رفتار گله ای سرمایه گذاران بازار سهام چین

This paper examines the existence and prevalence of investor herding behaviour in a segmented market setting, the Chinese A and B stock markets. It is the first study to detail the difference in herding behaviour across A and B markets. The results indicate that investors exhibit different levels of herding behaviour, in particular, herding strongly exists in the B-share markets. We also find that across markets herding behaviour is more prevalent at industry-level, is stronger for the largest and smallest stocks, and is stronger for growth stocks relative to value stocks. Herding behaviour is also more pronounced under conditions of declining markets. Over the sample period we are examining, herding behaviour diminishes over time. The results provide some indication to the effectiveness of regulatory reforms in China aimed at improving information efficiency and market integration.

Understanding the decision making process of various participants in the market has always been a challenging mission for academics and practitioners. Conventional theory of efficient markets asserts that markets are informationally efficient and investors form rational expectations of future prices, and that any new information entering into the market is instantaneously incorporated into expected prices in a homogenous manner. However, the efficient market hypothesis has been disputed both empirically and theoretically,1 and its major shortcomings in modelling real-life security returns have been noted by numerous past literature (see e.g., Shiller, 1989 and Summers, 1986). Behavioural economists however have attributed the imperfections in financial markets to various cognitive biases, human errors and responses. Herding activities among investors have been a popular behavioural explanation for the excess volatility and short term trends observed in financial markets. Investor herding causes prices to deviate from fundamental values and create implications for trading strategies and asset pricing models,2 thus it has received great attention in recent years. Human herding behaviour usually results from a tendency to imitate the actions of others. The definition proposed by Christie and Huang (1995, p31) is “individuals who suppress their own beliefs and base their investment decisions solely on the collective actions of the market, even when they disagree with its predictions”. One line of research explains the presence of herding behaviour among market participants through investor psychology, regarding herding among investors as irrational behaviour. Devonow and Welch (1996), for example, propose that investors disregard their own beliefs and follow other investors blindly due to an intrinsic preference for conformity with the market consensus. Pretcher (2001), on the other hand, seeks to explain herding behaviour from a neuroeconomic perspective. He argues that human herding behaviour, like other primitive instincts for survival, results from impulsive mental activity in individuals responding to signals from the behaviour of others, and these impulses are typically faster than rational reflections in emotionally charged situations. Meanwhile, another line of research treats herding behaviour as rational actions. Followers of this line of argument are of the view that there is an important link between rationality and emotion in decision making, in that psychological factors may be compatible with the optimizing behaviour of the agents. Bikhchandani and Sharma (2000) emphasize the distinction between intentional herding, which results from an obvious intent by investors to imitate others, and spurious herding, in which groups of investors facing similar decision problems and information sets take similar decisions.3Banerjee (1992) and Bikhchandani, Hirshleifer, and Welch (1992) argue that while individuals cannot access other investors' private information, such information is revealed through the actions taken by these investors. By observing prior investors' actions, invest/reject cascades can occur depending whether the first few investors all chose to buy/sell. Such herding externality results in inefficient outcomes and the cascades are also path-dependent and idiosyncratic, easy to reverse when more instructive public information enters the market. Yang (2011) argues that herd behaviour of investors is influenced by the precision of costly private signals and signal extraction. The evidence of herding behaviour among the market participants has direct implication to the market information efficiency as well as asset pricing behaviour. As a relatively new market growing at a staggering speed, the Chinese stock market with its unique macro- and microstructure features provides an interesting setting for the analysis of investor herding behaviour. Previous studies have found that when investors face few alternatives and heavy government involvement, they tend to speculate in the stock market, generating significant volatility (Green, 2003). Thus in a market which has traditionally been characterised by unsophisticated retail investors, heavy regulations and a lack of transparency, and which is now undergoing tremendous amounts of reform, the understanding of how investors behave amidst such process of transition is worthwhile and important. In this paper we investigate the presence of herding behaviour in the segmented Chinese mainland market, by examining the return dispersion of both A and B share markets. Utilising a recent and comprehensive dataset, we find no evidence of herding in A-share markets but significant evidence of herding in the B-share markets over the period 1999 to 2008. Our investigation shows that herding behaviour is more prevalent at industry level than at market level, among stocks with the largest and the smallest market capitalisation, among growth stocks than value stocks, and during periods of declining stock markets. Our results also reveal that the level of herding diminishes over the sample period, an indication to the effectiveness of regulatory reforms in China aimed at improving information efficiency and market integration. The remainder of the paper is structured as follows. The next section provides a review of relevant literature regarding herding behaviour. Section 3 presents the institutional background of the Chinese stock market. Section 4 describes the data that will be used in this study. Section 5 outlines the research design and the methodology. Section 6 reports the empirical results. Section 7 concludes.

In this paper, we examine the investment behaviour of market participants within the Chinese stock market, specifically with regard to their tendency to conform towards the market consensus, otherwise known as herding behaviour. The testing methodology used in this paper based on the approaches of Christie and Huang (1995) and CCK (2000), where equity return dispersions, measured by the cross-sectional standard deviation of returns, is used to detect herding behaviour among investors. In our empirical tests we adopt a modified version of the model, which corrects for the multicollinearity and autocorrelation problems presented in the dataset. Our results yield interesting findings about investor behaviour in the Chinese stock market. Tests conducted on the 10 year period from 1999 to 2008 reveal significant evidence of herding behaviour in the Shanghai and Shenzhen B-share markets, but no evidence of herding in the A-share markets. When the data is broken up into sub-periods however, it can be seen that while all four markets exhibited significant herding behaviour at the beginning of the decade, such behaviour has diminished over time and in the A-share markets investors now seem to make more rational investment choices. In addition, herding behaviour is found to be more prevalent at industry level than at market level, among the largest and smallest stocks, and for growth stocks relative to value stocks. We also find that herding behaviour is stronger during periods of market decline for both the A-share markets and the Shanghai B-share market. The results of the robustness tests show that herding behaviour persists even when changes in market liquidity are controlled for. These findings regarding investor herding behaviour have important investment and policy implications. First of all, the findings of different investment behaviours in the A-share and B-share markets suggest that it may not be appropriate to apply a universal asset pricing model to the markets. Secondly, in the B-share markets where persistent herding behaviour is documented, portfolio diversification strategies may not be as effective as they would in a herd-free market. That is, a greater number of securities are required to achieve the same level of diversification than an otherwise normal market. This may be an important implication for foreign fund managers wishing to trade in the Chinese stock market. In addition, to the extent that evidence of herding behaviour is indicative of relative market inefficiencies, our results suggest that the quality of information disclosure in the Chinese market has, generally speaking, improved substantially over the last decade. This may be a useful result for Chinese policymakers, showing that the reforms the Chinese stock markets underwent (and is still undergoing), such as those relating to the stabilisation of share prices, tightening of accounting standards and disclosure policies and the removal of restriction on foreign investment, have achieved desirable effects in the Chinese market. The findings also suggest that the Chinese government can better regulate its markets by targeting stocks of certain characteristics (e.g. stocks and industries which are more susceptible to herding) and under certain market conditions. It should be noted that this study employs an approach of herd detection which looks for evidence of a particular form of herding, namely herding towards the market consensus. The findings of this paper do not preclude the possibility that other types of herding behaviour exist in the Chinese stock market. Also, as Bikhchandani and Sharma (2000) point out, the current statistical specification used in empirical studies still requires further refinement in distinguishing true herding from spurious herding. Further research on herding using more powerful models will provide greater insights into the investment behaviour of market participants.