B0317
Title: Generalized Hadamard differentiability of the copula mapping and its applications in time series models
Authors: Natalie Neumeyer - University of Hamburg (Germany) [presenting]
Abstract: The dependence between components of multivariate time series can be modelled with copulas, and the empirical copula can be used as a nonparametric estimator. The empirical copula based on observations is of interest but also based on residuals in multivariate time series models with covariates. To show weak convergence of the empirical copula process the Hadamard differentiability of the copula mapping (which maps a joint cumulative distribution function to the corresponding copula) is a powerful tool in combination with the functional delta method. A generalization of the Hadamard differentiability of the copula mapping is stated, which allows deriving asymptotic expansions and weak convergence in situations where previous results are not applicable.
