and seminar participants at several universities and institutions for additional comments. All remaining errors are mine. Abstract Estimating the changes in the variance of wages features in two intensively researched areas of economics, one examining changes in residual inequality, the other investigating the link between college attendance and wage uncertainty. Both literatures typically neglect the effects of endogenous schooling choices and unobserved heterogeneity on observed changes in

more »

... iance. This paper develops a tractable empirical approach to both issues. The estimates suggest that selection bias matters for variance comparisons; accounting for selection can double or triple the variance differential between high school and college. The approach also allows residual inequality to be decomposed into two sources -unobserved heterogeneity (known to the agent making the schooling choice) and wage uncertainty (unknown to the agent). Results indicate that heterogeneity accounts for more of residual inequality for college attendees than for high school graduates, suggesting that previous estimates may exaggerate the deterrent effect of uncertainty on schooling choices.