Statistical Modelling of Temporal Dependence in Financial Data via a Copula Function

Filippo Domma (1) , Sabrina Giordano (1) , Pier Francesco Perri (1)

1 Department of Economics and Statistics

This paper is concerned with modelling dependence structure in a univariate financial time series using the concept of copula. We treat financial series as a first order Markov process. The Archimedean two-parameters BB7 copula is adopted to describe the dependence structure between two consecutive returns, whereas Log-Dagum distribution is employed to model the margins marked by skewness and kurtosis. A simulation evaluates the performance of ML estimates. Finally, we apply the model to daily returns and illustrate how the fitting can be improved when the ascertained dependence between consecutive returns is described through copula function.

Mots clés: Physical Sciences

Filippo Domma, Sabrina Giordano, Pier Francesco Perri. Statistical Modelling of Temporal Dependence in Financial Data via a Copula Function. Communications in Statistics - Simulation and Computation, 2009, 38 (04), pp.703-728. ⟨10.1080/03610910802645321⟩. ⟨hal-00514343⟩