Risk Measure Research Based On Stock Data

Financial risk will bring huge losses to countries and investors.Effective risk quantification can minimize risk and reduce economic losses.Therefore,quantitative research on risk is of great significance to both countries and investors.This thesis took the risk quantification of stocks as the research object.Firstly,the definition of risk and four ways to measure the risk were introduced,namely nominal value method,sensitivity method,volatility method and VaR method;then,four methods for calculating VaR have also been described,namely covariance matrix method,historical simulation method,Monte Carlo simulation method and measurement method based on GARCH model.Further,the adaptation environment and advantages and disadvantages of the four method have been given.At the same time,the Kupiec method to check whether VaR is in conformity with reality have been introduced.In addition,empirical analysis has been conducted based on the historical data which were randomly selected from large-cap stocks and small-cap stocks.Specifically,The thesis mainly used historical simulation method,covariance matrix method and GARCH model method to quantify the risk of small and large stocks.The results showed that the three methods can effectively measure the risk of stock data,but the measurement methods based on GARCH model were relatively conservative.In summary,the risk value of small-cap stocks was 6% of the investment amount,and the risk value of large-cap stocks was 2.5% of the investment amount.At the same time,the risk for small-cap stocks was about 2.4 times that of large-capital stocks.Finally,according to the results of empirical analysis,some corresponding investment suggestions have been proposed for investors.