Modeling and Forecasting Financial Volatilities:The Application of Conditional Autoregressive Range(WCARRX)Model

• Main Authors: I-Chieh Lin, 林依潔
• Other Authors: Eliza Wang
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/66462076763190810191

碩士 === 國立臺北商業技術學院 === 財務金融研究所 === 98 === The main prupose of this study is to apply the Conditional Autoregressive Range model(henceforth CARR)in modeling and forecasting the volatility of the Taiwan Stock Exchange Capitalization Weighted Stock Index(TAIEX). In contrast to the traditional return-based volatility model developed by Chou(2005). The sample used in this study covers 2233 daily observation, collecting from January 1,2001 to December 31, 2009. We define the first 1982 as in-sample data, and the remaining 251 observations are used as out-of-sample forest. In addition, we use the sum of squared intraday 5-minute returns as the proxy of realized volatility. In order to examine the return-volatility and volume-volatility relationship, we incorporate exogenous variables in the volatility equeation, including the lagged returns, absolute returns, turnovers, the number of trades, and trading volumes. The result shows that the lagged return, the lagged absolute return and the volume series provede extra explanatory powers over the conditional ranges in addition to the lagged ranges. In addition, we estimate the Weibull CARRX(WCARRX)model and compare the out-of-sample forecast of WCARRX with that of AR-GARCHX using different measures, such as the RMSE, MAE,and MAPE. The result shows that the WCARRX model dominates the AR-GARCH model, providing evidence in supporting the application of the WCARRX model in forecasting volatility.