证券市场日内流动性的综合度量、特征与信息含量  

Intra-day Liquidity in Security Markets:Synthetic Measurement,Characteristics and Information Contents

作　　者：曹迎春[1] 刘善存[1] 邱菀华[1] 

机构地区：[1]北京航空航天大学经济管理学院,北京100083

出　　处：《系统工程》2007年第3期1-9,共9页Systems Engineering

基　　金：全国优秀博士论文作者专项基金资助项目(200466);国家自然科学基金资助项目(70671006;70372011);信息管理和信息经济学教育部重点实验室开放基金资助项目(F0607-32)

摘　　要：证券市场的流动性是一个包含量、价、时的多属性概念,现有的度量方法只涵盖了流动性的部分属性.本文综合考虑股票交易数量、价格波动性和交易时间, 提出用单位时间内价格波动一单位所能吸收的交易量来度量日内流动性的新方法.进而用这种方法研究中国股票市场的流动性日内模式,建立对数自回归条件流动性（LACL）模型对流动性建模,探寻拟合和预测流动性的最佳模型并分析流动性的特征.实证结果表明：流动性表现出与传统指标类似的显著的日内模式; 残差项服从Weibull分布的LACL模型（WLACL）拟合日内流动性、买方流动性和卖方流动性序列的效果最理想; 流动性过程是一个均值回复过程, 当前时刻的流动性对未来期望流动性有影响, 其影响呈Weibull形式衰减.最后, 通过核平滑估计得出个股在不同时点的流动性分布, 应用相对熵度量流动性分布的变化作为市场在交易中吸收的信息含量的代理变量, 发现10：30～11：30市场吸收的信息含量最多.The Liquidity is a conception of multiple attributes including volume, price and duration, but existing measurements of liquidity only describes part of the three attributes. Taking the trading duration,trading volume and price volatility into account, this paper firstly introduces a new approach of measuring intra-day liquidity by taking the volume absorbed per volatility in one second as proxy. Secondly, we study the intra-day pattern of liquidity using this approach. And put forward Logarithmic Autoregression Conditional Liquidity （LACL） model to search for the best model to fit and forecast the intra-day liquidity and further analyze the characteristics of liquidity. Our empirical study on Pudong Development Bank shows a same significant intraday pattern as traditional one, and has found that WLACL model in which error term follows Weibull distribution is the best model to fit and forecast intraday liquidity, buy-initiated liquidity and sell-initiated liquidity. We have also found that the liquidity process is a mean-reverting process. The impact of a given liquidity on future expected liquidity will die out in Weibull form. Finally, we have got liquidity distribution at different time by kernel smoothing technology. Taking the change of liquidity distribution, which is measured by relative entropy as proxy of information content of market absorbed by trading, we found the information content by absorhed market is the largest in 10：30-11：30.

关 键 词：金融市场微观结构 流动性 UFH-GARCH模型 LACL模型 信息含量 

分 类 号：F830.9[经济管理—金融学]