Research On The Influence Of Skewness Factor On Portfolio Decision

There are many kinds of investment methods and different characteristics of financial assets,and portfolio decision-making refers to investing financial assets in different financial products in a certain allocation way,so as to improve income and disperse risk.The first portfolio model in the world is the mean variance model proposed by Markowitz in 1952.The mean variance model has attracted the attention of many scholars.After 70 years of development,experts and scholars all over the world continue to deepen the research of portfolio theory and put forward a variety of portfolio models such as multi factor model and high-order moment model of return.The skewness of the third-order moment in the high-order moment model is one of the important indicators in the financial market.The skewness of the yield represents the asymmetry of the left-right distribution of the yield of financial assets.The increase of the skewness will extend the yield on the right side of the coordinate axis to the positive direction,and the yield on the left side of the coordinate axis will gather near the origin,which can increase the upper limit of the yield and reduce the risk of large losses,It can better meet the needs of investors for income and downside risk.Based on this,this thesis takes the influence of skewness factor on portfolio decision as the theme,analyzes whether skewness can really effectively improve the investment utility of investors,and studies it from the following aspects: first,introduces the common methods of skewness measurement,namely,the traditional third-order moment method,quantile method,the hybrid method based on Cornish Fisher expansion and the time weighting method of skewness,The robustness of various measurement methods is compared and analyzed by graphic method and improved coefficient of variation method.Secondly,the skewness is introduced into the weight function as an important factor.At the same time,combined with the CRRA utility model in utility theory,the portfolio model is established,that is,the skewness model(SK model),the two factor characteristic variable model(ag-op model,etc.)and the comprehensive model composed of skewness and multi factor model.The BFGS algorithm in quasi Newton method is used to solve the substitution coefficient in the model.Third,select 20 representative constituent stocks from a shares,substitute the data of representative stocks into the model,establish the portfolio strategy,compare and test the yield,downside risk and "gambling" characteristics of each model,and analyze the advantages and disadvantages of each portfolio strategy.The main conclusions of this thesis are as follows: first,the robustness of the commonly used skewness measurement methods is compared through the graphical method and the improved coefficient of variation method.It is intuitively seen from the graphical method that the skewness measured by the quantile method has significant advantages over the skewness measured by the third-order moment method in resisting large amplitude oscillation.Further,it is calculated through the improved coefficient of variation method that the skewness robustness measured by the quantile method is improved by about 20% compared with the third-order moment method.Second,the empirical results of the model show that,with the equal weight model as the control group,the skewness model,the characteristic variable model and the comprehensive model have the earnings performance of the far exceeding equal weight model,and the three types of model strategies all reflect the gambling characteristic of the high frequency big increase.In terms of anti risk,the skewness model with skewness factor shows stronger anti risk ability.Its downside risk is reduced to close to the equal weight model,while the sharpe ratio and Sortino ratio are much higher than the equal weight model,and the frequency of large decline is also significantly lower than other models.Therefore,this thesis believes that the skewness model is a portfolio model with great investment value,The model SK(int)based on the skewness of quantile integral method is the most cost-effective model,which has high investment potential and research value.Taking stocks as the target of financial assets,this thesis selects 20 representative constituent stocks in a shares as the experimental data,involving many industries with large,medium and small cap.The empirical results are universal.In terms of time,it selects the latest transaction data from September 1,2008 to August 31,2021 for research,and takes the data from September 1,2018 to August 31,2021 as the test set for model back test,Based on the evaluation of income risk index,the reliability of empirical results is further improved.The innovation of this thesis is mainly reflected in: constructing a biased portfolio model;A two factor combination of asset growth rate and operating net interest rate is constructed as the characteristic variable,and the reciprocal of operating profit rate is also used as the characteristic variable for comparative experiment;For the data whose mean value is close to0 and is not suitable for the coefficient of variation to measure volatility,the coefficient of variation is improved.