Estimating Risk Aversion in C-CAPM Model: A Comparison between GMM and Multivariate GARCH-M

Abstract

Risk aversion, a useful indicator in the financial market, is generally modelled as a parameter using general equilibrium models with observable macroeconomic factors such as consumption-based capital asset pricing model (C-CAPM). In the estimation procedure, one way is to estimate the Euler equation directly using Generalised Method of Moments (GMM). The main problem of this method is to identify instruments in the estimation. Another method is to estimate no-arbitrage condition using multivariate generalised autoregressive conditional heteroskedasticity in mean model (GARCH-M). Until now, there is no evidence that which one is best in the performance of estimating the risk aversion. In this paper, we carry out a risk aversion estimating experiment and compare these two econometric methods using UK financial market as a test market. It shows that GMM is relatively superior to multivariate GARCH-M for predicting the risk aversion. In addition, it also reports certain ideal instruments for UK in the application of GMM method.