Credit Card Debt Puzzles

Two thirds of US households hold bank type credit cards, and most of them revolve high-interest card debt. Standard models cannot account for the observed co-existence of such debt with (i) substantial asset accumulation by retirement age, and (ii) substantial holdings of low-rate liquid assets. Hyperbolic discounting can resolve the former but not the second puzzle (Laibson, et al., 2000). Bertaut and Haliassos (2002) proposed a framework for the second puzzle where the person (or self) in charge of finances revolves debt in order to restrict the amount of available credit and thus control how much the shopper (self) can charge on the credit card. The current paper models shopper behavior, calibrates, solves, and simulates the accountantshopper model. It also compares behavior to a standard setup without selfcontrol problems, to one where a debit card is used instead, and to one with a lower credit card limit. Results suggest that the model can generate both types of co-existence, while also producing target credit card utilization rates, consistent with the Gross and Souleles (2002) empirical findings. JEL Classification: E210, G110.