Application of Competing Risks Models to Credit Risk Assessment

Abstract

Credit risk arises from the debtor’s possible failure to meet the terms and conditions of the credit contract. As a result, the bank does not receive a particular payment stipulated by the contractual provisions. Credit risk usually equates with the credit taker’s insolvency. Hu and Cheng (2015) note the shortage of studies devoted to other kinds of credit risks competing with the risk of default and their influence on the evaluation of the probability of default. In the article, a default and an early repayment are considered to be competing risks. Two approaches were used to research the intensity of competing risks: evaluation of cause-specific hazard and sub-distribution hazard respectively. The interpretation principles within the results acquired by the use of either method have been discussed. For either of the approaches, proper regression models have been set up, alongside conducting the sensitivity analysis. The results have been duly compared. The empirical study employed a sample of 5000 sixty-months’ credits granted by one of the Polish financial institutions. Application characteristics of the credit takers have been used in regression models as covariates.

Keywords

• fine-gray models
• sensitivity analysis of cox models
• probability of default
• early repayments