The Application Of Portfolio Selection Models In China's Captial Market

With the full implementation of the "Guiding Opinions on Regulating the Asset Management Business of Financial Institutions" jointly issued by the Central Bank,the China Banking Regulatory Commission,the Securities Regulatory Commission and the SAFE,the asset management industry has entered a new stage of development,and banks have been able to invest directly through wealth management subsidiaries In the equity market,the scope of institutional investors has been further expanded,and growing of institutional investors has continued to advance in the A-share market.Due to the relatively large amount of funds held by institutional investors,portfolio investment is generally adopted in equity investment to achieve compliance and reduce investment risks.According to the statistics of the public fund of Tian Tian Fund Web in March2020,most of China's fund products are still subjective transactions(1163 existing stock funds and 2871 mixed funds,of which only 190 are quantitative investment strategies).The scientific application and promotion of investment portfolio optimization methods can help improve the investment returns of China's asset management industry,reduce investment risks,and bring higher satisfaction to investors.Hence,research on portfolio optimization is an important topic which has both theoretical research value and practical application value.Since Markowitz proposed modern portfolio optimization theory based on mean and variance,scholars and the industry have improved the portfolio optimization theory from two aspects.The first aspect is the characterization of risk.Derived from the simple variance risk characterization method proposed by Markowitz,the semivariance method,the value at risk and the conditional value at risk are tried to solve the variance characterization risk such as the distribution of returns.Assumption is too strong or computational complexity is high.The second aspect is to start from the perspective of optimization methods.On the basis of a single-period unconstrained static portfolio optimization method,add corresponding constraints,and then derive a multi-period dynamic portfolio optimization method,an anti-inflation portfolio optimization method,and an international Optimization of investment portfolios.This thesis carried on a systematic research on the application of portfolio selection models in China's capital market from four aspects,which are organized as follows:First of all,the mean-variance portfolio optimization method(MV),meanabsolute deviation portfolio optimization method(MAD)and maximum expected absolute deviation portfolio optimization method(MXM)The portfolio model,the na?ve portfolio model,and the hybrid model are compared and analyzed.The addition of out-of-sample tests on the basis of traditional portfolio ranking and comparison alleviates the over-fitting problem that may be caused by single-period optimization methods,and provides more practical reference data for the selection of different portfolio optimization methods.This thesis divides the monthly return data of the Ashare and bond markets between 2007 and 2017 into 5 data sets,each spanning 6 years.The earliest data set starts in January 2007,and the latest The data set start time is January 2011,the start time of each data set is one year later,and the last year data of each data set is used as out-of-sample validation data.The empirical results show that all portfolio optimization methods have higher end-period returns than market portfolios,and the period-end yields and??ratios show a model of MVmulti-period MAD model>multi-period MV model.The research results in this thesis show that,under multiple circumstances,multi-period MXM is still the preferred portfolio optimization method in China's capital market.Thirdly,we study the method of optimizing the all assets portfolio against inflation.First,we conduct an empirical test on China's large-scale assets(stocks,commodity futures,bonds,real estate)to verify the effects of inflation on different assets.The empirical results show that bonds do not have the ability to hedge inflation;most commodity futures have the ability to resist inflation risks positively;the rate of return on house prices can hedge inflation risks;the stock industry index does not have the ability to hedge inflation.Subsequent anti-inflation portfolio research shows that in a combination with only commodity futures and real estate,the characteristics of the yield and variance of commodity futures are not as good as those of real estate,resulting in a lower weight in the portfolio.After the introduction of gold,bonds and stocks into the portfolio,the weight of real estate remains the largest.The empirical results in this thesis show that it is difficult to construct stock portfolios to outperform inflation in China through portfolio optimization methods.Last of all,we carry on research on the optimization methods of international portfolios.This thesis compares the mean-variance method(MV),the minimum maximum absolute deviation(MXM),and the equal weight method(EQW).On the basis of the CSI 300 Index and the SSE 50 Index,the monthly income data of the S?P 500 Index,the Dow Jones Industrial Index,and the FTSE All-Share Index from January 2009 to December 2018 are used to build a rolling window data set to compare different portfolio optimization Method and consider the impact of exchange rates on total variance.The empirical results show that no matter what model is used,the inclusion of foreign low volatility stock indexes can improve the??ratio of the portfolio.In terms of return performance,after hedging exchange rate risk,the MXM portfolio optimization method performs better than the mean-variance method(MV).We make several contributions to the literature,which can be described as following.Our first contribution is to use both China Mainland's capital markets datasets and international equity datasets to analyze the portfolio optimization method of minimizing the the maximum of expected absolute deviation,as well as other portfolio optimization methods.Our second contribution is to use China Mainland's equity datasets to analyze the results of multi-period portfolio optimization methods,such as multi-period mean-variance,multi-period mean-absolute deviation,multiperiod mean-maximum of expected absolute deviation,providing suggestions to investors.Our third contribution is to analyze nearly all kinds of assets' inflation correlation,especially China's short term bonds.Then we use kinds of optimization methods,including polynomial goal programming method to calculate the anti-inflation portfolios.This article also has some shortcomings.The first shortcoming is that in the analysis of single-period models,we come to the conclusion that MXM>MAD>MV based on our sample.We don't know that under what condition this conclusion is valid.The same shortcoming exists in the analysis of multi-period models.The second shortcoming is that the correlation between assets return and inflation rate may vary according to different economic status.This shortcoming may cause the difference between our research and previous researchs by scholars.