On Unstable Beta Risk and Its Modelling Techniques for New Zealand Industry Portfolios

Abstract

This paper adds to the literature a study on the time-varying beta risk of the New Zealand industry portfolios. The previous analyses of three major modelling techniques are extended to include the stochastic volatility model and the Schwert and Seguin approach based on the stochastic volatility model. Evidence generated clearly indicates that the betas of all the NZ industry portfolios are also unstable. It is found that, in the case of in-sample forecasting, the stochastic volatility model is the optimal technique, while the GARCH model is most favoured for out-of-sample forecasting, unlike prior work on other countries which suggests that the Kalman filter approach is preferred.