Realized Copulae
Matthias R. Fengler1, Ostap Okhrin2
1Fachbereich fuer Mathematik und Statistik, University St. Gallen, St. Gallen, Switzerland; 2Ladislaus von Bortkiewicz Chair of Statistics, Humboldt Universitaet zu Berlin, Berlin, Germany

This paper considers realized copulae as the possible way of modelling time varying dependency using copulae. Estimating integrated volatility together with the daily correlation by well-known methods of realized, by averaging and kernel volatility and covariances. Using Hoeffding-Lemma we propose several methods of estimation realized copulae. The efficiency has been compared with the time varying copulae based on daily data estimated through rolling window and local change point detection procedure.

Keywords: Copula; High-frequency data

Biography: Ostap Okhrin, born 1984, is a Juniorprofessor at the Ladislaus von Bortkiewicz Chair of Statistics at the Department of Economics and Business Administration at the Humboldt-University Berlin since April 2008. His research interests are dimension reduction methods, multivariate distributions, estimation and multivariate model selection, copula based distributions. His most recent work is dealing with the modelling of credit risk and the statistical analysis of financial risk.