| With the rapid development of society,internationalization of the economy,globalization,financial events occur frequently.The measurement of financial risks is of great significance.Considering extreme events,the tail risk metric is becoming more urgent.China’s financial data has its unique characteristics,and the distribution of which distribution has been widely discussed and concerned.Under the condition of determining the hypothesis distribution,it is also the core concern to accurately measure the risk.This paper discusses and analyzes these two main line problems,and selects China’s stock market data for empirical research to verify that the assumed distribution and selected risk measurement methods are accurate and effective.First,in terms of yield distribution.Considering the stock market yield data in China,it has the characteristics of peak,thick tail and asymmetry.Comparing the three distributions of normal,symmetric Laplace and asymmetric Laplace,qualitative probability maps and quantitative K-S tests are used.The conclusion is that the effect of fitting the rate of return data from high to low is asymmetrical Laplace distribution,symmetric Laplace distribution,and normal distribution.The maximum likelihood estimation is used for the parameters of the asymmetric Laplace distribution,and the explicit expressions of the three parameters σ、u、λ are obtained.Second,in terms of risk measurement.Based on the VaR value,we review the tail condition expectation TCE,the tail condition variance TCV,and propose a new tail variance TV.It is assumed that an explicit expression is given by the definition of four metrics on the basis of the asymmetric Laplace distribution.And analyze the calculation of the four risk measures in the portfolio.Continue to analyze the sensitivity of the three parameters σ、u、λ,and conclude that the sensitivity of the four metrics is slightly different,but the overall is high,and the measurement risk can be combined.Finally,in terms of evidence.12 stocks were selected from the Wind database,and the closing price of the past 10 years was converted to logarithmic yield data.After qualitative and quantitative analysis,the asymmetric Laplace distribution is chosen as the hypothesis distribution.Four risk metrics VaR,TCE,TCV,TV and three parameters σ、u、λ,actual values and predicted values are given.The results show that the difference is small.It is proved that the asymmetric Laplace distribution is very suitable for fitting the stock market in China.Yield data. |
