (World Bank 2004). Alternatively, aid can be fungible, which is said to be the case if the recipient has the ability to use aid for purposes other than those intended by the donor (McGillivray and Morrissey 2001). If this is the case, a donor might allocate a given amount of aid to augment the recipient government’s own expenditure in a particular area, but the recipient allocates part or all the funds elsewhere, or all the funds to the area intended by the donor but decreases its own expenditure in that area. The net result is that total expenditure in the intended area, donor and recipient, increases but by less than the amount of aid provided. This, like a pro-rich bias in recipient-government expenditure, does not necessarily mean that aid has not had an impact on poverty, positive or otherwise, simply that it cannot necessarily be inferred from an empirical association between a particular form of aid and poverty, however poverty relevant the former might be.
An important driver of poverty is economic growth, the year-on-year growth in real per capita national income. Ravallion and Chen (1997), Dollar and Kraay (2002), and Kraay (2006) are among the better known studies that examine this link. There is some debate over the process through which growth can reduce poverty. Some argue that high rates of growth per se are negatively associated with poverty: countries that achieve such growth will in general have fewer people living in poverty than would otherwise be the case. Others argue that growth is a necessary, but not sufficient, condition for poverty reduction, and that it is sufficient if the incomes of the poor grow as national incomes grow for growth to be termed pro-poor.14 A measure of the extent to which economic growth results in poverty reduction is the growth elasticity of poverty (GEP) (Bourguignon 2004). The GEP is defined as the percentage change in the incidence of poverty in response to a one percentage (p. 697) increase in per capita income. The GEP will understandably vary across countries, with estimates ranging anywhere from -2.0 to -5.0 (Ravallion and Chen 1997; Bruno Ravallion, and Squire 1998; Bhalla 2002; Adams 2003, 2004). This in turn suggests that a 10 percent increase in economic growth, for example, will lead to a decrease in the proportion of people living in extreme income poverty in the country in question of anywhere between 20 and 50 percent. A widely used global GEP is -2.0 (Collier and Dollar 2001, 2002).
Bourguignon (2004) demonstrates in a theoretical context that the GEP depends on changes in per capita income levels and the distribution of relative incomes or income inequality. Son and Kakwani (2004) find using empirical analysis that the GEP is smaller in countries with low levels of per capita income and in countries with higher levels of inequality. De Janvry and Sadoulet (2000) also report a positive empirical association between the GEP and levels of secondary schooling.
The preceding discussion points to two ways that the impact of aid on poverty can be assessed empirically. The first is to look at the association between aid and economic growth in recipient countries, after controlling for the impact of other variables on the latter, and then draw inferences about the numerical extent of this association for poverty reduction using an appropriate GEP. This essentially amounts to multiplying one coefficient, showing by how much a one-unit change in aid changes the growth rate, by another, the GEP. Looking for an association between aid and growth assumes that most forms of aid will have some impact on economic growth in these countries, positive or otherwise. While not all forms of aid are intended to drive growth, most will have some sort of relationship with it, as a cursory glance of the different types of aid listed previously would suggest. A numerical association between aid and growth might tell us if the former impacts on the latter, and if so, this is useful for advocacy purposes. But by suppressing all the causal channels that lead to this association, it does not tell us why or how this association emerges and as such is not especially useful for devising policies aimed at greater poverty reduction through aid.
The second way that the impact of aid on poverty can be empirically assessed is to look at the former’s impact on the GEP or on known drivers of the GEP. If aid impacts on growth of per capita income then it must, by definition influence its level. Aid might have some impact on inequality through an impact on governance, given that good governance can involve the introduction of more inclusive social policies. It might also have some impact on the other known driver of the GEP, secondary schooling, notwithstanding the fungibility issue. Information on an empirical association between aid and the GEP, if there is indeed one, could then be combined with actual growth rates in developing countries to provide an estimate of the extent to which aid has impacted poverty. Put differently, one would obtain an aid-GEP coefficient, which would then be multiplied by an observed economic growth rate to provide an estimate of the extent to which aid has reduced the number of people living in income poverty. Deriving such a coefficient from an analysis of aid and drivers of the GEP would appear to be a more cumbersome task as it would involve information on the coefficient between each driver and the GEP itself.
(p. 698) The discussion thus far has overlooked an obvious point. If it is reasonable to look for a broad association between aid and economic growth, then it would also appear reasonable to look for such an association between aid and income poverty, say using a poverty headcount statistic. This investigation would seek to determine the association between aid and poverty after controlling for the impact on the latter of variables other than aid. Put differently, it would seek to identify what the level of poverty would have been in the absence of aid using historical data. This exercise is subject to the same core criticism of aid and growth studies: it can at best tell us if aid reduces poverty and by how much, but not how. Such an exercise would provide the same basic information as looking at aid and the GEP. It would result in an aid-poverty coefficient, which would be multiplied by the number of people living below the corresponding poverty line to obtain the impact of aid on poverty. Of course it could be the case that aid has no impact on poverty, in which case the aid-poverty coefficient would be zero.
It was mentioned earlier that aid is not by and large given directly to poor people. Some aid is, in addition to a share of humanitarian and emergency relief aid. Conditional cash transfers and microfinance are among a number of forms of aid that can be given directly to people, which include those living in income poverty. Comprehensive aggregate data on these microlevel flows are hard to obtain, but it has been estimated that microfinance has accounted for as much as 10 percent of total aid in any given year. Both have the potential to impact income poverty, to a much greater degree than humanitarian and emergency relief. As such, there are strong grounds for looking for associations between these types of aid and poverty. Poverty impacts of these types of aid cannot be assessed through country-level or cross-country analysis discussed previously. They can, however, be assessed using evidence from field experiments, akin to randomized control trials (RCTs) (Banerjee and He 2003). The challenge in using the results from these field experiments is having a sufficiently large number of them to be able to make judgments about the overall or aggregate impact of aid on poverty.
Going Beyond Simple Descriptive Statistics: A Literature Survey
The discussion of the previous section identifies four areas of research that might provide insight into the impact of foreign aid on global poverty. They are research that looks for (1) an empirical association between aid and poverty, seeking to identify what the level of poverty would be in the absence of aid; (2) an empirical association between aid and growth, seeking to identify what the rate of the latter would be in the absence of the former so that inferences can be drawn about the impact of aid on poverty; (3) an empirical association between aid and an appropriate GEP, or on drivers of the GEP, (p. 699) so that this information can be combined with economic growth rates to provide an estimate of the contribution of aid to poverty reduction; and (4) the impacts on poverty of what might be called microinterventions supported by aid, identified through the application of field experiments akin to RCTs. Literature falling into these four areas of research is now considered.
Aid and Poverty
A surprisingly small but growing literature has emerged that examines whether aid impacts income poverty using the econometric analysis of cross-country data. In general, the empirical evidence is encouraging. Asra et al. (2005), Bahmani-Oskooee and Oyolola (2009), and Alvi and Senbeta (2012) all find empirical support for aid reducing headcount poverty.15 Conversely, Chong et al. (2009) find there is, on average, no association between growth and poverty reduction in developing countries but some weak evidence that foreign aid is effective at reducing poverty in countries with better institutions.16
Three other interesting findings from these studies emerge. Firstly, Asra et al. (2005) find that there are diminishing returns to aid. In other words, they find that aid is effective at reducing poverty up to a threshold, beyond which the marginal impact of aid starts to fall. They estimate that this threshold is where aid accounts for 26 percent of a recipient’s gross national income (GNI). This finding is in concordance with many studies examining the impact of aid on economic growth discussed later. Secondly, in addition to the poverty (headcount) rate, Alvi and Senbeta (2012) find that foreign aid is associated with a decline in the poverty gap index and squared poverty gap index. These measures reflect the depth and severity of poverty, as well as its incidence, implying that aid is effective at reaching the poorest in developing countries and reducing inequality among the poor.17 Thirdly, Alvi and Senbeta (2011) find that multilateral aid and aid grants do better in reducing poverty than bilateral aid and aid loans. While multilateral aid is often focused more on poorer countries and freer from political and strategic motives in its allocation, the reasons why it is more effective at reducing poverty are unclear. The finding that grants are more effective than loans would appear to indicate that loans have not been used for activities that benefit the poor in recipient countries.
Aid and Growth
In looking at the literature on aid and growth, it is instructive to distinguish between studies published in the late 1990s and earlier. The earlier studies were largely inconclusive, either finding that aid increased, decreased, or had no impact on growth or a key driver, savings. If an overall conclusion is possible from this literature is it that aid had no impact on growth in recipient countries (McGillivray et al. 2006).
(p. 700) The late 1990s marked a significant change in the aid-growth literature, commencing with the influential study of Burnside and Dollar (2000). Burnside and Dollar’s study and most that followed tended to use better data, theory, and empirical techniques than those that preceded them. There have been numerous comprehensive and objective surveys of this more recent aid-growth literature, which consists of scores of studies. These surveys include Morrissey (2001), Clemens et al. (2004), Dalgaard et al. (2004), Addison et al. (2005), McGillivray et al. (2006), and Feeny and McGillivray (2011). All point to the literature concluding that growth would have been lower in the absence of aid. More precisely, they point to almost all studies published since the late 1990s drawing this conclusion and no methodologically valid study concluding that growth would be higher in the absence of aid.18
This, of course, does not mean that there are no contrary voices emanating from empirical research circles. Doucouliagos and Paldam (2008) conduct a meta-analysis of the aid-growth literature and find that aid has failed to stimulate income growth. Mekasha and Tarp (2011) look closely at the Doucouliagos and Paldam analysis and suggest that its results are not robust owing to a number of technical issues. Rajan and Subramanian (2008) could not find any association between aid and growth. This widely cited paper has been extremely influential in policy circles and has led some observers to conclude that it is evidence that aid has been ineffective in promoting growth. What is interesting is that Rajan and Subramanian, like Doucouliagos and Paldam (2008), simply could not find a link between aid and growth, which is not to say that such a link does not exist. These studies certainly did not conclude that aid decreases growth in developing countries. That so many other studies have found the link suggests it does in all probability exist and that the observers who want to use the Rajan and Subramanian study as evidence that aid does not work are placing too heavy an emphasis on a single study.
Collier and Dollar (2001, 2002) provide an innovative contribution to the aid impact literature. They quantify the impact of aid on poverty, via its contribution to economic growth in recipient countries and derive what is termed a “poverty efficient” foreign aid allocation. Such an allocation is that which maximizes global poverty reduction subject to a budget constraint, which is simply the total amount of aid available for allocation globally in any particular year. Collier and Dollar find that aid is positively associated with per capita income growth in recipient countries, subject to the quality of the policy regime in them. Based on aid, policy quality, and poverty levels for a sample of 106 developing countries, Collier and Dollar (2002) estimate that aid has lifted 10 million people out of extreme income poverty each year. They also conclude if aid were better directed toward poorer countries with better policies, the number would rise to 19 million per year. The intuition behind this second finding is that by allocating more aid to countries with better policies and, as a result, less to countries with inferior policies, its contribution to global growth is larger, and for this reason its contribution to global poverty reduction is also larger.
Feeny and McGillivray (2011) conduct a broadly similar analysis, albeit focused solely on economic growth. They look at the amounts of aid to recipient countries that (p. 701) are “growth efficient,” which maximizes the impact of aid on these countries’ per capita income. Feeny and McGillivray find that the growth-efficient aid amount, averaged across recipient countries, is in the vicinity of 20 percent of recipient country GDP. It is possible to extend the findings of Feeny and McGillivray (2011) to calculate an average contribution of aid to recipient country GDP growth. Such a calculation is based on the aid-growth coefficients obtained and the average level of aid relative to GDP developing countries have received since 1970. There calculation suggests that per capita economic growth in these countries would be roughly one percentage point lower in the absence of aid. This number can be combined with the global GEP used by Collier and Dollar (2001, 2002) of -2.0, and data on the number of people living in extreme poverty between 1980 and 2008, the longest period for which data on global poverty are currently available. This combination suggests that aid has lifted 6 million per year or 170 million people in total out of extreme income poverty during this period. This number, which should be treated with healthy skepticism and no more than a rough approximation, is not terribly dissimilar to the estimate of 10 million people per year provided by Collier and Dollar (2002).
Aid on the GEP
There appear to be no studies that look at the relationship between aid and GEPs. A number of studies have, however, looked at the abovementioned GEP determinants, those being inequality, secondary schooling, and per capita income levels. As noted, if aid increases growth in per capita income then it must by definition increase the level of this income. The survey presented earlier of the aid-growth literature leads one to conclude therefore, that aid has a positive impact on the GEP by increasing the level of per capita income. Surveying studies of aid and inequality points to little consensus, unfortunately. Bjørnskov (2009) and Herzer and Nunnenkamp (2012) find that foreign aid is associated with increasing income inequality. Shafiullah (2011) finds that aid actually reduces income inequality in a sample of developing countries while Chong et al. (2009) find aid has no discernible impact. Increases in human capital include improvements in education and health. Few studies have looked at the impact of aid on schooling, and those that have, looked at enrollment per se and not specifically at secondary schooling. They also look only at the impact of education aid and are therefore subject to the criticism noted previously concerning the poverty impact of aid to specific sectors. Of these studies, Dreher et al. (2008) find that aid provided to the education sector of recipient countries is successful at increasing school enrollments. Michaelowa and Weber (2007) find a positive association between aid devoted to education and school enrolment and completion rates. These findings notwithstanding, it is unlikely that these studies offer much insight into the impact of aid on poverty through its impact on the GEP.
Other studies have focused on the impact of aid on the composition of public expenditures. Gomanee, Girma, and Morrissey (2005a) finds that aid improves welfare by (p. 702) increasing the share of public spending devoted to the social sectors. Whether this is good for poverty reduction remains to be seen, given the earlier comment about pro-rich biases of these expenditures in many developing countries. Mosley et al. (2004) provide empirical findings that might, however, provide insights into the impact of aid on the GEP. They find that aid increases the pro-poor orientation of public expenditures in developing countries.
Field experiments tend to find that aid has an important and direct impact on well-being in developing countries (Banerjee and He 2003). Studies have found that foreign aid interventions have assisted in achieving some notable global health outcomes. For example, Levine and the What Works Working Group (2004) document 17 successful health interventions that have been funded by aid. For example, the global eradication of smallpox, controlling tuberculosis in China, eliminating polio in the Americas, reducing maternal mortality rates in Sri Lanka, controlling river blindness in Africa, preventing diarrheal disease in Egypt through oral rehydration programs, controlling trachoma in Morocco, reducing guinea worm disease in Africa and Asia, and eliminating measles in Southern Africa. These findings are all well and good and suggest that aid might have reduced poverty in dimensions other than income. Unfortunately, however, field experiments seem not to have looked, or if so not very often, at the impact of aid-supported microinventions on income poverty. As such this literature would appear not to be informative with respect to the specific interest of this chapter.
This chapter sought to objectively evaluate the controversial relationship between foreign aid and poverty reduction in developing countries. Three approaches were undertaken to achieve this goal. The first approach was to analyze descriptive statistics and to examine the simple empirical relationship between foreign aid flows and poverty levels through time. The second approach was to examine the analytics of aid and poverty by highlighting the various ways in which foreign aid can potentially impact poverty. Finally, the third approach was to review the increasing number of empirical studies that have examined the aid-poverty relationship. Whether it is possible to conclude that foreign aid reduces poverty in developing countries was a specific focus of the chapter.
The descriptive statistics highlight the trends and simple empirical relationship between aid and poverty. They demonstrate that, over the period 1980 to 2008, aid volumes have trended upward while extreme poverty has tended to fall. This is true if aid is measured in nominal, real, or per capita terms. While this is encouraging from an aid supporter’s point of view, relationships between aid and poverty do vary across time (p. 703) periods and geographic regions, and the measures of aid and poverty employed also vary, making it very difficult to draw any robust conclusions. Interpreting the simple bivariate relationship is also difficult. Higher levels of aid being associated with lower levels of poverty could be explained by donors providing more assistance to less poor countries rather than poverty reduction arising as a result of the aid itself. Crucially, the descriptive statistics are unable to provide the counterfactual—or the levels of poverty that would have prevailed in the absence of foreign aid.
In analyzing the analytics of foreign aid and poverty, it is recognized that aid rarely gets provided directly to the poor in developing countries and this has important implications for how the impact of aid on poverty is assessed. Any relationship will be indirect. Regression analysis can help inform us of the impact of aid on income poverty, as well as on known drivers of poverty such as economic growth and improvements in health and education. Importantly, foreign aid could affect the growth elasticity of poverty—or the amount of poverty reduced from a given increase in growth. Unfortunately while regression-based approaches can reveal whether aid is effective at reducing poverty in developing countries, they are unable to reveal how aid achieves this objective, limiting the insights for policymakers.
A thorough and objective reading of the academic literature is encouraging. There is increasing evidence that aid has reduced poverty from empirical studies. Results from these studies suggest that the impact is fairly modest, with aid pulling between 6 and 10 million people out of extreme income poverty per year. Given that the latest data reveal that 1.22 billion people still live in poverty, foreign aid has reduced global poverty by only a very small percentage. Further, these figures suggest that the $4.3 trillion that has been provided in aid between 1960 and 2011 has at best pulled 520 million out of poverty. This equates to a cost in excess of $8,000 in aid to pull a person out of poverty. Whether this is value for the money is subjective and should be evaluated against other ways of reducing poverty. It is certainly of immeasurable importance for those that have been successfully pulled out of poverty, and it is a noteworthy achievement of the international community. It also indicates that increasing the poverty-reducing impacts of aid remains a worthy and important challenge.
We end this chapter with two suggestions for future research on the impact of aid on income poverty. The first concerns the use of poverty data in econometric analysis. It was noted that there has been surprising little use of these data. The data on poverty can always be questioned on the grounds of measurement error and cross-country comparability. But these comments can be made of much of the data used in cross-country empirical analysis and do not alone invalidate the use of poverty data. Much more use of these data would therefore appear to be warranted. It should indeed be a priority for future research. The literature needs more studies of the type discussed in the “Aid and Poverty” section that further develop the methods used by them to provide more informative and empirically rigorous results. The second direction is based on a recognition of one of the principal limitations of these studies: whether they can tell us if aid has impacted poverty and to what extent; yet they cannot identify the mechanisms or channels through which the former impacts the latter. It involves, firstly, seeking to clarify (p. 704) which drivers of poverty-reduction aid can drive and, secondly, seeking to establish the impacts of aid on these drivers. This need not involve purely empirical research. If it is possible to establish these impacts, this will not only tell us the impact of aid on poverty but also guide the design and implication of pro-poverty-reduction aid policies.
Adams, Richard, Jr. 2003. “Economic Growth, Inequality and Poverty: Findings from a New Dataset.” World Bank Policy Research Working Paper No. 2972. Washington DC: World Bank.Find this resource:
Adams, Jr. Richard. 2004. “Economic Growth, Inequality and Poverty: Estimating the Growth Elasticity of Poverty.” World Development 32(12):1989–2014.Find this resource:
(p. 706) Addison, Tony, George Mavrotas, and Mark McGillivray. 2005. “Development Assistance and Development Finance: Evidence and Global Policy Agendas.” Journal of International Development 17(6):819–36.Find this resource:
Alvi, Eskander and Aberra Senbeta. 2012. “Does Aid Reduce Poverty?” Journal of International Development 24(8):955–76.Find this resource:
Arvin, B. Mark and Francisco Barillas. 2002. “Foreign Aid, Poverty Reduction and Democracy.” Applied Economics 34(4):2151–56.Find this resource:
Asra, Abuzar, Gemma Estrada, Yangseon Kim and M. G. Quibria. 2005. “Poverty and Foreign Aid: Evidence from Cross-Country Data.” Economics and Research Division, Working Paper Series No. 65. Manila: Asian Development Bank.Find this resource:
Bahmani-Oskooee, Moshen and Maharouf Oyolola. 2009. “Poverty Reduction and Aid: Cross-country Evidence.” International Journal of Sociology and Social Policy 29(5):264–73.Find this resource:
Banerjee, Abhijit and Ruimin He. 2003. Making Aid Work. Cambridge, MA: MIT Press.Find this resource:
Bauer, Peter. 1966. “Foreign Aid: An Instrument for Progress?” Pp. 1–30 in Two Views on Aid to Developing Countries, edited by Peter Bauer and Barbara Ward. London: Institute for Public Affairs.Find this resource:
Bauer, Peter. 1991. The Development Frontier: Essays in Applied Economics. Hemel Hempstead: Harvester-Wheatsheaf.Find this resource:
Bhalla, Sirjit. 2002. Imagine There’s No Country: Poverty Inequality and Growth in the Era of Globalization. Washington DC: Institute for International Economics.Find this resource:
Bjørnskov, Christian. 2009. “Do Elites Benefit from Democracy and Foreign Aid in Developing Countries?” Journal of Development Economics 92(2):115–24.Find this resource:
Boone, Peter. 1996. “Politics and the Effectiveness of Foreign Aid.” European Economic Review 40(2):289–329.Find this resource:
Bourguignon, Francois. 2004. “The Growth Elasticity of Poverty Reduction: Explaining Heterogeneity across Countries and Time Periods.” World Bank Working Paper. Washington, DC: World Bank.Find this resource:
Bruno, Michael, Martin Ravallion, and Lyn Squire. 1998. “Equity and Growth in Developing Countries: Old and New Perspectives on the Policy Issues.” Pp. 98–112 in Income Distribution and High Growth, edited by V. Tani and K-Y Chu. Cambridge, MA: MIT Press.Find this resource:
Burnside, Craig and David Dollar. 2000. “Aid, Policies and Growth.” American Economic Review 90(4):847–68.Find this resource:
Chong, Alberto, Mark Gradstein and Maria-Cecilia Calderón. 2009. “Can Foreign Aid Reduce Income Inequality and Poverty?” Public Choice 140:59–84.Find this resource:
Clemens, Michael. Steven Radelet and Rikhil Bhavani. 2004. Counting Chickens When They Hatch: The Short-Term Effect of Aid on Growth, Center for Global Development Working Paper Number 44. Washington, DC: Center for Global Development.Find this resource:
Collier, Paul and David Dollar. 2001. “Can the World Cut Poverty in Half? How Policy Reform and Effective Aid Can Meet International Development Goals.” World Development 29(11):1787–1802.Find this resource:
Collier, Paul and David Dollar. 2002. “Aid Allocation and Poverty Reduction.” European Economic Review 46(8):1475–1500.Find this resource:
Dalgaard Carl, Henrik Hansen and Finn Tarp. 2004. “On the Empirics of Foreign Aid and Growth.” Economic Journal 114(496):191–216.Find this resource:
De Janvry, Alain and Elisabeth Sadoulet. 2000. “Growth, Poverty and Inequality in Latin America: A Causal Analysis 1970–94.” Review of Income and Wealth 46(3):267–87.Find this resource:
(p. 707) Dollar, David and Aart Kraay. 2002. “Growth Is Good for the Poor.” Journal of Economic Growth 7(3):195–225.Find this resource:
Doucouliagos, Hristos and Martin Paldam. 2008. “Aid Effectiveness on Growth: A Meta Study.” European Journal of Political Economy 24:1–24.Find this resource:
Dreher, Axel, Peter Nunnenkamp and Rainer Thiele. 2008. “Does Aid for Education Educate Children? Evidence from Panel Data.” World Bank Economic Review 22(2):291–314.Find this resource:
Easterly, William. 2006. The White Man’s Burden: Why the West’s Efforts to Aid have done So Much Ill and So Little Good. London: Penguin.Find this resource:
Feeny, Simon and Mark McGillivray. 2011. “Scaling-Up Foreign Aid: Will the ‘Big Push’ Work?” The World Economy 4(1):54–73.Find this resource:
Feyzioglu, Tarhan, Vinaya Swaroop and Min Zhu. 1998. “A Panel Data Analysis of the Fungibility of Foreign Aid.” World Bank Economic Review 12(1):29–58.Find this resource:
Franco-Rodriguez, Susanna, Mark McGillivray and Oliver Morrissey. 1998. “Aid and the Public Sector in Pakistan: Evidence with Endogenous Aid.” World Development 26:1241–50.Find this resource:
Friedman, Milton. 1970. “Foreign Economic Aid: Means and Objectives.” Pp. 63–78 in Jagdish Bhagwati and Richard Eckaus. Foreign Aid. Harmondsworth: Penquin Books.Find this resource:
Gang, Ira and Haider Khan. 1991. “Foreign Aid, Taxes and Public Investment.” Journal of Development Economics 24:355–69.Find this resource:
Gomanee, Karuna, Sourafel Girma, and Oliver Morrissey. 2005a. “Aid, Public Spending and Human Welfare: Evidence from Quantile Regressions.” Journal of International Development 17(3):299–309.Find this resource:
Gomanee, Karuna, Oliver Morrissey, Paul Mosley, and Arjan Verschoor. 2005b. “Aid, Government Expenditure, and Aggregate Welfare.” World Development 33(3):355–70.Find this resource:
Heller, Peter. 1975. “A Model of Public Fiscal Behaviour in Developing Countries: Aid, Investment and Taxation.” American Economic Review 65:429–45.Find this resource:
Herzer, Dierk and Peter Nunnenkamp. 2012. “The Effect of Foreign Aid on Income Inequality: Evidence from Panel Cointegration.” Kiel Working Paper No. 1762. Kiel: Kiel Institute for the World Economy.Find this resource:
Kosack, Stephen. 2003. “Effective Aid: How Democracy Allows for Development Aid to Improve the Quality of Life.” World Development 31:1–22.Find this resource:
Kraay, Aart. 2006. “When is Growth Pro-Poor? Evidence from a Panel of Countries.” Journal of Development Economics 80(10):198–227.Find this resource:
Levine, Ruth and the What Works Working Group. 2004. Millions Saved: Proven Successes in Global Health. Washington, DC: Center for Global Development.Find this resource:
McGillivray, Mark. 2003. “Aid Effectiveness and Selectivity: Integrating Multiple Objectives in Aid Allocations.” DAC Journal 4(3):23–36.Find this resource:
McGillivray, Mark and Oliver Morrissey. 2001. “Fiscal Effects of Aid.” UNU-WIDER Discussion Paper No.2001/61. Helsinki: United Nations University World Institute for Development Economics Research.Find this resource:
McGillivray, Mark, Simon Feeny, Nils Hermes and Robert Lensink. 2006. “Controversies over the Impact of Development Aid: It Works, It Doesn’t, It Might, but That Depends …” Journal of International Development 18(7):1031–50.Find this resource:
McGillivray, Mark, David Fielding, Sebastian Torres, and Stephen Knowles. 2013. Does Aid Work for the Poor? Geelong: Deakin University.Find this resource:
Mekasha, T. J. and Fin Tarp. 2011. “Aid and Growth: What Meta-Analysis Reveals.” UNU-WIDER Working Paper No. 2011/22. Helsinki: United Nations University World Institute for Development Economics Research.Find this resource:
(p. 708) Michaelowa, Katharina and Anke Weber. 2007. “Aid Effectiveness in the Education Sector: A Dynamic Panel Analysis.” Pp. 357–86 in Theory and Practice of Foreign Aid, edited by S. Lahiri. Amsterdam: Elsevier.Find this resource:
Morrissey, Oliver. 2001. “Does Aid Increase Growth?” Progress in Development Studies 1:37–50.Find this resource:
Mosley, Paul, John Hudson, and Arjan Verschoor. 2004. “Aid, Poverty Reduction and the New Conditionality.” Economic Journal 114(496):F217–43.Find this resource:
Moyo, Dambisa. 2009. Dead Aid: Why Aid Is Not Working and How There Is Another Way for Africa. New York: Farrar, Straus and Giroux.Find this resource:
OECD. 2008. Is It ODA? Factsheet. November. Paris: OECD.Find this resource:
OECD. 2012. International Development Statistics Online. Paris: OECD.Find this resource:
Pack, Howard and Janet-Rothenberg Pack. 1990. “Is Foreign Aid Fungible? The Case of Indonesia.” Economic Journal 100:188–94.Find this resource:
Pack, Howard and Janet-Rothenburg Pack. 1993. “Foreign Aid and the Question of Fungibility.” Review of Economics and Statistics 75:258–65.Find this resource:
Rajan, Raghuram G. and Arvind Subramanian. 2008. “Aid and Growth: What Does the Cross-country Evidence Really Show?” Review of Economics and Statistics 90(4):643–65.Find this resource:
Ravallion, Martin and Shaohua Chen. 1997. “What Can New Survey Data Tell Us about Recent Changes in Distribution and Poverty?” World Bank Economic Review 11:357–82.Find this resource:
Riddell, Roger. 2007. Does Foreign Aid Really Work? Oxford: Oxford University Press.Find this resource:
Sachs, Jeffrey. 2006. The End of Poverty: Economic Possibilities for Our Time. London: Penguin.Find this resource:
Shafiullah, Muhammad. 2011. “Foreign Aid and Its Impact on Income Inequality.” International Review of Business Research Papers 7(2):91–105.Find this resource:
Son, Hyun, H. and Nanak Kakwani. 2004. “Economic Growth and Poverty Reduction: Initial Conditions Matter.” International Poverty Centre Working Paper No. 2. New York: United Nations Development Program.Find this resource:
UN. 2003. Indicators for Monitoring the Millennium Development Goals: Definitions, Rationale, Concepts and Sources. New York: United Nations.Find this resource:
Ward, Barbara. 1966. “The Decade of Development: A Study in Frustration.” Pp. 31–58 in Two Views on Aid and Development, edited by Peter Bauer and Barbara Ward. London: Institute of Public Affairs.Find this resource:
World Bank. 2004. World Development Report 2004: Making Services Work for Poor People. New York: Oxford University Press.Find this resource:
World Bank. 2012. World Development Indicators Online Database. Washington DC: World Bank.Find this resource:
(1.) ODA is defined as financial flows to developing countries that are: (1) undertaken by the official sector; (2) have economic development and welfare as their main objective; and (3) provided on concessional financial terms (any loan must have a grant element of at least 25 percent) (OECD 2008). Aid that is overtly military in nature is excluded from ODA.
(2.) These numbers are in constant 2011 prices. If current price data are used, ODA increased from $4.7 billion in 1960 to $165 billion in 2011. These numbers refer to disbursements of ODA, less repayments of concessional loans.
(3.) The international development and related literatures abound with works that fall into the pro- and anti-aid positions. Some date back many decades. Not all look specifically at poverty but at aid from a general development-effectiveness positions that consider impacts on economic growth, health, education, governance, and the like. Works that are highly critical of aid include Bauer (1966, 1991), Friedman (1970), Easterly (2006), and Moyo (2009), while those that are supportive albeit often with caveats include Ward (1966), Sachs (2006), and Riddell (2007).
(4.) This statistic has been obtained by, firstly, taking the average ODA to ratio of each developing country during the period 1970 and 2010 and, secondly, taking the average of these receipts across all developing countries for which GDP data are available. Requisite data were taken from OECD (2012) and World Bank (2012).
(5.) These countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Korea, Luxemburg, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom, and the United States (OECD 2012).
(6.) The international community’s strategy to achieve the Millennium Development Goals relies heavily on the ability of aid to contribute to achievements in health and education and, to this extent, reductions in poverty in these nonincome quality of life dimensions.
(7.) Studies comprising this relatively small literature include Boone (1996), Kosack (2003), Mosley et al. (2004), Gomanee et al. (2005a, 2005b), Michaelowa and Weber (2007), and Dreher et al. (2008). Results remain a little mixed but tend on balance to support the view that aid is positively associated with higher achievements in health and education.
(9.) The data used to construct Figure 30.1 and those in the remainder of this chapter have been obtained from OECD (2012) and World Bank (2012). The poverty data plotted in Figure 30.1, and those showing global poverty levels, are for the sum of people living below each poverty line in all low- and middle-income countries. The World Bank (2012) only reports these aggregates, and equivalent regional numbers, from 1981. Currently, the latest such data are for the year 2008.
(10.) This is also the case for measuring total aid (in absolute terms, irrespective of population) in constant price terms, both for all countries combined, those in South Asia and in sub-Saharan Africa.
(12.) Excluded from the data used to plot Figure 30.8 is a handful of small population countries for which per capita aid flows were tremendously high and distorted the overall relationship depicted. That noted, a negative relationship between aid and poverty exists with these countries included.
(13.) Of all the components of ODA, food and humanitarian aid, which includes emergency relief, tends to be more predominantly allocated directly to poor people. But these forms of aid have constituted roughly 8 percent of total ODA over the period 1960 to 2011 (OECD 2012). Other types of aid, which are not formally considered forms of aid in its reporting, that are given directly to poor people are referred to later in the chapter.
(14.) Bourguignon (2004) estimates that economic growth explains 26 percent of the variation in changes in poverty headcounts. Kraay (2006), conversely, finds that virtually all of the cross-country variation in changes in poverty is due to cross-country changes in growth.
(15.) These studies recognize that the aid variable is likely endogenous due to simultaneity bias. While aid might reduce poverty, it is also the case that more aid is allocated to poorer countries. The studies use different techniques to control for the endogeneity of aid.
(16.) Arvin and Barillas (2002) examine the causality between aid, democracy, and poverty. They find results vary greatly by region, with aid reducing poverty in East Asia and the Pacific but having a detrimental impact in low-income countries. Unfortunately the so-called measure of poverty employed by this study is simply the level of GDP per capita.
(18.) That the aid-growth literature seems to be in agreement that growth would be lower in the absence of aid does not imply that there is no disagreement among the researchers involved. There is intense disagreement over contingencies to which the aid-growth relationship is subject. Burnside and Dollar (2000) concluded that aid only contributed to growth in countries with “good” economic policies, broadly defined. Subsequent studies struggle to replicate this finding, with others instead pointing to contingencies such as structural vulnerability, political stability, democracy, and climate (McGillivray 2003; Clemens et al. 2004; McGillivray et al. 2006).