CMStatistics 2018: Start Registration
View Submission - CFE
Title: Monitoring cointegration in a system of homogeneous cointegrating regressions Authors:  Etienne Theising - University of Cologne (Germany) [presenting]
Martin Wagner - University of Klagenfurt (Austria)
Dominik Wied - University of Cologne (Germany)
Abstract: The aim is to extend a previous procedure for detecting a structural change in a system of homogeneous cointegrating regressions. Therefore, we construct test statistics based on the residuals of fully modified OLS estimators to account for error serial correlation and regressor endogeneity and to obtain nuisance parameter free limiting distributions. In particular, we consider the pooled FM-OLS estimator for homogeneous panel cointegration. We, however, focus on the finite $N$ case and consider only large $T$ asymptotics. We examine four different detectors along with `` self-normalized '' versions of those and additionally compare them to their equivalents based on residuals of the classical FM-OLS estimator applied to each cointegrating regression separately to assess the relative performance in terms of small size distortion under the null of no structural change, and power and detection time under various alternatives.