Fifty Years of Mincer Earnings Regressions

Abstract

The Mincer earnings function is the cornerstone of a large literature in empirical economics.This paper discusses the theoretical foundations of the Mincer model and examines theempirical support for it using data from Decennial Censuses and Current Population surveys.While data from the 1940 and 1950 Censuses provide some support for Mincer’s model, datafrom later decades are inconsistent with it. We examine the importance of relaxing functionalform assumptions in estimating internal rates of return to schooling and in accounting fortaxes, tuition, nonlinearity in schooling, and nonseparability between schooling and workexperience. Inferences about trends in rates of return to high school and college obtainedfrom our more general model differ substantially from inferences drawn from estimates basedon a Mincer earnings regression. Important differences also arise between cohort-based andcross-sectional estimates of the rate of return to school. In the recent period of rapidtechnological change, widely used cross-sectional applications of the Mincer model producedramatically biased estimates of cohort returns to schooling. We also examine theimplications of accounting for uncertainty and agent expectation formation. Even when thestatic framework of Mincer is maintained, accounting for uncertainty substantially affects rateof return estimates. Considering the sequential resolution of uncertainty over time in adynamic setting gives rise to option values, which fundamentally changes the analysis ofschooling decisions. In the presence of sequential resolution of uncertainty and optionvalues, the internal rate of return - a cornerstone of classical human capital theory - is not auseful guide to policy analysis.

Keywords: returns to schooling, Mincer model, earnings regressions

JEL Classification: C31