Operations Strategy Case Study

Case Title:

Assessment of Retail Credit in a Private Bank with the help of ‘Discriminant Analysis

Abstract:
After the recent US Financial crisis 2008, recession has made a strong impact especially on the banking and financial services. As a result of this, many people lost their jobs and could not even repay their loans or debts. This lead to a rise in the default rates which in turn resulted in a rise in Non Performing Asset (NPA) of the banks. Hence, banks also started monitoring their retail accounts continuously to avoid any further delinquency.

Karunya Bank, one of the private banks based in Bangalore, also noticed that its NPA has increased in the last few quarters. To deal with the matter, George Mathew, a research executive of the bank wanted to know the repayment behaviour of the customers. For this, he collected crucial information like age, income, number of dependents and years of marriage of the customer. Based on their repayment track record, he classified the customers into two categories – ‘Clean’ and ‘Defaulter’. He used Discriminant Analysis as a tool to know the repayment behaviour of the customers. The analysis shows that the bank was correct in choosing the independent variables i.e., age of the customer, income of the customer, number of dependents and years of marriage. The bank could even classify the customers into default and non-default. Age and income of the customer are the two important variables which helped the bank in explaining the repayment behaviour. Ultimately, the bank was able to succeed in reducing the risk caused due to defaulters.

Pedagogical Objectives:

• To use the discriminant analysis
• To select the important variables that discriminate
• How a discriminant function can help the bank in scrutinising new applicants on the basis of discriminant score.

Keywords : Business Research, Statistical Technique, Descriminant Analysis, dependent Variables, Independent Variables, Descriminant Function, Descriminant Score, Descriminant Function Coefficients, Wilds Lambda, Eigenvalue, Standard Canomical Descriminant Function Coefficients