Several New Copula Models Based on Autoregressive Condition

In studying dependence structure of portfolio, we need to identify the marginal distributions of univariate time series. The Copula-GARCH models introduced by WEI Yan-hua et al. (2004) and Patton allow for time-varying volatility but neither time-varying skewness nor time-varying kurtosis. In this paper, we propose several new copula models based on autoregressive conditional higher-moments volatility characteristics of financial time series. The conditional higher-moments volatility and asymmetrical higher-moments volatility effects are introduced to the marginal distributions of dependence copula model. These specifications are more consistent with the characteristics of financial data.