REIT volatility prediction for skew-GED distribution of the GARCH model

摘要

This study investigates how specification of return distribution for REIT influences the performance of volatility forecasting using three GARCH models (GARCH-N, GARCH-ST and GARCH-SGED). Daily prices on the REIT provide an empirical sample for discussing and comparing relative ability to accurately out-of-sample volatility, given the growth potential of REIT markets in the United State from the perspective of global investors. Empirical results indicate that the GARCH-SGED model is superior to the GARCH-N and GARCH-ST model in forecasting REIT volatility in the United State, for all forecast horizons in which model selection is based on MSE or MAE. Meanwhile, the DM-tests further confirm that volatility forecasts using the GARCH-SGED model are more accurate than those generated using the GARCH-N and GARCH-ST model in all cases. These findings demonstrate the significant influences of both skewness and tail-thickness on the conditional distribution of returns.