The University of Liverpool
Department of Mathematical Sciences
Division of Statistics and Probability
SEMINAR
NONPARAMETRIC TESTS FOR CONDITIONAL SYMMETRY IN DYNAMIC MODELS
(Coauthored with J.C. Escanciano)
M. A .Delgado
Universidad Carlos III de Madrid
Wednesday, 1st March 2006, 2pm
The Whittaker Room (211)
Abstract:
This article proposes omnibus tests for the symmetry of conditional
distributions around a parametric function in a dynamic context. The
underlying conditional distribution is unknown, and conditional moments may
not exist, or may depend on the explanatory variables, allowing for
conditional heteroskedasticity and heterokurtosis of unknown functional
form. The testing procedure is based on suitable functionals of the
multiparameter empirical process of residuals and explanatory variables.
Since the resulting test statistics are not asymptotically pivotal, the
test is implemented with the assistance of a bootstrap procedure. Large
sample properties of the tests are formally justified under fairly weak
regularity conditions on the parametric specification of the center of
symmetry and the serial dependence structure of the data. Finite sample
properties are examined by means of a Monte Carlo experiment.
Keywords and Phrases: Omnibus tests; Symmetry tests; Conditional
distributions; Time series; Empirical processes; Bootstrap.
Following the talk, tea and biscuits will be available in Room 304
ALL WELCOME
********************************************************
Ingrid Harper
University of Liverpool
Department of Mathematical Sciences
Division of Statistics and Probability
Mathematical Sciences Building
Peach Street
Liverpool L69 7ZL
Tel: 0151 794 4751
Fax: 0151 794 4754
********************************************************
|