Better credit, better schools: Credit quality, race, and discrete-choice residential sorting

Studies in public finance have long established the positive relationship between housing prices and the quality of local public schools, and demonstrate that less educated, lower income, and minority households tend to live in lower quality public school districts. Previous research also finds that minority families who in theory could afford to move into higher quality school districts often do not. This study examines minority consumption of school quality by estimating how credit quality, a key yet often ignored component of housing affordability, may additionally contribute to differences in residential sorting patterns for minority families. The study employs confidential mortgage industry data that provides information on household race, financial attributes, credit quality indicators and property data.; This study employs several quantitative modeling techniques to answer key research questions. While descriptive chapters employ basic regression technique, the main structural model employs a discrete choice framework for analyzing residential choice among minority groups in Southern California, and is estimated using Generalized Method of Moments. Two important policy-related questions are examined in this dissertation: (1) Is minority status significantly correlated with key determinants of mortgage pricing, such as credit quality, after accounting for differences in financial attributes, and how does this ultimately impact housing affordability? (2) Do differences in the distribution of credit quality explain disparities in minority consumption of school quality? The research questions provide an analytical framework to guide the quantitative model design. First, I investigate the role of race and financial status on credit quality and mortgage pricing. I find significant differences in credit quality for minorities, even when controlling for financial attributes and other factors that directly impact credit quality. I also find that minorities pay significantly higher mortgage origination fees than whites, even when controlling for financial attributes. Second, results for the discrete choice model are used to analyze factors driving residential sorting patterns. I find that credit quality indicators are significantly correlated with some school quality indicators, and that the inclusion of credit quality in a model of residential sorting does help to explain some of the observed disparities in minority consumption of school quality.