Time to default analysis in personal credit scoring
Time to default analysis in personal credit scoring

Summary/Abstract: Credit scoring is a technique mainly used in making consumer credit decisions. Traditional credit scoring systems aim to estimate the probability that an applicant will default. However, for the financial institution, not only if but also when the creditor defaults is important. The aim of this paper is to analyse the usefulness of survival methods, especially the Kaplan-Meier estimator, in the process of making decisions of credit granting, as well as in monitoring credit portfolios. The focus is put on the differences in the influence of particular characteristics used in scoring models on the probability of default and the probability of early repayment during credit life-span. The survival analysis approaches to credit scoring were tested on 60-month personal loan data from one of the Polish financial institutions. All loans were observed until 24 months after inception, or until default or earlier repayment if one of these happened earlier. In the first part of analysis, early full repayments qualified as censored data, and in the second as failures.

• Details
• Contents