Identification Of Stock Price Volatility And Measurement Of Its Risk Of The Listed Company

This paper measures the risk of market volatility of four industry indices inShanghai stock exchange and some important industry indices in Shenzhen stockexchange with the POT model that can well facet the extreme value of financial timesseries.At first, this article illustrates the concept and classification of the risk andintroduces the evolution of the methods of the financial market risk measurement. Inthis paper, we mainly talk about market risk. Then this paper gives a whole descriptionon the volatility of the log-return of the mentioned indices above using the standarddeviation, and then decomposes the market risk of the indices with the volatilitydichotomy theory. The risk is decomposed into systematic risk and nonsystematic riskto better grasp the market volatility trend. Usually market volatility cannot be directlyobserved, but the typical dynamic modeling method--ARCH model can well capture themarket volatility and thick tail of the financial time series, when the distributionhypothesis and form and parameters of the model are optimized. This paper models andmeasures the return and volatility of the mentioned indices with suitable models.Before using the POT model, we make some necessary tests for the log-return ofthe mentions indices above. Test results show that the log-return series are notindependent, but stable, and accords with necessary conditions for using POT model.About the threshold selection, because the threshold selection is not only a statisticalproblem but also a financial issue, and different investors usually choose differentthresholds. So we can’t determine the threshold purely according to statistics theories.This paper gives up the thick tail distribution and normal distribution intersectionmethod and the kurtosis method, and Chooses mean experience excess method todetermine the threshold, and then estimates the corresponding value at risk.This paper measures the market risk mainly with the POT model. At the same timeas contrast, we also give a simple introduction for ARCH，fractile quantile estimates andBMM, and at last give the value at risk based on them.The article makes goodness-of-fit test for POT through estimating the excessvolume of the mentioned indices with GPD, and tests the prediction failure rate throughKupiec test. Test results show that POT model can grasp the extreme volatility characters,and the prediction failure rate is controlled near the level of significance. This means the goodness-of-fit and prediction accuracy of POT is higher.