New Research Suggests Household Leverage Plays Important Role in Recessions
November 11, 2009
Increases in household leveraging from 2002 to 2006 are connected to virtually all of the decline in the consumption of durable goods during the early stages of the recession, said Amir Sufi, associate professor of finance.
Meanwhile, beginning with the third quarter of 2008, reductions in credit card limits help explain the reduction in spending during and after the financial crisis, he said during the Becker Brown Bag Series, sponsored by the Becker Center on Chicago Price Theory, at Harper Center on November 11. “Together, the growth in household leverage and credit card utilization rates can explain the entire drop in auto sales, while the growth in household leverage explains only 20 percent of the increase in unemployment,” Sufi said.
The mortgage default crisis is primarily concentrated in U.S. counties that experienced high growth in household leverage before the recession, he said. Furthermore, housing prices have plummeted at least 40 percent in those counties while they continue to rise in counties that experienced low growth in household leverage before the recession, Sufi said.
“Think of high-leverage-growth counties as places like Florida and California and low-leverage-growth counties as places like Texas,” he said. “Of course, there is within-state and across-state variation, but this should give you some sense of the counties I’m talking about.”
From 2006 through the third quarter of 2008, household leverage can explain almost all variation in defaults, housing prices, auto sales, and new housing building permits across counties in the severity of the economic downturn, Sufi said. It also serves as a fairly reliable indicator of unemployment, although unemployment — more a national economic gauge — also rose slightly in some counties with low growth in household leverage, he said.
“However, if you just focus on the third quarter of 2008 through the second quarter of 2009, which is the most severe part of the recession in terms of unemployment and GDP decline, you seem to see an equivalent collapse in auto sales and unemployment in both the high- and low-levered-growth counties,” Sufi said.
A reduction in credit card limits offers a possible explanation of why economic conditions worsened across all counties starting in the third quarter of 2008, he said. “What’s pretty surprising is that before the third quarter of 2008, as mortgage defaults are rising and there seems to be real trouble in the economy, there is no direct evidence that credit card limits are declining,” Sufi said. “If anything, there is expansion as credit card companies seem to be extending more and more credit to households. More importantly, there is no differential trend in the high-leverage-growth and low-leverage-growth counties.”
However, in the third quarter of 2008, credit card companies massively reduced credit limits across counties, he said. “If this is a proxy for a more general trend in the economy of credit availability – and the Federal Reserve Loan Officer Survey suggests a massive tightening of credit standards right around the third and fourth quarters of 2008 – it seems there is something about this credit availability story that matters,” Sufi said.
The research of Sufi and others on this phenomenon provides “suggestive evidence” that household leverage plays an important role in recessions, he said. “It’s not conclusive evidence that it is the main driver,” Sufi said. “But it at least opens questions about why household leverage has not been given more attention in the macroeconomic literature. If this type of work accomplishes anything, it is that maybe there is some evidence that these channels warrant a closer look.”
Pietro Biroli, a second-year economics student in the PhD Program, attended Sufi’s talk to get a better understanding of why the recession occurred. “There were a few implications for policy and a few implications for research in general,” Biroli said. “Among them was the idea that household leverage and lending-to-income ratios should be embedded more into the models and decisions we make. It’s not complicated and it’s a clear point that is missing from most of the policies and models we have.”
— Phil Rockrohr