Modelování a predikce akciových výnosů prostřednictvím modelů typu GARCH

Abstract

The goal of this diploma work is to determine whether for modelling and prediction of volatility on the stock exchange is preferable GARCH model or EGARCH model. The analysis is performed on stock market of Great Britain, Germany and Switzerland. The diploma work is structured into five chapters including introduction and conclusion. The second chapter defines the volatility of yields on financial assets. Stock indices are characterized and three stock exchange are described, London Stock Exchange, SIX Swiss Exchange and Frankfurt Stock Exchange. The third chapter deals with time series, statistical method of estimations, description of volatility model and model assessment criteria. The fourth chapter is the key part of the diploma work. It analyses specific financial time series, calculates characteristics of the position and variability of logarithmic yields and undertakes a normality test. The estimation is conducted by means of the GARCH and EGARCH models using the maximum credibility technique. The prediction of volatility which is evaluated the RMSE criterion. In conclusion, this part compares and evaluates the results in terms of selected criteria.