alternative interpretation of the evidence

The present paper by Finn Tarp is, in his own words, mainly a review of issues that have been in focus in the foreign aid literature, including articles penned by himself and collaborators. The paper is long and rich, and I will restrict my comments to the section (four) on the aid-growth nexus where he discusses estimation problems and findings in recent contributions. The characteristics of the more advanced studies based on cross-country regression analysis and panel data are that: • Other variables than aid that may affect growth are controlled in the regressions along the lines of recent empirical growth studies. • Interdependence between the growth regressors (multicolinearity) is controlled. • Simultaneity, i.e. the possibility that recipient country growth af-fects the allocation of aid (reverse causality), is controlled. • The possibility that aid has different effects in different recipient country environments is considered (e.g. through the use of fixed-effect estimations or dummy variables). • Robustness tests are carried out, i.e. checks of whether results are sensitive to small changes in the estimation specification, variable definitions, proxy variables used, countries included, time period covered, etc. 1. What does the evidence show? In the paper, Tarp argues that, “Overall, the view that aid works in promoting growth and development has gained ground in recent ye-ars in the academic literature…. ” (p. X) and that “the single most common result in modern aid-growth literature is that aid has a posi-tive impact on per-capita growth ” (p. X). I am inclined to take issue with these claims. My assessment of the literature leads me to conclude that none of the recent (or earlier) stu