Mortgage default rates and borrower race

Journal of Real Estate Research, The,  Sep/Oct 1999  by Anderson, Richard,  VanderHoff, James

• E-mail
• Print
• Link

Conclusion

Our analysis of conventional mortgage data confirms the results of previous analyses of FHA mortgages: Black households have higher marginal default rates. Further we find no evidence of higher profitability on loans to Black borrowers but find evidence of lower equity for Black borrowers. These results are not consistent with racial discrimination in mortgage lending. We conclude further research on costs associated with default and with trigger events is needed to understand why minority borrowers have higher default rates.

The authors thank Michael Grossman for valuable comments. Also, Ron Anderson and an anonymous referee provided suggestions that substantially improved this article. VanderHoff thanks the Morris Beck Foundation and the Rutgers University Research Council for support.

Endnotes

1 Discrimination refers to "uneconomical discrimination": discrimination that causes expected profitability of loans made to one group to exceed that of loans made to other groups. This usage does not include economic discrimination, which lenders use to distinguish between good and bad risks. See Peterson (1981) and Becker (1993).

1 We use default to indicate the property is claimed by the lender, not the technical definition that a mortgage is delinquent in payments.

3 These econometric critiques are made by several researchers. For examples, see Brueckner (1994a, b), Cappozza (1994), Ross (1994, 1998) and Yezer, Phillips and Trost (1994) and other articles in the special issues of Cityscape (February, 1994) and the Journal of Real Estate Finance and Economics (November, 1994) devoted to discrimination in real estate finance.

1 See BCGH (I 994a, b) and Brueckner (I 994a).

5 Several studies of default rates on conventional mortgages have been published but these do not include the borrower race in the model (see Phillips, Rosenblatt and VanderHoff, 1995).

6 We analyze Black default because preliminary analysis indicated that other minority borrowers had a small percentage of loans and did not default at a rate statistically different from White borrowers. Also, the literature has predominately focused on Black default rates. See the articles in Cityscape (February, 1994) and the Journal of Real Estate Finance and Economics (November, 1994).

7 The value of options depends on interest rates, the variability of house prices and interest rates (see Kau, Keenan and Kim, 1993).

8 For example, racial discrimination in employment markets could lessen default costs associated with restrictions on employment opportunities.

9 When the omitted variables are correlated with race, the coefficient on the race variable will be biased towards zero, indicating no racial difference in default rates.

11 The rent free occupancy of the house during the foreclosure proceeding is included as a negative transaction cost.

11 The standard errors are biased in the linear model due to heteroskedasticity and predicted value of the default probability may lie outside the (0,I) interval problems (see Greene, 1993). BCGH (1994a, b) discuss the effect of dependents on default probability.