Time to Default in Credit Scoring Using Survival Analysis
Time to Default in Credit Scoring Using Survival Analysis

Summary/Abstract: Credit risk assesment has been dominated by logistic and probit regression techniques. As the use of credit scoring has expanded over the past 20 years, concerns have been raised about whether its use may unfairly affect minorities. The aim of this paper is to investigate the behavioral of gender variable in different credit scoring models and what are the benefits of using this variable. The first result shows that females have a probability of surviving 22 months of 75%; conversely, for the male group, the probability of surviving the same time is slightly more than 75%. Adding education as variable, we observe that male with university degree recorded a survival probability of 90% after 20 months, male with high school 75% while female with high school only 50%. The hazard for female is in average 1.12 times the hazard for males. The results show that the single and divorced females survive more than males in the same marital status.

• Details
• Contents