ABSTRACT
The main goal of this article is to investigate empirically the Kalman approach to estimate the time-varying beta parameter as a systematic investment risk market in Poland, Czech Republic, and Hungary. In our research, we investigate the assessments of beta on the basis of seven specifications of time-varying beta for the 12 largest companies listed on the Warsaw Stock Exchange (Poland), 7 on Prague Stock Exchange (Czech Republic), and 11 on Budapest Stock Exchange (Hungary). The obtained results are compared with the estimates received on the basis of Sharpe’s linear model. Estimations are made using the maximum likelihood method for monthly data in the period 2005–2017. We are presenting the ranking of the used specifications according to three criteria of goodness of fit and the matrix of correlation coefficients between the results of these specifications. The results show that the Kalman filter estimators outperform the others.
Supplementary material
Supplemental data for this article can be accessed on the publisher’s website
Notes
1. BEKK model—this acronym comes from the names of Yoshi Baba, Rob Engle, Denis Kraft, and Ken Kroner who wrote the paper synthesizing the work on multivariate ARCH models.
2. Thomson Reuters Czech Republic Total Return Index. This index was used instead of PX-GLOB to ensure compatibility between the Hungarian and Polish market, WIG is calculated as total return index for Polish market.
3. Thomson Reuters Hungary Total Return Index.
4. Findings state these issues have no significant impact on estimation procedure.