Title: On spectral analysis and forecasting performance of regime switching GARCH-MIDAS models
Authors: Jie Cheng - Keele University (United Kingdom) [presenting]
Abstract: The general expressions of autoregressive functions and spectra for regime-switching GARCH-MIDAS models under certain conditions are constructed. Simulation studies of theoretical spectral density functions of these models are presented. The results show the different effect of switching characteristic, long-term and short-term GARCH components and how it governs the distributions of the volatility. These results can be regarded as a starting point to propose some specification tests. We also extend a previous result on forecast evaluation to the case of a regime-switching GARCH-MIDAS model and the theoretical results provide a new insight to forecasting performance evaluation, especially for regime-switching models.