Font Size: a A A

The Research On The Return Distribution, Volatility Characteristics And Risk Measurement Of Stock Market

Posted on:2008-02-15Degree:DoctorType:Dissertation
Country:ChinaCandidate:M YuFull Text:PDF
GTID:1119360242965288Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Since finance plays the role as the vitals maintaining the operation of a country's economy, and the bridge connecting economies of different countries as well, financial market and financial system becomes the core of economy. Governments and investment institutions all over the world are thereby trying to ensure the stability of financial markets and financial institution. As a result, constituting a valid prewarning system for financial risk is badly needed. Financial market is a high-risk industry. And the risk of financial market and financial institutions is closely related to the safety of financial system, macro economy of the country or even the world, as the economic integration and globalization develops. Therefore risk management turns more and more important to every country, every institution and even the whole society. Chinese government has also paid great attention to risk regulation for financial market, especially for stock market.With the continous enlargement of the Chinese stock market in scale in the last 15 years, it becomes more and more important to the economic development. However, Chinese stock market is still a developing market which has many immature characteristics, such as great volatility, with its stock price rise and drop too fast. Then, studying on the characteristics of return distribution and volatility of Chinese stock market by econometric models is certain to help investors to avoid risk, and CSRC to do regulations. Thus government, researchers and investors have paid more and more attentions to the research on return distribution and volatility of Chinese stock market.As the stock indices reflect the whole market, analysis of Stock indices is helpful for understanding the volatility and risk of the market, without the effect of single stock. The widely-used indices of Chinese stock market include Shanghai Composite Index and Shenzhen Component Index. In order to make comparative study of stock markets home and abroad so as to find the rule of return distribution, volatility and risk of both Chinese market and developed markets, this paper also investigates some of the most important indices of the world, including UK'FTSE100 index, America's S&P500 index, Japan's NIKKEI index,German's DAX index,France's CAC40.First, the paper makes a comparative study among ARCH Models. Then the ARCH models are used to evaluate the volatility of China's stock market so that the conditional heteroskedasticity, leverage effect, long memory are investigated. And the evaluation effects of China's Stock Market's volatility by different ARCH models from different point of view are compared. In addition, by using two mixed distribution models, the leptokurtosis and fat-tail of stock return are examined. Based on the above analysis, an improved Laplace distribution model is proposed. Furthermore, the root generating non-Gaussian return distribution of the stock market ? is expounded from the perspective of both economy and capital market.To combine the function of both distribution and volatility models, the ARCH Models are adapted that introduce error item based on t and GED distribution. With maximum likelihood estimation, the volatility of stock market is forcasted. Meanwhile, the forecasting effects of different ARCH Model are compared. Concerning the facts that fat-tail results in the underestimation of risk by VAR, this paper investigates the fitness between the VAR caculated by ARCH Models based on different distributions and empirical data.Since Leptokurtosis, skewness and fat-tail are always used to depict market risk, this paper introduces the improved Laplace distribution to risk management of stock portfolio. Empirical results show, improved Laplace distribution and asymmetric Laplace distribution are similar for estimating Leptokurtosis, while improved Laplace distribution is better for estimating skewness.In order to analyze the correlation between return distribution and volatility, the thesis uses ARCH-M Models to simulate the return series of China's Stock Market and developed markets ao as to discover the relation between investment return and time-varying risk. Comparative study is made on risk premium coefficients and investor's risk tolerance between the stock market in China and developed markets. In addition, the best measurement for time-varying risk of securities markets home and abroad is investigated on ARCH-M domain.Finally, as an important improvement to ARCH-M Models, the thesis introduces time-varying VaR to GARCH-M Models, and GARCH-M-VaR models are obtained. Then a maximum likelihood method for GARCH-M-VaR based on BHHH algorithm is porposed. With the GARCH-M-VaR models proposed, the empirical research on stock markets in different countries is made. At the end of this part, a comparative study is made between GARCH-M-VaR models and GARCH-M Models, and some interesting and creative results are found.
Keywords/Search Tags:Return Distribution, Return Distribution Volatility, ARCH-M Model, Risk Measurement, Value at Risk
PDF Full Text Request
Related items