基于Mike-Farmer委托驱动模型的研究  

Mike-Farmer order-driven model

作　　者：顾高峰[1] 任飞[1] 蒋志强[1] 周炜星[1] 

机构地区：[1]华东理工大学商学院,上海200237

出　　处：《上海理工大学学报》2011年第5期457-472,共16页Journal of University of Shanghai For Science and Technology

基　　金：国家自然科学基金资助项目(11075054;10905023;71101052);上海市晨光计划人才资助项目(CG201032)

摘　　要：Mike-Farmer微观模型是功能强大的委托驱动模型,能再现很多经典的统计规律.本文介绍了Mike-Farmer委托驱动模型的构建过程,Mike-Farmer委托驱动模型生成的收益率,发现收益率在不同时间尺度下遵循幂律分布,服从负三次方定律.以Mike-Farmer委托驱动模型为平台,进行收益率幂律分布和波动率聚簇效应的成因研究,发现收益率的幂律分布和市价订单委托价格的概率分布相关,而波动率的聚簇效应与订单委托价格时间序列的时间记忆性保持一致性.最后简要介绍了模型的应用前景.Mike-Farmer microscopic model is a powerful order-driven model. It can reproduce many stylized facts. In the paper, the Mike-Farmer model was introduced in detail. It was found that the distributions of returns obtained from the model at different timescales can be modeled as the power-law distribution in the tails, whose exponents are close to the well-known cubic law. The reasons for the power-law distribution of returns and the clustering effect of volatilities were then studied and it was concluded that power-law tails are caused by the power-law tail in the distribution of market order prices and the clustering effects are related to the long memory in the time series of submitting order prices. The application of the model was briefly presented.

关 键 词：金融物理学 委托驱动模型 收益率幂律分布 波动率聚簇效应 

分 类 号：F830.91[经济管理—金融学]