Mnb One Credit-Card Portfolio

Abstract

A credit-card company must value portfolios of customers based on their future earnings. The payment characteristics of customers serve to classify them into states. This case can be the basis for discussing state dynamics over time in a Markov process.

Excerpt

UVA-QA-0613

MNB ONE CREDIT-CARD PORTFOLIO

The decision had to be made tomorrow. Sabrina Rapley, MNB One's CFO, had requested that MNB One's credit-card company acquire a portfolio of customers from another credit-card company. Believing that MNB One had gained as many new customers as possible via organic growth, the CFO viewed acquisition as the only way to continue to grow.

Karel Duguid was the lead analyst on the team charged with making a decision on which portfolio of one million customers to purchase. The team had whittled the possible portfolios down to two potential targets. The first was the credit-card portfolio of the major retail chain Markus Nayman, which was redirecting its core operations away from credit cards. It was attractive to MNB One because it comprised lower-risk, established customers. The second candidate was the credit-card portfolio of Sirtem, a competing credit-card company, which had just been placed in receivership, forcing it to sell its credit-card business. Sirtem had acquired customers solely through the Internet, and most of its customers were young, college-age consumers for whom Sirtem's card was their first credit card. Even though Sirtem was having financial difficulties, many customers in its portfolio were current in their payments and thus very attractive to MNB One. Both portfolios looked good, though MNB One would buy only one.

To sell off the portfolios, the two credit-card companies had provided a random, normalized sample of 10,000 of their customers so that potential purchasers could evaluate them. Duguid had his analysts look at the samples of both companies. The analysts used the samples to determine the monthly average contribution that each type of customer would provide and the current profile of customers. MNB One believed, however, that it had a superior approach to customer relations. So Duguid's analysts took even larger samples of customers for each portfolio from MNB One's customer data base, in each case matching the customer characteristics for both Sirtem and Markus Nayman. This method was used to determine probabilities for customer transitions over time.

MNB One's Card-Customer Partitioning

. . .

Keywords: Markov Process, credit card banking financial services dynamic models transition matrices