"Development assistance began in earnest in the 1960s. Soon, thereafter, it became a widely researched topic, and the evaluation of aid as a vehicle for poverty reduction continues today. The AEL is huge and important. Recently, three meta-analyses have reviewed the accumulated research results from this literature –Doucouliagos and Paldam (2006, 2007a, 2008). Together, these meta-analyses reach the sad result that aid has failed in its primary mission. The purposes of this paper are to survey the AEL, to identify and discuss patterns in this literature and to review the findings from the three meta-analyses. Moreover, we explore the process through which research is conducted in the AEL.
A careful analysis of the AEL reveals a highly significant reluctance bias. If an AEL researcher finds several results he/she will be reluctant to report those that suggest that aid causes harm. Rather, the most significantly positive result is likely to be selected as the key finding for a given aid effectiveness study. Considering the issues, this is not surprising, but it is an impediment to uncovering the true effects of aid.
When this tendency is combined with the widespread practice of polishing one's findings so that the published results are clear and statistically significant, research can fail to converge for decades. It is remarkable that our meta-analyses could find no evidence of aid effectiveness, even though 74% of the published aid-growth effects are positive.
The reader may be wondering how common these publication selection biases are in empirical economics in general. Obviously, the AEL deals with a highly emotional subject, where special interests align with everyone's desire to do ‘good’. Nonetheless, it is also common for meta-analyses to detect asymmetries in funnel plots, reflecting a similar selection of economic results. Dozens of areas of empirical economics have, by now, been investigated for signs of publication selection, and the majority of these contain evidence of substantial or severe biases (Doucouliagos and Stanley, 2007). Meta-analysis is a very useful tool and provides a unique perspective on empirical literatures. It allows us to assess if the results of a literature are converging, whether results suffer from publication selection, and whether there exists an underlying genuine empirical effect.
Our three meta-analyses of the AEL have failed to find evidence of a significantly positive effect of aid. Consequently, if there is an effect, it must be small. Development aid is an activity that has proved difficult to do right. Even though such a negative conclusion may cost some support for aid in the short run, it may ultimately prevent aid fatigue in the longer run. Our findings underscore the need to find more effective ways to employ development aid."
ansonsten mal kleine Donations für Software-Entwicklungen
