Abstract
The assumption of stable beta coefficient within the context of the market model, and the consequent use of ordinary least squares. estimation, has been widely criticized. In this study an alternative market model is offered in which beta is allowed to vary over time. Varying parameters regression is used to estimate the variable beta and the estimates are compared to those produced by ordinary least squares in terms of accuracy of predictions. The results indicate that neither model shows a clear advantage over the other. Even when a large change in beta was simulated, the differences in prediction errors were rather small though slightly in favor of the varying parameters model.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call