Estimating Credit Scoring for Individuals

Abstract

The present paper presents a scoring model for individuals who require an application for funding through a bank loan [3], [4], [8-10]. We were taken into account 28 criteria for calculating the total score of customer and we have determined the degree of indebtedness based on National Bank of Romania regulations [1], [12]. We have estimated the probability of default using logistic regression. Following analysis and simulations on a sample of 80 applicants, of which 42 are already accepted for credit, there were established five risk classes (I, II, III, IV, V) and indebtedness as a percentage of income eligible. We determined the regression coefficients, using a solver in Excel and the data entry (factors) were: Indebtedness determined by income and loan amount; Number of products held in the bank; Number of dependent persons; Time worked in last job (in months).
Keywords: Bank loan, Banking system,Credit risk management, Logistic regression, Scoring.