Estimation of Bank Credit Refunding Failure Using Logit Regression (in Persian)

The current crises and also diminishing profitability in a banking system are often resulted from lack of efficiency in credit risks management and control. The most important instrument that banks need to adopt for monitoring and management of credit risk is customers’ ranking system. This study tries to employ Logit regression in a practical model for ranking the customers and to estimate the probable credit default in Parsian Bank. For this reason, certain data pertaining to the customers’ previous and present records such as professional stability, collateral, income and some other major factors are collected in order; then the credit default probability was measured for each customer through ranking and scoring. The results showed that the non-default probability for the credits paid to the customer has positive relationship with variables such as the value of collateral received from the customer, their monthly income, the status of real-estate ownership of the credit-applying customer, the age of applicant, and also their professional and educational status. On the other hand, the aforementioned probability is negatively related to “the amount of credit paid to the customer” as well as to the “repayment duration”.