Too Many Shocks Spoil the Interpretation

42 Pages Posted: 6 Apr 2020

See all articles by Adrian Pagan

Adrian Pagan

University of Sydney; Australian National University (ANU) - Centre for Applied Macroeconomic Analysis (CAMA)

Tim Robinson

University of Melbourne - Melbourne Institute: Applied Economic & Social Research

Multiple version iconThere are 2 versions of this paper

Date Written: April 2, 2020

Abstract

We show that when a model has more shocks than observed variables the estimated filtered and smoothed shocks will be correlated. This is despite no correlation being present in the data generating process. Additionally the estimated shock innovations may be autocorrelated. These correlations limit the relevance of impulse responses, which assume uncorrelated shocks, for interpreting the data. Excess shocks occur frequently, e.g. in Unobserved-Component (UC) models, filters, including Hodrick-Prescott (1997), and some Dynamic Stochastic General Equilibrium (DSGE) models. Using several UC models and an estimated DSGE model, Ireland (2011), we demonstrate that sizable correlations among the estimated shocks can result.

Keywords: Partial Information; Structural Shocks; Kalman Filter; Measurement Error; DSGE

JEL Classification: E37; C51; C52

Suggested Citation

Pagan, Adrian and Robinson, Tim, Too Many Shocks Spoil the Interpretation (April 2, 2020). Melbourne Institute Working Paper No. 2/20, Available at SSRN: https://ssrn.com/abstract=3567444 or http://dx.doi.org/10.2139/ssrn.3567444

Adrian Pagan

University of Sydney ( email )

Rm 370 Merewether (H04)
The University of Sydney
Sydney, NSW 2006 2008
Australia

Australian National University (ANU) - Centre for Applied Macroeconomic Analysis (CAMA) ( email )

Tim Robinson (Contact Author)

University of Melbourne - Melbourne Institute: Applied Economic & Social Research ( email )

Level 5, FBE Building, 111 Barry Street
161 Barry Street
Carlton, VIC 3053
Australia

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
61
Abstract Views
390
Rank
486,005
PlumX Metrics