Agent-based Modeling For Multiple Order Books And Market Dynamics Analysis In Stock Index Futures And Stock Markets

The investors in Chinese stock market have shown typical irrational behaviors and hot spot chasing behavior.On the one hand,they provide important liquidity to the market by submitting limit orders.However,on the other hand,due to the irrational behaviors and homogeneous behaviors,they may also cause shocks on market liquidity,triggering large volatility in market prices and even complicated market dynamics.Therefore,it is of great importance to explore the influence of investors' irrational factors on such limit order markets as China.The existing microstructural theories typically assume that order submission decisions are based on rational expectations framework,hence ignoring the irrational behaviors of investors and separating asset allocation process from order submission decision process.Therefore,there are extremely few multiple order books models that can manifest the characteristics of real stock markets.With the method of agent-based modeling,this thesis builds a multiple order books model to depict the adaptive asset allocation behaviors of investors in multi-assets allocation,the irrational behavior biases in multi-assets sales,and the cross-market trading behaviors of investors,with the aim of exploring the influence of these investor behaviors on the dynamics of futures and stock markets.First of all,as for the characteristic of “following others” of investors in the process of asset allocation,this thesis models investor heterogeneous beliefs and adaptive asset allocation behavior,and builds a multiple order books model that is designed to allocate cash asset based on the realized return of stocks.Research shows that adaptive asset allocation leads to the significantly positive correlation in order flows(diagonal effect);meanwhile,there is an inverse U-shaped relation between the intensity of choice of adaptive asset allocation and the commonality effect in liquidity and return of multiple order books.In addition,with chartist beliefs,a high intensity of choice of investors' adaptive asset allocation may result into market meltdown and thus loss of investors' wealth.Next,concerning the irrational behavior biases of investors revealed in multi-assets sales and in order to further improve investors' adaptive asset allocation,this thesis builds a multiple order books model that can directly make stock allocation based on the forecast of investors.Meanwhile,this model built depicts the behavioral bias of investors in multi-assets sales,that is,disposition effect.Research suggests that disposition effect decreases market pricing efficiency,lowers market volatility,and increases market liquidity,thereby resulting in wealth loss of investors.In addition,disposition effect also significantly affects investors' order submission behaviors in limit order markets.Finally,in order to explore the role of stock index futures in stock market crash,this thesis has established a stock index on the basis of the built multiple order books stock model,and by relying on the index formed,built and calibrated an agent-based stock index futures model.Research suggests that speculators provide liquidity for the stock index futures market,while both arbitragers and hedgers significantly decrease the volatility of stock market and increase its pricing efficiency.No matter when the stock market is in smooth running or faces extreme volatility and even collapse,stock index futures have played a positive role in stabilizing the stock market,and show no asymmetric cross-market impact on the stock market.