Marc Hallin
Marc Hallin - Université Libre de Bruxelles, Belgium
Marc Hallin holds a PhD in Sciences & Mathematics from the Université libre de Bruxelles (1976). His main research interests are in mathematical statistics (asymptotics, time series, and rank-based inference) and econometrics (high-dimensional time series and factor models). He is the author or coauthor of about 200 publications in mathematical statistics and econometrics. He has been Editor-in-Chief of the International Statistical Review, and is Editor-in-Chief of Statistical Inference for Stochastic Processes. He is also on the editorial boards of the Journal of the American Statistical Association, the Journal of Econometrics, the Annals of Computational and Financial Econometrics, the Journal of the Japan Statistical Society, and the Annales de l¹Institut de Statistique de l’Université de Paris. A Fellow of the Institute of Mathematical Statistics (I.M.S.), the American Statistical Association (A.S.A.), and the International Statistical Institute (I.S.I.), he is member of the Classe des Sciences of the Royal Academy of Belgium.
Research stay at UC3M: DEPARTMENT OF STATISTICS (DIC 2017 - JUN 2017)
Project:
My present research interests are mainly related to (i) measure transportation and its applications in the definition of multivariate distribution and quantile functions, and their empirical counterparts, involving ranks and signs; (ii) dynamic factor models in the analysis of high-dimensional time series. (i) Unlike the real line, the p-dimensional real space Rp is not canonically ordered for p>1. Hence, such concepts and statistical tools as distribution and quantile functions, ranks, and signs, are not canonically defined. In Chernozhukov et al. (2016, Annals of Statistics, to appear), new concepts of ranks and signs are proposed, based on measure transportation ideas. Those concepts enjoy all the properties one is expecting from ranks and quantiles---among them, distribution-freeness under absolutely continuous distributions. The results there, however, require finite second-order moments---an assumption which is unfortunate when dealing with ranks and quantiles. I am presently working on establishing much stronger result---e.g., Glivenko-Cantelli, Donsker and iterated logarithm---by adopting the geometric approach of McCann (1995, Duke Mathematical Journal), which does not involve any costs nor any moments. (ii) High-dimensional time series are ubiquitous in such areas as financial and macro-econometrics. In the past few years, I have coauthored a series of papers on the so-called general dynamic factor mode, which encompasses all other time-series factor models in the literature. My recent activity in the area deals with a dynamic factor model approach to the challenging analysis of volatilities in high dimension, and with a dynamic factor model approach to panels of functional data. On both topics, I am sure to find a good fit with statisticians at Carlos 3 and econometricians as well.
