Monitoring cyclical processes. A non-parametric approach

E. Andersson

Journal of Applied Statistics, 2002, vol. 29, issue 7, 973-990

Abstract: Forecasting the turning points in business cycles is important to economic and political decisions. Time series of business indicators often exhibit cycles that cannot easily be modelled with a parametric function. This article presents a method for monitoring time-series with cycles in order to detect the turning points. A non-parametric estimation procedure that uses only monotonicity restrictions is used. The methodology of statistical surveillance is used for developing a system for early warnings of cycle turning points in monthly data. In monitoring, the inference situation is one of repeated decisions. Measurements of the performance of a method of surveillance are, for example, average run length and expected delay to a correct alarm. The properties of the proposed monitoring system are evaluated by means of a simulation study. The false alarms are controlled by a fixed median run length to the first false alarm. Results are given on the median delay time to a correct alarm for two situations: a peak after two and three years respectively .

Date: 2002
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (17)

Downloads: (external link)
http://www.tandfonline.com/doi/abs/10.1080/0266476022000006685 (text/html)
Access to full text is restricted to subscribers.

Related works:
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:taf:japsta:v:29:y:2002:i:7:p:973-990

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/CJAS20

DOI: 10.1080/0266476022000006685

Access Statistics for this article

Journal of Applied Statistics is currently edited by Robert Aykroyd

More articles in Journal of Applied Statistics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst (chris.longhurst@tandf.co.uk).