واکنش بازار کوتاه مدت پس از انجام معاملات متوقف در بازار بورس چینی

In this paper, we study the dynamics of absolute return, trading volume and bid–ask spread after the trading halts using high-frequency data from the Shanghai Stock Exchange. We deal with all three types of trading halts, namely intraday halts, one-day halts and inter-day halts, of 203 stocks in Shanghai Stock Exchange from August 2009 to 2011. We find that absolute return, trading volume, and in case of bid–ask spread around intraday halts share the same pattern with a sharp peak and a power law relaxation after that. While for different types of trading halts, the peaks’ height and the relaxation exponents are different. From the perspective of halt reasons or halt durations, the relaxation exponents of absolute return after inter-day halts are larger than those after intraday halts and one-day halts, which implies that inter-day halts are most effective. From the perspective of price trends, the relaxation exponents of excess absolute return and excess volume for positive events are larger than those for negative events in case of intraday halts and one-day halts, implying that positive events are more effective than negative events for intraday halts and one-day halts. In contrast, negative events are more effective than positive events for inter-day halts.

Financial markets are complex systems characterized by emerging extreme events [1]. It is a meaningful thing to answer the question how financial market dynamics are affected when the system undergoes an extreme event, such as financial crash, interest rate shock, large price change or trading halt. In recent years, statistical physics are applied to understand these financial markets dynamics, which are discovered sharing the feature of power law distributions. Early works were done by Lillo and Mantegna, who focused on relaxation dynamics of aftershocks after a crash. They researched the 1-min logarithm changes of Standard and Poor’s 500 index during 100 trading days after the Black Monday and discovered the decaying patterns in the rate of aftershocks larger than some threshold, as Omori’s law after earthquakes [2] and [3]. Weber et al. further found after crash period is characterized by the Omori process on all scales [4]. They studied the 1-min return series of three different data sets: S&P 500 index in the 15 000 trading minutes after Black Monday on 19 October 1987; quotes of 100 most frequently traded stocks at NYSE, NASDAQ and AMEX in two months after the crash on 27 October 1997; return series of General Electric stock in three months after 11 September 2001. In addition, the relaxation dynamics of aftershocks after large volatilities rather than large crashes were also investigated and similarly decay as a power law [5] and [6]. Furthermore, the aftershocks dynamics after US Federal Open Market Committee meetings that will announce interest rate change was described as an analog of the Omori law [7]. It is an example of aftershock dynamics that is clearly due to an external event. Apart from the dynamic of occurrence rate of aftershocks, another related topic is the relaxation dynamics of some financial measures after large price change/large bid–ask spread change. Zawadowski et al. examined high frequency data from NYSE and NASDAQ to conclude that volatility, volume and in case of the NYSE bid–ask spread increase sharply at the large intraday price change and decay according to a power law [8] and [9]. Ponzi et al. studied the dynamics of the bid–ask spread and the mid-price after a sudden variation of spread, and then found that the spread decays as a power law to its normal value [10]. Moreover, the order flow measures, such as the volume of different types of orders and the volume imbalance, were also discovered to increase before extreme events and decay slowly as a power law [11] and [12]. These dynamics can also be viewed as a type of switching phenomena in financial markets [13], [14] and [15]. These researches illustrate the scale-free behavior of trading volume both before and after the end of a trend [13] and [15]. A significant sudden jump of the volatility and then a power law decay can also be found for both microtrends in the German DAX future time series and macrotrends in the daily closing price time series of all S&P 500 stocks [14]. However, the discovery of Jiang et al. is a bit of different, that is, the volatility dynamics both before and after large fluctuations are symmetric in time scales of minutes, while asymmetric in daily time scales [16]. These analyses reveal that the asymmetry is mainly induced by exogenous shocks, whose precursory and relaxation dynamics in social systems have been studied by Refs. [17], [18], [19], [20] and [21]. So far researchers have studied the dynamics around financial crash, large price changes or interest rate changes with the technique of statistical analysis in high-resolution data. This paper will study another kind of extreme events commonly occurred in stock markets, namely trading halts, from this point of view. Trading halt is one of microstructure mechanisms in equity market designed to temporarily stop trading during the period of extremely price movement or of the announcement of significant events. Some financial papers [22], [23], [24], [25], [26], [27], [28], [29], [30], [31] and [32] have studied the effects of trading halts, such as price discovery, liquidity and volatility, from the perspective of information dissemination and transaction cost. While no unified pattern has been discovered and few studies focus on comparing the dynamics around different types of trading halts. Therefore, we attempt to find the unified behavior of different financial measures and to compare the relaxation dynamics around different trading halts. In this work, we investigate the relaxation dynamics of several financial measures around different types of trading halts in Chinese stock market. This paper is organized as follows. The next section provides data description. Section 3 characterizes the dynamics of cumulative return. Section 4 investigates the relaxation dynamics of absolute return, trading volume and bid–ask spread after trading halts. Section 5 studies the power law behavior after trading halts. Conclusions are provided in Section 6.

In this article, we investigate the short-term market reaction after trading halts in Chinese stock market. By analyzing the dynamics of average cumulative return, we find trading halts prevent the sustained rising or falling and play a certain role in price discovery. While for different types of trading halts, the stabilities of cumulative return in after-halt period are different. The average cumulative returns after intraday halts are most volatile, while those after one-day halts are most stable. This difference may relate with the complexity of the information contained in the halt announcement. By further analysis, we conclude that absolute return, trading volume, and in case of bid–ask spread around intraday halts all share the same pattern with a sharp peak and a power law relaxation after that, which is consistent with the typical signature of an exogenous shock [21]. While for different types of trading halts, the peaks’ height and the relaxation exponents are different. From the perspective of price trends, the peak of positive events is higher than that of negative events for all these three measures. This can be explained by traders’ asymmetric psychology. Furthermore, the relaxation exponents of excess absolute return and excess volume for positive events are larger than those for negative events in case of intraday halts and one-day halts. This means that positive events are more effective than negative events for intraday halts and one-day halts. In contrast, the relaxation exponents of excess absolute return and excess volume for positive events are smaller than those for negative events in case of inter-day halts. This means that negative events are more effective than positive events for inter-day halts. From the perspective of halt reasons or halt durations, intraday halts show the highest peak, and one-day halts show the lowest peak. In addition, the relaxation exponents of excess absolute return after inter-day halts are larger than those after intraday halts and one-day halts, which implies that inter-day halts are relatively more effective than intraday halts and one-day halts. These financial dynamics give us a better understanding of the market reaction after different types of trading halts. Also, comparative analysis based on these patterns can help supervisors to easily assess the effects of different trading halts and improve suspension rules. Of course, there are more things worthy of further study. For example, now we can only speculate the origin of these differences between different types of trading halts. It may be related to behavioral trading or information content around the trading halts. This needs to be further conformed. More importantly, researchers or investors can employ these different patterns to develop different “trading halts strategies”. Which type of halts is suitable for momentum strategies? Which type of halts is suitable for contrarian strategies? Are these strategies profitable? All these need to be further tested.