This paper considers the interpretation of "Mincer rates of return." We test and reject the Mincer model. It fails to track the time series of true returns. We show how repeated cross section and panel data improves the ability of analysts to estimate the ex ante and ex post marginal rate of...