Identification Failure in the Intertemporal Consumption CAPM

Using a general equilibrium representative agent framework developed by Lucas (1978), Hansen and Singleton (1983) studied the behaviour of asset returns and consumption growth with maximum likelihood techniques. The model was rejected, but their work is commonly cited for estimates of the underlying parameters governing risk aversion. This paper studies the robustness of the Hansen and Singleton results to alternative specifications of the model and extensions of the data set. We find that estimates of the risk aversion parameter vary widely across data sets and instrument specifications because the Hansen-Singleton model is nearly nonidentified due to the unpredictability of consumption growth. A natural identifying assumption equates quot;consumption betaquot; with the coefficient of relative risk aversion, and leads to estimates which are much better behaved