Research On Risk Measurement Of Open-end Funds Based On Improved Historical Simulation Method

The open-end fund market in China has undergone years of development and has rapidly expanded in size.Due to its numerous types and flexible investment characteristics,it is widely welcomed by the investment market.However,due to the uncertainties in the investment process,such as market changes,it is necessary to have a relatively accurate control of the risks of the fund,in order to avoid or reduce losses caused by risks,and also to guide investors in making more rational investment decisions.Therefore,researching models that can accurately measure the risk of open-end funds in China can not only supplement the theoretical methods of risk management,but also provide technical guidance to investors in reality.This paper firstly introduces the development of risk management and the classification of risk characteristics of open-end fund,and expounds the current situation and existing problems of traditional fund risk management in China.Then,it introduces the definition,statistical properties,and selection of window width of nonparametric kernel density estimation,GARCH related model,and the calculation methods of VaR.At the same time,it provides models for calculating VaR using three traditional historical simulation methods.In order to further improve the prediction effect,combining the relevant theories of nonparametric kernel density estimation,we will perform kernel density estimation on the yield data of the research object.The paper also provides a method of replacing a single kernel function with a combined kernel function,and provide two improved historical simulation methods.Finally,the Kupiec failure rate test method is used to test the accuracy of VaR estimation by various historical simulation methods.In the empirical study on the risk measurement of open-end funds,first of all,through the set screening conditions for fund information,the daily cumulative net worth of 15 funds is selected as sample data,including YinHua Small-Medium Cap Mixed Funds.After logarithmic processing of the data,one of the funds is empirically analyzed using three traditional historical simulation methods,it is found that each model performs differently when estimating VaR at different confidence levels.Secondly,we use the historical simulation method improved by kernel density estimation to conduct empirical analysis on the sample funds,and its estimation results for VaR are better than those of the previous three traditional historical simulation methods.Then,the combined kernel function of Gaussian kernel function and Exponential kernel function is used to replace the single Gaussian kernel function into the improved historical simulation model for empirical analysis.The estimated results of this model for VaR are more satisfactory both in terms of covering the number of days of real losses and the gap between estimated losses and real losses.Therefore,this more accurate method is used to conduct empirical analysis on the remaining 14 funds and obtain their average VaR and conduct economic analysis.At the same time,Kupiec failure rate method is used to test the effectiveness of the VaR estimation using the historical simulation method improved by the combined kernel density.Finally,it is concluded that this method is a model for accurately estimating the risk of open-ended funds.Based on the research on the risk measurement of open-end funds,this paper proposes corresponding suggestions for the prevention and control of fund risks from the perspectives of regulatory authorities,fund companies,and fund retail investors.