This study looks at well-known finance puzzles and examines whether they can at least in part be explained by the fact that most of the existing finance research has failed to account for real estate as a major asset class. In Chapter 1 of this dissertation, I examine whether the inclusion of returns to residential and commercial real estate can improve the empirical performance of asset pricing models and whether the systematic risk associated with holding real estate carries a positive risk premium. The empirical evidence in this chapter shows that the inclusion of returns to real estate can drastically improve the empirical performance of a variety of linear asset pricing specifications. There is also strong evidence that the systematic components of both residential and commercial real estate risks are priced by the market, indicating that investors have to be compensated for holding stocks whose returns are positively correlated with the value of their real estate holdings.
Chapter 2 of this dissertation examines the financial portfolio choices of households as a function of their exposure to real estate risk. Using micro level panel data, I find that after controlling for wealth, income, and other-possibly unobserved-household characteristics, exposure to real estate, and in particular to owner-occupied housing, reduces the percentage of financial assets held in stocks and other risky financial assets. Higher mortgage balances, on the other hand, are associated with higher shares of financial assets held in the form of risky financial assets, indicating that the amount of equity accumulated in the home matters for households' financial decisions.
The research set forth in this dissertation shows why models that fail to account for an asset class as important as real estate have difficulties in explaining observed asset prices and portfolio patterns.