Mortgage default rates and borrower race

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

• E-mail
• Print
• Link

The origination data is more fully described in Phillips and VanderHoff (1994) and the payments data is more fully described in Phillips, Rosenblatt and VanderHoff (1995). A relatively small number of loans, about 1100, were in both data sets and include data on borrower characteristics, These data include loans from twenty states.

"The data set also included information on adjustable rate mortgages (ARMs) but these loan are not included in the analysis. The FRM and ARM default models are significantly different and the ARM model did not identify borrower race as a factor affecting default probability. This result may stem form the low number of minority borrowers who choose the ARM.

11 Loans are classified as a default if payments and loan balances are not paid. A loan in which payments were continued after a period of nonpayment would not be classified as a default.

16 This use of annual observations to analyze the probability of default is an application of discrete time methods, as discussed by Allison (1982).

" Because EQUITY is measured when payments were stopped not when the deed to the property was transferred, its estimated coefficient provides insights into the reasons for initiating the process that leads to default not the conditions that exist at loan origination or the end of the foreclosure process,

19 This index created by Haurin, Hendershott and Kim (1991) combines house price data from the American Chamber of Commerce, Coldwell Banker and the National Association of Realtors.

11 We estimate LTV as a function of borrower characteristics. This estimate and house price provide an estimate of initial loan amount.

21 The probit model estimates do not indicate the magnitude of the effect of a variable on the probability. Analysis of the linear model indicates that the additional variables reduce the estimated effect of BLACK on default probability by 12%.

21 We use actual values of EQUITY, not the predicted value used in the default model.

22 We recognize that loan profitability depends not only on equity but on the recovery rate on the equity and on collection costs. However, our data does not include measures of the factors required to make a more exact measure of profitability.

References

Allison, P. D., Discrete-Time Methods for the Analysis of the Event Histories, In S. Leienhardt (ed.), Sociological Methodology, San Francisco: Jossey-Bass, 1982.

Ambrose, B. W. and C. A. Capone, Jr., Do Lenders Discriminate in Processing Defaults?, Cityscape, 1994, 2:1, 89-98.

Berkovec, J. A., G. B. Canner, S. A. Gabriel and T. H. Hannan, Race, Redlining, and Residential Mortgage Loan Performance, Journal of Real Estate Finance and Economics, 1994a, 9:3, 26394.

-., Mortgage Discrimination and FHA Performance, Cityscape, 1994b, 2:1, 9-24.

Becker, G. S., Nobel Lecture: The Economic way of Thinking, Journal of Political Economy, 1993, 101, 385-409.

Brueckner, J. K., Unobservable Default Propensities, Optimal Leverage and Empirical Default Models: Comments on 'Bias in Estimates of Discrimination and Default in Mortgage Lending: The Effects of Simultaneity and Self-Selection,' Journal of Real Estate Finance and Economics, 1994a, 9:3, 217-22.