The Cross-Section of Expected Housing Returns
63 Pages Posted: 18 Jan 2019
Date Written: January 8, 2019
This paper performs a large-scale empirical asset pricing analysis of the cross-section of residential real-estate returns. Using monthly housing returns for 9,831 different zip codes across 178 Metropolitan Statistical Areas (MSAs), we estimate, for each MSA, a multifactor model with systematic housing-market risk (U.S. and local MSA) and idiosyncratic zip code-specific housing risk. We find that U.S. and MSA housing risks are positively priced in 26% and 22% of the MSAs, respectively. The evidence that MSA-level housing-market risk is priced in roughly a fifth of all MSAs runs counter to the common belief that the U.S. housing market is locally segmented. We also find that idiosyncratic risk is positively priced only in 22% of the MSAs, suggesting that the under-diversification of households' real estate portfolios is not widely priced. In the last part of the paper, we link MSA variation in the pricing of risk to MSA fundamentals. We find that illiquidity is important for the pricing of the U.S. housing-market risk, while homeownership increases the probability that MSA-level risk is positively priced. Idiosyncratic risk is more likely to be positively priced in MSAs with less undevelopable land and lower liquidity, indicating that under-diversification is more binding when households face fewer housing supply constraints and more illiquidity.
Keywords: Expected Housing Returns, Idiosyncratic Risk, Systematic Risk, Market Segmentation
JEL Classification: G12, R30
Suggested Citation: Suggested Citation