Doctoral Dissertations

Date of Award

12-2000

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Economics

Major Professor

Matthew N. Murray

Abstract

This study analyses the question of gender discrimination in mortgage lending. The federal fair lending regulations prohibit discrimination in granting credit with respect to race, gender, marital status, color, religion, age or receipt of public assistance. If discrimination exists in the mortgage market, it will keep creditworthy applicants from accessing home ownership, which represents the principal mean of capital and wealth accumulation.

During the last two decades, studies regarding the problem of discrimination in mortgage credit have principally focused on the race issue. Race appeared to have, in most empirical studies, a significant impact on the outcome of mortgage application, with, in every instance, higher rejection rates for minorities than for non-minorities. Very few studies found interest in factors other than race affecting the distribution of mortgage loans. One of the variables mostly ignored in the analysis of discrimination in mortgage lending seems to be the one related to gender bias, raising the question whether sex discrimination in the mortgage lending market no longer represents a significant problem.

The data used to examine the impact of gender on the mortgage lending market in this study were obtained from the 1996 Home Mortgage Disclosure Act (HMDA). Mortgage applications and outcomes in six Metropolitan Statistical Areas (MSAs) (Atlanta, GA; Austin TX; Memphis, TN; Boston, MA; Chicago, IL; and New York, NY) were analyzed, using a model of mortgage lending incorporating applicant and loan characteristics available in HMDA data. The study undertook both an MSA and a cross regional comparison (South-North), in order to account for socio-economic and cultural differences across MSAs and across regions.

Due to some limitations of the HMDA data, particularly the unavailability of information about the applicant's credit history, this study used a particular sampling method, the matched-pair method, similar but somewhat different from the one used by the Federal Reserve System. This statistical sampling method allowed the obtaining of exact matches of male and female applicants in terms of income levels and loan amounts requested. The results of probit regressions on the matched-pairs data sets were compared to those obtained using unmatched data sets in order to assess whether close matching of male and female applicants allows a better use of HMDA data as an instrument for fair lending regulations screenings.

The comparative analysis of these results suggested that the matching process makes a sensible difference in the gender variable's ability to predict mortgage lenders' action. The empirical results indicated that once male and female applicants are exactly matched (in terms of income and loan amount requested), for any income group, little differentiation in the outcome of their mortgage loan application would be linked to gender.

Moreover, the findings suggested that variables such as race, loan amount, income, mortgage type, and purpose could be predictors of mortgage lender's decision only for low and median income applicantsIn contrast with several findings in the literature discussing racial discrimination in mortgage lending, the results of this study asserted that an applicant's nonwhite status is not a deterrent to obtaining a mortgage loan. Moreover, the grouping of the observed MSAs into regions uncovers little geographical differences in mortgage decision. In sum, once a mortgage loan applicant is in the high-income group (over $75,000), none of the explanatory variables used in the present study seems to play any significant role into predicting lenders' action.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS