Study On Efficiency Of Trading Halts In Chinese Stock Market Based On Agent-based Computational Finance

As an important price-stabilization mechanism, the trading halt mechanism canproduce high efficiency of information revelation, low price volatility and removeorder imbalanced. It has been widely used by the main markets all over the world.However, the implementation effect of the trading halts is widely debated. Thisdissertation uses the statistical analysis method for extreme events and the method ofagent-based computational finance to investigate the effectiveness of the trading haltsin Chinese stock market.Firstly, this dissertation uses the trading and trading halts data of Shenzhenstocks which are the component stocks of Shanghai&Shenzhen300Index from Aug2009to Aug2011to investigate the effects of the trading halt mechanism on Chinesestock market. By creatively borrowing the statistical analysis methods for extremeevents, this paper investigates the dynamics of price discovery and market volatilityduring trading halts and separate all halts into positive and negative halts. The resultsshow that the cumulative return changes considerably when resumption, and tends toa stable value subsequently. The trading halts suppress prices continue to increase ordecline and contribute to price discovery. We also find that the absolute return has asignificant peak around the trading halts, followed by a slow relaxation in bothpositive and negative events. After resumption, the price volatility has significantpower-law decay property. This result indicates that the trading halts lead to morevolatile, and it is unable to achieve the aim of stabilizing the market.Secondly, in order to find out what makes the change of the market bought bythe trading halts mechanism, this paper analyzes the reason from the perspective oforder imbalance and replaces the trading halts of removing the order book to clean upthe order imbalance. And then we simulate the trading halts mechanism on artificialstock market based on Mason platform and hope to find out the effect of orderimbalance. When analyzing the data generated by the simulation, we still use themethod employed in empirical analysis. The results show that the agent-basedcomputational model simulating the market change brought by the trading halts verywell. The results of simulation are the same with the empirical analysis results. Trading halts really make the absolute return have a significant peak around thetrading halts, followed by a slow relaxation with significant power-law decay property.Hence, we can analyze the reason based on agent-base computational finance. Weconclude that it is the order imbalance made the price volatility around the tradinghalts has the peak and power-law decay property.