Assessing Extremal Dependence in Equity Markets

• Activity Number: 308
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Jose Faias*+ and Miguel de Carvalho and António Rua
• Companies: Universidade Catolica Portuguesa and Ecole Polytechnique Fédérale de Lausanne, Swiss Federal Institute of Technology and Banco de Portugal
• Keywords: Asymptotic independence ; Multivariate extreme theory ; Tail dependence ; Empirical likelihood ; Risk modelling ; Sectoral tail comovements

Abstract

In recent years there has been an increasing interest in modelling dependence in heavy tail phenomena such as the latest turbulence episodes in financial markets. The evidence of asymptotic independence in the financial data has led to the need of rethinking risk modelling and inference tools for multivariate extremes. In this paper we propose an inference scheme for assessing extremal dependence of several pairs of variables, in the context of asymptotic independence. Our approach is based on the fact that the problem of interest can be rewritten as an empirical likelihood problem for comparing the means of different populations, where such means represent the Hill's estimate of the coefficient of tail dependence. A triangular array representation allow us to obtain a nonparametric Wilks' theorem, and we apply the method to assess extremal dependence in equity markets.