Time-Varying Behaviour of Sector Beta Risk – The Case of Poland

The problem of proper beta (measure of systematic risk) estimation is crucial both for academic considerations and financial market practice purposes. There is a group of empirical studies that questioned the assumption of beta time-invariance, while only some of them tried to model the process of beta time-variation. Basing on previous research, we apply the state-space methodology, which was found to be the most relevant. We focus our attention on the Polish stock market and five sector indices. Unlike other studies, we estimate our models using three different data frequencies (daily, weekly and monthly), while holding the estimation period fixed. The results indeed show the dependence on data frequency; however, in most cases, the persistence parameter is close to unity, which indicates long-lasting shocks to beta.