【作者】 王晨；

【导师】 陈守东；

【作者基本信息】 吉林大学 ， 数量经济学， 2010， 博士

【摘要】 金融风险度量的核心是价格波动性的估计和预测。本文以金融市场波动的区制关联性与风险度量为题,在对金融市场波动性模型的理论与方法综述的基础上,对金融市场波动的相关性、区制关联性、风险度量问题进行了研究。首先,以我国股市波动为研究对象,考虑了交易制度的变迁,采用不同频率不同阶段的数据样本,运用马尔可夫区制转移模型系统分析了我国股市波动区制在风险识别上的应用。其次,对2005年我国汇率形成机制的市场化改革和证券市场的股权分置改革以来的日度数据样本,基于单变量马尔可夫区制转移模型对股市同利率、汇率波动的区制关联性进行分析和比较,并采用了两种方法度量市场间波动的区制相关性。第三,基于二元向量区制转移模型,分别分析了汇率、利率与股市波动率的区制关联性并对其显著性进行了检验,进而比较了波动性预测的效果。最后,以我国股市波动为例对VaR度量结果进行了系统的比较与分析。本文对我国金融市场波动性分析与风险度量的创新应用是:采用了单变量和二元向量的马尔可夫区制转移模型,按交易制度分阶段对我国股市波动的区制性和风险识别进行了分析;研究了股市和汇率波动的区制关联性在波动预测和风险度量方面的应用;并对股市波动VaR度量的双向影响进行了深入研究。这些研究对我国金融市场风险的度量与预警以及金融市场的多样化和国际化均具有重要的理论和现实意义。更多还原

【Abstract】 China’s stock market, bond market, foreign exchange market, money market has gradually improve and strengthen the securities market, the successful reform of non-tradable shares, the formation mechanism of exchange rate marketization reform deeply, various financial product development and innovation of China’s financial markets, to enjoy unprecedented development. Financial market volatility system analysis and characterization, risk measurement is an important problem of risk management, use econometric model on the financial market fluctuations process and risk degree of empirical analysis, research situation of China’s financial market turmoil has important significance and characteristics.Financial risk measurement is the core of the estimated price volatility and prediction. Based on the model of financial market volatility theory and method, on the basis of the review of the financial market fluctuations, the correlation relationship risk measurement system, the problem is studied. From two aspects, one is to study the single market fluctuations cycle system, depicting the market operation cycle and the corresponding risk characteristics, the other is established between multiple market system, and describes the impact between market conduction and interconnected.The specific contents and main conclusions are as follows:The first chapter, includes the introduction. Mainly discussed the background and significance of topic, For the domestic and foreign relevant literatures were reviewed, And the risk management method and measure technology for review and overview, The paper gives content and structure, Clearly this in order to solve the problem and innovation.The second chapter, financial market volatility model theory development and methods. This chapter presents the text involved in the empirical research method. The first day is one of the financial market volatility model research progress, The second section of GARCH model and SV model theory and method, The third section presents markov switching model theory and method.The third chapter, China’s stock market volatility of system analysis and empirical research. In China’s stock market volatility for research object, by choosing different frequency of different sections of data sample, markov system of different structure and markov switching model estimation method is different, and the system are analyzed in detail fluctuations in risk identification system, the application that: (1) on the research of data weeks, considering the system structure of the model, the ARCH effect and non-normal sex are not significant. (2) for our market monthly and weekly return volatility Gibbs sequence, sampling method is the maximum likelihood estimation method for system changes reflect risk has better recognition. (3) in our country after 1995 stock market fluctuations on data and data of the system, the system is more suitable to two regime. The year of 1995 is China’s trading system is the important points and temporal variation of structural change significantly breakpoint. Gibbs (4) sampling estimation method of China’s stock market for the whole week yields sample data and better fitting.The fourth chapter, the Chinese stock market fluctuations in currency exchange market, with the analysis of system and empirical research. In this paper based on single variable markov switching model with money, foreign exchange market volatility system analysis and comparison, the two methods of measurement and stock markets and foreign currency market volatility and the regime correlation. Select our July 2005 to the exchange rate, the interest rate since the length of the same sample with the stock market volatility data, an empirical research. By comparison, rates faster appreciation and raising interest rates as the control policy for the stock market fluctuation, and the high variance, from exchange rate fluctuations regime correlation is the interest rate volatility area has greater impact, namely currency controls. Specific found on Chinese stock market respectively, and the interest rate volatility in the area of relevance, (1) stock and interest rate volatility of high variance area from correlation coefficient sequences from low variance area correlation coefficients sequence, (2) stock and interest rate volatility regime correlation probability correlation coefficient sequences original sequence correlation coefficients. About the Chinese stock market and the rate of exchange rate fluctuations regime correlation, (1) the relationship exchange rate fluctuation rate compared with the stock market fluctuations can make high variance correlation zone; (2) when selecting the appropriate models to estimate the structure, the stock market and the high rate of exchange rate fluctuations regime variance correlation coefficients also can make area than low variance correlation coefficient sequences, (3) stock and currency volatility multi-member regime probability correlation coefficient sequences original sequence correlation coefficients.Chapter 5, based on vector system of Chinese stock market volatility markov switching model of forecasting and compared. Based on the binary vector markov switching model are analyzed, the stock exchange, the interest rate volatility and the system and inspection and compared in this basis, the four models to predict the stock market volatility structure comparison, mainly as follows: (1) to found in July 2005, since it’s data, exchange rate volatility degree of area and the variance of stock market volatility relationship between system by significant test, the average rate of exchange rate fluctuations of market system changes of variance volatility over 80% of the vector, both for stock markov switching model of the sequence of volatility fitting effect is good. (2) on July 2005 to the degree of data, since the interest rate volatility of the stock system and mean variance volatility multi-member regime relevance not through significant test, interest rate volatility of the stock market fluctuation multi-member regime average rate of no significant effect of variance area, two vector markov switching model of stock market volatility and forecasting effect are inferior to exchange rate and the stock market volatility vector markov switching model. This chapter (3) from the perspective of relevance that area from July 2005 to type in the price of the monetary policy and exchange rate, the interest rate on the stock market tool to predict the effects of volatility multi-member regime, significant correlation. (4) structure of exchange rate fluctuations average rate for the stock market volatility multi-member regime and the variance of vector multi-member regime area of our country, and markov switching model of different sample data, degrees, estimated interval prediction effect, the stability is good.Chapter 6, based on the nonlinear model of Chinese stock market volatility VaR measurement and compared. Based on Chinese stock market volatility VaR measurement results from the basic theory, VaR by choosing different frequency interval samples and data to construct VaR model back-test system for the Chinese stock market is the most representative of the Shanghai index in 95%, 97.5% and 99% VaR model measurement results of the comparison and analysis, we find that for the degree of data, (1) in volatility estimates on the choice of model, considering the ARCH effect and non-normal sex assumptions on the efficiency of the VaR model can have certain. (2) for the Chinese stock market volatility VaR measurement system, the ARCH effect weakens when adopting markov switching model, the measurement results of the model of MS variance are basically the same as GARCH model. (3) for our market volatility VaR measurement system, using vector model including asymmetric coefficient of the ARCH effect, and are not significant in the sample under the condition of the shorter length measurement results, it can return through inspection, relative to other model has obvious advantages. For weeks, (1) degree of data on Chinese stock market volatility VaR measurement, considering one-way wave SWARCH model for risk, and MS variance model results are basically the same, (2) for the Chinese stock market volatility VaR measurement can be empty in margin trading mechanism as low-class system should be considered for the risk of two-way fluctuation, MS variance and SWARCH model of measurement results are good model.Based on the system of financial market fluctuations in collecting, sorting out the area of relevance theory and empirical risk measurement and, on the basis of the research literatures in Chinese stock market as samples, the main research with foreign exchange market and money market fluctuations, and the system based on VaR measurement and compared. Based on the method of comparison on the main work is: (1) for the Chinese stock market volatility and prediction of the single variables on markov switching model and the MLE or GIBBS sampling estimation method in identifying the fluctuation of the different areas. (2) for our stock exchange in the currency market fluctuation, and the correlation analysis, the dual vector system markov switching model of mean and variance multi-member regime area of two kinds of structure, and examines the relationship between their multi-member regime. (3) in the study of China’s stock market volatility and foreign exchange, money markets, not the relevance to the original sequence of correlation analysis, probability analysis and forecast for the thought that the area between the market research, and then used for making risk measurement analysis.Based on China’s financial market volatility, system analysis and correlation of the main risk measurement results and innovative applications are: (1) through the SWARCH model and parameters of the model, MS variance MLE estimation and Gibbs sampling method in the financial market volatility description on the comparison of difference and the quality, there is significant in Chinese stock market transition phases and characteristics. (2) in China, foreign currency market shares and the operation characteristic, the three measures market operation cycle and the main area of the relationship between, By selecting and measure liquidity to markets and volatility of dynamic indexes, the study found that the exchange rate of the stock market volatility variance explained, and then makes a stronger ability of model prediction ability greatly improved. (3) on Chinese stock market fluctuations in currency exchange, and the correlation analysis, the dual vector switching model of mean or variance regime structure, and tests the relationship between market volatility regime of significant. (4) through the segment and choose a different frequency of data transfer system, vector model has better in VaR measurement, can apply the effect after the financial crisis of the volatile market conditions of reality. And according to the risk degree and the Chinese financial market operation characteristic of the important facts, countermeasures and Suggestions are experienced in China, further turbulence in financial markets and economic growth cycle lays the foundation of relevance.更多还原