We compare the classification of business cycle for the euro area into recessions and expansions as defined by the CEPR committee with that obtained using alternative methods; in particular, we analyze the accuracy of dating criteria based on nonlinear methods, like those used for computing recession probabilities.
We show that the CEPR chronology is not fully in line with classifications implied by alternative algorithms based on GDP only, thus confirming that the CEPR committee considers more variables.
Multivariate approaches can replicate the business cycle features identified by the CEPR. We find that the CEPR chronology is indeed more consistent with the latent common component, estimated using an econometric factor model.
