Partial Identification of the Return to Schooling

Juergen Meinecke | Australian National University

Although the return to schooling has been studied (too) many times, it is not, ostensibly, a well-defined object. In this paper, I present an minimalist discrete dynamic programming model in which a person chooses his optimal schooling level. Out of this model grows a specific definition of the return to schooling. Also, the model results in a few predictions about the monotone comparative statics of the optimal schooling choice. I then connect the economic model to an econometric partial identification model. I demonstrate how the return to education is partially identified given a mainstream survey data set that includes information on schooling and earnings. In the end, I estimate an informative upper bound on the return to schooling.