|
Abstract:
|
Copulas provide a powerful tool for analyzing multivariate data, as they allow modeling the dependence structure of random variables independently of their marginal models. Although copulas have been in use in the applied statistical literature for more than twenty years, covariate adjustment for copulas has been considered only recently. Patton [2006, Modeling Asymmetric Exchange Rate Dependence, Int. Economic Review, 47(2)] described the conditional copula framework for those situations in which additional information is present through covariates. In the current work, we consider nonparametric estimation of the functional relationship between the copula parameter and the covariate(s). First order asymptotic expressions for bias and variance of the nonparametric estimator are derived and computational issues related to bandwidth selection are discussed through a simulation study.
|
