# 16-088/VI (2016-10-20)

Tino Berger, University of Goettingen, Germany; Lorenzo Pozzi, Erasmus University Rotterdam, The Netherlands
Global business cycle, dynamic factor model, Bayesian, model selection
JEL codes:
C52, C32, F44

We investigate the presence of international business cycles in macroeconomic aggregates (output, consumption, investment) using a panel of 60 countries over the period 1961 - 2014. The paper presents a Bayesian stochastic factor selection approach for dynamic factor models with predetermined factors. The literature has so far ignored model uncertainty in these models as common factors (i.e., global, regional or otherwise) are typically imposed but not tested for. We focus in particular on the existence of a global business cycle as the literature has, in our opinion unjustifiably, taken for
granted its existence. In contrast to the literature, we find no evidence to support its presence.