流动性风险与市场风险的集成风险度量研究

【作者】 李徐；

【导师】 张金清；

【作者基本信息】 复旦大学 ， 金融管理与金融工程， 2008， 博士

【摘要】 本文研究的是如何在同一个框架内度量流动性风险与市场风险,即流动性风险与市场风险的集成风险度量。本文使用copula函数构建了流动性风险与市场风险的集成风险度量模型,旨在探讨如下三个问题:(1)如何度量单项资产中由流动性风险因子与市场风险因子驱动的集成风险。(2)如何度量资产组合中由多组流动性风险因子与市场风险因子驱动的集成风险。(3)如何在上述研究的基础上对集成风险度量模型进行了动态分析。在构建集成风险度量模型之后,本文结合中国A股市场数据进行实证检验,度量了不同流通规模股票的集成风险,并将本文模型与现有模型进行了比较。本文共分为9章:第1章绪论。开篇提出问题,然后介绍本文的主要创新点和难点,并对研究方法和框架进行总体规划。第2章文献评述。首先对集成风险的概念进行解析,然后对流动性风险度量、市场风险度量以及两者集成度量的现有研究进行评述,并对基于copula函数的集成风险度量研究及其相关金融研究进行综述。第3章引入copula函数度量流动性风险与市场风险的集成风险,其中运用copula函数的关键在于怎样选择最优copula函数。第4章对股票进行集成风险分析。把流动性风险与市场风险确定为中国A股的主要风险,并采用半参数方法分别为流动性风险因子与市场风险因子建模。建模之后,从理论和实证两个方面对股票流动性风险因子和市场风险因子的相依性进行了考察,从而为集成风险度量做准备。第3章和第4章构成了集成风险度量的两块基石:一块是集成风险度量的方法——最优copula函数理论;另一块是集成风险度量的原动力——市场需求,即中国A股风险因子的现实情况。第5章至第7章则是本文的核心部分:依次解决上述需要探讨的三个问题,并进行实证检验。第5章考察单项资产的流动性风险与市场风险的集成风险度量。第5.1节着重集成风险度量模型的构建,首先建立一次变现条件下的集成风险度量模型,然后以预期变现期和变现策略两个维度,分别讨论固定变现策略前提下的集成风险度量问题和固定预期变现期前提下的集成风险度量问题。在第5.2节基于中国A股的实证研究中,度量了不同规模公司股票的集成风险。与集成风险度量方法相比,传统VaR方法将低估或者高估风险;而对于个股,只有选择最优变现期或最优变现策略才能最小化集成风险。第6章考察资产组合的集成风险度量。本章把第5章的模型从单只股票推广到由两资产组成的资产组合中。在资产组合层面上,集成风险既可以表现为由不同市场风险因子驱动的集成风险,又可以表现为由不同流动性风险因子与市场风险因子联合驱动的集成风险。于是,在第6.1节构建资产组合的集成风险度量模型之后,第6.2节度量了只考虑市场风险的资产组合集成风险,第6.3节度量了兼顾流动性风险与市场风险的资产组合集成风险。经实证检验,与传统方法相比,资产组合集成风险度量模型可以刻画资产组合中各风险因子的“厚尾有偏”特征和风险因子之间的复杂相依性。如果不考虑流动性风险,集成风险将被低估,而如果简单的加总流动性风险,集成风险将被高估。第7章是对集成风险度量的动态分析。第5章和第6章是在静态框架下度量集成风险,而本章试图把集成度量模型推广到动态情形下。第7.1节介绍基于copula函数的多维时间序列模型。第7.2节根据本文的集成风险度量要求,构建基于copula函数的集成风险动态度量模型,并进行了实证研究。与静态框架相比,动态度量模型由于考虑了时变性,更能够反映近期股票流动性风险因子与市场风险因子的变化。第8章对本文进行回顾和展望。首先对本文的主要结论进行回顾,从理论与实证两个方面对本文模型和现有风险度量模型进行系统的分析比较。然后基于本文研究对金融风险管理提出若干建议,并总结本文研究的不足,最后对集成风险度量的前景进行展望。更多还原

【Abstract】 What the thesis studies is how to measure the liquidity risk and the market risk in the same flame,namely integrated risk measurement of incorporating liquidity risk and market risk.This article uses the copula function to construct the integrated risk measurement model,aiming at the following three problems:(1) How to measure individual asset’s integrated risk driven by liquidity risk factor and market risk factor? (2) How to measure portfolio’s integrated risk driven by many groups of liquidity risk factors and market risk factors?(3) How to extend the integrated risk measurement model in the dynamic frame on the basis of the above research.After constructing the integration risk measurement model,integrated risks of different size firms’ stocks are calculated in empirical study on Chinese A-shares market.And the results calculated by our model are compared with those calculated by traditional models.This thesis consists of nine chapters:Chapter One is the introduction.It puts forward the question,then introduces this article’s main innovation and difficulty,and carries on the overall plan for the research technique and frame.Chapter Two is literature review.First it analyzes the concept of integrated risk, and then reviews the existing research on liquidity risk measurement,market risk measurement,as well as their integrated measurement and summarizes the study on integrated risk measurement and its related financial studies based on copula function.Chapter Three introduces copula function to measure integrated risk of incorporating liquidity risk and market risk.How to choose optimal copula function is the key to applying copula function.Chapter Four is the analysis of stocks’ integrated risk.It identifies liquidity risk and market risk as the main risks of China’s A-shares,and adopts the semi-parametric method to construct the model for liquidity risk factor and market risk factor.After modeling,the dependence of liquidity risk factor and market risk factor of stock are studied from the theoretical and empirical aspects,so as to prepare for integrated risk measurement.Chapter Three and Chapter Four constitutes two footstones of integrated risk measurement.One is the method of integrated risk measuremen??timal copula function theory;the other is the motivity of integrated risk measurement,the market demand that is the current situation of risk factor of China’s A-shares.Chapter Five to Chapter Seven is the hard core of this thesis which resolves the above three problems and carries on empirical test.Chapter Five studies the individual asset’s integrated risk measurement of incorporating liquidity risk and market risk.Section 5.1 focuses on the model construction of integrated risk measurement.First of all,it constructs the integrated risk measurement model under one-time liquidation.Then,it discusses respectively the problem of integrated risk measurement on the premise of given liquidation strategy and the problem of integrated risk measurement on the premise of fixed expected liquidation.In section 5.2,in empirical study on Chinese A-share market, integrated risks of different size firms’ stocks are calculated.Compared with the integrated risk measurement,risk measures are underestimated or overestimated by traditional VaR approaches.For individual share,optimum liquidation period or optimum liquidation strategies need to be determined in order to minimize their integrated risks.Chapter Six explores integrated risk measurement of portfolio.It extends the model of Chapter Five from individual asset to two-asset portfolio.In the portfolio level,integrated risk represents integrated risk driven by different market risk factors, and integrated risk jointly-driven by different liquidity risk and market risk factors as well.So after constructing the model of integrated risk measurement of portfolio in Section 6.1,Section 6.2 calculates portfolio integrated risk only in view of market risk. Section 6.3 calculates portfolio integrated risk in consideration of liquidity risk and market risk.Compared with the traditional method,it is empirically tested that the model of integrated risk measurement of portfolio can describe the heavy-tailed and skewed characteristic of various portfolio risk factors,and the dependence among the risk factors.Integrated risk will be underestimated if we do not consider liquidity risk and integrated risk will be overestimated by the way of adding liquidity risk to market risk directly.Chapter Seven is the dynamic analysis of integrated risk measurement.Chapter Five and Chapter Six measure integrated risk in the static frame,however,this chapter attempts to generalize the integrated risk measurement model into dynamic condition. Section 7.1 introduces the multivariate time series model based on copula function.In terms of the demand of integrated risk measurement in the thesis,Section 7.2 constructs the dynamic integrated risk measurement model on the basis of copula function,and does the empirical research.In comparison with the static frame, dynamic measurement model can reflects the change of liquidity risk factor and market risk factor of the recent situation more owing to considering the time variant.Chapter Eight gives a review and outlook of the thesis.At first,it reviews the main conclusions,making comparison between the model in the paper and the existing integrated measurement model from the theoretical and empirical aspects. Afterwards,it offers several proposals for financial risk management based on the article’s study,and makes a summary of research insufficiency of the thesis.Finally,it expects the future of integrated risk measurement.更多还原