The Value of a Private Education: Differential Returns and Selection on Observables

The academic value of a private education versus a public education is explored. This study first attempts to see whether there are differential returns to a private education based on ability level. It also intends to non-parametrically control for the selection on observables of the decision to attend private school. Using 8th graders from the National Educational Longitudinal Study of 1988, we find that the there is a positive effect on standardized math test scores using OLS. Based on quantile regression results, average students benefit most from a private education. Low-achieving and high-achieving students do not benefit as much. The particularly low return to a private education for high achievers suggests most of them would succeed at any school. Average treatment effects on the treated (ATET) are twice as large and average treatment effects (ATE) are more than three times as large as ordinary least squares estimates. These results suggest that perhaps private schools are actually not selecting the best students, but rather selecting average students and adding value to them.