We investigate the issue of pervasive credit constraints among US households. There is considerable debate about the incidence of constraints and whether the observed low borrowing in some groups of the population arises from low demand or from denial of credit. Using information on unsecured borrowing from the Consumer Expenditure Survey for (1988-1993), the paper estimates credit constraints and shows how these differ with household characteristics. It finds that around 31% of households are constrained, with young college educated households being the most constrained. Moreover, the low level of borrowing observed among black households is shown to be a demand rather than a supply effect. Copyright 2007 , Oxford University Press.