Abstract
The beta of a stock is important in a variety of contexts, ranging from the cost of capital, asset-pricing theory, to hedging using index derivatives. It is common to measure betas by estimating the market model using straight ordinary least square (OLS) regression in obtaining beta estimates. This assumes that betas are constant, despite strong economic arguments in favour of time-varying betas.
In this article, we test for time-varying betas in the context of a market model with Generalised Auto Regressive Conditional Heteroscedasticity (GARCH) errors, using the modified Kalman filter of Harvey et al. (1992). The null of beta constancy is rejected for 52 per cent of stocks. This has significant implications for portfolio diversification and hedging.
Get full access to this article
View all access options for this article.
