Impact of Foreclosure Laws on Mortgage Loan Supply and Performance

Agarwal et al. (2003) look at the effects of exemption law on consumer bankruptcy; Agarwal et al. (2005) focus on the impact of state exemption laws on small business bankruptcy. Their regression models of consumer/small business finance situation on state exemption laws do not control for regional factors (include only unemployment rate).

Deficiency judgment grants lenders stronger bargaining power to extract concessions from borrowers in delinquency.

Foreclosure laws changed over time. Table 1 shows the foreclosure during our sample period 1999–2012. For example, Hawaii moved away from a power of sale state to a judicial foreclosure requirement state from the end of 2012. We thank the suggestions from anonymous referee.

Holmes (1998) uses the geography discontinuity approach to empirically test the right-to-work laws in different states on business activity. Ho and Pennington-Cross (2006) apply a similar cross-state identification strategy to test the effects of the anti-predatory lending laws on credit flows.

Collins et al. (2011) investigate the impact of foreclosure laws on loan outcomes using the MSA-level loan data covering 22 MSAs in the US. Mian et al. (2014) and Gerardi et al. (2013) use the 10-mile zip code border shared by states as the border identification strategy to delineate states with different foreclosure laws.

Lenders can also charge different fees or set different interests with different law environments. Our paper focuses on the loan supply. We thank the suggestions from anonymous referee.

Details describing loan selection criteria and data overview are documented by Freddie Mac in “Single Family Loan-Level Dataset: General User Guide”.

Adjustable Rate Mortgages (ARMs), initial interest, balloons, any mortgages with step rates, Relief Refinance Mortgages, Federal Housing Administration/Veterans Affairs (FHA/VA) Guaranteed Rural Housing (GRH), and HUD-Guaranteed Native-American mortgages (Section 184) are excluded. First, we consider only fixed rate mortgages to keep interest rate risks relatively exogenous in the models. Second, the government sponsored mortgages often enjoy special exemptions or concessions, (for example, deficiency judgments are prohibited on FHA loans and strongly discouraged on VA loans). The inclusion of exotic loans, and/or loans with concessionary terms could further distort the results, and could create more endogeneity to the empirical results.

The Freddie Mac data report only the first three digits of the five-digit postal code of the mortgaged property. Therefore, we refer the zip codes hereinafter in this paper to the three-digit zip codes.

Our paper focuses on impact of foreclosure laws under different market environments. The credit market has significantly changed after 2008, especially when Lehman Brother collapsed on September 15, 2008, followed by Freddie and Fannie go into the conservatorship of soon after that. We thank the suggestions from anonymous referee.

The 90-day delinquent measure is commonly used to identify loan default (see e.g., Quercia et al. 2014 and Agarwal et al. 2012). A default threshold based on 60-day delinquent has also been used as an alternative default measure in the prior research (Keys et al. 2009; Keys et al. 2010; Jiang et al. 2014; Conklin 2017). We use both default measures in our study, and the results remain largely unchanged.

The aggregate loan observations in the three foreclosure states are 3,713,934 (judicial states), 8,814,319 (right of redemption states), 5,415,778 (“non-recourse states), respectively, and the aggregate nationwide sample is 15,626,550.

Appendix 3 presents detailed description for each sample.

See Cox (1972), Cox (1975), Deng et al. (1996), Deng (1997) and Clapp et al. (2006) for more discussions regarding the estimation of the cox proportional model.

Loans with borrower’s FICO score lower than 500 are excluded.

The calculation details (standards) are documented in Single Family Loan-Level Dataset: General User Guide by Freddie Mac.

A borrower is defined as a first-time homebuyer, if he (she) is an individual who purchases and resides in a mortgaged property as a primary residence, and has no ownership interest (sole or joint) in another residential property during the three-year period preceding the date of the purchase of the mortgaged property (See Single Family Loan-Level Dataset: General User Guide by Freddie Mac).

Further information of the calculation of default option is presented in Appendix 2.

Due to space constraint, we only report the regression results of the impact of judicial foreclosure on loan amount over the full period: 1999–2012 (Table 6). Results of the impacts of the other two legal variables on loan amount for the sub-periods are summarized in Tables 7 and 8.

There are 3,546,396, 7,294,756, 4,613,248, and 12,317,399 loan-month observations in the “judicial” sample, the “right of redemption” sample, the “non-recourse” sample, and the nationwide sample, respectively.

Cutts and Merrill (2008) provide detailed discussions of the “foreclosure days” calculation. A breakdown of timeline of the foreclosure by state is attached in Appendix 1.

State-level foreclosure days are included in the judicial process analyses.

The empirical results in the judicial analyses (unreported) show that foreclosure days do not affect borrower’s default decisions.

We replicate the robustness tests using the subsamples of non-repurchased single-family home loans and find the results that are similar to those presented in Tables 18 and 19. The results are not reported in the paper due to space constraint.