A Comparison of Alternative Forecast Models of REIT Volatility

UNCP Author/Contributor (non-UNCP co-authors, if there are any, appear on document)
Dr.. Zhixin "Richard" Kang, Associate Professor (Creator)
The University of North Carolina at Pembroke (UNCP )
Web Site: http://www.uncp.edu/academics/library

Abstract: This study compares the relative performance of several well-known models in the forecasting of REIT volatility. Overall our results suggest that long-memory models (ARFIMA & FIGARCH) provide the best forecasts. Using either a large sample or some statistically justified small subsamples, we find that long memory models consistently outperform their short-memory counterparts (GARCH & Stochastic Volatility models) over a variety of forecast horizons. We also find that asymmetric models (EGARCH & FIEGARCH) are the worst performers among all models. Our study complements and extends a recent study of Cotter and Stevenson (2008) which demonstrates the usefulness of long-memory models in modeling REIT volatility. We conclude that in addition to modeling REIT volatility, long-memory models should also be adopted to forecast REIT volatility.

Additional Information

The Journal of Real Estate Finance and Economics 42.3 (2011)
Language: English
Date: 2011
REIT Volatility, Forecasting , Short-Memory Model, Long-Memory Model

Email this document to