Default Risk on Home Mortgages: A Further Test of Competing Hypotheses

Abstract

The test of these hypotheses is based on the estimation of a mortgage default equation and the results support the equity model of default. JK use this finding to argue that an efficiently designed homeownership subsidy program should provide for a subsidization of borrower equity rather than borrower monthly payments. It is then pointed out that this is exactly opposite to the subsidy pattern used under the FHA Section 235 homeownership program that was operative during the late 1960s and early 1970s. The purpose of this comment is to suggest that the empirical results which JK offer in support of the equity model of default are subject to question. It is, therefore, unwise to base policy recommendations on one particular view of the mortgage default process. The implied criticism of previous government homeownership programs on these grounds may also be misplaced. ' Consider the coefficients of the variables listed under location characteristics in Table 1. From the estimated coefficients of the SMSA variables, we find that an increase in the rate of new starts; an increase in the fraction of pre- 1940 housing; a decrease in city size; higher rates of population growth; and higher income per capita all significantly reduce expected default rates. Location in the city center results in significantly higher expected default than location in the suburban or rural areas. Finally, location in a neighborhood with blighted structures increases the estimated probability of default. These effects could reflect the interaction of housing