Forecasting with Equilibrium-correction Models during Structural Breaks
Jennifer Castle,
David Hendry and
Nicholas Fawcett
No 408, Economics Series Working Papers from University of Oxford, Department of Economics
Abstract:
When breaks occur, equilibrium-correction models (EqCMs) based on cointegration face forecasting problems. We investigate approaches to alleviate forecast failure following a location shift, including updating, intercept corrections, differencing, and estimating the future impact of an 'internal' break during its progress. Although updating can lead to a loss of cointegration when an EqCM suffers an equilibrium-mean shift, we show that updating can help when collinearities are changed by an 'external' break and the EqCM itself remains constant. Both mechanistic corrections help compared to just retaining a pre-break estimated model, but an estimated model of the break process could outperform. Throughout, we apply the approaches to the much-studied example of EqCMs for UK M1, and compare with updating a learning function as the break evolves.
Keywords: Forecasting; Location Shifts; Equilibrium-Correction Models (search for similar items in EconPapers)
JEL-codes: C1 C53 (search for similar items in EconPapers)
Date: 2008-10-01
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (5)
Downloads: (external link)
https://ora.ox.ac.uk/objects/uuid:ab8455d8-9ed5-4561-b19f-b4dac1e1d5ee (text/html)
Related works:
Journal Article: Forecasting with equilibrium-correction models during structural breaks (2010) 
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:oxf:wpaper:408
Access Statistics for this paper
More papers in Economics Series Working Papers from University of Oxford, Department of Economics Contact information at EDIRC.
Bibliographic data for series maintained by Anne Pouliquen ( this e-mail address is bad, please contact ).