Econometrics and Finance Seminar: Marc Hallin

R-Estimation in Semiparametric Dynamic Location-Scale Models

Marc HALLIN, ECARES, Université libre de Bruxelles

(with D. La Vecchia)

Ranks offer a flexible, powerful, and too often neglected alternative to quasi-likelihood methods in econometrics.Here we propose rank-based estimation (R-estimators) as a substitute for Gaussian quasi-likelihood and standard semiparametric estimation in time series models where conditional location and/or scale depend on a Euclidean parameter of interest, while the unspecified innovation density is an infinite-dimensional nuisance. Applications include linear and nonlinear models, featuring either homo- or heteroskedastic conditional distributions (e.g. conditional duration models, AR-ARCH, discretely observed diffusions with jumps, etc.). We show how to construct R-estimators achieving semiparametric efficiency at some predetermined reference density while preserving root-n consistency and asymptotic normality irrespective of the actual density. Contrary to the standard semiparametric estimators (in the style of Bickel, Klaassen, Ritov, and Wellner), our R-estimators neither require tangent space calculations nor innovation density estimation. Numerical examples illustrate their good performances on simulated data. A real-data analysis of the log-return and log-transformed realized volatility of the USD/CHF exchange rate concludes the talk.