Asymmetry in the Business Cycle: Friedman's Plucking Model with Correlated Innovations

Abstract

Recent research has shown that unobserved components (UC) models can, under certain conditions, be estimated without imposing the common zero-correlation restriction between the permanent and transitory innovations. Estimates applying this model to US real GDP suggest that US output experiences considerable permanent movements. This result is in stark contrast to the findings based on estimates of zero-correlation models which suggest that fluctuations in output are primarily transitory. If the transitory component of US real GDP is asymmetric, particularly during recessions, then the linear UC model may over-emphasize permanent movements due to the dominance of expansions in the data. This paper produces and estimates an unobserved components model that allows for asymmetric transitory movements and for correlation between all the innovations. The asymmetry is modeled using Markov-switching in the transitory component, in the spirit of Kim and Nelson's (1999) version of Friedman's plucking model. The findings suggest that both permanent movements and asymmetric transitory shocks are important for explaining post-war output fluctuations in the U.S.