서울시 주택담보대출의 신용위험 스트레스 테스트

A Stress Test to Estimate the Credit Risk of Residential Mortgages in Seoul

Based upon 45,464 housing loans of a commercial bank, the purpose of this study is to build up a credit risk model of 11 independent variables with binary logistic regression and then estimate the credit risk levels under the stressful situations of macro-economy. Two dependants stand for credit risk: delinquency and bad loan ratios. Stresses upon macro-economy are schemed in two ways: one is the sharp declines of market value of loan security by 5%, 10%, 15% and 20% and the other is the steep rise of market interest rates by 25%, 50%, 75% and 100%. We generated random numbers one million times for each independent variable proper for its mean, standard deviation and distribution then insert them into equation to predict two ratios of delinquency and bad loan under the stresses. Hard landing of housing value by 5%, 10%, 15% and 20% entails the soaring up delinquency rate from 1.56% of the base scenario to 1.59%, 1.63%, 1.68% and 1.73%, respectively. The ratios of bad loan are predicted to increase from 0.68% up to 0.71%, 0.74%, 0.78% and 0.83% corresponding to the scenarios of security value loss. Similarly, the rapid hike-up of market interest rates makes mortgage debt service heavier thereby degenerating the quality of loans tremendously: the delinquency rates are expect to go from the current level of 1.56% to 2.60%, 4.48%, 7.72% and 12.49% by the increase of 25%, 50%, 75% and 100%, each. Bad loan ratios are also exacerbated from 0.68% to 1.42%, 2.94%, 5.94% and 11.35% in the same scenarios. Overall, the current average of LTV ratio of 37.4% is not believed to worsen the credit risk of Seoul mortgages, at least in the near future. This is despite of the tight figures of 52.8 years of borrowers' average ages and the premise of absence of unexpected turmoil of macro-economy.