We study a large-scale quasi-experiment in the Brazilian banking sector characterized by an unexpected and macroeconomically relevant increase in lending by commercial government banks. Using credit registry data, we find that this intervention led to a reduction in lending rates, but it did not lead to a change in private banks’ credit supply. Firms reliant on government banks experienced a substantial increase in debt, and government banks faced a large increase in loan defaults driven by indebted firms. We find a small increase in employment at the firm level, suggesting limited direct benefits of the intervention. At the regional level, we find that branch presence cannot explain credit growth due to cross-market borrowing. Once we account for this channel, we find real effects at the regional level that are substantially larger than those at the firm level, emphasizing the general-equilibrium effects of large-scale interventions.