Research On The Effectiveness Of The Market Risk Measurement Model In China's Industries

Under the background of the new normal economy,the operating environment of financial institutions is facing the dual changes of market structure and financial system.The trend of marketization is gradually increasing.Risk as a basic property of the market has increasingly attracted the attention of financial institutions.In the report of the Nineteenth Congress,President Xi Jinping emphasized the importance of a sound financial supervision system and the bottom line of not taking systemic financial risks.To this day,the tenth anniversary of the economic crisis in 2008 has passed.Here,we should sum up the lessons learned from the financial crisis,and we should also be prepared to prevent financial risks.In this process,we need to continuously strengthen the supervision of market risks and improve the accuracy of risk assessment.In this context,this paper studies the market risk of China's capital market.This article classifies China's financial assets according to industry standards.According to the Shanghai-Shenzhen 300 industry index,we have selected the following industry categories,including:electrical equipment,automobiles,franchised retail,aviation,electric power,building materials,and pharmaceuticals,a total of 28,spanning time For 2012 to 2017.Through the descriptive statistical analysis of the data,the distribution of the returns of China's industrial financial assets shows a state of peak and heavy tail,requiring us to pay more attention to the tail risk in the next step of risk measurement.The measurement model used in this paper is divided into two frameworks,namely VaR system and ES system.VaR refers to the value of insurance,its concept is intuitive,the calculation method has a wide range of simple applications,and is a risk measurement method that is explicitly required by the Basel Accord.ES is the expected loss.Its main consideration of tail risk is theoretically superior to VaR,but it is more difficult in practical applications.Both methods of measurement have their merits.This article combines two systems to measure risk.Under the VaR system,GARCH model,GJR-GARCH model,EWQR model and CAViaR model are introduced in this paper.In the first two cases,we substituting four distributions of returns respectively:normal distribution,t distribution,Hansen-spiking t distribution,and generalized error distribution.Through these four distributions,we can better match the fluctuations in real financial asset returns.The CAViaR model does not assume the distribution of the model.Under the ES system,we introduced GARCH model and GJR-GARCH model and EWQR model.As above,we need to introduce four yield distributions.In the empirical analysis,we substituted 28 industries into the above measurement models.Since the true VaR value is unknown,we use the indirect test method to test the model.We use the UC test and DQ test method for the five CAViaR models.For GARCH model,GJR-GARCH model and EWQR model,we used UC test,DQ test,DV test and? test.The original hypothesis of the above tests is that the model is set correctly.Rejecting the original hypothesis implies a model error.The empirical analysis results are as follows.Under the VaR system and the ES system,the most effective models are the EWQR model,the GARCH-t,and the GJR-GARCH-t model.Under the VaR system,Adaptive model performance is more effective.The results show that the distribution of the returns of our industry financial assets is non-normal and has the characteristics of spikes and thick tails.Since the GARCH-t and GJR-GARCH-t models are more effective,the current yield fluctuations are t-distribution.Investors and regulators need to grasp the distribution of the current fluctuations in the rate of return and consider the risk of the tail when carrying out risk measurement.Only in this way can we better grasp the expected loss.At the same time,the empirical results show that EWQR is more suitable for the analysis of China's industrial financial risk measurement.The model has superiority in theory and has good practical significance in application.However,in the five CAViaR models,only Adaptive is more effective,and the other four groups are less effective.Although it is theoretically superior to VaR,its practical application does not apply to financial assets of the Chinese industry.