Abstract: What is the impact of granular credit risk on banks and the economy? We quantify single-name exposure risk in bank portfolios by applying a novel empirical strategy to an administrative loanlevel dataset from Norway. Exploiting the fat-tailed properties of the loan-share distribution, we use the granular instrumental variable approach to show that idiosyncratic borrower risk survives aggregation within banks’ portfolios. These granular credit shocks spill over from affected banks to firms, reducing investment and raising default risk among non-granular borrowers, with sizeable consequences for the real economy.