Testing for international business cycles: A multilevel factor model with stochastic factor selection

The empirical literature on common international business cycles has largely ignored model misspecification in estimated factor models as the various are typically imposed but not tested for. This paper proposes a Bayesian stochastic selection approach for multilevel models. procedure is applied to three-level dynamic with global factor, six regional factors and three development level factors. We estimate using real GDP growth data panel of 60 countries over period 1961?2017. find robust evidence presence cycle, four (Europe, North America, Latin America Asia) two (industrial emerging market economies). suggests that both geographical proximity important dimensions cycle synchronization should be considered simultaneously, point previously made existing literature.