Abstract:
The risk premium puzzle is even worse than previously reported if housing is also taken into consideration next to equity. While housing premia are only moderately smaller than equity premia, they are significantly less volatile and the Sharpe ratio of housing is significantly larger. Hence, three question arise: i) are existing approaches to explain the equity premium puzzle also capable of explaining even larger Sharpe ratios than previously required, ii) can return rates and volatilities of various assets be differentiated, and iii) can different Sharpe ratios between the two risky assets be matched. We analyze these questions, next to business cycle statistics, by including housing into seminal approaches to solve the risk premium puzzle in production economies. Non-disaster economies with habit formation, capital adjustment costs and limited factor mobility fail to generate a Sharpe ratio of housing of the empirically observed size and do not explain co-moving economic activity. A basic model with time-varying disaster risk can reproduce the large Sharpe ratio of housing. Moreover, the model can explain different means and volatilities of the risky assets, economic activity comoves and the model explains the volatility ratio of business investments, residential investments and house prices. However, the model does not allow to disentangle the Sharpe ratios of the risky assets and premia on equity remain too involatile.
