Are GARCH Models Appropriate for Analysing Volatility Structures in Fundamental Valuations of the OMXS30?
Abstract: This thesis investigates the volatility structures found in forward-looking fundamental valuations of the Swedish stock index OMXS30. The evaluated data constitutes daily observations of P/E ratios based on twelve months earnings estimates during the period 2009-01-02 until 2018-10-18. The analysis is conducted by applying a GARCH modelling framework to a log-return transformed return series derived from the raw data. This thesis reveals that the underlying data exhibit commonly observed properties found in financial time series with the most prominent ones being volatility clustering (i.e. heteroscedasticity) and leptokurtic behaviour. Parameter efficiency in the maximum likelihood estimation procedure is evaluated using five different distributional assumptions for the GARCH model innovations, namely: Normal distribution, Student-t distribution, skewed Student-t distribution, Generalised error distribution and the skewed Generalised error distribution. The final model choice entails a symmetric GARCH(1,1) model with innovations assumed to be generated from a skewed student-t distribution. Finally, this model proves to be sufficient in describing the volatility structure found in the return series.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)