1、 2010 届本科毕业论文 外文翻译 姓 名： X X X 学 号： 2006XXXXXXX 系 别： 金融与贸易系 专业班级： 06 级国际经济与贸易 X 班 英文原版： The Impact of Introduction of QFII Trading on the Lead and Volatility Behavior: Evidence for Taiwan Index Futures Market Wen-Hsiu Kuo Department of Business Administration National Cheng Kung University and Depart
2、ment of Finance Ling Tung University 1, Lingtung Road, Nantun 408 Taichung City, Taiwan whkuomail.ltu.edu.tw Shih-Ju Chan Department of Business Administration Kao Yuan University 1821,Chung-Shan Rd., Lu-Chu Hsiang Kaohsiung County 821,Taiwan 3. Methodology 3.1. Cointegration test and vector error
3、correction model First, we test the market efficiency hypothesis (MEH) in Taiwan index futures market by examining whether the cointegrated relationship (i.e., long-run equilibrium relationship) among futures, spot prices and several macroeconomic factors exists before and after the opening up of fu
4、tures market to QFII. Given that the five variables are integrated of order one, the cointegration test proposed by Johansen and Juselius (1990) is performed. If there are cointegrated relationships among futures, spot prices and several macroeconomic factors, then we suggest that some market ineffi
5、ciency exists in Taiwan index futures market. Second, for cointegrated series, Granger causality tests need to be performed in the corresponding VECM framework according to the Granger Representation Theorem proposed by Engle and Granger (1987). This study employs the VECM to examine whether the lea
6、d-lag relationship between the futures and spot markets differs for the pre- and post-QFII periods. To control effects of macroeconomic factors on the relationship between the futures and spot markets, we incorporate the macroeconomic factors into the VECM. Therefore, this paper adopts the following
7、 VECM7 framework with five variables to study the lead-lag relationship between the futures and spot markets for the pre-QFII, post-QFII, and whole periods, respectively Additionally, in order to test the impact of structural change due to the introduction of QFII on the short term dynamics and long
8、-run error correction term between stock index and stock index returns, a dummy variable (dt) is introduced into Eqs. (1) and (2) for the whole period.8 The modified model may be specified as follows: where dt is a dummy variable that takes on a value of 1 if observation t lies within the post-QFII
9、period, otherwise 0. If the dummy is statistically significant, then the introduction of QFII has an impact on the lead-lag relationship between futures and spot markets. 3.2. Switching GJR-GARCH (1,1) and standard GJR-GARCH (1,1) models In analyzing the impact of the opening up of the futures marke
10、t to QFII on the level and nature of futures price volatility, there are two issues that need to be addressed. First, does the existence of QFII trading in itself have any effect on volatility? Second, if the existence of QFII trading affects volatility, how does it? To address the first issue, a sw
11、itching GJR-GARCH (1,1) model is employed to examine the impact of the existence of QFII trading on the level and nature of futures price volatility for the whole sample period. Lee and Ohk (1992) present the modified GARCH model, which imposes an autoregressive structure on conditional variance and
12、 captures the change in the level and slope of time-varying volatility using dummy variables. This modified model is called the switching GARCH model. But the switching GARCH model is connected with the shortcoming that it assumes a symmetric response to news and fails to account for observed asymme
13、try in the market. Glosten, Jagannathan, and Runkle (1993) propose a GJR-GARCH model, which can capture the asymmetric impact of shocks on volatility. Hence, in the spirit of Lee and Ohk (1992) and Glosten, Jagannathan, and Runkle (1993), we propose the conditional variance equation of GJRGARCH (1,
14、1) with a dummy variable dt. As stated in the previous section it is necessary to remove the influences of macroeconomic factors on futures market volatility by incorporating control variables into the mean equation. Consequently, the modified model, switching GJR-GARCH (1,1), can be specified as follows:
