Decomposing the co-movement of the business cycle: a time-frequency analysis of growth cycles in the euro area

This article analyses the frequency components of European business cycles using real GDP by employ-ing multiresolution decomposition (MRD) with the use of maximal overlap discrete wavelet transforms (MODWT). Static wavelet variance and correlation analysis is performed, and phasing is studied using co-correlation with the euro area by scale. Lastly dynamic conditional correlation GARCH models are used to obtain dynamic correlation estimates by scale against the EU to evaluate synchronicity of cycles through time. The general findings are that euro area members fall into one of three categories: i) high and dynamic correlations at all frequency cycles (eg France, Belgium, Germany), ii) low static and dy-namic correlations, with little sign of convergence occurring (eg Greece), and iii) low static correlation but convergent dynamic correlations (eg Finland and Ireland).