While disentangling the returns to schooling from the returns to unobserved ability is difficult, estimating the returns to college quality is harder still. No good instruments have been found for college quality, and the sample sizes of twins are often too small to obtain accurate estimates of the returns to college quality. 6 A notable exception is Berhman et al. (1996), who find that, after controlling for family background characteristics using twins, there are significant returns to attending colleges of higher “quality” in several observable dimensions. By using data on pre- MBA wages, we are able to distinguish how the average effect on wages varies across the.
