VaR Estimation And Inspection Based On Shanghai And Shenzhen Stock Data

Risk management has constantly undergone a transformation in the past years.As a new method of measuring financial risk, value at risk(VaR) has played an increasingly important role in commercial financial transactions. It is originally a response to the financial disasters, whereas nowadays VaR has been applied to many derivatives due to its consistent development, which changes the way people used to deal with risks to a great extent, and can help financial institutions and individuals avoid risks. In May 2005, the reform had a great influence on stock markets. The thesis aims to undertake research into earnings ratio per day in the Shanghai and Shenzhen stock markets over the past ten years. Owing to different characteristics in different stages, the earnings ratio is divided into three stages.We regard the earnings ratio as the object of study to examine its distribution,stability and ARCH effect. GARCH model, TGARCH model EGARCH and historic simulation are established in suitable periods. VaR of the two stock markets will be evaluated in different periods through the four ways. We use Kupiec to examine whether the results of VaR work. Finally, we use the Bayesian approach to estimate the GARCH model parameters under different prior information and conclude that the improved model is better for data fitting, and the calculation of value at risk(VaR) accuracy is higher, the different duration of priori information accumulation effects have different simulation results.