تحقیقات تجربی از رفتار توده وار در بازار بورس چینی: شواهد حاصل از تجزیه و تحلیل رگرسیون
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|17426||2010||14 صفحه PDF||سفارش دهید||8363 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Global Finance Journal, Volume 21, Issue 1, 2010, Pages 111–124
This study examines the herding behavior of investors in Chinese stock markets. Using a least squares method, we find evidence of herding within both the Shanghai and Shenzhen A-share markets and no evidence of herding within both B-share markets. A-share investors display herding formation in both up and down markets. However, we cannot find herding activity for B-share investors in the up market. By applying quantile regression analysis to estimate the herding equation, we find supporting evidence of herding behavior in both A-share and B-share investors conditional on the dispersions of returns in the lower quantile region.
In the recent finance literature, empirical analysis of herding behavior has received considerable attention in studies examining the grouping behavior of investors. The importance of investigating herding behavior stems from the fact that investors, following the actions of others, tend to form a collective decision that, in turn, drives stock prices away from their underlying fundamental values. The resulting divergence between market price and fundamental value offers arbitragers an opportunity to reap excess profits. A long-run consequence of this herding behavior may lead to greater instability and inefficiency if the market correction fails to make the market price and the fundamental value converge. Numerous papers have been devoted to the study of herding activities in global markets.1 For instance, by applying firm-level data to examine whether investors in global markets have a tendency to exhibit herd behavior, Chang, Cheng, and Khorana (2000) find significant evidence of herding in South Korea and Taiwan and partial evidence of herding in Japan. However, there is no evidence of herding for market participants in the U.S. and Hong Kong. The evidence indicates that herding behavior is more likely taking place in emerging markets. Following the same approach, Demirer and Kutan (2006) test whether investors in Chinese markets, in making their investment decisions, are following market consensus during periods of market stress rather than private information. Their test results find no evidence of herd formation, suggesting that market participants in Chinese stock markets make investment choices rationally. However, a recent study of Chinese stock markets by Tan, Chiang, Mason, and Nelling (2008) reports that herding occurs under both rising and falling market conditions. This herding phenomenon is more profound in A-share investors.2 Thus, the evidence on herding behavior in the Chinese markets is inconclusive. This paper attempts to provide new empirical evidence that helps to resolve the mixed findings of herding behavior in Chinese markets. Our study is also motivated by the inadequacy of using aggregate data in analyzing herding behavior. In particular, previous studies of Chinese stock market behavior customarily follow the classifications A-share vs. B-share markets or Shanghai stock exchange vs. Shenzhen stock exchange. While this conventional approach offers a general direction for herding activity in terms of market classification, its drawback is that it fails to provide precise information that explains behavioral changes conditional on a particular market condition. The approach used by Christie and Huang, 1995, Chang et al., 2000 and Gleason et al., 2004 attempts to argue that the formation of herds is more likely to be present during periods of market stress, since investors are more likely to suppress their own beliefs and use market consensus during large changes in price. The testing methodology thus suggests that equity return dispersions are sensitive to aggregate market returns squared, especially during periods of market stress. If applying least squares estimators produces a negative coefficient on the market return squared term, it would suggest the existence of herding behavior. It is generally recognized that least squares estimators are based on the mean of the conditional distribution of stock return dispersions. Such a model specification is inconsistent with behavior involving stressful environments, since the tail information has not been addressed. The innovation of this paper is to examine the data conditional on different quantiles and test the behavioral relation between stock return dispersions and aggregate market movements with different quantile distributions. An additional benefit from using quantile regression is that some of the statistical problems, such as errors in variables, sensitivity to outliers, and non-Gaussian error distribution, can be alleviated ( Barnes & Hughes, 2002). The remainder of this paper is organized as follows. Section 2 presents the methodology used to detect herding behavior. Section 3 describes the data. Section 4 reports evidence of herding behavior based on a least squares estimator by organizing Chinese stock data into aggregate level, A-share and B-share groups, and four sub-markets. Section 5 presents a quantile regression method and applies it to estimate the herding equation. Section 6 concludes the paper.
نتیجه گیری انگلیسی
This study examines the herding behavior of investors in the Chinese stock markets. Estimations are based on daily data classified into aggregate market, A-share and B-share groups, and four sub-markets: SHA, SHB, SZA, and SZB. The test results consistently show that investors in the aggregate market display herding behavior. Dividing the data into A-share and B-share markets offers evidence that A-share investors consistently display herding behavior, but B-share investors do not reveal such a phenomenon. We examine possible behavioral changes by using different market conditions. The statistical results on A-share investors indicate that herding is present in both up and down markets. However, we are unable to find evidence of herding for B-share investors on days when the market is up. We further test the herding equation by employing a quantile regression model, which is based on the distributions of the dependent variable (return dispersions) and the estimations are made by using the sample points conditional on a specific quantile. The evidence shows that herding behavior is more prevalent at the median and lower tails of the quantile distributions of the return dispersions. This behavior holds true for the evidence derived from the Chinese aggregate stock market, the A-share markets group (SHA and SZA) and the B-share markets group (SZB and SZB), and individual sub-markets (SHA, SZA, SHB, and SZB), although the evidence for the SZB is less significant compared with the other cases. This finding casts some doubt on the empirical conclusion that A-share market investors display herding activity, while B-share market investors do not. By using a quantile regression procedure, this study finds that B-share investors consistently exhibit herding behavior in the quantiles from the 10% to 50% levels on days when stock market returns are up. Thus, the mixed results in the literature are mainly due to the fact that in empirical estimations, the model fails to capture the asymmetric responses of market returns and ignores the distributional information from the quantile regression approach.