The Great Recession was one of the largest macroeconomic events of the post-war period, but there is still significant disagreement about its causes. Economists such as Mian and Sufi (2014) have argued that the collapse of household net worth was a primary driver of the decline in household demand and employment. However, Chodorow-Reich (2014) and others have emphasised that the financial crisis reduced the supply of credit to firms, who then cut investment and employment. Finally, theoretical work like Eggertsson and Krugman (2012) has focused on the contraction in the supply of credit to households and subsequent decline in household demand.
The role of household credit supply in the Recession has important implications for modelling and policy. But there has been relatively little empirical evidence on how much household credit actually affected employment. In my job market paper (Mondragon 2014), I provide new evidence on this channel using rich data on the flow of household credit to US counties.
The effect of supply shocks to household credit on employment
Learning about the effects of supply shocks to household credit is difficult due to standard endogeneity issues. To solve this problem, I exploit the distress of a large healthy lender, Wachovia, as a natural experiment. The distress at Wachovia is useful because it resulted from Wachovia’s purchase of a large lender with toxic assets – Golden West Financial – in 2006.
- My empirical strategy relies upon Wachovia differentially contracting household access to credit during the Crisis.
I use public application-level data on the near-universe of US mortgages to study how Wachovia’s distress affected household access to credit. Figure 1 shows the probability a loan application is originated at Wachovia relative to applications at other lenders in the same county. This is estimated by high-, middle-, and low-income group for home purchase, home improvement, and refinance loans. Strikingly, Wachovia’s origination behaviour was average leading up to the Crisis, but during the Crisis Wachovia was significantly less likely to originate a loan, especially for low- and middle-income applicants.
Figure 1. Probability a loan application originates at Wachovia by home purchase, improvement, and refinance and by income groups
Because of this contraction, counties exposed to Wachovia experienced larger declines in the flow of household credit. This resulted in lower house prices, house sales, and non-housing expenditures. Using exposure to Wachovia as an instrument for the flow of household credit, I estimate the causal effect of supply-driven declines in household credit on employment.
- I find that a 10% reduction in the flow of non-refinance mortgages from 2007-2010 reduced employment by 3%, driven entirely by losses in construction and non-tradeables.
This result is robust to controlling for other important shocks such as subprime lending, household leverage, house prices, industry shocks, and firm credit. This result shows employment was very responsive to household credit shocks, but the significance of this channel depends critically on the size of the shock itself.
Quantifying the household credit channel
To gauge the aggregate contribution of the household credit channel, I construct a measure of the shock to each county. I then calculate the direct employment losses due to the household credit shock. I show that under general conditions, it is only necessary to measure and weigh the shocks from the lenders in a county in order to construct the shock to that county. Building on Greenstone et al. (2014), I exploit the fact that many lenders operate in multiple counties. This allows me to recover the common shock from each lender through its effect on credit quantities across counties. I then weigh each shock by the lender’s market share to construct a measure of the shock to a county.
- Using my measure of the shock to household credit supply to each county, I estimate that the aggregate direct effect of supply shocks to household credit reduced employment by over 4%, or about 60% of the observed decline.
This effect is large. While it ignores cross-county spillovers, including aggregate price effects, it suggests the household credit channel was a quantitatively important cause of aggregate employment losses.
This has a number of important implications. My results suggest that household demand depended critically on the flow of credit. So, it is necessary to include these effects when accounting for the costs of the Global Crisis. Similarly, policy actions to expand the supply of credit to households and to support household demand would likely have helped avoid some of the severe employment losses. My work also suggests that there are important frictions to substitution in the household credit market. Given the large effects of shocks to household credit that I estimate, it is critical to further understand the sources and implications of these frictions.
Chodorow-Reich, G (2014), “The Employment Effects of Credit Market Disruptions: Firm-level Evidence from the 2008-2009 Financial Crisis,” Quarterly Journal of Economics 129, no. 1: 1-59.
Eggertsson, G and P Krugman (2012), “Debt, Deleveraging, and the Liquidity Trap: A Fisher-Minsky-Koo approach,” Quarterly Journal of Economics, 127(3): 1469-1513.
Greenstone M, A Mas, and H Nguyen (2012), “Do Credit Markets affect the Real Economy? Quasi-experimental evidence from the Great Recession and “Normal” Economic Times,” NBER Working Paper No. 20704.
Mian, A and A Sufi (2014), “What Explains the 2007-2009 Drop in Employment?,” Econometrica Vol. 82, No. 6, November, 2197-2223.
Mondragon, J (2014), “Household Credit and Employment in the Great Recession,” working paper.