Study On High-Dimensional Futures Portfolio Risk Management Based On Family-GARCH And R-Vine Copula

The core content of asset portfolio risk management is the identification and evaluation of risks,and it avoids risks through the arrangement of portfolio weights.For the four typical characteristics of'high-dimensionality','futures','portfolio' and'risk',this paper constructs a portfolio risk assessment model based on the family GARCH model and R-Vine Copula theory.This paper optimizes the weight arrangement on the basis of quantifying the portfolio risk,so that we can achieve the purpose of managing portfolio risk and provide reference for decision-making in risk management activities.In terms of model construction,this paper describes the marginal distribution of multi-dimensional variables based on the ARMA-fGARCH model and the skewed t-distribution,and proposes a selection method for intra-fGARCH models based on the parametric test ideas.In order to solve the problem of dependence structure analysis under high dimensional conditions,this paper applies R-Vine Copula theory to describe the high-dimensional distribution and simulates its joint distribution by means of Monte Carlo simulation.For the measurement of portfolio risk,VaR is used as the benchmark,and the calculation method is corrected cause of future's bi-directional trading feature.Then the mean-VaR portfolio optimization model is constructed.Finally,we use random search to obtain the portfolio efficient frontier.Risk manager can select the most appropriate risk estimation model and change the security weights according to the optimization model.In such way,we construct a high-dimensional futures portfolio risk model.In empirical research,this paper selects all 18 futures contracts that meet the data screening conditions in China's futures market since 2010 to form a high-dimensional futures portfolio,and applies the model proposed in this paper to estimate the portfolio risk,and makes corresponding tests to analyze the appropriate application scope of our model in portfolio risk management.As well,we study the features of China's futures market.This article finally draws the following empirical conclusions:1.According to its own characteristics,all futures varieties of China's futures market have inconsistent performance on the characteristics of 'fat-tail','aggregation of volatility' and 'leverage effects'.2.In the dependence structure of China's futures market,there is no 'center'with a large weight of dependence.3.The model proposed in this paper is a valid model for portfolio risk assessment at the 1%level,but there is a slight underestimation of risk.4.Compared with other models,the model proposed in this paper is more conservative in the risk appetite regions of high-risk preference.Therefore,portfolio risk managers tend to choose this model to manage high-risk-preferred portfolios.