Abstract

Class Title: Copulae and time varying non Gaussian Dependency Structures Time-varying and non-Gaussian dependencies for multivariate time-series are in demand for many economic models. Current models available suffer from the misfortune of dimensionality or restrictive assumptions on the parameters and the distribution. New promising classes of models are copulae that allow for non-exchangeable and non Gaussian dependency structures with a small number of parameters. For Hierarchical Archimedean Copulae (HAC) a novel adaptive estimation technique based on Local Likelihood Approximation (LPA) for the parameters and of the structure of HAC in a time varying context is presented. Typical applications are in the financial field but also more recently in the spatial analysis of climate parameters. An analysis of time varying dependency structure of stock indices and exchange rates reveals periods with constant and turmoil dependencies. The economic significance of the suggested modelling is evaluated using Value-at-Risk of a portfolio and the correction of the implied correlation smile of, for example, CDO ’s.