| This dissertation is a contribution to a line of research that commenced with the seminal paper by Hansen and Singleton (1982). Here, I link the actions of the financial market agents to their goal of smoothing consumption intertemporally, and see how the asset pricing model explains the data.;This research extends the intertemporal capital asset pricing model of Lucas (1978) to incorporate time nonseparability and examines several moments of asset returns under different parameterizations of the utility function to find out which features of the data are being explained by the model. Of primary focus is the path of the habit persistence parameter as more complexities are introduced in the data characterization.;A monthly data set from January 1959 to December 1994 containing personal consumption expenditure, return on assets, and the risk free return is used.;I apply the Semi Non Parametric (SNP) method of Gallant and Tauchen (1989) to the trivariate series of gross real riskfree rate, gross rate of return on assets, and growth rate of real consumption expenditures to provide a statistical description of the data. SNP is a nonparametric method for the estimation of density functions using polynomial series expansion.;I then apply the Efficient Method of Moments (EMM) approach of Gallant and Tauchen (1996) to estimate the parameters of the structural model. The asset pricing equations are numerically solved using the quadrature based method of Tauchen and Hussey (1991). EMM uses the expectation under the structural model of the score from the SNP model as the vector of moment conditions. The optimized EMM objective function can be used to test that the structural model is correctly identified. If a model fails the test then examination of the elements of the minimized score can reveal the features of the data the structural model cannot describe. |
