Financial forecasting with time series econometrics Garch models for capital markets

This thesis is concerned with a quantitative analysis of Financial Market Forecasting Methods. The thesis guides through and examines state-of-the-art ARCH models for financial forecasting and lays the foundations for their understanding. The thesis goes beyond the current academic discussion, as it asks questions about practical value of the important models and their interpretation and provides answers.; Before going into an analysis of current advanced time series modeling, we need to put classic time series knowledge into perspective. We first address basic random walks and their properties. The concepts of stationarity and ergodicity are being introduced as a foundation of time series integration techniques. These prerequisites allow then to formulate autoregressive and moving average processes, as they evolve into the ARMA concept.; The main chapter compares and evaluates the different GARCH approaches, interpretations and enhancements that have evolved since the late 1980s. We elaborate on the most important models and give examples for their field of application. Literally dozens of different variants of GARCH models have been proposed and tested in a vast research literature.