An Empirical Study of the Anomailies on Taiwan Stock Spot and Futures Markets

• Main Authors: Ying-Sing Liu, 劉映興
• Other Authors: Jack J.W. Yang
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/32355336837696815790

博士 === 雲林科技大學 === 管理研究所博士班 === 97 === This study emphasizes on the effect correlation of five issues: time trend and seasonal effect, transaction tax reduction, turnover rate, firm-size, and book-to-market equity ratio. We will adopt the empirical data of Taiwan futures market covering the recently entered contracts of Taiwan Stock Index futures (TX), Electronic Sector Index futures (TE) and Finance Sector Index futures (TF) from February 1, 2002 to June 31, 2007 by observing the daily and intraday 5-minute interval trading data. Meanwhile, the common stock monthly trading data, TAIEX daily and monthly trading data from Taiwan Stock Exchange Market listed companies, during the period from June of 1982 to February of 2006 are also used. The mixture of distributions hypothesis (MDH), an important theory in the futures market, was proposed in discussion of the relationship between the futures return volatility and the trading volume. First, we employ the Flexible Fourier Form (FFF) to examine the time trend and seasonal effect of trading volume, and adopts TGARCH-cum-volume model to test the volatility-volume relation. Second, we prove that the return volatility also contains the time trend and seasonal factors. We employ the VAR-cum-FFF model to discuss Granger causality of daily return volatility and trading volume of the index futures at Taiwan futures market. Furthermore, we also apply FFF to the return volatility and volume activity in a bivariate equation structural framework to test the effect of transaction tax reduction for the return volatility and trading activity. Major empirical results indicate that tax reduction has a boosting effect on speculative trading activity, while futures return volatility has a significant correlation to time trend and seasonal factors. In the stock market, we employ Schwert and Sequin''s (1990) market model and dual time-varying beta market model to test the fact that the systematic risks of long-term turnover-sorted portfolios are non-stationary. We also adopts the Fama and French’s (1993) three-factor model and the dual time-varying beta’s three-factor- GARCH(1,1) model to examine the sensitivity and response orientation of three- factor betas of style portfolios to the sum of intra-monthly conditional market volatility in the bull and bear market months. Major empirical results indicate that there are positive turnover-size effect, reverse firm-size effect and reverse book-to- market effect in Taiwan Stock Market. Furthermore, only the constant market beta of the highest turnover-sorted portfolio appears to be asymmetric in bull and bear market months. The time-varying coefficient of higher turnover-sorted portfolios is significant in bear market months. Further, three-factor’s betas are sensitive to the sum of intra-monthly conditional market volatility, and the time-varying betas can explain part of the average return for most style portfolios, and finally the conditional heteroscedasticity does exist in the residual returns of the model.