中国证券市场信息结构与信息风险测度研究

发布时间：2018-04-03 17:59

本文选题：市场微观结构　切入点：信息结构　出处：《天津大学》2012年博士论文

【摘要】：信息是投资者制定投资策略的根本依据，直接影响了金融市场资产价格的发现过程以及资产的价格行为。对信息的深入研究，是开展证券市场微观结构理论研究的基石与必经之路。基于一个系统性的视角，就信息在不同时间尺度下的多维度结构差异，及其对投资者行为作用机理的区别与联系，本文分别深入的研究了低频与高频状态下信息结构的测度与信息风险的度量问题。如下为研究的主体内容摘要：中国股票市场日内信息结构的测度研究：在日内时间维度下，基于信息融入资产价格的非瞬时性，应用隐马尔科夫模型对不可观测的股票信息状态建模，并通过转移概率矩阵刻画信息状态在时间维度上的动态关联性。通过实证验证了模型的信息识别能力，且发现中国股票市场信息效应具有聚集性的特点。实证还估计了样本的信息状态与信息强度，进一步的，，通过信息状态转移概率矩阵，推断我国股票市场的信息融入速率。中国股票市场日内高频信息风险度量研究：基于日内信息组成结构的时变特性，推导时变知情交易概率的非参数计算表达式，并构建测度日内高频信息风险的方法。应用此方法设计实证验证了模型能够实时捕捉日内不断变化的信息风险状态,且对由有毒信息流引起的资产价格突变具有预测功能。中国股票市场日间信息结构的测度研究：在日间的时间频度下，基于投资者能够从每日的交易数据推断日间变化的交易类型组成结构的假设，构建了一个允许知情和未知情交易到达率时变且可预测的GARCH结构信息模型。应用该模型设计实证，研究了中国股票市场投资者学习市场交易信息并调整其交易行为的动态过程。并在此基础上构建日间时变的知情交易概率指标，进一步的研究真实金融市场中信息结构在日间频度下的运动变化模式。中国股票市场日间信息结构的判断与信息风险度量研究：首先，通过将对称订单流冲击引起的交易应用于信息模型的构建，解决以往模型蕴含订单数值特征与市场实际不相符的问题。然后，基于新模型的理论基础模拟买卖订单到达数据并绘制其频数分布图，据此判断个股信息结构的类型与适用的模型。最后，理论分析交易者的日度学习行为与形成稳定信息结构的关系，阐述低频状态下资产信息风险测度的原理，并应用新模型实现对个股信息风险程度的度量。
[Abstract]:Information is the fundamental basis for investors to formulate investment strategies, which directly affects the discovery process of asset prices and the price behavior of assets in financial markets.The further study of information is the cornerstone and only way to carry out the research on the microstructure theory of securities market.Based on a systematic perspective, the difference and relationship between the multi-dimensional structure of information at different time scales and the mechanism of its action on investors' behavior are discussed.In this paper, we study the measurement of information structure and the measurement of information risk in low frequency and high frequency respectively.The following is a summary of the subject of the study:Research on the Measurement of Intra-Day Information structure in Chinese Stock Market: under the dimension of intra-day time, based on the non-instantaneous integration of information into asset prices, the invisible Markov model is used to model the unobservable stock information state.The dynamic correlation of information state in time dimension is described by transition probability matrix.The information recognition ability of the model is verified by empirical analysis, and it is found that the information effect of Chinese stock market has the characteristics of agglomeration.The paper also estimates the information state and the information intensity of the sample, and further, through the information state transition probability matrix, infer the information integration rate of the stock market in our country.Research on intraday high-frequency information risk measurement in Chinese stock market: based on the time-varying characteristics of intraday information structure, the non-parametric expression of time-varying informed trading probability is derived, and a method to measure intraday high-frequency information risk is constructed.By using this method, it is proved that the model can capture the changing information risk state in real time, and it can predict the sudden change of asset price caused by toxic information flow.A study on the Measurement of daytime Information structure in Chinese Stock Market: based on the assumption that investors can infer the composition of daytime trading types from daily trading data,A GARCH structure information model with time-varying and predictable arrival rates of informed and unknown transactions is constructed.By using this model, the dynamic process of investors learning trading information and adjusting their trading behavior in Chinese stock market is studied.On the basis of this, we construct the time-varying probability index of informed transaction between day and day, and further study the movement mode of information structure in the real financial market under the daytime frequency.A study on the judgment of Day Information structure and the Measurement of Information risk in Chinese Stock Market: firstly, the transaction caused by the shock of symmetric order flow is applied to the construction of information model.To solve the problem that the model contains the numerical characteristics of the order and the market reality does not match.Then, based on the theory of the new model, we simulate the arrival data of purchase and sell orders and draw its frequency distribution map, according to which we can judge the type of information structure of individual stock and the applicable model.Finally, the relationship between the daily learning behavior of traders and the formation of stable information structure is analyzed theoretically, and the principle of asset information risk measurement under low frequency is expounded, and the new model is applied to measure the degree of information risk of individual stock.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：F832.51;F224

【共引文献】