Credit Risk Assessment and Credit Default Behaviour in Sri Lanka.

Abstract

This research investigates effectiveness of credit risk assessment in identifying credit default behaviour of bank borrowers. Accordingly, this research aims to examine the credit risk assessment variables used in credit risk assessment, effectiveness of credit risk assessment and formulate a model to assess the credit risk - probability of default. Research findings of Fatemi & Fooladi (2006), Kalapodas & Thomsan (2006), Richard et al. (2008) and Chen et al. (2015) emphasized that even though credit risk assessment is an important function in the field of banking and finance there are several challenges further there is a research gap in identifying how effectively credit risk assessment can identify default loan behaviour. In the conditions of increasing government sector non-performing loan balance, increasing pressure in NPL provisioning, concentration of Sri Lankan NPL among few banking entities it is high time to effectively assess credit risk to over com credit decision problem. Researcher has examined 106 credit events which included 53 credit default and 53 non default facilities using Pearson Chi square, Cramer‘s V and independent sample T test to examine the bivariate statistical relationship between individual independent variables with credit default behaviour and observe if there is any statistically significant difference between default and non-default bank borrowers. Researchers have used binary logistic regression approach to identify effectiveness of credit risk assessment and found financial and loan variables are statistically significant and developed a model in line with logistic regression approach to determine probability of default with accuracy level up to 86.5%. Further it was identified that in line with in line with EXP (B) values obtained by increase in 1 unit of selected credit risk variables - financial, loan, borrower specific and repayment specific oddness of non-default is increased by 26.864, 3.187, 0.744 and 1.723 respectively.