Aid effectiveness: research, policy and unresolved

This art icle was downloaded by: [ 68.33.167.238] On: 16 August 2014, At : 07: 48 Publisher: Rout ledge I nform a Lt d Regist ered in England and Wales Regist ered Num ber: 1072954 Regist ered office: Mort im er House, 37- 41 Mort im er St reet , London W1T 3JH, UK Development Studies Research: An Open Access Journal Publicat ion det ails, including inst ruct ions for aut hors and subscript ion informat ion: ht t p:/ / www.t andfonline.com/ loi/ rdsr20 Aid effectiveness: research, policy and unresolved issues a M.G. Quibria a Depart ment of Economics, Morgan St at e Universit y, Balt imore, MD, USA Published online: 08 Aug 2014. To cite this article: M.G. Quibria (2014) Aid effect iveness: research, policy and unresolved issues, Development St udies Research: An Open Access Journal, 1:1, 75-87 To link to this article: ht t p:/ / dx.doi.org/ 10.1080/ 21665095.2014.922890 PLEASE SCROLL DOWN FOR ARTI CLE Taylor & Francis m akes every effort t o ensure t he accuracy of all t he inform at ion ( t he “ Cont ent ” ) cont ained in t he publicat ions on our plat form . Taylor & Francis, our agent s, and our licensors m ake no represent at ions or warrant ies what soever as t o t he accuracy, com plet eness, or suit abilit y for any purpose of t he Cont ent . Versions of published Taylor & Francis and Rout ledge Open art icles and Taylor & Francis and Rout ledge Open Select art icles post ed t o inst it ut ional or subj ect reposit ories or any ot her t hird- part y websit e are wit hout warrant y from Taylor & Francis of any kind, eit her expressed or im plied, including, but not lim it ed t o, warrant ies of m erchant abilit y, fit ness for a part icular purpose, or non- infringem ent . Any opinions and views expressed in t his art icle are t he opinions and views of t he aut hors, and are not t he views of or endorsed by Taylor & Francis. The accuracy of t he Cont ent should not be relied upon and should be independent ly verified wit h prim ary sources of inform at ion. Taylor & Francis shall not be liable for any losses, act ions, claim s, proceedings, dem ands, cost s, expenses, dam ages, and ot her liabilit ies what soever or howsoever caused arising direct ly or indirect ly in connect ion wit h, in relat ion t o or arising out of t he use of t he Cont ent . This art icle m ay be used for research, t eaching, and privat e st udy purposes. Term s & Condit ions of access and use can be found at ht t p: / / www.t andfonline.com / page/ t erm s- and- condit ions I t is e sse n t ia l t h a t you ch e ck t h e lice n se st a t u s of a n y give n Ope n a n d Ope n Se le ct a r t icle t o con fir m con dit ion s of a cce ss a n d u se . Development Studies Research, 2014 Vol. 1, No. 1, 75–87, http://dx.doi.org/10.1080/21665095.2014.922890 Aid effectiveness: research, policy and unresolved issues M.G. Quibria* Department of Economics, Morgan State University, Baltimore, MD, USA (Received 20 November 2013; accepted 7 May 2014) Downloaded by [68.33.167.238] at 07:48 16 August 2014 This paper provides a critical review of the recent research on aid effectiveness. It argues that there is an enormous disjunction between research on aid effectiveness and current policies and practices: in particular, recent empirical research efforts have spawned a large body of work that is rife with controversies and insiders’ drama, but sheds little light on important policy issues regarding allocation, design and delivery of foreign aid. The paper argues that a convergence of the two universes – research and policies – is essential both for a sophisticated understanding of the underlying issues and for formulating appropriate policies and practices for aid effectiveness. Keywords: aid effectiveness; poverty; growth; selectivity; aid policy 1. Introduction Does foreign aid help economic development? Although the empirical economics literature on aid effectiveness, which investigates the macroeconomic impact of foreign aid on the economic development of poor countries, has grown briskly – both in volume and in econometric sophistication – it has shed surprisingly little light on this question. Aid has in many instances been accompanied by rapid economic growth and brisk poverty reduction; but in others, it has been accompanied by deteriorating economic outcomes. Given the vast diversity of empirical economic outcomes across countries, summarizing this experience in the form of a robust statistical relationship has proved extremely difficult and contentious. While some authors have concluded that foreign aid is effective, others have reached the diametrically opposite conclusion. Yet still others have gone on to find a common ground between the two conclusions: even if aid is usually ineffective, it can be effective under some special circumstances. In view of the different readings of the evidence, the economics profession has become sharply divided between those who are optimistic about the impact of foreign aid and those who are pessimistic. The body of research on aid effectiveness, which was built around the cross-country regression framework, has been highly aggregative and narrowly focused. The state of the current literature was succinctly summarized by Rajan and Subramanian (2011), two leading contributors to the subject: ‘This literature does not provide robust evidence of either a positive or negative correlation between foreign aid inflows and the economic growth of poor countries.’ The present paper argues that the empirical research on aid effectiveness has few coherent and robust findings either to inform or to guide policy, and that aid policies and practices have often been influenced by defunct research that has proven conclusively wrong. To establish this argument, this paper first provides a brief critical review of the state of the empirical research that has underpinned much of the present-day foreign-assistance policies of donor agencies, and then discusses the policy issues that have featured in current aid-effectiveness deliberations. This juxtaposition makes it clear that there is a big disjunction between the empirical economic research and the policies and practices of aid. The paper is organized as follows: Section 2 discusses the background to the current debate on aid effectiveness and places the ongoing analytical and policy issues into perspective; Section 3 provides a synoptic review of the current research on aid effectiveness and highlights both the current debates and the disparate and conflicting findings; Section 4 reviews some of the salient policy issues associated with the current aid policies and practices in light of the research on aid effectiveness; and finally, based on our preceding discussions, Section 5 offers some concluding observations on the state of foreign aid empirics and policies. *Email: mgquibria.morgan@gmail.com © 2014 The Author(s). Published by Routledge. This is an open-access article distributed under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/3.0/, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The moral rights of the named author(s) have been asserted. 76 Downloaded by [68.33.167.238] at 07:48 16 August 2014 2. M.G. Quibria The context Research on foreign aid has always been marked by heated controversies. In the early days of foreign aid, research focused on the impact of aid on domestic saving, which then was considered the most critical determinant of economic development. While optimists held that foreign aid would lead to a dollar-for-dollar increase in savings, pessimists argued that foreign aid would lead to unproductive government expenditure, corruption and the crowding out of private savings. Experiences varied across countries. However, the ‘average’ results from the regression literature indicate that the truth lies inbetween these two extremes: while aid does increase savings, it does not increase savings dollar-by-dollar by the amount of aid. This literature, which is considered crude by today’s technical standard, focused on savings and investment as a measure of aid effectiveness; this reflected a narrow perspective, because savings–investments are a means, not the end, of economic development. The next round of research focused on the relationship between aid and economic growth. An important earlier contributor was Boone (1996), who was applauded for his technical econometric innovation – being the first to introduce the so-called instrumental variable method to this analysis – as well as for his forceful argument. He attributed to political-economic reasons the ineffectiveness of aid to raise growth. Boone reasoned that in a society where the political elite dominate the masses, aid is no more than an income transfer to the elite group. That transfer only increases the consumption of the rich to the exclusion of the poor, as the latter has no effective representation in the polity. Boone’s conclusions apparently flounder in the face of evidence to the contrary: many aid-recipient countries have made significant strides in poverty reduction in the last 30 or so years. His conclusions, which were a broad-brush generalization of the conditions in developing countries, nevertheless, resonated with many economists as well as policy-makers in developed countries who were skeptical about the impact of aid. Although by the mid-1990s there was a considerable volume of empirical work on the macroeconomic impact of aid (for a review, see Hansen and Tarp 2000), it received little attention outside the academic community. The intellectual isolation of this period was broken by two papers (the second being the published version of the first) by Burnside and Dollar (BD) (1997, 2004)1 that suggested that (1) aid is generally ineffective in promoting growth and (2) the impact of aid on growth is positive in countries with a good policy environment. This conditional effectiveness is indicated by the significant and positive coefficient on the ‘aid*policy’ interaction in the growth regression.2 The term ‘good policy’ has been used by BD to indicate macroeconomic soundness. It is an index of fiscal, monetary and trade policy indicators – more precisely, it is a linear combination of the budget surplus, inflation and trade openness. The appeal of the BD proposition of conditional aid effectiveness is its reductive simplicity; it encapsulates the issue of aid effectiveness into a simple success criterion: an index of sound macroeconomic policies that is easily monitored and fully consistent with the prevailing orthodoxy of the ‘Washington consensus’. The principal policy conclusion that has been widely drawn by the international donor community from the BD work is selectivity: aid should be allocated only to countries with good policies. The principle of selectivity has gradually emerged as the conventional wisdom and the operating framework for aid allocation by international development agencies. The message of selectivity was further reinforced in the high-profile World Bank (1998) report Assessing Aid, which went beyond the quantitative results: it added further ‘evidence’ from the World Bank’s extensive evaluation studies. This emphasis on selectivity is, however, neither new nor novel. Almost four decades ago, Bauer, a fervid critic of foreign aid, suggested that aid be allocated ‘more selectively both politically and geographically’ (1966, 32). Bauer (1984, 60–61) exhorted donors to be ‘deliberately discriminating’ and recommended that aid be given only to those governments that ‘promote the welfare of the people’ by ‘effective administration, the performance of the essential tasks of government and the pursuit of liberal economic policies’. However, what gave the argument of BD its added cogency was its apparent rigorous empirical grounding. Since its publication, the BD analysis came under intense scrutiny, which slowly chipped away at the conclusions of these researchers. However, even if their evidence is taken at face value, it is not clear what conclusion one can reasonably infer from it. If aid has been more effective in some countries, does that imply that replicating the same conditions elsewhere will bring forth firstorder improvements in aid efficiency? However, many of these conditions are not simply replicable. Some political and legal institutions in poor countries are historically rooted and shaped by political, historical and social constraints peculiar to the individual country; consequently, change can only be gradual or it could require substantial investments of human, physical and financial resources that are beyond the immediate fiscal capacity of the country (Quibria 2014). 3. The research controversies BD research on aid effectiveness has stimulated a sizeable growth in the empirical literature on aid effectiveness. This literature has both examined the robustness of the BD empirical findings and advanced a number of alternative hypotheses on aid effectiveness. Development Studies Research Downloaded by [68.33.167.238] at 07:48 16 August 2014 3.1. Empirical results The inquiry into whether the BD conditional effectiveness proposition is robust has proceeded at three different levels. First, a number of authors – for example, Dalgaard and Hansen (2001), Hansen and Tarp (2001), Hudson and Mosley (2001) and Lensink and White (2001) – have undertaken similar empirical investigations by relying on the growth-regression framework. Their inquiries, however, have differed in terms of regression models, data sets, and the use of estimators; and none of these models have yielded a significant interaction effect between the BD policy index and aid. Second, Easterly, Levin, and Roodman (2004) – who re-estimated the BD model with an updated and extended dataset (a longer timeframe and greater country coverage) – could also not find any statistically significant aid–policy interaction term. Finally, Roodman (2007), who has conducted a further set of robustness tests, noted that the BD (2000) result is not robust with alternative plausible definitions of aid, policies, and growth. Although the BD analysis has been largely debunked by economists, it continues to wield considerable influence on aid donors. Recent years have seen the development of a set of alternative hypotheses on aid effectiveness. First, the empirical works by such authors as Hadjimichael et al. (1995), Durbarry, Gemmell, and Greenaway (1998) and Lensink and White (1999) showed that aid contributes positively to growth but its marginal contribution is subject to diminishing returns.3 This result was further confirmed by Hansen and Tarp (2001). They formulated a unified empirical framework that allows for various types of nonlinearities in the aid–growth relationship, such as quadratic aid and policy along with aid–policy interaction. They found that when the empirical relationship introduces nonlinearity in the impact of aid, it drives out the significance of the aid*policy interaction effect. This implies that aid has a positive effect on growth, although the effect seems to taper off as the volume of aid increases. This result is not conditional on the existence of ‘good policy’. Roodman (2007), who carried out an extensive set of robustness tests on the cross-country aid–growth relationships, found this result to be robust on a number of counts. Second, the above hypothesis of unconditional aid effectiveness was further confirmed by Clemens et al. (2012), who found that on average aid has been effective – with a modest and positive effect on growth and investment – when aid is economic. Clemens et al. divided foreign assistance into three categories: (1) emergency and humanitarian aid, which is likely to be negatively correlated with growth; (2) aid that impacts growth only in the long run, such as aid to support democracy, the environment, health and education and (3) aid that could stimulate growth within four years, such as various types of policy-based lending, 77 investments in infrastructure and aid for productive sectors such as industry and agriculture. Clemens et al.’s analysis focused on the last type, which accounts for almost half of all foreign assistance. They found that there is a positive causal relation between this type of aid and economic growth, albeit one that is subject to diminishing returns. This finding is not conditional on the recipient’s quality of institutions or policies. The paper reported that the statistical results are robust and relatively free from the econometric estimation problems typical of growth regressions. Third, Guillaumont and Chauvet (2001) argued that a country’s structural vulnerability (to external shocks) has a significant impact on the effectiveness of aid. Adding a ‘vulnerability’ variable4 to the BD formulation, they found that policy, aid and vulnerability all have a significant impact on growth. They found that the aid–policy interaction term is no longer significant but that aid is more effective when structural vulnerability is high. Aid flows help to promote growth – or to contain negative growth – in countries that are structurally vulnerable to external shocks. The Guillaumont and Chauvet (2001) hypothesis regarding the role of structural vulnerability to external shock on aid effectiveness was further confirmed by Collier and Dehn (2001), who found that the interaction term involving the change in aid and the change in export prices is significant. Fourth, though Guillaumont and Chauvet (2001) focused on the adverse effects of external economic volatility, not all socioeconomic volatility is externally induced. For instance, political instability in developing countries can generate considerable economic volatility and depress economic growth (Alesina et al. 1996; de Haan and Siermann 1996; Fosu 2002). This instability in turn is likely to have important implications for the aid effectiveness. Chauvet and Guillaumont (2004) conjectured that political instability can affect the productivity of foreign assistance in two opposite ways. On the one hand, aid can mitigate some of the negative consequences of political instability, such as economic disruptions and individual suffering. In this regard, foreign aid can be highly productive. On the other hand, since political instability implies uncertainty in the position of the ruling government, foreign assistance is likely to be deployed inefficiently. In particular, Chauvet and Guillaumont (2004) showed a highly significant and negative coefficient for interactions between aid and a measure of political instability. Although ongoing political instability can adversely affect aid effectiveness, Collier and Hoeffler (2004a, 2004b, 2004c) argued that foreign aid can play a significant role in preventing conflicts as well as improving prospects for post-conflict peace and prosperity. Their research suggested that the three most important risks for civil wars related to the level of per capita income, its rate of Downloaded by [68.33.167.238] at 07:48 16 August 2014 78 M.G. Quibria growth and the dependence on natural resource exports. They argued that aid, by substantially increasing growth in poorer countries, provides extra security benefits. Along with poverty alleviation, risk reduction offers a strong rationale for enhanced foreign assistance to riskprone, poorer countries. Similarly, Collier and Hoeffler argued that aid can play an important role in the process of rebuilding and development in the aftermath of conflict. These countries need massive investments in basic infrastructure but often lack adequate fiscal resources to undertake these investments. However, post-conflict societies are also usually characterized by a ‘high-corruption equilibrium’ (Tirole 1992; McGillivray 2003) and fragile civil administration, weaknesses that can potentially undermine the productiveness of aid. Nevertheless, the empirical results of Collier and Hoeffler suggest that on balance, aid has actually been highly effective in stimulating growth in post-conflict societies. Fifth, Dalgaard, Hansen, and Tarp (2004) highlighted the role of non-economic structural factors on aid effectiveness. In particular, they found that geographical and climate-related factors had a significant impact on growth (both directly and through the aid*tropics interaction effect). In general, they maintained that geographically challenged countries display a lower level of effectiveness, a fact that should be factored into the aid-allocation calculus. Sixth, a set of recent studies by Rajan and Subramanian (2007, 2011) argued that aid is simply ineffective (the unconditional aid-ineffectiveness hypothesis): the first paper suggested that foreign aid reduces the efficiency of manufacturing investment by adversely affecting governance and thereby limiting the growth of manufacturing exports that have been the traditional engines of growth; the second paper suggested that the beneficial impact of aid can be significantly nullified by the inevitable erosion of competitiveness (of the tradable sectors) caused by aid inflows. They argued that this happens due to the realexchange rate overvaluation associated with any large windfall, i.e. the so-called Dutch disease. Though the empirical research of Rajan and Subramanian was careful and rigorous, they seemed to overstate the importance of the Dutch disease. Even if Rajan and Subramanian were correct in believing that aid can erode competitiveness in certain sectors, it is not clear why governments cannot counter this adverse effect by appropriate fiscal and monetary policies. Furthermore, while developing countries typically produce far below capacity, the symptoms of the Dutch disease arise when countries produce close to their production possibilities’ frontiers. Moreover, if foreign assistance is directed toward improving the productive capacity of the economy (through investments in infrastructure, education, institutions and health), this productivity increase could potentially offset the loss of competitiveness resulting from the Dutch disease (Adam and Bevan 2006). A recent study by Fielding and Gibson (2012) on sub-Saharan Africa found that the long-run impact of foreign aid on the real-exchange rate was far from uniform across countries – including some measure of real-exchange rate depreciation in some countries. Sixth, the bulk of the aid-effectiveness studies focused on the aid–growth nexus. There are, of course, exceptions. There are few studies that focused on poverty – albeit in indirect and somewhat perfunctory ways. One such example is BD (1998), who by following their earlier works on aid–policy interactions (BD 1997, 2000), ran a set of regressions that replaced economic growth as the dependent variable with infant mortality as a proxy for poverty: a choice that is not clear, since other more obvious (income/consumption) measures of poverty are available. Not surprisingly, BD’s (1998) findings were completely analogous to their studies on the growth-effects of aid. Specifically, they found that aid reduces infant mortality, but only under good economic management.5 Since BD did not conduct any sensitivity analysis, the robustness issue remains: how would this relationship change if poverty is measured in different ways? A more interesting, though perhaps equally questionable, story emerged from the work of Collier and Dollar (2001, 2002). They asked an apparently simple question: what is the ‘poverty-efficient’ aid-allocation rule? That is, how should we allocate aid to bring about a maximum reduction in global poverty? This set of papers has received less attention than some of the earlier research of BD (1997, 2000); however, they should not be viewed in isolation. In essence, they complement the earlier papers to complete their story about aid and development. According to Collier and Dollar, the impact of aid on poverty depended on two factors: (1) its impact on per capita income growth, which is subject to diminishing returns and (2) the relationship between per capita income growth and poverty reduction. The first can be derived from the BD aid–growth relationship, and the second by the growth elasticity of poverty.6 Here, Collier and Dollar made a heroic and patently unrealistic assumption7 that the growth elasticity of poverty is two for all countries. The optimal allocation was obtained by equating the marginal productivity of aid in terms of poverty reduction across recipient countries. As the growth elasticity of poverty is uniform and constant across countries, this implied that the optimal allocation would be obtained when the marginal contribution of aid to economic growth is equal across countries. Their empirical results suggested that (1) the existing allocation of aid is grossly inefficient and (2) if aid allocation had followed the Collier–Dollar efficiency principle,8 the poverty reduction impact would double. Development Studies Research Downloaded by [68.33.167.238] at 07:48 16 August 2014 Finally, in recent years, econometric works have included meta-analyses – the regression of regression analyses – to synthesize the results from the existing body of empirical studies. These meta-analyses control for heterogeneity among studies. One putative advantage of a meta-analysis is that it can potentially overcome the subjectivity of the traditional literature surveys and provide a more systematic and objective quantitative assessment of the existing body of findings. Surprisingly, even these types of studies, which are supposed to provide more objective analyses, have contributed little to resolving the controversies. Consider, for instance, two such studies by Doucouliagos and Paldman (2009) and Mekasha and Tarp (2011): while the former failed to find any significant impact of foreign aid on growth, the latter found an impact that is both positive and statistically significant. 3.2. Final observations Recent years have seen a huge proliferation of econometric studies in all areas, including aid effectiveness – thanks to the easy availability of high-powered microcomputers, sophisticated software and new innovative econometric techniques. However, as the foregoing brief review of these studies of aid effectiveness suggests, this empirical literature has yielded few robust conclusions to inform practical policies (Rajan and Subramanian 2008). This literature, which is based almost exclusively on growth regressions, has many shortcomings. First, much of the empirics are based on ad hoc specifications with little or no rigorous theoretical underpinnings. Second, while recent years have seen a progressive growth in econometric complexity, they have not yielded greater clarity in understanding. The mechanics of the process seems to have largely overtaken thinking and reflection. Third, despite improvements, growth regressions are still fraught with myriad of technical econometric issues such as parameter heterogeneity and endogenous regressors, measurement errors, influential observations and error correlation: a host of issues that undermine reliability (Temple 1999). As the macro aid-effectiveness literature has failed to make any meaningful contribution to the understanding of the intricacies of aid effectiveness, it has led many to explore alternative, micro approaches. One such approach is the evaluation of aid projects, programs and policies through randomized control trials (RCTs). Under idealized conditions that seldom exist in reality, RCTs can overcome some of the methodological shortcomings of the macro aideffectiveness literature and provide impact evaluation of micro-level aid interventions. However, the RCT approach has its own limitations. First and foremost, it cannot analyze the effect of an economy-wide policy change such as trade liberalization. Second, it is now well known that there is a ‘micro–macro’ paradox, which suggests 79 that the success at the project level does not ensure success at the macro level. Even if all the projects are successful it does not mean that they will ensure success at the macroeconomic level because of the so-called fungibility issue – i.e. aid money being used for purposes other than those earmarked. Third, it provides information only about the average impact, not about when and how it works. In other words, RCTs provide little information about the underlying causal mechanisms. Finally, even with their considerable expense, RCTs provide at best local knowledge that may not apply to other contexts: there is no reason to suppose what works in one place will work elsewhere. This issue has come to be known as the problem of external validity. All this has prompted the suggestion that the secret of aid effectiveness is more likely to be revealed by trial and error than by RCTs (Deaton 2013). In short, the existing ‘rigorous’ empirical literature appears to have hit a wall: it offers little illumination beyond providing statistical codification of the obvious: foreign aid has been effective in some countries and ineffective in others. This begs the question whether this line of analysis should be abandoned in favor of in-depth longitudinal studies of individual countries that can bring to the fore important country-specific historical, social, political and cultural factors, often glossed over by cross-country regression analysis. It can be argued that the salient issues of aid effectiveness for Nepal, for example, can only be gleaned through in-depth studies of Nepal – and not by context-less generic regressions using data from a hundred plus countries. To derive the maximum benefits from such studies, they should be conducted within a common analytical framework and be informed by economic theory, history and solid empirical evidence. However, as noted by Ranis (2006), such studies, which are few and far between, should be an integral part of the future research agenda on aid effectiveness. 4. Some policy issues The following section highlights some policy issues that are associated with the design and delivery of foreign development assistance. This section examines the implications of recent research on policy issues and, in particular, how the design and delivery can be further informed and improved by research. 4.1. Growth versus poverty reduction Poverty reduction has been accepted as the overarching development objective of the international development community. This has been formalized in the Millennium Development Goals (MDGs),9 which were adopted by the United Nations (UN) at the Development Summit of the UN in 2000. Despite the acceptance of poverty reduction Downloaded by [68.33.167.238] at 07:48 16 August 2014 80 M.G. Quibria as the overarching objective of international development assistance, the bulk of the current research does not reflect this concern: most research in this area is more concerned with economic growth than poverty reduction. This neglect partly reflects a bias of the economics profession in favor of growth empirics and partly reflects the general perception that growth and poverty reduction are essentially coterminous. The latter is reflected in such statements as ‘the aid bureaucracies [these days] define their final objective as “poverty reduction”, [which is] (today’s more politically correct name for “growth”)’ (Easterly 2003, 34). However, the distinction between growth and poverty reduction is not a trivial one. The examples of India and Bangladesh are instructive. As Sen (2011) argued, while Bangladesh has half the income of India in per capita terms, it outperforms India in almost all social and human development indicators. This disjunction between economic growth and human and social indicators suggests that it is important to focus directly on poverty and human development. However, the few studies that explored the aid–poverty links, as discussed in the preceding empirical section, took a crude analytical approach. As discussed earlier, of these studies, the Collier and Dollar framework for optimal allocation of foreign aid for global poverty reduction is noteworthy. Yet, it suffers from a number of analytical shortcomings. First, poverty reduction is not a function of economic growth only: it is also influenced by many other factors, such as human and social investments.10 The currently accepted multidimensional concept of poverty, which has been encapsulated in the MDGs, goes beyond the traditional view that equates poverty with low income. Collier and Dollar’s approach to deriving a ‘poverty-efficient’ aid-allocation rule fails to recognize both the multidimensionality of poverty and the role social and human investments play in poverty reduction. Second, the Collier–Dollar formulation of poverty-efficient allocation was not the appropriate analytical framework if the objective was to attain predefined poverty reduction targets (within a given timeframe) in all developing countries, as envisioned in the MDGs. According to the millennium compact, the fundamental basis for allocating aid across countries should be MDG assessments11 and Poverty Reduction Strategy Papers (PRSPs).12 In reality, as Sachs (2005, 270) noted, the MDGs were chronically underfunded. Third, as noted by Collier and Dollar (2002), if aid allocation were not politically constrained with ad hoc limits on allocations to large countries, then the poverty-efficient allocation would imply overwhelmingly favorable disbursements to India, with its better policies and a higher incidence of poverty. However, such an allocation rule conflicts with the notion of inter-country equity, as envisioned in the millennium compact.13 In light of the above, there are reasons to be skeptical about the extent to which the insights on aid effectiveness can be transferred from growth to poverty reduction. However, it should be obvious that, depending on the quality and composition, an amount of development assistance can have distinctly different impacts on economic growth, as contrasted from poverty reduction. Therefore, if poverty reduction is the overarching objective, the empirical analysis should be framed in such a way that it speaks directly to the question of poverty reduction. 4.2. Defining good policies and institutions The selectivity proposition, popularized by BD and the World Bank (1998), is anchored in a notion of ‘good’ policies/institutions. While few would quarrel with the fact that good policies and institutions contribute to aid effectiveness, there is little agreement on what constitutes good policies and institutions. Dollar and his collaborators used a list of different indices to explore policies and institutions in their aid-effectiveness studies. This list of indices includes the BD policy index, which is essentially a proxy for sound macroeconomic policy; the index of economic management, which is a combination of the BD policy index and the international country risk guide (ICRG) index, a la Knack and Keefer (1995); the Kaufmann, Kraay, and Zoido-Labotan (2002) (KKZ) index of governance14 and the World Bank’s country policy and institutional assessment (CPIA) index.15 As we noted earlier, even though the conditional aid-effectiveness proposition is intuitively plausible, it has also proven to be statistically fragile. Despite persistent criticisms, CPIA has been used by the World Bank to allocate concessional aid resources.16 Allocating aid by CPIA punishes countries that are the most developmentally challenged. As Dalgaard, Hansen, and Tarp (2004) have noted, there is a strong correlation between countries with poor CPIA and countries in the tropics. Thus, using the CPIA to allocate aid punishes countries with unfavorable initial conditions. It is possible to conflate climate-related problems with poor CPIA ratings and the willingness to reform. Second, as the CPIA is likely to be endogenous, it cannot be meaningfully used for forecasts and policy simulations. Aside from the above specific objections against the CPIA, there are some general conceptual and methodological issues that apply to all such indices. First, it has been noted that popular indices of governance and institutions such as the ICRG and the KKZ are largely measures of outcomes (Glaeser et al. 2004) and not ‘deeper characteristics’ of institutions, in the sense of North (1981).17 As such, these indicators are poor surrogates for institutional quality, and unsuitable for exploring the causal relationships between institutions and growth. Third, the most common indices of good institutions are subjective assessments. In the case of the CPIA, it Downloaded by [68.33.167.238] at 07:48 16 August 2014 Development Studies Research reflects the perceptions of the World Bank bureaucracy; in the case of the ICRG and the KKZ, they are based largely on surveys of domestic and foreign investors. In these surveys, the respondents are asked to provide their views on the safety of their investments or their ratings on the ‘rule of law’. Given that the survey results are an aggregation of individual views, the indices essentially reflect the investors’ perception, and not an objective assessment of the institutional framework. Fourth, the indices implicitly take the reform agenda embodied in the Washington and post-Washington consensus as the benchmark. This reform agenda is largely ahistorical: it represents a ‘one-size-fits-all’ model that does not take into account any particular country’s circumstances. These indices imply that, irrespective of its stage of economic development or its position in its historical trajectory, a country would benefit from minimizing its distance from the Washington and post-Washington consensus. However, for a poor country, attaining the best-practice institutions or first-best policies is not feasible – nor even perhaps desirable. Recent development experiences of high-performing Asian economies offer three lessons in this respect.18 First, there is no unique route to desirable institutional outcomes. The process of institutional development is gradual, path dependent and endogenous.19 Institutions need to be suitable to local conditions; the experience of China in this respect is instructive. It did not achieve its growth miracle by implementing the Washington and postWashington consensus; it achieved its growth miracle by implementing policies and institutions (which created economic incentives and fostered market competition) that were appropriate to local conditions. To give an example, China has achieved some measure of effective private property rights through unique institutional innovations despite the absence of any de jure private property rights until very recently (Qian 2003). Rather than privatize land and industrial assets, the Chinese government implemented novel institutional arrangements, such as the household responsibility system and township and village enterprises (TVEs). Under the household responsibility system, land was ‘assigned’ to individual households according to their size. In TVEs, formal ownership rights were given not to private hands but to local communities (townships or villages). Local governments had a vested interest to ensure the prosperity of these enterprises as their equity stake in TVEs generated revenues directly for them. In the economic and political environment of China, property rights were effectively more secure under direct local government ownership than they would have been under a private property-rights legal regime. According to Rodrik (2005), the efficiency loss incurred due to the absence of private control rights was probably outweighed by the implicit security guaranteed by local government control. 81 Second, a transition from a low- to high-growth trajectory typically combines orthodox and unorthodox institutional practices (Rodrik 2005). Again, the experience of China is illustrative: China provided market incentives through a two-track system of reform that combines elements of orthodoxy with unorthodox practices. Its reform in agricultural liberalization, property rights, and trade liberalization is far from comprehensive. For example, China did not achieve the benefits of trade liberalization through a comprehensive program of tariff reductions but by creating a cluster of special economic zones. Third, to accelerate growth, large-scale institutional reform is neither necessary nor feasible. Indeed, wellknown historical episodes of growth acceleration have been achieved through gradual experimentation (Rodrik 2005). Examples from recent economic history include Korea in the 1960s experimenting with deregulation of the currency and the real interest rate; China in the 1970s proceeding gradually with experimental liberalization and India in the 1980s dismantling some anti-business practices. In short, it is not appropriate to compare the institutions of a poor developing country to the ‘first-best’ policy-institutions of advanced countries. As Dixit (2004) noted, it is neither necessary nor possible to create Western-style institutions from scratch. He recommends incremental improvements – working with the existing alternative institutions and building on them. This, of course, presupposes a good understanding of the various institutions of governance: how they function and interact with each other. In sum, good polices and institutions for aid effectiveness are not clearly and unambiguously defined: they are context specific and path dependent. There is no single set of ‘ideal’ policies and institutions, the benchmark against which the performance of all countries can be precisely measured; there is no single template that can be mechanically applied to all countries, irrespective of their economic constraints and stages of development. This behooves donors to take a more flexible – and less doctrinaire – approach to policies and institutions. 4.3. Ex post versus ex ante conditionality An interesting econometric result from BD (1997), which received little attention and was largely neglected in practice, was that aid has no influence on policy reform. This finding is further corroborated by such studies as Alesina and Dollar (2000), Botchway et al. (1998) and Killick, Gunatilaka, and Marr (1998). A succinct summary of these results is given by the World Bank (1998) report, Assessing Aid, which noted that there is ‘surprisingly little relationship between the amount of aid and policies’ (p. 47): there exists ‘a mountain of literature [that] conclude[s] with skepticism about the ability of conditionality Downloaded by [68.33.167.238] at 07:48 16 August 2014 82 M.G. Quibria to promote reform in countries where there is no strong local movement in that direction’ (p. 51). However, in practice, the World Bank does not pay heed to its own advice that policy conditionality does not work.20 There are many reasons why policy conditionality is ineffective. First, there is often a divergence of views between the donor and the recipient regarding the program. This divergence can relate to both primary issues, such as the content of the program, and secondary issues, such as the means, sequence or timeframe for achieving the program. Second, policy conditionality often fails because of the dynamic time inconsistency problem. The recipient government may agree to a reform program prior to receiving aid but renege on the promise after receiving aid (as incentives change after aid disbursement). As the interactions between the donor and the recipient are both dynamic as well as asymmetric, the issue of conditionality is more than designing an incentive-compatible contract in a static principal-agent framework. Third, the current structure of incentives on the donor side may also have some adverse impact on the final attainment of conditionality. Existing incentive systems in donor agencies place a high value on aid disbursement, even if it means some connivance at the failure of conditionality. Similarly, such failures may arise from individual compassion, because of the so-called ‘Samaritan’s dilemma’: there is a keen desire to help the poor in aid-recipient countries. However, while the poor may benefit from conditionalities in the long run, there is often a trade-off between relatively low short-term gains against potentially higher long-term benefits. This can lead aid agencies to overlook the non-fulfillment of policy actions in poor countries (Kanbur 2006). The ineffectiveness of policy conditionality has elicited two different types of views. The first view is that conditionality works in theory but not in practice because of flawed application. According to this view – associated with Mosley, Harrigan, and Toye (1995), among others – conditionality would work if it was properly designed and conscientiously implemented. This requires that conditionality be simple; breaches of conditionality be punished consistently and reforms must be country owned.21 However, the concept of country ownership has remained fluid: it sometimes refers to the commitment of the whole recipient society – the government, civil society and the private sector; sometimes, it refers only to the commitment of the government. Given the plasticity of the concept, Buiter (2004) finds this an ‘unhelpful’ and ‘misleading’, concept ‘whose time has gone’. However, this criticism notwithstanding, donor agencies repeat ad nauseam the importance of country ownership.22 To confer greater ownership to recipient countries, the World Bank, the IMF and other international donor agencies now develop their country programs around PRSPs.23 The second type of view is that traditional ex ante policy conditionality does not work and should be replaced by ex post policy conditionality24 (which is tantamount to selectivity).25 That is, aid should be given to countries based on ex post policies. If pursued consistently, selectivity will ensure a superior outcome. In a ‘repeated game’, as long as the donor consistently rewards aid to countries that demonstrate good policies, it will elicit good behavior from the recipient. Gunning (2000) listed the four objections against selectivity (or ex post conditionality). First, selectivity by definition excludes aid to countries with poor governance and unsound policies;26 consequently, poor people living in those countries who could potentially benefit from foreign assistance suffer. Second, countries with good policies are able to generate adequate domestic and foreign private investments and hence can do so without foreign assistance. Third, selectivity makes aid-allocation contingent on the definition of good policies. While some aspects of good policy may be objectively defined, others involve subjective judgments; given this subjectivity, there is little consensus on good policies, leading to donor–recipient bargaining. Fourth, selectivity may conflict with ownership. This happens when donors attempt to provide detailed ‘multidimensional’ definitions of ‘good policies’, which recipients often find inconsistent with their development objectives. Gunning (2000) considered the first two objections unsustainable. With respect to the first objection, he argued that poor people in poor countries do not benefit from foreign assistance when the quality of governance is questionable. However, one way to circumvent this problem is to assist the poor through NGOs. With respect to the second objection, Gunning argued that even if policies are good, poor countries do not metamorphose overnight into developed countries. In the interim period, when domestic savings and foreign private investments are inadequate, foreign aid has to play a key role in the transformation process. In sum, ex ante policy conditionality is largely ineffective in practice, so is selectivity (which is now commonly used in conjunction with process conditionality), contrary to donors’ original expectation. Conditions exogenously imposed rarely succeed. This led many, including Kanbur, Sandler, and Morrison (1999) and Ranis (2006), to argue that for effectiveness, the recipients need to be given full autonomy over aid allocation, project implementation and policy formulation. Moreover, aid from all agencies should be pooled and allocated as lump-sum transfers to recipient countries. 4.4. Measurement: outcomes versus policies How should aid be allocated across countries? How does the current practice of selectivity, which is based on assessments of government policies, compare to an alternative Downloaded by [68.33.167.238] at 07:48 16 August 2014 Development Studies Research based on assessments of outcomes as a measure of country performance? The same debate extends to the issue of conditionality. Outcome-based conditionality should be distinguished from policy conditionality. In the former, donors focus on results in terms of impact and outcomes, rather than on inputs, activities and outputs.27 There are pros and cons for each choice. The main argument in favor of policybased conditionality vis-à-vis outcome-based conditionality is that the former is easier to observe and monitor and has greater incentive effects. Policies are more directly controllable by governments and their implementation can be more easily monitored. On the other hand, outcomes are not under the full control of governments; they reflect a variety of influences, including negative exogenous shocks. Moreover, there is often a large time lag between policy decisions and outcomes such as economic growth and poverty reduction. This combination of time lags and weak links between policies and outcomes can further dilute incentives for governments to undertake positive policy actions. The main argument for outcome-based conditionality vis-à-vis policy-based conditionality is that it promotes greater ownership and accountability. Some observers argued – for example, Gunning (2000) – that the present practice of a detailed assessment of the entire policy environment is unnecessary and undermines ownership. As donors should be more concerned with outcomes, governments should be given a free hand to choose their policies. This freedom helps promote ownership of policies and strengthens accountability, which contributes to greater private sector confidence. All types of conditionalities are imperfect in the sense that they do not ensure a first-best outcome. Drazen and Fischer (1997) identified three reasons for such failures: first, government policies are imperfectly observable; second, results are not fully determined by policies but are also influenced by luck and third, governments have varying degrees of competence, which cannot be readily distinguished ex ante. In addition, there is a lot of uncertainty – as well as imperfect information – regarding the ‘results chain’ that tracks the causal consequence of a development intervention, from inputs and activities to outputs, outcomes and impacts. Under outcome-based conditionality, donors should focus on impact and outcome indicators. However, the implementation of such conditionality is fraught with practical difficulties. The results indicators commonly suggested for outcome-monitoring are GDP growth, changes in poverty and changes in child mortality; however, except for growth rates, current data on poverty and mortality are not always readily available. Second, as most of the outcome indicators are likely to change gradually, any meaningful impact assessment can only be done after an interval of a few years, and such assessments 83 may reward or punish the current government for the actions of the previous government. Given these difficulties, outcome-based conditionality that purports to monitor longer term impact and mediumterm outcome indicators may be supplemented by output and other intermediate results indicators. Depending on the availability and accuracy of different types of indicators, the optimal choice may need to include a mixture of impact and outcome indicators – that is, intermediate and final results.28 5. Concluding remarks This paper juxtaposes the empirical research on aid effectiveness against current policy concerns. In recent years, this chasm between the empirical research and the policies and practices seems to have widened further with the former trending more toward obfuscation and obtuse econometrics and away from substantive policy issues and practices. Given this divergent concern of research and policy, one cannot help wonder whether the two are marching to different drummers and whether they will ever come to converge. Needless to emphasize, such a convergence of the two universes are essential both for a sophisticated understanding of the underlying issues and for devising appropriate policies and practices for effective use of foreign assistance. This paper argues that empirical research on foreign aid must be reframed. The focus of the research needs to go beyond the current obsession about the ‘average’: does foreign aid work on ‘average’? Or what is the ‘average’ effect of a particular aid intervention? Research needs to focus on why, how and when foreign aid has worked in particular societies. Only by finding the mechanisms and processes that explain why and how aid works in a particular society will it be possible to design and deliver foreign aid effectively. This discovery would require going beyond the narrow analytics of cross-country growth regressions or for that matter RCTs. As a tool, cross-country growth regressions have proved to be too coarse to capture the complex mechanisms and processes undergirding aid effectiveness. Similarly, RCTs have their own shortcomings. They focus exclusively on the average impact; they provide little or no light on causality and they also lack external validity. All these reasons also make RCTs an unreliable analytical basis for formulating robust strategy and policies at the macro level. An effective aid policy requires countryspecific insights, which can be gleaned only from in-depth country studies that capture the flavor and texture of individual countries – in particular its institutions and politics29 (Deaton 2013) – nuances that are lost in mechanical manipulation of data. Currently, there is a huge disjunction between research and practice, similar to what transpired in other sciences in 84 M.G. Quibria earlier times. In his magisterial history of cancer research, Mukherjee (2010) noted that little interactions took place prior to the 1960s between those who studied cancer in the laboratory and those who treated cancer in the clinic: ‘The two conversations seemed to be occurring in sealed and separate universes.’ Researchers and the community of practice in foreign aid seem to similarly inhabit two separate universes. However, as the history of biomedical sciences suggests, the prospects of breakthroughs in intractable diseases are greatest when there is a tight feedback mechanism between research and practice. The field of foreign aid is no exception to this general rule. 4. 5. 6. Downloaded by [68.33.167.238] at 07:48 16 August 2014 Acknowledgments This review paper, which draws on my earlier writings on the subject, has benefited from the comments of Tetsu Ito, Frank Harrigan, Rana Hasan, Haider Khan, Salim Rashid, Chen Si, T.N. Srinivasan, Eisuke Suzuki as well as of the participants of an ADB colloquium on development effectiveness. I have also benefitted enormously from conversations and research collaboration with Nurul Islam and Antu Mushid over the years. The comments and suggestions of three anonymous referees and Emma Giberthorpe, the chief editor of this journal, are gratefully acknowledged. Finally, I thank Chickie Custodio, Vincent Conti and Cherry Zafaralla for their valuable research, editorial and secretarial assistance for the paper. 7. 8. Notes 1. 2. 3. This empirical research, which utilized a data set that covered 56 countries over the period 1970–1993, was based on a set of regression equations that took the general form: GDP growth (per capita) = other terms + b. aid + c. (aid*policy) + d. aid2 + error. Burnside and Dollar find that none of the regression coefficients are significant in their full sample. However, when they exclude some ‘outlier’ observations, the aid-squared term becomes insignificant. Next they drop the aidsquared term and experiment with a modified regression equation: GDP growth (per capita) = other terms + b. aid + c. (aid* policy) + error; this exercise renders the coefficient of aid insignificant but makes the coefficient of aid–policy interaction term highly significant. This diminishing return – which arises from bottlenecks in the physical and human capital infrastructure – possibly reflects the absorptive capacity constraint, an idea that dates back to Millikan and Rostow (1957), RosensteinRodan (1961) and Chenery and Strout (1966). The telltale signs of the absorptive capacity constraints are often manifest in the annual portfolio performance reviews (which identify various implementation issues) of the donor agencies. However, one needs to be careful not to attribute all implementation delays to absorptive capacity constraints. Some of these implementation problems may also stem from the cumbersome policies, procedures and practices across donor agencies that put an enormous demand on the scarce administrative resources of the poor countries. Easterly (2002) provided an interesting account of the heavy transaction costs of foreign aid for recipient countries. 9. 10. 11. 12. Guillaumont and Chauvet included four components in their vulnerability index: instability of agricultural income (a proxy for natural disasters), volatility of export earnings, the long-term trend in the terms of trade and the initial population. Chauvet and Guillaumont (2004) have subsequently argued that as political instability is widespread in the developing world, the discussion of aid effectiveness should explicitly consider political instability. The index of economic management used in this regression is given by: management = −1.8 + 0.65 × ICRG + 5.4 × Fiscal–1.4 × Inflation + 2.1 × Open, where ICRG is a measure of institutional quality strength that includes property rights, absence of corruption and quality of the bureaucracy; Open is the Sachs–Warner measure of trade openness; Inflation is the rate of increase of the price level and Fiscal is the budget surplus relative to GDP. As Srinivasan (2001) argued, the growth elasticity of poverty – which expresses a relationship between two endogenous variables of economic growth and poverty – is neither a stable nor a ‘deep’ parameter (in the sense of being related to technology and preferences, Lucas (1976)). Moreover, Lucas has demonstrated, the use of such parameters for policy simulations leads to misleading results. Collier and Dollar also assumed that donors have absolutely no influence on recipients’ policies. This assumption, which simplifies the algebra, does not accord with reality. The growth–aid relationships posited by BD (2000) differed from that in Collier and Dollar (2001, 2002). In BD (2000), the coefficient on the estimated aid 2 term was statistically insignificant, whereas in Collier and Dollar (2001, 2002), this term was negative and statistically significant. However, without this negative coefficient on the aid 2 term, Collier and Dollar’s poverty-efficient rule does not yield an interior solution. The MDG-concept of poverty is multidimensional: it is expressed in terms of a number of goals and indicators, which include eradication of extreme poverty and hunger; achievement of universal primary education; promotion of gender equality and empowerment of women; reduction of child mortality; improvement of maternal health; combating HIV/AIDS, malaria and other diseases and ensuring environmental sustainability. While the goals represent laudable benchmarks to address global poverty, they have been arbitrarily set: it is not clear why all developing countries should follow an identical path to poverty reduction regardless of their considerable differences. Mosley, Hudson, and Verschoor (2004) constructed a propoor (public) expenditure (PPE) index that is a weighted average of the proportions of GDP spent on poverty-related activities such as health and education. They argued that this PPE index (which reflects human and social investments by the government) was a key determinant of poverty reduction and went on to develop an alternative formulation to Collier-Dollar’s poverty-efficient allocation of aid. There has been an outpouring of MDG-related reports in the last few years from international organizations. Most of these are country-level reports produced by the UN; since 2004, the World Bank (in collaboration with the IMF) has issued an annual report called The Global Monitoring Report, which tracks progress in MDGs at the global level. To receive development assistance, low-income countries (with a few exceptions, such as India) are required to prepare national poverty reduction strategies. These PRSPs are prepared by governments with the assistance of the World Bank and IMF staff. The PRSP of a country Development Studies Research 13. Downloaded by [68.33.167.238] at 07:48 16 August 2014 14. 15. 16. 17. 18. 19. typically catalogs its macroeconomic, structural and social policies; reviews its programs to promote growth and reduce poverty and estimates external financing needs. According to the World Bank and the International Monetary Fund (2005), PRSPs are guided by five core principles: they should be (1) country-driven, involving broad-based participation by civil society and the private sector; (2) results-oriented, based on outcomes that benefit the poor; (3) comprehensive in addressing the multidimensional nature of poverty; (4) partnership-oriented, involving the participation of development partners – bilateral, multilateral and nongovernmental and (5) long-term in perspective. There is an inherent equity-efficiency trade-off in aid allocation. Allocations that support the neediest may not be the speediest in global poverty reduction. No matter what the objective is – be it growth or poverty reduction – this dilemma between equity and efficiency persists. BD (2004) revisited the empirical aid–growth relationship, employing the KKZ index of governance. The KKZ index is an amalgam of a large number of subjective assessments of institutional quality primarily made by institutional investors. The CPIA index has 16 components in 4 categories: macroeconomic policies, structural policies, public sector management and social inclusion. Prior to 2004, the CPIA had 20 items in four different categories. In 2004, there was a review of the CPIA that led to the deletion of some items and to the streamlining and combining of others (World Bank 2005) for more information. In recent years, the World Bank has undertaken further reviews of the index, leading to some changes of the process within the Bank, but not in the content of the index. The actual allocation formula followed by the World Bank, which has been adopted by regional development banks with some marginal modifications, is both complex and convoluted. Roughly, allocation of aid per capita for a country is largely based on its ‘performance’ rating, though some weight is also given to its per capita income. The performance rating is derived from the country’s CPIA and portfolio performance scores (the weights being 0.80 and 0.20, respectively). This weighted average is multiplied by a ‘governance factor’, which is essentially derived from the scores of the governance items in the CPIA. As is evident from the allocation formula, the CPIA – in particular, governance factors – drives the allocation process. See World Bank (2004). According to North (1981, 201–202), institutions are as a ‘set of rules, compliances procedures, and moral and ethical behavioral norms designed to constrain the behavior of individuals in the interest of maximizing the wealth or utility of principles’. In other words, the institution is the overarching framework of rules and constraints that regulates the interactions among the individuals. Quibria (2006) argued that the empirical relationship between governance and growth is not as watertight as it is conventionally assumed. Drawing on a set of crosscountry growth regressions for developing Asia, he demonstrates that rapid economic growth in Asia has not necessarily gone hand in hand with superior governance. This point was forcefully made by North (2000). He argued that even if we did have it right for one economy it would not necessarily be right for another economy and even if we have it right today it would not necessarily be right tomorrow … we do know a good deal about the institutional foundations of successful 85 development … . What is still missing is how to get there. The key is the way path dependence will constrain the process of institutional and economic change. In other words, the context matters for institutional innovation. 20. 21. 22. 23. 24. 25. 26. Conditionality in the traditional sense refers to ex ante policy conditionality; i.e. policy and institutional reform conditions attached to loan disbursements by international financial institutions. However, in recent years, there has been a shift in emphasis toward ‘process conditionality’, which links lending to changes in the process. The process now involves the participation of NGOs and local communities in PRSPs. The putative objectives of process conditionality are to foster greater accountability of the government, minimize corruption and inculcate respect for human rights. The operations evaluation department of the World Bank identified four key leadership criteria for country ownership: (1) the locus of initiative must be in the government; (2) key policy-makers must be intellectually convinced; (3) there must be evidence of public support from the top political leadership and (4) there must be broad-based stakeholder participation. Fostering country ownership thus entails extensive consultation between the government and other segments of society, including civil society and the private sector. In addition to eliciting new ideas, knowledge and opinions, this consultation can help to promote a consensus on the strategy. As the definition of ownership is largely subjective, an assessment of ownership has, by necessity, remained subjective. Despite the rhetoric, donors have consistently undermined ownership by maintaining various types of controls over the design and implementation of reform programs. This is contrary to the recommendation of Stiglitz (1999), who argued that the donors’ role should be limited to that of economic advisors who apprise countries of the prevailing views. However, most recipient countries view PRSPs as a vehicle for accessing donor money than a declaration of ownership for various reasons. First, PRSPs continue to be largely donor-driven in countries where the domestic capacity to formulate such a strategy is lacking (Easterly 2006). Second, even where such capacities exist, the PRSP exercise is often an act of ‘ventriloquism’ than an expression of national economic determination (Van de Walle 2005). Recipients write what donors want to hear – they highlight the programs and strategies that the donors favor and likely to fund. Ex post policy conditionality is often coupled with process conditionality, which focuses on participation: openness, transparency and inclusive nature of the polity of recipient countries. Assessment of participation can be subjective and imprecise, because they tend to understate the value of indigenous institutions (such as local government institutions and civil society) vis-a-vis internationally visible NGOs for ensuring transparency and accountability (Barder and Birdsall 2006). In its traditional sense, (ex ante) conditionality is a dual of selectivity. Under conditionality, a country receives aid on the basis of a promise to undertake a stipulated set of policy actions: it entails a set of prior actions before the loan is disbursed. While conditionality relates to ex ante reform, selectivity relates to ex post reform – aid is made available ex post based on the success of the reform. Thus, selectivity excludes countries with well-meaning, enlightened leaders who have the will but not the 86 27. Downloaded by [68.33.167.238] at 07:48 16 August 2014 28. 29. M.G. Quibria institutional capacity to address governance issues (Barder and Birdsall 2006). Inputs refer to the financial, human and material resources used for a development intervention; for example, the budget used for constructing schools or health centers. Outputs refer to products, goods and services that result from a development intervention; for example, the number of schools built or the number of health centers opened. Outcomes refer to intermediate indicators of results, such as the number of students graduated from school and the number of visitors to the health centers. And finally, impact refers to long-term consequences of the intervention; for example, improvements in health and educational indicators. As it is difficult in practice to distinguish between the medium-term outcomes and long-term impacts, they are often lumped together under the heading of outcomes. Barder and Birdsall (2006) recommended a hands-off approach to aid allocation to poorer countries, based on evidence of progress on the ground – measured in terms of outcomes rather than intermediate inputs (as the European experience suggests, there are serious problems with shifting toward intermediate indicators, see Adam et al., 2004). This type of arrangement would afford recipient institutions more flexibility, autonomy and space for institutional experimentation. According to Deaton (2013), the key to understanding aid effectiveness lies in the relationship between aid and politics, as political and legal institutions play a central role in fostering an environment conducive to prosperity and economic growth. References Adam, C. S., and D. L. Bevan. 2006. “Aid and the Supply Side: Public Investment, Export Performance, and Dutch Disease in Low-Income Countries.” World Bank Economic Review 20 (2): 261–290. Adam, C. S., G. Chambas, P. Guillaumont, S. Guillaumont Jeanneney, and J. Willem Gunning. 2004. “PerformanceBased Conditionality: A European Perspective.” World Development 32 (6): 1059–1070. Alesina, A., and D. Dollar. 2000. “Who Gives Aid to Whom and Why?” Journal of Economic Growth 5 (1): 33–63. Alesina, A., S. Ozler, N. Roubini, and P. Swagel. 1996. “Political Instability and Economic Growth.” Journal of Economic Growth 1 (2): 189–211. Barder, O., and N. Birdsall. 2006. “Payments for Progress: A Hands-Off Approach to Foreign Aid.” Working Paper no. 102. Washington, DC: Center for Global Development. Bauer, P. T. 1966. “Foreign Aid: An Instrument for Progress?” The Freeman 16: 25–34. Bauer, P. T. 1984. Reality and Rhetoric: Studies in the Economics of Development. Cambridge, MA: Harvard University Press. Boone, P. 1996. “Politics and the Effectiveness of Foreign Aid.” European Economic Review 40 (2): 289–329. Botchway, K., P. Collier, J. W. Gunning, and K. Hamada. 1998. Report by a Group of Independent Persons: External Evaluation of the ESAF. Washington, DC: IMF. Buiter, W. H. 2004. “Country Ownership: A Term Whose Time Has Gone.” Remarks prepared for the development policy forum, “conditionality revisited”, organized by the World Bank, Paris, France, July 5. Burnside, C., and D. Dollar. 1997. “Aid, Policies, and Growth.” Policy Research Working Paper no. 1777. Washington, DC: World Bank. Burnside, C., and D. Dollar. 1998. “Aid, the Incentive Regime, and Poverty Reduction.” Working Paper no. 1937. Washington, DC: World Bank. Burnside, C., and D. Dollar. 2000. “Aid, Policies and Growth.” American Economic Review 90 (4): 847–868. Burnside, C., and D. Dollar. 2004. “Aid, Policies and Growth: Revisiting Evidence.” Policy Research Working Paper no. 3251. Washington, DC: World Bank. Chauvet, L., and P. Guillaumont. 2004. “Aid and Growth Revisited: Policy, Economic Vulnerability and Political Instability.” In Annual World Bank Conference on Development Economics, Europe, 2003: Toward Pro-Poor Policies – Aid, Institutions and Globalization, edited by N. Stern, B. Tungodden, and I. Kolstad, 95–110. Washington, DC: World Bank. Chenery, H. B., and A. M. Strout. 1966. “Foreign Assistance and Economic Development.” American Economic Review 56 (4): 679–733. Clemens, M. A., Radelet, S., Bhavnani, R. R., and Bazzi, S. 2012. “Counting Chickens When They Hatch: Timing and the Effects of Aid on Growth.” Economic Journal 122 (561): 590–617. Collier, P., and J. Dehn. 2001. “Aid, Shocks, and Growth.” Working Paper no. 2688. Washington, DC: World Bank. Collier, P., and D. 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. Collier, P., and D. Dollar. 2002. “Aid Allocation and Poverty Reduction.” European Economic Review 46 (8): 1475–1500. Collier, P., and A. Hoeffler. 2004a. “Violent Conflict.” In Global Problems: Global Solutions, edited by B. Lomberg, 129– 156. Cambridge: Cambridge University Press. Collier, P., and A. Hoeffler. 2004b. “Greed and Grievance in Civil War.” Oxford Economic Papers 56 (4): 663–695. Collier, P., and A. Hoeffler. 2004c. “Aid, Policy and Growth in Post-Conflict Societies.” European Economic Review 48 (5): 1125–1124. Dalgaard, C., and H. Hansen. 2001. “On Aid, Growth and Good Policies.” Journal of Development Studies 37 (6): 17–41. Dalgaard, C., H. Hansen, and F. Tarp. 2004. “On the Empirics of Foreign Aid and Growth.” Economic Journal 114 (496): 191–216. Deaton, Angus. 2013. The Great Escape: Health, Wealth, and the Origins of Inequality. Princeton, NJ: Princeton University Press. Dixit, A. K. 2004. Lawlessness and Economics. Princeton, NJ: Princeton University Press. Doucouliagos, H., and M. Paldam. 2009. “The Aid Effectiveness Literature: The Sad Results of 40 Years of Research.” Journal of Economic Surveys 23 (3): 433–461. Drazen, A., and S. Fischer. 1997. “Conditionality and Selectivity in Lending by International Financial Institutions.” Paper presented at a symposium in memory of Michael Bruno, Jerusalem, November 23–24. Durbarry, R., N. Gemmell, and D. Greenaway. 1998. “New Evidence on the Impact of Foreign Aid on Economic Growth.” CREDIT Research Paper 98/8, University of Nottingham. Easterly, W. 2002. “The Cartel of Good Intentions: The Problem of Bureaucracy in Foreign Aid.” Journal of Policy Reform 5 (4): 223–250. Easterly, W. 2003. “Can Foreign Aid Buy Growth?” Journal of Economic Perspectives 17 (3): 23–48. Easterly, W. 2006. The White Man’s Burden. New York: The Penguin Press. Easterly, W., R. Levine, and D. M. Roodman. 2004. “New Data, New Doubts: A Comment on Burnside and Dollar’s ‘Aid, Downloaded by [68.33.167.238] at 07:48 16 August 2014 Development Studies Research Policies, and Growth’.” American Economic Review 94 (3): 774–780. Fielding, D., and F. Gibson. 2012. “Aid and Dutch Disease in Sub-Saharan Africa.” WIDER Working Paper 26. Helsinki: World Institute for Development Economics Research. Fosu, A. K. 2002. “Political Instability and Economic Growth: Implications of Coup Events in Sub-Saharan Africa.” American Journal of Economics and Sociology 61 (1): 329–348. Glaeser, E. L., R. La Porta, F. Lopez-de-Silanes, and A. Shleifer. 2004. “Do Institutions Cause Growth?” NBER Working Paper no. 10568. Cambridge, MA: National Bureau of Economic Research. Guillaumont, P., and L. Chauvet. 2001. “Aid and Performance: A Reassessment.” Journal of Development Studies 37 (6): 66–92. Gunning, J. W. 2000. “The Reform of Aid: Conditionality, Selectivity and Ownership.” Paper presented at the SIDA conference on “aid and development”, Stockholm, January 20–21. Haan, J. de, and C. L. J. Siermann. 1996. “Political Instability, Freedom, and Economic Growth: Some Further Evidence.” Economic Development and Cultural Change 44 (2): 339–350. Hadjimichael, M. T., D. Ghura, M. Muhleisen, R. Nord, and E. Ucer. 1995. “Sub-Saharan Africa: Growth, Savings, and Investment, 1986–93.” Occasional Paper no. 118. Washington, DC: International Monetary Fund. Hansen, H., and F. Tarp. 2000. “Aid Effectiveness Disputed.” Journal of International Development 12 (3): 375–398. Hansen, H., and F. Tarp. 2001. “Aid and Growth Regressions.” Journal of Development Economics 64: 547–570. Hudson, J., and P. Mosley. 2001. “Aid Policies and Growth: In Search of the Holy Grail.” Journal of International Development 13 (7): 1023–1038. Kanbur, R. 2006. “The Economics of International Aid.” In Handbook of The Economics of Giving, Reciprocity, and Altruism, edited by S. Christophe-Kolm and J. MercierYthier, Chapter 26, 1559–1587. Amsterdam: North-Holland. Kanbur, R., T. Sandler, and K. Morrison. 1999. The Future of Development Assistance: Common Pools and International Public Goods. Washington, DC: Johns Hopkins Press for the Overseas Development Council. Kaufmann, D., A. Kraay, and P. Zoido-Lobatón. 2002. “Governance Matters II: Updated Governance Indicators for 2000-01.” Policy Research Working Paper no. 2772. Washington, DC: World Bank. Killick, T., R. Gunatilaka, and A. Marr. 1998. Aid and the Political Economy of Policy Change. London: Routledge. Knack, S., and P. Keefer. 1995. “Institutions and Economic Performance: Cross-Country Tests Using Alternative Institutional Measures.” Economics and Politics 7 (3): 207–227. Lensink, R., and H. White. 1999. “Is There an Aid Laffer Curve?” CREDIT Research Paper 99/6. University of Nottingham. Lensink, R., and H. White. 2001. “Are There Negative Returns to Aid?” Journal of Development Studies 37 (6): 42–65. Lucas, R. E. Jr. 1976. “Econometric Policy Evaluation: A Critique.” Carnegie-Rochester Conference Series on Public Policy 1 (1): 19–46. McGillivray, M. 2003. “Aid Effectiveness and Selectivity: Integrating Multiple Objectives into Aid Allocations.” WIDER Discussion Paper no. 2003/71. Helsinki: World Institute for Development Economics Research. Mekasha, T., and F. Tarp. (2011). “Aid and Growth: What MetaAnalysis Reveals.” WIDER Working Paper 22. Helsinki: World Institute for Development Economics Research. Millikan, M. F., and W. W. Rostow. 1957. A Proposal: Key to Effective Foreign Policy. New York: Harper and Row. Mosley, P., J. Harrigan, and J. Toye. 1995. Aid and Power. 2nd ed. London: Routledge. 87 Mosley, P., J. Hudson, and A. Verschoor. 2004. “Aid, Poverty Reduction and the ‘New Conditionality’.” Economic Journal 114 (496): F217–F243. Mukherjee, S. 2010. The Emperor of All Maladies: A Biography of Cancer. New York: Scribner. North, D. C. 1981. Structure and Change in Economic History. New York: Norton & Co. North, D. C. 2000. “Institutions and the Performance of Economies over Time.” Keynote address, second annual global development conference, Tokyo, Japan, December 10–13. Qian, Y. 2003. “How Reform Worked In China.” In In Search of Prosperity: Analytical Narratives on Economic Growth, edited by D. Rodrik, Chapter 11, 297–333. Princeton, NJ: Princeton University Press. Quibria, M. G. 2006. “Does Governance Matter? Yes, No or Maybe: Some Evidence from Developing Asia.” Kyklos 59 (1): 99–114. Quibria, M. G. 2014. “Governance in Developing Asia: Concepts, Measurements, Determinants and Paradoxes.” ERD Working Paper 388. Manila: Asian Development Bank. Rajan, R., and A. Subramanian. 2007. “Does Aid Affect Governance?” American Economic Review 97 (2): 322–327. Rajan, R., and A. Subramanian. 2008. “Aid and Growth: What Does the Cross-Country Evidence Really Show?” The Review of Economics and Statistics 90 (4): 643–665. Rajan, R., and A. Subramanian. 2011. “Aid, Dutch Disease, and Manufacturing Growth.” Journal of Development Economics 94 (1): 106–118. Ranis, G. 2006. “Toward the Enhanced Effectiveness of Aid.” Center Discussion Paper no. 938. New Haven, CT: Economic Growth Center, Yale University. Rodrik, D. 2005. “Growth Strategies.” In Handbook of Economic Growth, edited by P. Aghion and S. Durlauf, Chapter 14, 967–1014. Amsterdam: North-Holland. Roodman, D. M. 2007. “The Anarchy of Numbers: Aid, Development and Cross-Country Empirics.” World Bank Economic Review 21 (2): 255–277. Rosenstein-Rodan, P. N. 1961. “International Aid for Underdeveloped Countries.” Review of Economics and Statistics 43 (2): 107–138. Sachs, J. D. 2005. The End of Poverty: How We Can Make It Happen in Our Lifetime. New York: Penguin Books. Sen, A. 2011. “The Quality of Life: India vs. China.” The New York Review of Books. http://www.nybooks.com/articles/archives/ 2011/may/12/quality-life-india-vs-china/?page=2 Srinivasan, T. N. 2001. “Growth and Poverty Alleviation: Lessons from Development Experience.” Working Paper no. 17. Tokyo: ADB Institute. Stiglitz, J. 1999. “The World Bank at the Millennium.” Economic Journal 109 (454): F577–F597. Temple, J. 1999. “The New Growth Evidence.” Journal of Economic Literature 37 (1): 112–156. Tirole, J. 1992. “Persistence of Corruption.” Institute of Policy Reform Working Paper no. 55. Washington, DC: Institute of Policy Reform. Van de Walle, N. 2005. Overcoming Stagnation in Aid-Dependent Countries. Washington, DC: Center for Global Development. World Bank. 1998. Assessing Aid: What Works, What Doesn’t, and Why. Oxford: Oxford University Press. World Bank. 2004. IDA’s Performance-Based Allocation System: IDA Rating Disclosure and Fine-tuning the Governance Factor. Washington, DC: World Bank. World Bank and International Monetary Fund. 2005. 2005 Review of the PRS Approach: Balancing Accountabilities and Scaling up Results. Washington, DC: World Bank.