B0996
Title: Robust approaches for change point detection in panel data
Authors: Alexander Duerre - TU Dortmund (Germany) [presenting]
Roland Fried - TU Dortmund University (Germany)
Abstract: In panel data we usually observe a large number of individuals over a time period. One often assumes stability of the model over time. There are different proposals to test this assumption. However, up to now little attention has been paid to robustness, although a small fraction of outliers can lead to a false rejection of the null-hypothesis or even mask a change point. We therefore propose alternative robust procedures for different situations with short-range serial and possibly cross-sectional dependence. Some simulations compare these methods with existing nonrobust tests.
Journal