This paper uses micro data from four OECD countries (the United States, Spain, Italy, and the Netherlands), to assess the determinants of household debt holding and to investigate whether or not credit constraints are important for household debt holding. We extend the existing literature in important ways. First, we present comparative evidence for four countries at the micro level, where we rely on household panel data for two countries; we are thus able to control for unobserved heterogeneity via individual household effects and to track changes in household behaviour over time. Second, by making data across countries as comparable as possible, we can explore the importance of the differences in institutional settings for debt incidence, debt outstanding and credit constraints. We also explore the implications for debt holding from consumption models, including a numerically solved precautionary savings model. We find that inter-country differences are substantial and remain even after controlling for a host of observable characteristics. This points to institutional differences between the countries being important.