Nonparametric Dynamic Conditional Beta

This article derives a dynamic beta representation using a Bayesian semiparametric multivariate generalized autoregressive conditional heteroskedasticity (GARCH) model. The conditional joint distribution of excess stock returns and market excess returns is modeled as a countably infinite mixture of normals. This allows for deviations from the elliptic family of distributions. Empirically, we find the time-varying beta of a stock nonlinearly depends on the expected value of excess market returns. The nonlinear dependence is robust to different GARCH specifications as well as more factors in the model. In highly volatile markets, beta is almost constant, while in stable markets, the beta coefficient can depend asymmetrically on the expected market excess return. We extend the model to several factors and find empirical support for a three-factor model with nonlinear factor sensitives.