Detecting multiple mean breaks at unknown points in official time series

In this paper, we propose a computationally effective approach to detect multiple structural breaks in the mean occurring at unknown dates. We present a non-parametric approach that exploits, in the framework of least squares regression trees, the contiguity property of data generating processes in time series data. The proposed approach is applied first to simulated data and then to the Quarterly Gross Domestic Product in New Zealand to assess some of anomalous observations indicated by the seasonal adjustment procedure implemented in X12-ARIMA are actually structural breaks. © 2008 IMACS. Published by Elsevier B.V. All rights reserved.