The Determinants of Subprime Mortgage Default and Prepayment: The Implications for Financial Regulations and Predatory Lending Open Access
Downloadable ContentDownload PDF
Beginning in 2007, a steep surge in U.S. subprime mortgage defaults devastated the U.S. and global financial markets. Lack of appropriate regulation has been cited as a principal reason for the ensuing financial crisis. In this dissertation, we investigate the major determinants of subprime mortgage default and prepayment with a focus on regulatory issues. We address three primary research questions: (1) Do state-level anti-predatory lending laws (APLs) significantly affect borrowers' decisions to default and prepay in the subprime mortgage market? (2) Can the option-based model fully explain subprime mortgage default and prepayment?, and (3) What are the impacts of state APLs and other variables on the pricing of subprime mortgages? In order to find the answers to these research questions, we introduce a new comprehensive database on subprime mortgages and legal indices of state-level APLs. We employ a multinomial logistic regression model to estimate the probabilities of the two competing risks of subprime mortgage termination - default and prepayment. Our findings suggest that state APLs on prepayment penalties would be effective in curbing subprime mortgage default and excessive pricing. However, state APLs on points and fees are found to increase the probability of default and pose a risk of possible unintended negative consequences. The empirical results also suggest that the option-based model plays a vital role to explain subprime default and prepayment. In addition, other predictor variables representing loan and borrower characteristics are found to be important. Finally, state APLs on points and fees are shown to be effective in curbing the excessive pricing of subprime lending. The main contributions of this study are: (1) the empirical model explicitly takes into account the effects of state-level APLs on the subprime default and prepayment decision, (2) this dissertation is the first to date to introduce the points and fees variable into the empirical model to estimate the competing risks of mortgage termination, and (3) uses a comprehensive and representative database at a national level to minimize selection bias. The findings from this study can provide the baseline for future research and ongoing discussion on financial regulatory reform.