House Prices, Investors, and Credit in the Great Housing Bust (Job Market Paper)
I study the role of investors in stabilizing housing markets during the Great Housing Bust. Using transaction-level housing data, I distinguish between two types of investors that were active during this period: large corporate investors and small household investors. I estimate that following a mortgage credit contraction, house prices fell by 30 percent more in markets where household investors absorbed larger shares of house purchases. To rationalize this result, I build a heterogeneous agent model of the housing market featuring both types of investors. I show that equilibrium house prices fall sharply following a mortgage credit contraction when household investors are required to absorb falling housing demand. In contrast to corporate investors, household investors are sensitive to changing credit conditions and the illiquidity of housing assets. Prices must fall to generate sufficiently large returns to compensate previously indebted homeowners for the increase in borrowing required to invest in additional housing.
Conditionally Accepted at American Economic Journal: Macroeconomics
Recent work has shown that neighborhoods with better schools tend to have higher house prices. In this paper, we investigate how the link between house prices and school quality affects intergenerational mobility. We build a dynamic heterogeneous agent model where individuals can choose to live in different neighborhoods whose house prices are determined in equilibrium. School quality in each neighborhood comes from property tax revenue. Simulations from the calibrated model show that neighborhoods with worse schools have lower mobility and create poverty traps. We then evaluate policies that can improve mobility such as equal school funding and housing vouchers.
Revised May, 2020
Awarded the A. R. Bergstrom Prize in Econometrics in 2019
Fluctuations in house prices can generate large movements in household expenditure. However, empirical work exploring this relationship must deal with the endogeneity problems associated with using house prices as a regressor. A popular instrumental variables strategy exploits cross-sectional variation in house prices as predicted by the local housing supply elasticities of Saiz (2010). As an alternative, I introduce a Bartik instrument for house prices that consists of the interaction between the pre-existing local supply of housing characteristics and broad changes in the relative demand for those characteristics. I show that the instrument is a strong predictor of house price growth in both the cross-section and through time. I then use household panel data on non-durable expenditures to estimate the elasticity of consumption with respect to local house price growth. I report precise estimates in the range of 0.10 to 0.15, which correspond to marginal propensities to consume out of housing wealth of 1.2 to 1.8 cents in the dollar. These estimates are robust to controls for aggregate fluctuations, local business cycles, and local industry and demographic composition. In contrast, I show that the traditional housing supply elasticity instrument produces inconsistent estimates when confronted with these same controls. Thus, the Bartik instrument succeeds in generating plausibly exogenous variation in house prices when housing supply elasticity instruments may fail. When decomposing variation in the Bartik instrument, I find that the identified consumption response to house prices is largely driven by times and locations where house prices varied the most: during the 2008 recession and in the Western US.
Research In Progress
Unemployment Risk by Age and Industry During a Pandemic Recession (with Murat Ozbilgin)
Beliefs over Housing Booms and Busts: Evidence from the Great Recession (with Christos A. Makridis)
The Purchase Price Premium for New Houses (with Don Schlagenhauf)
Optimal Monetary Policy and the Exchange Rate (with Christie Smith)