The lax underwriting in non-prime mortgage markets is widely perceived as one cause of the recent difficulties in the housing market. Policymakers are currently considering moves such as enforcing more careful underwriting to provide additional discipline to mortgage markets. This research explores the possibility of another approach to supplement or replace some of these efforts, namely the use of policy to create incentives for Fannie Mae and Freddie Mac (together, the GSEs) to help "check" behavior in non-prime markets. The hypothesis is that the GSE Act affordable housing goals have increased GSE focus on targeted loan purchases, which in turn has led prime market lenders to compete more aggressively for borrowers on the margin between prime and subprime credit quality. As a consequence, these marginal borrowers will be more inclined to take prime mortgages rather than higher-cost subprime loans. We test this hypothesis and find empirical support for it. We observe a negative relationship between the growth in GSE market share and the growth in subprime market share over time, and find that the impact of the GSEs on subprime lending tends to be stronger in high-minority neighborhoods, where subprime lending has been concentrated and growing the fastest. Simulations show that a 10 percent increase in GSE market share (for example, from 20 to 22 percent) can cause 45,000 borrowers using prime instead of subprime loans a cost savings of about $1.7 billion. These results suggest that the GSEs, regardless of their post-conservatorship form, should continue to devote attention to serving underserved populations and suggest that significant welfare benefits will accrue. (Contains 3 figures, 8 tables, and 24 footnotes.)