Household Credit and Employment in the Great Recession

How much did the contraction in the supply of credit to households contribute to the decline in employment during the Great Recession? To answer this question I use variation in shocks to mortgage credit supply across U.S. counties. First, I non-parametrically identify lender-specific shocks, which I then aggregate into a county-level shock. I show this shock can help account for the household credit channel, but direct estimates are biased. I use two related sources of exogenous variation to correct this bias. First, I exploit a county's exposure to a large and previously healthy lender as a natural experiment. The instrumental variable estimates confirm that direct estimates are biased, and that shocks to household credit supply had large effects. A one standard deviation decline reduces employment by 3 percent, the flow of home purchase credit by 7 percent, refinance credit by 20 percent, and home improvement credit by 5 percent. These effects are very persistent, providing cross-sectional evidence that recoveries after financial shocks are slow. Second, I identify additional lenders likely to satisfy the exclusion restriction and recover effectively identical estimates. These results imply the direct effects of the household credit channel account for at least 20 to 30 percent of aggregate employment losses. This shows shocks to the supply of credit to households were significant contributors to the severity of the Great Recession and subsequent slow recovery.