Federal mandates and mortgage supply : regression discontinuity analyses of the community reinvestment and GSE Acts.

Abstract

In this dissertation, I provide evidence of the causal impact on mortgage supply of the Community Reinvestment Act (CRA) and the "Government-Sponsored Enterprises (GSE) Act", laws requiring banks and the GSEs (Fannie Mae and Freddie Mac), respectively, to help improve credit access for low-income households and neighborhoods. While financial markets evolved rapidly since the early 1990's, I use discontinuities in the laws' eligibility rules to identify their effects. To implement the analyses, I use a census of mortgage applications collected under the Home Mortgage Disclosure Act. Overall, these programs appear to have had limited impact. I first analyze CRA's effect on mortgage lending in targeted neighborhoods: census tracts with a median family income (MFI) under 80% of MSA MFI. The regression discontinuity (RD) estimates suggest an overall credit supply shift of at least $6 billion ($2007) from 1994 and 2002 in targeted neighborhoods. In addition to CRA's direct effect on bank lending, I also find that unregulated institutions lend more in targeted tracts ("crowd-in"). Further analysis suggests that information spillovers from increased bank lending helps generate crowd-in. In Chapter 2, I examine CRA's effect on home purchase mortgage lending to households with income under 80% of the MSA MFI. In both Chapters 1 and 2, I find CRA's impact is concentrated in the largest MSAs, where enforcement is most intense. The RD estimates indicate that CRA caused a 6% increase in large MSA bank home purchase lending at the cutoff.

(cont.) Unlike in Chapter 1, there is no theoretical basis for crowd-in and none is found. Nor do I find that banks crowd-out unregulated institutions. Finally, I measure the impact of one of the three goals established under the GSE Act. Under this goal the GSEs target census tracts with MFI under 90% of MSA MFI. The RD estimates suggest this goal led to a 3-4% increase in GSE purchases, and increased GSE-eligible originations by 2-3% at the cutoff. Unlike previous research, I find no evidence that the GSEs crowd-out FHA and subprime loans. The results imply a lower bound of the goal's impact of $2.4 billion between 1997 and 2002.