Factors affecting college attainment and student ability in the U.S. since 1900

Abstract

We develop a dynamic lifecycle model to study long-run changes in college completion and the relative ability of college versus non-college students in the early twentieth century. The model is disciplined in part by constructing a historical time series on real college costs from printed government documents dating to 1916. The model captures nearly all of the increase in attainment and ability sorting between college and non-college individuals between the 1900 to 1950 birth cohorts. Time variation in college costs, the college earnings premium, and the precision of ability signals all play a critical role for explaining different data moments and time periods, primarily through their interaction with binding borrowing constraints. Our quantitative results imply that attainment is broadly driven by the interaction of changing real college costs and the rising earnings premium, while ability sorting is driven by the earnings premium and increasing precision of ability signals.