Please use this identifier to cite or link to this item:
e Azevedo, Joao Valle
Koopman, Siem Jan
Rua, Antonio
Year of Publication: 
Series/Report no.: 
Tinbergen Institute Discussion Paper 03-069/4
This paper proposes a new model-based method to obtain a coincident indicator for the business cycle. A dynamic factor model with trend components and a common cycle component is considered which can be estimated using standard maximum likelihood methods. The multivariate unobserved components model includes a stationary higher order cycle. Also higher order trends can be part of the analysis. These generalisationslead to a business cycle that is similar to a band-pass one. Furthermore, cycle shifts for individual time series are incorporated within the model and estimated simultaneously with the remaining parameters. This feature permits the use of leading, coincident and lagging variables to obtain thebusiness cycle coincident indicator without prior analysis of their lead-lag relationship. Besides the business cycle indicator, the model-based approach also allows to get a growth rate indicator. In the empirical analysis for the Euro area, both indicators are obtained based on nine key economic timeseries including gross domestic product, industrial production,unemployment, confidence indicators and interest rate spread. This analysis contrasts sharply with earlier multivariate approaches. In particular, our more parsimonious approach leads to a growth rate indicator for the Euro area that is similar to the one of EuroCOIN. The latter is based on a more involvedapproach by any standard and uses hundreds of time series from individual countries belonging to the Euro area.
Band-pass filter
Coincident indicator
Dynamic factor model
Kalman filter
Leading indicator
Unobserved components time series model
Phase shift
Document Type: 
Working Paper

Files in This Item:
450.61 kB

Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.