Consider a labor market for men and women with two types of workers in each group, high ability $($H types$),$ and low ability $($L types$).$ H types among men and women have a productivity of $70$ and L types among men and women have a productivity of $30.$ Among women $2/3$rds are H types, and $1/3$rd are L types, and among men $12$ are H types, and $12$ L types. Suppose affirmative action for women reduces the cost of college for all women to $10.$ What happens to the average wage gaps between men and women? Recall that employers know the shares of H and L types, but cannot observe individual types.