Title: GLS estimation and confidence sets for the date of a single break in models with trends
Authors: Eric Beutner - Vrije Universiteit Amsterdam (Netherlands)
Yicong Lin - Maastricht University (Netherlands) [presenting]
Stephan Smeekes - Maastricht University (Netherlands)
Abstract: The aim is to present the asymptotic results on generalized least squares (GLS) estimation of time series models with a single break in level and/or trend at some unknown date where the errors are assumed to be stationary. As GLS estimators, relying on the unknown inverse covariance matrix of the errors, are usually infeasible, we estimate the unknown inverse matrix after doing the modified Cholesky decomposition. Then, we construct feasible GLS estimators. Based on inverting a sequence of likelihood-ratio tests, we construct valid confidence sets for the timing of the break and compare these with existing methods. Extensive Monte Carlo simulations demonstrate that our proposed method has good finite sample properties in terms of either the coverage rates or the lengths of the confidence sets.