Do Liquidity Constraints and Interest Rates Matter for Consumer Behavior? Evidence from Credit Card Data

Using a panel of about 24,000 individual bankcard accounts from several large U.S. card issuers (January 1995-January 1998), Gross and Souleles test whether consumer debt responds to changes in credit supply in ways the canonical Permanent-Income Hypothesis (PIH) rules out. They find that an increase in a cardholder’s credit limit generates an immediate and significant rise in debt, with an average “marginal propensity to consume out of liquidity” (MPC, dDebt/dLimit) of 10-14 percent – a response that is far larger (almost 50 percent) for accounts already near their credit limit but remains significant even for accounts starting well below it, which the authors read as consistent with buffer-stock models of precautionary saving rather than only currently-binding liquidity constraints. They also find that debt is sensitive to the account’s own interest rate, with a long-run elasticity of about -1.3, larger for interest-rate decreases than increases (which the authors argue helps explain the popularity of temporary promotional “teaser” rates) and smaller for accounts near their credit limit. A substantial minority of borrowers simultaneously carry card debt and low-yielding liquid assets, a pattern the authors say conventional models cannot easily rationalize. The results rely on an unusually rich set of controls (credit-risk scores, account fixed effects) and instruments (exogenous timing rules in issuers’ credit-supply policies) to address the endogeneity of credit-limit and interest-rate changes, and the paper’s own conclusion frames the results as evidence that credit-card borrowers are an important channel for the transmission of monetary policy and other credit-supply shocks, while noting that the companion question – how issuers themselves adjust credit supply in response to monetary policy – is left for future work.

Nicholas S. Souleles | Macro Paper Warehouse

Do Liquidity Constraints and Interest Rates Matter for Consumer Behavior? Evidence from Credit Card Data