异质从众行为与股价波动的Agent仿真研究

Simulation of Stock Price Volatility Based on Agent-based Heterogeneous Herd Behavior

已有研究尚未明确证券市场从众行为与过度波动这一金融异象的关系,针对忽略从众行为本身也存在着差异、未将成交量纳入分析框架等问题,从行为金融学视角出发,构建基于Agent的异质从众行为人工市场仿真平台。首先通过在模型中设定6种不同的从众概率,推演了不同从众行为对证券市场股价波动的异质性影响;其次验证了异质从众概率下价格波动与市场成交量关联的强弱差异。仿真结果表明:异质性从众行为带来不同程度的股价波动,但完全从众行为下无市场交易;随着从众行为概率的提高,股价均值不断上升,成交量对于价格波动的贡献率也不断增加,影响的持续性也在增加;适当的从众行为可以增加证券市场活力,但较高的从众行为将会推高市场的风险,降低市场的有效性。

The existing research has not yet clarified the relationship between the herd behavior and excessive volatility. In order to solve the problems of neglecting herd behavior itself, such as differences and not incorporating trading volume into the analytical framework, this paper constructs an agent-based artificial market simulation platform for heterogeneous herd behavior from the perspective of behavioral finance. Firstly, by setting six different probabilities in the model, the heterogeneity effect on stock volatility was deduced. Secondly, the difference relevance between the price volatility and the market volume under the heterogeneous herd probability was verified. The simulation results show that the heterogeneous behavior leads to different levels of stock volatility, but there is no market trading under the complete herd behavior;as the probability of herd behavior increases, the average of stock price, the contribution rate of volume to price volatility and the persistence of influence are all increasing;proper herd behavior can increase the vitality of the stock market, but higher herd behavior will push up the market risk and reduce the effectiveness of the market.