God, bad, or ugly: Does it really matter?
- Each year, donor countries spend billions of Euros on development cooperation. Not surprisingly, a large strand of research has emerged which examines the impact of development cooperation. A sub-discipline within this strand of the literature deals with the question of whether the impact or effectiveness of development cooperation depends on the quality of the recipient country's policy and institutional environment. Over hundreds of studies have assessed this question at the macro level. In so doing, most of these studies test whether a potential effect of aid on the growth of a recipient country’s gross domestic product (GDP) is conditional on the country's policy and institutional environment. However, even after decades of research and hundreds of studies, no conclusive result has been found. One of the main reasons for the inconclusive state of the literature is that most macro-level studies have to deal with a high risk of endogeneity, treat aid as nothing but a pure income transfer, and rely on low-quality GDP data. To solveEach year, donor countries spend billions of Euros on development cooperation. Not surprisingly, a large strand of research has emerged which examines the impact of development cooperation. A sub-discipline within this strand of the literature deals with the question of whether the impact or effectiveness of development cooperation depends on the quality of the recipient country's policy and institutional environment. Over hundreds of studies have assessed this question at the macro level. In so doing, most of these studies test whether a potential effect of aid on the growth of a recipient country’s gross domestic product (GDP) is conditional on the country's policy and institutional environment. However, even after decades of research and hundreds of studies, no conclusive result has been found. One of the main reasons for the inconclusive state of the literature is that most macro-level studies have to deal with a high risk of endogeneity, treat aid as nothing but a pure income transfer, and rely on low-quality GDP data. To solve these three methodical issues, some authors have started to change the analytical focus from the macro to the micro level. Thus, these authors assess the determinants for the performance of individual development projects instead of the determinants for an effect of aid on GDP. Yet, even though the number of studies focusing on the micro level has increased steadily over the last few years, the state of the literature on the determinants for the performance of development projects still contains multiple highly relevant research gaps. The present thesis seeks to address three of these research gaps. The first research gap addressed by this thesis is related to the specific type of development cooperation. So far, nearly all existing studies focus on projects by Multilateral Development Banks. Research on the determinants for the performance of bilateral development projects is still rare. Thus, even though donors pledge to implement effective development projects, there are hardly any micro-level studies on bilateral projects. So far, only three studies use a sample which includes bilateral projects. Yet, none of the three studies assess the determinants for the performance of bilateral technical development projects. The first paper in the present thesis (GIZ paper) seeks to address this research gap by assessing the determinants for the performance of projects by the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), a bilateral state-owned aid agency active in the area of technical cooperation. The results of the paper indicate that some but not all of the existing theoretical arguments can be extended to bilateral technical projects as well.. For example, the level of market interventions in the recipient county only affects the performance of financial development projects, while the recipient country’s government capacity affects both technical and financial development projects. The paper also indicates that effects of determinants may vary among project sectors. The paper also highlights a dilemma of technical development cooperation. The countries with low government capacity are usually the ones most in need of technical cooperation projects. But, at the same time, they are also the countries in which these projects have the poorest performance The second research gap addressed by this thesis is related to one specific factor in the policy and institutional environment of recipient countries, namely corruption. This determinant is often cited as essential for project performance but has gained surprisingly little coverage in empirical studies. The few existing studies on the effect of corruption on project performance are inconclusive. Some find a statistically significant correlation, while others do not. Furthermore, so far, all existing studies use corruption perception indices as a measurement for corruption, despite the fact that these indices have well-known deficits when it comes to this research topic. One of these deficits is that such indices do not distinguish between different forms of corruption, even though it is likely that the effect of corruption will vary depending on the type of development project and form of corruption. The second paper in this thesis (Corruption paper) seeks to address this inconclusive state of the research while focusing on one specific form of corruption, namely bribery between private firms and public officials. The paper finds a small but statistically significant correlation between the corruption level and the performance of World Bank projects. The systematic effect of corruption on project performance confirms the need to consider the risk of corruption in the design and during the implementation of projects. Nonetheless, the relatively small effect of corruption and the low pseudo R-squareds advise not to overestimate the relevance of corruption for project performance. At least for the project level, the paper finds no indication that corruption is a primary obstacle to aid effectiveness. The third research gap addressed by this thesis is related to one specific sample, namely recipient countries of the International Development Association (IDA). The question of whether the policy and institutional environment affects project performance is of particular relevance for these countries, given that the World Bank's ratings on a country's policy and institutional environment decide how much IDA resources it receives. One core justification of such an allocation system is that it helps to steer more resources to places where they are most effective. However, so far, there is no conclusive empirical evidence for this statement. The only study specifically focusing on this topic, a study by the Independent Evaluation Group of the World Bank from 2010, has essential methodological limitations. The third paper of this thesis (CPR paper) seeks to address this research gap by testing whether a more refined analysis confirms the assumption of previous studies that the policy and institutional environment of IDA-recipient countries, measured by the Country Policy and Institutional Assessment ratings, has an effect on the performance of World Bank projects. Overall, neither the main regression models nor any of the robustness tests indicate a substantial correlation between the policy and institutional environment and project performance. Only for Investments Loans is the coefficient large enough to assume some effect. The overall results not only contradict the results of previous studies, but also raise strong doubts around one of the core justifications for the allocation system of the IDA. All three papers rely on a statistical large-N analysis of the performance ratings of individual development projects. These ratings are usually assigned based on the final evaluation of a project and indicate the merit or worth of an activity. The merit or worth of an activity itself is measured by criteria like relevance, effectiveness, and efficiency. In the case of the two papers on World Bank projects, the needed data stem from different databases of the World Bank. The relevant data for the GIZ paper are gathered from internal evaluation reports of the GIZ. Logistic regressions are applied as the main analytical tool. Overall, the three papers show that the policy and institutional environment of recipient countries matters for project performance, but only to a small degree and under certain circumstances. This result highlights that many researchers and practitioners tend to overestimate the role that the policy and institutional environment of recipient countries plays in project performance. Furthermore, the thesis shows that authors of future studies should consider possible interactions between project- and country-level determinants whenever possible, both in their theoretical arguments and statistical models. Otherwise, the debate on the determinants for project performance is at risk of degenerating into a statistics tournament without any connection to reality.…