Housing as a Measure for the Long-Run Risk - Job Market Paper - PDF

I evaluate the effects of long-run consumption growth risk and housing consumption risk on asset prices. Current asset values are affected by the risk-return tradeoff in the long-run. Housing plays an important role in the economy. As an asset, it is particularly sensitive to long-run risk-return trade off; as a consumption component, it accounts for one fifth of the total expenditures in non durable goods and services. The investment horizon for housing is usually distant in the future. Investors fear shocks that can affect the value of their house for a long period of time. Such shocks affect substantially the services obtained from the house and its price as an asset as well. I use a non-separable utility function with non-housing consumption and consumption of housing services, which generates an intertemporal composition risk, besides the traditional consumption growth risk. The composition risk has effects for the valuation of cash flow growth fluctuations far into the future due to the persistence of consumption growth. I provide a closed form solution for the valuation function despite the non-separability. This allows me to quantify the price of risk in the long-run with inputs from vector autoregressions. I evaluate the different exposure to long-run risk of a cross section of portfolios of securities, and characterize the price of risk for different investment horizons. The model also explains the spread of the returns to different portfolios sorted in book to market and housing returns, at different investment horizons.

GMM Estimation of an Asset Pricing Model with Habit Persistence (with Hugo Garduño), 2005- PDF - PS

In this paper we estimate and test for the first time the theoretical model presented in Campbell and Cochrane (1999). This has been the most succesful paper recently in Asset Pricing. And we do it in three different market settings. First we estimate assuming complete markets with aggregate data from the NIPA accounts. Second, we assume that the non-stockholders are not marginal, hence do not matter in pricing (limited participation). Finally we drop the assumption of complete insurance and we estimate the model with incomplete markets. We estimate these two last settings with household level data on consumption and asset holding from the Consumer Expenditure Survey (CEX) database. We find evidence that a complete markets model is better able to explain average returns, whereas a model that includes limited participation of agents in the stock market and incomplete consumption insurance among individuals is better able to explain the equity premium and does so with a lower value of the RRA coefficient than a model with complete markets.