An unobserved components model that yields business and medium-run cycles

We generalize the unobserved components (UC) model to allow the permanent component to have different dynamics than the transitory components when decomposing U.S. economic activity using a multivariate UC model of (log) output, consumption, and investment. We find that these proposed dynamics in the permanent component are statistically significant and distinct from those of the transitory components. Our approach provides an alternative explanation for the growth cycles identified by Comin and Gertler (2006) that is related to the cyclical movements in technology, in a framework consistent with the Beveridge and Nelson (1981) decomposition.