عدم تقارن تاثیر قیمت معاملات سازمانی در بازار سهام چینی

The asymmetric price impact between the institutional purchases and sales of 32 liquid stocks in the Chinese stock market in 2003 is carefully studied. We analyze the price impact in both drawup and drawdown trends with consecutive positive and negative daily price changes, and test the dependence of the price impact asymmetry on the market condition. For most of the stocks, institutional sales have a larger price impact than institutional purchases, and a larger impact of institutional purchases exists only in a few stocks with primarily increasing tendencies. We further study the mean return of trades surrounding institutional transactions, and find that the asymmetric behavior also exists before and after institutional transactions. A new variable is proposed to investigate the order book structure, and it can partially explain the price impact of institutional transactions. A linear regression for the price impact of institutional transactions further confirms our finding that institutional sales primarily have a larger price impact than institutional purchases in the bearish year 2003.

The study of large trades is of crucial importance for both market risk estimation and personal investments, since large trades generally have a strong impact on the stock price and consequently increase investors’ costs [1], [2], [3], [4], [5], [6], [7], [8] and [9]. Meanwhile, growing evidence shows that large trades play a major role in trading in stock markets, which represent a large fraction of the market’s total trading volume [10], [11], [12], [13] and [14]. Therefore, there have been a variety of studies focusing on the price impact of large trades and the factors may influence their impact. Large trades generally refer to those transactions executed with large number of shares. Kraus and Stoll first studied block trades traded in blocks of 10 000 or more shares in the New York Stock Exchange, and found that block purchases have a larger permanent impact than block sales [15]. Keim and Madhavan studied block trades in the US markets, and found that there exist significant differences between the temporary and permanent impacts of buyer-initiated and seller-initiated trades [10]. Gemmill collected the 20 largest customer purchases and sales for each share in the London Stock Exchange, and found that the impact of purchases is larger than the impact of sales, both temporarily and permanently [16]. An alternative way of investigating large trades is to study institutional trading, since institutions generally have large amounts of capital and their transactions have, on average, a large number of shares. Chan and Lakonishok analyzed the price impact of the entire sequence of institutional trades in the New York and American Stock Exchanges, and found that the overall price impact after purchases and sales displays an intriguing asymmetry [17]. Chiyachantana et al. analyzed the institutional trading in 32 international stocks, and revealed that the price impact asymmetry remarkably depends on the stock market condition: the price impact is much higher for institutional purchases in the bull market, while institutional sales have a larger price impact in the bear market [18]. Explanations appear in the literature to account for the price impact asymmetry between large purchases and sales. Saar used a trading model to investigate the behavior of institutional investors, and claimed that the information difference between buys and sells may cause the asymmetry [19]. However, this information explanation is hard to test from empirical measurement. Frino et al. and Gregoriou attributed the price impact asymmetry to the bid-ask bias near open and close trades [13] and [20]. This price impact asymmetry can still be observed not using the opening or closing prices. Anderson measured the price impact using the transactions surrounding block trades, and related price effects to changes in order book depth [21]. To the best of our knowledge, few studies have been conducted on the price impact asymmetry of institutional trading in the Chinese stock market. The main purpose of this paper is to test if the price impact asymmetry between institutional purchases and sales exists in the Chinese stock market. We will further attempt to study the dependence of the price impact asymmetry on the detailed condition of the Chinese stock market. Though it is still not clear what really causes the price impact asymmetry, a new variable introduced to facet the volumes and gaps of different price levels may offer a description of the order book structure surrounding institutional transactions. More remarkably, this variable can partially explain the price impact of institutional trading. The rest of the paper is organized as follows. Section 2 introduces the data we analyzed and their summary statistics. Section 3 studies the price impact of institutional trading, and examines the asymmetry between the price impacts of purchases and sales. In Sections 4 and 5, we investigate the returns of trades and the order book structure related variable surrounding institutional transactions. Section 6 provides a regression model of explaining the price impact of institutional trading. Section 7 concludes.

In this paper, we have studied the price impact of institutional trading of 32 liquid stocks in the Chinese stock market in 2003, and find an asymmetric behavior between the institutional purchases and sales. Our studies of the Chinese stock market confirm the finding that institutional sales have a larger price impact than institutional purchases in the bare market, however with a relatively smaller impact asymmetry than other stock markets [13], [18] and [21]. To test the possible dependence of the price impact asymmetry on market condition, we divide the price time series into drawup and drawdown trends. We observe that institutional purchases have a larger impact than sales only in four stocks with primarily increasing tendencies. This may provide a weak evidence of large price impact of institutional purchases in the market maintaining an upward trend. We further investigate the mean trade-by-trade return surrounding institutional transactions, and observe an abundant phenomenon: a run-up (run-down) appears prior to institutional purchases (sales); the mean trade-by-trade return surges at the time when the institutional transaction occurs; a price reversal is observed immediately after institutional transactions; institutional trading has a permanent effect on the stock price later after the price reversal. Remarkably, the asymmetric behavior is also observed before and after institutional transactions, showing different mean trade-by-trade returns surrounding institutional purchases and sales. A new variable CC is proposed to investigate the order book structure surrounding institutional transactions. This order book structure related variable shows a behavior very similar to that of the mean trade-by-trade return, and large (small) trade-by-trade returns are accompanied by large (small) CC. The close relation between them may suggest that this variable could be regarded as one of the explanatory variables for price impact. Furthermore, the larger values of CC surrounding institutional sales may partially explain the larger impact of institutional sales on the stock price. A linear regression model is built to explain the price impact of institutional trading taking into account four explanatory variables, i.e., the capitalization of a firm, the order book structure related variable, a dummy variable denoting the order sign, and the previous volatility. The positive coefficient of the dummy variable provides further evidence that institutional sales have a larger price impact than purchases in the bearish year 2003. The coefficient of variable CC is positive and statistically significant, indicating that this variable is essential in explaining the price impact. Similar linear regressions of the price impact of institutional purchases and sales are performed separately, and the coefficient of CC for institutional sales is larger than that for purchases which further confirms the larger impact of sales on the stock price.