Jahrbücher f. Nationalökonomie u. Statistik (Lucius & Lucius, Stuttgart 2006) Bd. (Vol.) 226/1 Causal Returns to Education: A Survey on Empirical Evidence for Germany By Anton L. Flossmann and Winfried Pohlmeier, Konstanz∗ JEL C21, J24, J31 Returns to education, potential outcome approach, instrumental variables, unconfoundedness, control function approach. Summary This paper surveys the empirical evidence on causal effects of education on earnings for Germany and compares alternative studies in the light of their underlying identifying assumptions. We work out the different assumptions taken by various studies, which lead to rather different interpretations of the estimated causal effect. In particular, we are interested in the question to what extend causal return estimates are informative regarding educational policy advice. Despite the substantial methodological differences, we have to conclude that the empirical find- ings for Germany are quite robust and do not deviate substantially from each other. This also holds for the few