期现货市场间信息溢出效应研究

Analysis of information spillover effect between futures market and spot market

采用2000年7月10日到2006年6月30日铜期货与现货价格的日度数据,运用GARCH 模型估计了两个市场的上涨和下跌的条件 VaR,并利用基于回归的线性-Granger 因果检验,基于核函数的均值-Granger 因果检验,波动率-Granger 因果检验及风险-Granger因果检验方法分析了期、现货两个市场间的信息溢出效应.这是首次采用基于核函数的检验统计量来研究期货与现货间的关系,由于本方法可以使用所有的滞后阶数,从而使得检验效力更强.考虑到期货市场买空卖空的运行机制,还提出了上涨 VaR 和极端上涨风险溢出两个新的概念.实证结果表明,期货市场与现货市场间存在双向的信息溢出效应,进一步的分析表明从期货市场到现货市场的信息溢出要显著强于从现货市场到期货市场的信息溢出.

This paper adopts GARCH model to estimate the conditional VaR in both downside and upside direction. The paper further utilizes the linear Granger causality test based on regression, Granger causality test in mean, Granger causality test in variance and Granger causality test in risk based on kernel function to study information spillover effect between the futures market and the spot market of copper, using daily data of copper price of two markets ranging from July 2000 to June 2006. This paper uses the kernel based statistical test to explore the correlation between the futures market and the spot market for the first time. Because all the lagged orders are used, the test statistics has a relatively strong power. Considering there are long and short in the futures market, we also introduce two new conceptions which are the upside VaR and upside risk spillover. Our findings indicate that there exists extremely significant two-way information spillover effect between the futures market and the spot market of copper. Further analysis reveals that the spillover effect from the futures market to the spot market is more significant than that from the spot market to the futures market.