The paper investigates the existence of "super pro-poor" policies-that is, policies that directly influence the income of the poor after accounting for the effect of growth. It uses a dynamic panel estimator to capture both across- and within-country effects, and a Bayesian-type robustness check...