Analyzing U.S. Output and the Great Moderation by Simultaneous Unobserved Components

Abstract

This paper seeks to determine the causal interaction between structural trend and cycle innovations in an unobserved components framework of aggregate output. For the purpose of identification, I propose allowing for shifts in volatility. This strategy provides good estimation precision when applied to U.S. industrial production. In the early 1980s, predominance of cycle shocks gives way to strong negative spillovers of trend impulses, consistent with real business cycle theories. The coincident reduction of macroeconomic volatility was mainly caused by pronounced dampening of transitory disturbances. This is in accordance with an important role of macroeconomic policy in explaining the Great Moderation.

Trend-Cycle Decomposition: Implications from an Exact Structural Identification

“…When applied to US GDP data we find that the results for 1947Q1 to 2012Q3 are more consistent with a model where cycle shocks enter trend, rather than when trend shocks enter cycle. There is some support for this result in the existing literature in the two-regime model for industrial production of Weber (2011). Furthermore, all of the models consistent with the data (BN,MNZ and RC2T) imply that smoothed cyclical fluctuations are many times larger than filtered cycles, reflecting that filtered estimates are not reliable indicators of business cycles.…”

“…We show the difficulties in obtaining a structural form identification for the interactions between shocks to trend and shocks to cycle in the general state-space form of the BN decomposition as provided in MNZ, also noted by Proietti (2006) and Weber (2011), and how these may be resolved with assumptions adopted from JvN. By way of illustration we apply these to US GDP and show that the data supports the interpretation of cycle shocks influencing trend rather than the alternative that trend shocks influence cycle.…”

On the identification of multivariate correlated unobserved components models

“…In general, our results can provide a useful basis for growing strands of literature that apply and develop correlated UC models, e.g. Morley (2007), Sinclair (2009), Startz & Tsang (2010, Weber (2011), Klinger & Weber (2014).…”

Trend–cycle–seasonal Interactions: Identification and Estimation

“…As noted in the Introduction, a growing literature provides empirical evidence that the trend (permanent) and cycle (transient) components of economic time series are correlated. As discussed by Weber (2011), the economic rationale for such correlation can include real business cycle theories, nominal rigidities, hysteresis, policy responses to temporary shocks, and so on. Estimates of the correlation between the innovations of the trend and cycle for output 4 or related series (such as employment) are negative and relatively close to −1; for example, MNZ, Sinclair (2010), Weber (2011), Dungey et al (2015.…”