We use administrative data from the U.S. Census Bureau to estimate the causal effects of loose credit conditions on firm employment and worker earnings. To obtain quasi-random variation in firms’ exposure to credit booms, we exploit the segmentation of high-yield (BB+ rated) versus investment grade (BBB- rated) firms in credit markets. Loose credit conditions generate cyclical fluctuations in employment: high-default risk firms create jobs during the credit boom, but then experience financial distress and destroy these jobs during the ensuing bust. We show that these firm-level boom-bust dynamics are transmitted to individual workers. To obtain quasi-random variation in workers’ exposure to boom-induced job creation, we exploit the importance of parental connections in determining where labor market entrants are first employed. We find that recent high-school graduates with parents at high-yield (BB+) firms can more easily find high-paying jobs during credit booms, compared to graduates with parents at investment-grade (BBB-) firms. But ten years later, graduates whose parents were at BB+ firms have substantially lower earnings. The magnitude of these negative long-term effects is comparable to the effect of entering the labor market during a recession. Our results suggest that loose credit market conditions lead firms to create short-lived jobs that impede workers’ long-run accumulation of human capital.