Application of multiple discriminant analysis credit scoring models, for credit card consumers - the case of Barclaycard Kenya

Abstract

Credit risk analysis is a process that allows financial institutions to minimize the amount of follow-up on late payment and loan default to be performed. In order to reduce credit card default risk at Barclaycard Kenya and other credit card lenders in Kenya, this study investigates the suitability of multiple discriminant analysis model in differentiating between good and bad credit card holders. Secondary data comprising of 100 good and 100 bad card holders was collected from existing customers application forms. The classification of an applicant as good or bad payer is based on characteristics and behavior of the person. Variables such as age, annual income and number of credit cards held were analyzed to create constituency by credit analysis. Discriminant analysis technique is applied using statistical information related to the variables of the study to discriminate good credit risks from bad credit risks with an aim of application in the evaluation of new credit card applicants. From the analysis, it emerged that discriminant analysis can identify groups differences existing in predetermined groups. However, some variables such as sex, nationality, town and annual income were found to be weak discriminants. On overall, Mda technique is applied successfully therefore recommended for evaluation of new credit card applicants in Kenya.