Discussion Papers, Center for European, Governance and Economic Development Research 300
According to a growing body of empirical literature, global shocks have become less important for business cycles in industrialized countries and emerging market economies since the mid-1980s. In this paper, we analyze the question of what might have caused a decoupling from the global business cycle: the smaller size of the global shocks or a reduced sensitivity of national business cycles to these shocks? To this end, we employ a large scale hierarchical dynamic factor model that decomposes the growth rates of GDP, consumption, and investment for 106 countries over 1961-2014 into a global, a group-, and a country-specific factor, as well as an idiosyncratic component. The factor loadings and conditional variances are allowed to vary over time according to random walk processes. Instead of assuming that the parameters change, we test for time variation using a Bayesian stochastic model specification search. Our results confirm a reduction in the importance of the global business cycle for the vast majority of our countries. However, the sensitivity of most countries to global or group-specific shocks as measured by the factor loadings has not changed over time. Instead, the magnitude of the global shocks relative to group-specific and country-specific shocks has decreased, resulting in a lower relevance of global shocks for national cycles.
global business cycle dynamic factor model time-varying parameter stochastic volatility model selection