Modeling the Dynamic Dependence Structure in Multivariate Financial Time Series

• Activity Number: 267
• Type: Contributed
• Date/Time: Tuesday, August 9, 2005 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economics Statistics Section
• Author(s): Mihaela Serban*+ and John Lehoczky and Anthony E. Brockwell and Sanjay Srivastava
• Companies: Carnegie Mellon University and Carnegie Mellon University and Carnegie Mellon University and Carnegie Mellon University
• Keywords: copula ; correlation ; GARCH ; portfolio ; volatility

Abstract

The dependence structure in multivariate financial time series is of great importance in portfolio management. By studying daily return histories of 17 exchange-traded index funds over a period of approximately eight years, we identify important features of the data and propose two new models to capture these features. The first is an extension of the multivariate BEKK model, which includes a multivariate-t type error distribution with different degrees of freedom. We demonstrate that this error distribution is able to accommodate levels of heavy-tailed behavior and thus provides a better fit than models based on a multivariate-t with a common degree of freedom. The second model is copula-based and can be regarded as an extension of the standard/generalized DCC model (Engle 2002; Cappiello et. al. 2003) to a Student copula. Model comparison is carried out using criteria, including the AIC and BIC. We also evaluate the two models from an asset-allocation perspective, constructing optimal portfolios based on the Markowitz theory. Our results indicate the copula model performs better than the extension of the BEKK model.