The Returns to College Education: A Reassessment of Heterogeneous Treatment Effects

Stefanie Lightner, University of Minnesota
Andrew Halpern-Manners, University of Minnesota

With the continued growth in demand for college-educated labor and accompanying expansion in the share of students who expect to attend a four-year college, we reassess the evidence on the uniformity in returns to college attendance. Neoclassical economic arguments imply that the returns to college vary as a function of the likelihood of successfully completing a degree; those who are better prepared academically to meet the challenges of college are also more likely to enjoy the rewards of finishing. In contrast, Brand and Xie show that the benefits of a college education are generally largest among those who are the least likely to graduate. This finding is based on a model of heterogeneous treatment effects that requires strong assumptions about the “ignorability” of various pre-treatment variables, including non-cognitive skills. In this paper, we subject these assumptions to a series of robustness checks using data from the National Longitudinal Survey of Youth, 1979 and High School and Beyond.