Debt management is a serious and timely concern. The sub-prime mortgage problem in the US showed that the repercussions of poor debt management can extend into other sectors and markets, and have global economic consequences. Credit card debt by its nature merits similar concern.

Credit card debt is not secured by assets, nor is the financial status of card holders checked after initial issuance of the card.

"Credit card debt is no less severe than mortgage debt if the risk of credit debt is not well controlled," say Ying Zhao and Inseong Song of HKUST and their co-author Yi Zhao. "Risk type identification is therefore an important managerial issue for credit providers."

One approach promoted is to either convert delinquent customers to non-deliquent status or "fire" them, but the authors believe a more subtle approach could be better in the long run. Not everyone fails to repays because they cannot repay some low-risk customers may do so because of non-risk reasons such as oversight. These customers may be profitable in the longer term but if they find their accounts blocked due to delinquent payment, it may lower their preference for the card and they may stop using it.

The authors therefore develop a model to identify the more profitable customers from the less desirable ones by studying their repayment behaviors from the time they are issued the cards, as well as their credit history over time with the card issuer.

The model, as applied to 1,500 cardholders of a Hong Kong bank, effectively identifies low- and high-risk consumers. Moreover, the authors show that credit card companies can secure higher revenues if they selectively target high-risk delinquent customers, rather than blocking all delinquents on the first missed payment or, as is commonly done in the industry, waiting three months to block them.

The targeted strategy blocks the high-risk customers from using their card until they meet the minimum repayment, but not low-risk ones. The authors argue this is less costly to card issuers than letting the high-risk customers continue to use the card for two more months.

The improvement in profits from this strategy varies depending on the repayment a consumer would make if he or she chose not to be delinquent, and the probability they would choose to stay with the credit card. The larger likelihood a customer would stay, the more revenues could be earned from these customers, by as much as 20 per cent.

Managers can use the findings to better control for risk, the authors say.

"Our study enables credit card companies to identify high-risk consumers and potentially profitable consumers. As a result, the companies can be more flexible and effective in their credit supply policies, and develop effective marketing strategies by targeting specific segments of their consumers," they say.

"Our study also helps credit card companies understand the effects of credit policies on consumer delinquency in terms of different consumer segments. We show that the companies can develop targeted card policies to control consumer risk and reduce consumer default."