A study of mortgage underwriting criteria and urban mortgage credit availability

Helen Ahada Stanford, University of Pennsylvania


This study investigates the hypothesis that mortgage underwriting criteria may be relaxed to lend to first-time, lower-income, urban households--Philadelphia--without significantly increasing risk. The study considers several private- and public-sector policy issues relevant to mortgage underwriting criteria. First, placed in the context of homeownership markets: the decline of homeownership; the impact of affordability of homeownership for income- and capital-constrained households; and the changing demographics of particularly the potential first-time homebuyer. A related policy issue is the transformation of mortgage markets that may further discourage certain potential homebuyers: the standardization of mortgage underwriting, encouraged by the growth of the secondary market; the changing role of savings and loan institutions; and the unbundling of financial services by many institutions, which creates distance between the lending institution and the household. Finally, the study discusses community reinvestment and lending without regard to race, ethnicity, or sex, and to lower-income households in the inner-city. The purpose of the study is to examine the viability of relaxed criteria loans by comparing ex ante underwriting--the determinants of loan-to-value--with ex post realizations--delinquency, default, and foreclosure. The analysis is performed with a data set that has not previously been studied--Pennsylvania Housing Finance Agency loans, funded by mortgage revenue bonds. The logistic regression ex ante results show that the probability of the borrower household receiving a high loan-to-value product increases in the following situations: a first-time homebuyer, a member of an ethnic minority, holders of high consumer debt, and the purchasers of row or townhouses in a low-to-moderate-income and minority area. The ex post default model results--default defined as loans greater than 60 days overdue--showed that first-time buyers and households of African American and Hispanic origin was not statistically significant to the event of default. The overall predicted probability of default was 45.4 percent. Changing the households to first-time owners, decreasing the mean income from $28,000 to \$20,000, purchasing a row or townhouse in a low-to-moderate-income area, resulted in decreasing the predicted probability of default to 40.5 percent. This and other study estimates suggest that it is feasible to shape loan products for the populations of concern, and in so doing, decrease the probability of certain risk exposures.

Subject Area

Banking|Finance|Urban planning|Area planning & development

Recommended Citation

Stanford, Helen Ahada, "A study of mortgage underwriting criteria and urban mortgage credit availability" (1995). Dissertations available from ProQuest. AAI9543148.