中国股市超高频数据反馈特征实证研究  

An Anatomy of the Feedback Characters:Evidence from China Stock Exchange High Frequence Data

作　　者：刘衡郁[1] 

机构地区：[1]上海财经大学,上海200433

出　　处：《运筹与管理》2008年第3期114-121,106,共9页Operations Research and Management Science

摘　　要：本文采用tick数据作样本,将分笔成交量分解为买入指令流和卖出指令流,依据向量自回归模型(VAR),研究1分钟和5分钟频段内收益率和净买入指令流二者之间关系,以揭示证券交易中的反馈特征。研究发现:1分钟系统存在正负反馈交易特征,5分钟系统存在负反馈交易特征;投资者交易行为存在粘性,频率越高,粘性行为越明显;当期反馈系统面临净买入指令流冲击时,收益率最初反应很大,随之减小,最后冲击效应趋于平稳。This paper adopts tick data as samples, studies the relationship between the 1-minute and 5-minute stock yields and the net buying order flow based on the VAR methodology, then discloses the feedback trading characters of the trader in the stock market. It is found that there are positive and negative feedback trading char- acters in the stock market when 1-minute yield is studied, only negative when 5-minute yield is studied. The be- haviour of traders at this period will affect the trader＇ behaviour at the next several periods. The higher the frequency is, the more periods will be affected. When the yield equation is affected by a shock of large buying order, the contemporaneous yield responses sharply, and the yields become calm gradually. The outcome of shock is that the calm yield is greater than the yield with no shock.

关 键 词：金融学 反馈特征 向量自回归 高频数据 

分 类 号：F830.91[经济管理—金融学]