Author(s): Fabrice Orlandi and Karl Pichelmann
A variety of statistical methods and econometric techniques can be used attempting to disentangle the non-cyclical trend component of a time series and its purely cyclical part. This paper serves the purpose to demonstrate the potential contribution from the use of unobserved components modelling techniques to decompose the EUR-11 unemployment series. In general, unobserved components models appear to be an attractive and quite flexible tool to discriminate between the cyclical and the trend component in unemployment; in particular, the multivariate version of the model allows to use information contained in the price series to assist the decomposition of the unemployment rate. Consequently, this specification is potentially closer to the NAIRU concept than univariate filtering techniques.