Revolvers for Self-Control

The paper deals with a newly discovered credit card puzzle. Many US households revolve a balance on high-interest credit cards while holding low-interest liquid or total safe assets that could be used to repay this balance. Such behavior seems to ignore obvious arbitrage opportunities and to violate fundamental notions of portfolio equilibrium. The main resolution proposed to date is by Lehnert and Maki (2000), who attribute such behavior to strategic default motives. They point to bankruptcy laws that allow households to discharge a large part of their unsecured debt. According to their argument, people who exhibit such patterns of behavior plan to declare bankruptcy and hold assets in amounts that can be salvaged after bankruptcy. Had they used those amounts to repay credit card debt, the funds would not be available after declaring bankruptcy. The authors support their argument by empirical evidence showing that this type of behavior is more prevalent in states with generous bankruptcy laws. We first take a closer look at the demographic characteristics and overall portfolios of households that revolve credit card debt using the 1995 and 1998 Surveys of Consumer Finances. We show that such behavior is quite widespread and not confined to those who are on the verge of default. Displaying such behavior when the overall household portfolio and circumstances do not indicate financial trouble imposes unnecessary interest costs and is hardly strategic behavior. We document that, in 1998, over 20% of U.S. households, representing almost one third of card holders, carried an outstanding balance on their bank-type credit card although they had sufficient liquid assets (checking, saving, and money market deposit account balances) to pay off the balance in full. Among such households, the median outstanding balance is at $850, while median liquid financial assets are several times this at $5,000. The median ratio of liquid assets to outstanding credit card balance is 4.6. This puzzling behavior is not a reflection of low introductory interest rates on bank credit cards. The median interest rate for these households was 15 percent. We find econometrically that, although all age and education groups exhibit this behavior, it tends to be more common among younger households and those with a high school degree but no college degree. It is somewhat more common among the middle class, i.e. households with incomes between $25,000 and $100,000 than among either low-income or high-income households. We then argue that such behavior could well be motivated by the desire for self control (or spouse/children control) by households that tend to utilize (some target percentage of) their credit line once it is freed up. There is empirical evidence (by Souleles and Gross) for target utilization of credit lines even when increases in lines are exogenous, i.e. not initiated by customers. If households tend to adjust their spending and overall credit card balance to the size of their credit line, then repaying debt uses up liquid assets and restores the original credit card balance by encouraging more spending. This behavior is consistent with persistently revolving credit while maintaining liquid assets, and it need not be confined to households on the verge of declaring bankruptcy. We examine whether such behavior can be generated in the context of a calibrated model.