A Research On The Risk Measurement Of China Securities Investment Fund Based On VaR Methods

Recently, Funds especially the open-end securities investing funds in chinaare of fast development, great investment and widespread influence. There follows aquestion for these booming funds that how to efficiently manage and keep watch therisk of funds for avoidance of loss of funds after national fund market was open tothe world. The advent of national funds nearly was the moment of the beginning of anew foreign risk management tool named Value-at-Risk. It was said that Value-at-Risktriggered am evolution and create a new time in the field of risk management, andnow it has become the prevailing risk management tool.The paper utilizes VaR as a tool to measure the level of national open-endsecurities investing funds. It specifically chooses models of EWMA, historicalsimulation and monte carlo simulation with20days,250days and20daysrespectively for the measurement of potential loss value within a future day for everysample fund, and chooses the date from jan042011to sep282012as the test date tocheck the validation of the models utilizing Kupiec model. Another part of the paperis to calculate and analyse values of a portfolio from a fund using the model ofGARCH and to test the validation the model under the confidence level of95%and99%with the same test date above mentioned. eventually, it comes to the followingconclusions:1National open-ended equity fund serious exits the character of being leptokurtic and fat-tailed and most sample funds not exist the effect of ARCH.2Thevalues of VaR Fund360001,240005,240004,377010,519688are comparativelyhigher over the whole sample funds, and we cannot judge the risk situation of a fundjust by its certain investing type if VaR is reliable.3Confidence level has amimportant effect on the effectiveness of the model.4We calculate average value ofVaR of the portfolio under the confidence level95%and99%are3.111%and4.400%respectively, but the former are not significant.5Model of monte carlo is superiorthan other models according to the test results. we could deprive from conclusionsabove that the assumption distribution trend using analysis experience should beconsistent with the distribution in the prediction of VaR by investors or fundcompanies. GARCH model cant be used indiscriminately. Fund companies shouldselect the optimal model according to their VaRs so as to improve the ability of theprediction of the risk. They should also strengthen their portfolio risk measurementand risk prevention and management.