The Study On Risk Parity Model Based On Social Network And Timing Judgment

With the development of the economy in our country,the wealth of residents increases rapidly,which prompts residents to attach importance to the preservation and appreciation of wealth,and stimulates their investment demand.Because residents'risk preference is not high,it is particularly important to control the risk of investment and build a robust and decentralized portfolio,which can reduce the fluctuation of portfolio and enhance the anti-drop ability of portfolio in order to achieve the goal of the preservation and appreciation of wealth.It is of great significance to meet the diversified investment demand in the market.With the rapid development of financial market and the continuous innovation of financial products,the correlation between financial assets is increasing,and the strong connection between assets is not conducive to the decentralization of the portfolio.Therefore,from the perspective of social network,this paper constructs the association network between stocks by Affine Propagation clustering algorithm,and chooses the stocks corresponding to clustering centers in the clustering network as the target assets of the portfolio to disperse the correlation between the assets.Mean-variance model pioneered the modern portfolio research.The idea of quantifying risk and return in the model has far-reaching influence on the subsequent portfolio research.The Risk Parity model builds a robust portfolio by using the quantitative measurement.However,the Risk Parity model pursues the balanced contribution of asset risk,which is essentially allocating risk rather than return.Therefore,this paper adds timing judgment into the basis of risk parity to adjust the weight of portfolio.When the asset in the portfolio show upward or downward trend,it adds or reduces the weight of the asset to increase the return and Sharp ratio of the portfolio.This paper uses Affine Propagation clustering algorithm to select the target stock,and determines the optimal number of assets in the portfolio through silhouette coefficient and investment practice.During the sample period,this paper compares the investment result of the risk parity model and the risk parity model with timing judgment,and finds that the risk parity model with timing judgment has better allocation effect,both return and Sharp ratio,and it has lower maximum withdrawal.By comparing the investment of CSI 300 stock index in the same period,it is found that the portfolio by the two models can increase the investment return and obtain better Sharp ratio,so that investors can obtain greater returns when they bear the same risk.In order to make the model more practical,this paper divides the sample period every six month,and adjusts the weight according to the risk parity model in each cycle.At the same time,additional adjustments are made according to timing judgment during the holding period.In the dynamic adjustment of weights,the risk parity model with timing judgment is more effective than the classical risk parity model.In addition,this paper divides the sample period into oscillation interval,bull market interval and bear market interval according to the stock market cycle to test the robustness.This paper finds that under the three market conditions,the return of the portfolio determined by risk parity model with timing judgment is higher than that of the risk parity model,and the former portfolio has higher Sharp ratio and lower maximum withdrawal.In bull market,although the return of the portfolio by the two models is lower than the direct return of the CSI 300 index in the same period,the Sharp ratio is still higher than the index.In bear market and oscillation interval,the return and Sharp ratio of portfolios by the two models are higher than those of the index in the same period.From the perspective of decentralization and robustness,this paper constructs a portfolio model,which is innovative in the construction and application of the model.The model can be used in specific portfolio practice.In this paper,the clustering network is used to choose the target asset in portfolio,which can be used as a reference for investors to construct decentralized portfolio.Furthermore,on the basis of risk parity theory,this paper adds timing judgment and combines quantitative investment theory with traditional portfolio theory,which can provide reference for investment research and practice.Of course,there are also some shortcomings in this study,such as the choice of the asset and the choice of transaction price in the empirical model,which can be considered in future portfolio research and practice.