Measuring co-movements in the Euro area using a nonstationary factor model

Abstract

This article investigates to what extent business cycles co-move in the four largest euro area economies, using a large-scale database of nonstationary series for the euro area over the 1980:Q1 to 2003:Q4 period. We apply the methodology proposed by Bai (2004) and Bai and Ng (2004) to construct a coincident indicator of the euro area business cycle, based on the first common factor estimated from a dynamic factor analysis on the level of the variables. The indicator appears to be significantly close, from a statistical point of view, to the level of the euro area GDP in the most recent period. We also show that national developments are increasingly correlated to the indicator at the business cycle frequencies. We finally suggest a decomposition of GDP growth along the different stationary and nonstationary factors.