A comparison of forecast models of REIT volatility: GARCH model, AFIMA model, Markov switching model

In order to find the optical forecast model of REITs volatility, the paper uses the GARCH model, ARFIMA model and Markov switching model to analysis three REITs from the Hong Kongs Hang Seng market. Empirical results have shown that: Real volatility as the standard, the order of Predictive ability is that the Markov switching model, ARFIMA model, GARCH model, EGARCH, FIEGARCH asymmetric GARCH model; historical volatility as the standard, the order of the predictive ability is that the GARCH model, Markov switching model, ARFIMA model; to sum up, Markov switching model is the best forecast model for three Hong Kong REITs. This study provides effective information to the supervision of Hong Kong REITs and it is useful for development of China future REITs.