Quantile Treatment Effects of College Quality on Earnings: Evidence from Administrative Data in Texas

This paper uses administrative data on schooling and earnings from Texas to estimate the effect of college quality on the distribution of earnings. We proxy for college quality using the college sector from which students graduate and focus on identifying how graduating from UT-Austin, Texas A&M or a community college affects the distribution of earnings relative to graduating from a non-flagship university in Texas. Our methodological approach uses the rich set of observable student academic ability and background characteristics in the data to adjust the earnings distributions across college sectors for the fact that college sector quality is correlated with factors that also affect earnings. Although our mean earnings estimates are similar to previous work in this area, we find evidence of substantial heterogeneity in the returns to college quality. At UT-Austin, the returns increase across the earnings distribution, while at Texas A&M they tend to decline with one’s place in the distribution. For community college graduates, the returns relative to non-flagship four-year graduates are negative across most of the distribution of earnings, but they approach zero and become positive for higher earners. Our data also allow us to estimate effects separately by race and ethnicity, and we find that historically under-represented minorities experience the highest returns in the upper tails of the earnings distribution, particularly among UT-Austin and community college graduates. While we focus on graduates, we also show our estimates are robust to examining college attendees as well as to many other changes in the sample and to the estimation strategy. Overall, these estimates provide the first direct evidence of the extent of heterogeneity in the effect of college quality on subsequent earnings, and our estimates point to the need to consider such heterogeneity in human capital models that incorporate college quality.