Application of Copula Vector Autoregression Model in Multi Asset Portfolio Risk Management

Abstract

It is challenging to use the covariance models and single VAR models to capture these complicated features accurately, and it may result in biases of risk evaluation. Nevertheless, the Copula model is capable of efficiently describing multidimensional dependency structures. The dynamic interactions among the time series can be effectively characterized by Vector Autoregression (VAR) model. The consolidated Copula VAR model is not only able to reflect effectively the complex dependence relationships between assets in a multiple asset portfolio, but also flexibly capture the transmission and diffusion process of risk, and as such enhances goal accuracy of risk prediction as well as work efficiency of management. Firstly, this paper is organized as: in the second section introduces the theory of Copula model and VAR model, these two theoretical, current application status in financial risk management, and its integration necessity and advantages were analyzed. From the empirical results, build a Copula VAR model and test its performance of multi asset portfolio risk assessment by concentrating on how well the model is capable of capturing risk in extreme market circumstances. Compared with the traditional risk management methods, Copula VAR model proved to be better on risk finding out, risk transmission path analysis and risk early-warning.