Forecast accuracy after pretesting with an application to the stock market

Jan Magnus () and Dmitry Danilov
Additional contact information
Dmitry Danilov: Eurandom, Eindhoven University of Technology, The Netherlands, Postal: Eurandom, Eindhoven University of Technology, The Netherlands

Journal of Forecasting, 2004, vol. 23, issue 4, 251-274

Abstract: In econometrics, as a rule, the same data set is used to select the model and, conditional on the selected model, to forecast. However, one typically reports the properties of the (conditional) forecast, ignoring the fact that its properties are affected by the model selection (pretesting). This is wrong, and in this paper we show that the error can be substantial. We obtain explicit expressions for this error. To illustrate the theory we consider a regression approach to stock market forecasting, and show that the standard predictions ignoring pretesting are much less robust than naive econometrics might suggest. We also propose a forecast procedure based on the 'neutral Laplace estimator', which leads to an improvement over standard model selection procedures. Copyright © 2004 John Wiley & Sons, Ltd.

Date: 2004
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (6)

Downloads: (external link)
http://hdl.handle.net/10.1002/for.916 Link to full text; subscription required (text/html)

Related works:
Working Paper: Forecast Accuracy after Pretesting with an Application to the Stock Market (2002)
Working Paper: Forecast Accuracy after Pretesting with an Application to the Stock Market (2002)
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:jof:jforec:v:23:y:2004:i:4:p:251-274

DOI: 10.1002/for.916

Access Statistics for this article

Journal of Forecasting is currently edited by Derek W. Bunn

More articles in Journal of Forecasting from John Wiley & Sons, Ltd.
Bibliographic data for series maintained by Wiley-Blackwell Digital Licensing () and Christopher F. Baum ().