Brazil initiated a major credit expansion program through government banks in 2011. The program primarily targeted public sector workers with offers of payroll-backed loans. Using individual-level administrative data we find that the program led to a 15 percentage point rise in debt to initial income for public sector workers. We develop a new method for estimating workers' expected income growth, and show that “consumption smoothing” cannot explain the rise in consumer borrowing. Instead, the evidence supports “consumption binging”: less financially sophisticated workers borrowed more at high real interest rates, and experienced both higher consumption volatility and lower average consumption.

As the financial system expands to new clients and services, countries are promoting financial education, with unknown long-run returns. In 2011, we studied the short-run impact of a comprehensive financial education program through a randomized controlled trial with 892 high schools in Brazil. This paper uses administrative data for 16,000 students over the next nine years to measure the program's long-term impact. We find that treatment students are less likely to borrow from expensive sources or to make delayed loan repayments than control students. The program also caused students to shift from formal jobs to microenterprise ownership.