The Default of U.S. Mortgage Loans: Focusing on the Borrowers: or Lenders-friendly Policies

본 연구는 연체율이 급등하는 시점에서, 차입자 또는 대출자-우호 정책을 중심으로 주택담보대출 채무불이행을연구한다. 미국의 법원경매(Judicial foreclosure)와 소유권 회복제도(Redemption)를 차입자 우호정책으로, 소구 주택담보대출(Recourse mortgage loan)을 대출자 우호정책으로 구분하여 그 효과를 분석하였다, 소구 주택담보대출정책은 채무불이행 확률에 통계적으로 유의한 음(-)의 영향을 미쳤으나, 법원경매, 소유권 회복제도는 통계적 유의성을 확보할 수 없었다. 또한, 관측되지 않는 당월 DTI를 추정하여, 당월 실업률 변수와 비교하여, 채무불이행 모형에 대한 영향력 차이를 확인하였다. 당월 DTI 변수는 채무불이행에 통계적으로 유의한 양(+)의 영향을 미쳤으나, 당월 실업률 변수는 통계적으로 유의하지 않았다. 마지막으로 계량모형과 기계학습모형의 예측력을 비교하여, 계량모형의 예측력이 상대적으로 우수함을 검증하였다.The study focuses on analyzing the effects of borrowers or lenders-friendly policies on mortgage defaults during a period of high delinquency rates. Specifically, the study investigates the impact of judicial foreclosure, redemption, and recourse mortgage loans on mortgage defaults. The results showed that while judicial foreclosure and redemption were classified as borrower-friendly policies, they did not have a statistically significant effect on the probability of default. Recourse mortgage loans, on the other hand, were found to have a statistically significant negative effect on the default, meaning that they reduced the likelihood of borrowers defaulting on their loans. Additionally, the study compared the influence of the borrower’s current debt-to-income (current DTI) ratio and unemployment rate on the default model. The current month’s DTI variable was found to have a statistically significant positive effect on debt default, meaning that higher current DTI ratios increased the likelihood of borrowers defaulting. However, the current month’s unemployment rate variable did not have a statistically significant effect on debt default. Finally, the study compared the predictive power of an econometric model and a machine learning model, and found that the econometric model was weakly better in predicting mortgage defaults.