格兰杰因果关系检验可以定义为统计假设检验，用于确定时间序列中的一个。当学者们对概念股票进行研究时，最常用的是时间序列模型。这是因为股票价格波动和市场投资高度依赖于波动的持续时间，并考虑了不同的时间序列模型。Engle和Victor K. Ng(1993)使用了不同的时间序列模型来研究新闻在股票市场中产生的波动。根据Eun和Shim(1989)，可以认识到概念股票可能在不同的时区运行，每个时区有不同的开盘和收盘时间。正因为如此，才引起了投资者的极大关注(Hiemstra et al .， 1994)。
Impulse response function is a yet another function which can be considered important, as it gives the output in the presence of a brief input signal (Whitelaw, 1994). Thus an impulse response may refer to the reaction of any kind of dynamic system in response to one or another kind of external change. Thus, the impulse response describes the reaction of a specific system as a function of time. Impulse response function test will be done in order to examine the reasonable amount of assumption and to do the analysis of the co-integration equation.
According to the current literature, the Chinese concept stocks are volatile. Thus, it is necessary to understand the Chinese concept stocks and its relationship with the investors’ attention with the help of the ginger causality test, variance decomposition analysis and the impulse response function test. Thus the research hypothesis can be:
The volatility associated with the Chinese concept stocks and its relationship with the investors’ attention
One of the major challenges which has to be faced in the empirical analysis is the lack of a direct measurement of the foreign investor attention. There may be a few indirect proxy variables such as extreme returns, turnover of the concept stocks, the headlines and the advertising expenses. In order to do the research, the information available in Oriental Fortune Forum has been used. This is the most popular stock forum in China. A number of sample stocks have been collected in CSI 300 Index from January 2009 to December 2013. Another data used for comparison is from SSE Composite index from CSMAR of the same scope.
In the above figure, the relationship between the investor’s attention and the SSE Composite index has been analyzed. It can be observed that the two variables are moving broadly together. Also, whenever there is an increase in the number of posts, there is a rise in the concept stock price. On the basis of the monthly number of posts which were 190997 in the year 2009 where the value of the peak index is 3412, the corresponding number of posts was only 46101 in October 2012.
The granger causality test may be defined as the statistical hypothesis test for the purpose of determination of one of the time series. When the scholars are doing the investigation related to the concept stocks, the time series models are most commonly adopted ones. This is because the stock price fluctuations and the investments in the markets are highly dependent on the time duration of the fluctuation and different time series models have been taken into consideration. Engle and Victor K. Ng (1993) have use different kinds of time sequence models in order to study the fluctuation which is created by the news in the stock markets. According to Eun and Shim (1989), it can be recognized that the concept stocks may operate in different time zones, each having different opening and closing hours. Because of this, there is a huge investor’s attention (Hiemstra et al, 1994).