Estimation methods for the extreme value index
Jan Beirlant – Professor at KU Leuven (Belgium)

We briefly review the main trends in univariate estimation of the extreme value index and extreme quantiles. Specific themes such as bias reduction, adaptive choice of the number of extremes used in estimation, and the comparison of the underlying models for the different solutions are discussed. An application to hunting black swans in the European banking sector using extreme value analysis is provided.

Jan Beirlant obtained his PhD in Statistics at KU Leuven in 1984, and holds a professorship at the Department of Mathematics at KU Leuven university. From the start he worked at the construction and coordination of the Leuven Center for Statistics (LStat) which he chaired for three years. From 2003 to 2009 he was Dean of the Faculty of Science, and from 2009 till august 2013 he was vice-rector of KU Leuven responsible for the KU Leuven campus in Kortrijk.
His main research topic is on the extreme value methodology, studying the occurrence of rare events with large impact. He published over 100 papers in statistical research journals and published a book on “Statistics of Extremes: Theory and Applications” jointly with Y. Goegebeur, J. Segers and J.L. Teugels.