The Fixed-Effects Model in Returns to Schooling and Its Application to Community Colleges: A Methodological Note
By: Susan Dynarski, Brian Jacob, & Daniel Kreisman | December 2016
The purpose of this note is to develop insight into the performance of the individual fixed-effects model when used to estimate wage returns to postsecondary schooling. The authors focus attention on the returns to attending and completing community college. While other methods (instrumental variables, regression discontinuity) have been used to estimate the returns to four-year colleges, the individual fixed-effects estimator has primarily been employed in the analysis of community colleges. Using data from Michigan, the authors test the common-trends assumption that underlies the individual fixed-effects estimation strategy. They find that the common-trends assumption is violated in their sample. The authors suggest an estimation strategy that allows for individual-specific pre-trends in earnings. Including these trends substantially alters the estimated returns.