Font Size: a A A

The Calculation Of VaR Based On Different Distributions And Empirical Study

Posted on:2004-02-09Degree:MasterType:Thesis
Country:ChinaCandidate:L HeFull Text:PDF
GTID:2156360122967101Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Since the breakdown of the Briton Wood system, the exchange rate, interest rate, stock price and the commodity price in the global area has a high volatility. Especially in the 90's, the whole world financial market changed enormously because of the globalization of the world economy and the integration of the financial market. The acute competition in the financial market caused the innovation tide of the financial derivatives, and the relax of the control on it, which in turn fierced the competition in the world financial market. What's more, it becomes the intrinsic motivation and provides a favorable environment for the development of financial market. This spire like development aggrandize the uncertainty and volatility of the whole financial market. China's stock market, as a part of the whole world financial system also undergoes a world-shaking change.All those things above bring forward a question of how to effectively perform financial risk management, and VaR risk management tool emerges as times require. But the computation of the VaR in most cases is based on the assumption that the rate of return follows a normal distribution. This paper estimates Value-at-Risk using parametric distributions that capture systematic skewness and kurtosis in returns series. The distributions studied are normal distribution, student-t distribution, skewed student-t distribution and general error distribution. Besides this, considering the conditional Heteroskedasticity of the time serial in financial market, apply the GARCH model into the estimation of VaR. At last, do empirical study using some of those models in China's stock market combining with the analyzing of the impact of policy change to the risk of China's stock market. We get some interesting results. Firstly, the distribution of index return rate in China's stock is not normal distribution exactly. So, the calculated VaR based on normal distribution is not tenabe. Secondly, the rate of return in China's stock has auto regressive heteroskedasticity phenomena and it should be considered in the calculation of VaR. Last, student-t distribution and GED can better describe the distribution of return rate in China's stock market. In a word, all the conclusion above indicates that the method in this paper is valid and credible.
Keywords/Search Tags:Value-at-Risk, GARCH model, Student-t Distribution, General Error Distrib
PDF Full Text Request
Related items