Is the Hamilton Regression Filter Really Superior to Hodrick-Prescott Detrending? Extended Version

A heavily revised and streamlined version of this WP was published in Macroeconomic Dynamics (2024). The final authenticated version is available online at DOI: doi.org/10.1017/S136510052400018X (Open Access)

31 Pages Posted: 17 Sep 2022 Last revised: 2 May 2024

See all articles by Reiner Franke

Reiner Franke

University of Bremen - Department of Economics; University of Kiel - Institute of Economics

Jiri Kukacka

Charles University - Institute of Economic Studies; Academy of Sciences of the Czech Republic

Stephen Sacht

University of Kiel

Date Written: June 30, 2023

Abstract

A recent article by J.D. Hamilton from 2018 attracted a great deal of attention, not only because of its telling title, "Why you should never use the Hodrick- Prescott filter", but also because it offered an alternative approach to detrending, the Hamilton regression filter (HRF). His contribution was actually often read as a suggestion to replace the Hodrick-Prescott (HP) filter with HRF, which also means using and interpreting it in the same way as HP. The present paper challenges this view, at least with respect to quarterly business cycle data of aggregate output. Geared towards the US business cycle, it generates a great number of artificial data with a known trend and cyclical component and asks how well a given detrending method can recover the true decomposition. Moreover, besides the usual HP smoothing parameter λ = 1600 the investigation considers values λ* from previous work that are seven to twelve times higher. On the basis of three different statistics measuring a distance between the estimated and true trend, it turns out that HRF is distinctly out-performed by both versions of HP, while HP with λ* is systematically better than HP-1600.

Keywords: business cycles, smoothing parameter, trend concept, growth regimes

JEL Classification: C18, C32, E32

Suggested Citation

Franke, Reiner and Kukacka, Jiri and Sacht, Stephen, Is the Hamilton Regression Filter Really Superior to Hodrick-Prescott Detrending? Extended Version (June 30, 2023). A heavily revised and streamlined version of this WP was published in Macroeconomic Dynamics (2024). The final authenticated version is available online at DOI: doi.org/10.1017/S136510052400018X (Open Access), Available at SSRN: https://ssrn.com/abstract=4210446 or http://dx.doi.org/10.2139/ssrn.4210446

Reiner Franke

University of Bremen - Department of Economics ( email )

Bremen, D-28334
Germany
++49 421 218 7392 (Phone)
++49 421 218 4336 (Fax)

University of Kiel - Institute of Economics

Olshausenstrasse 40
24098 Kiel, 24098
Germany

Jiri Kukacka (Contact Author)

Charles University - Institute of Economic Studies ( email )

Opletalova 26
Prague 1, CZ-11000
Czech Republic

HOME PAGE: http://ies.fsv.cuni.cz/en/contacts/people/58305408

Academy of Sciences of the Czech Republic ( email )

Pod Vodarenskou vezi 4
Prague 8, CZ-18200
Czech Republic

HOME PAGE: http://www.utia.cas.cz/people/kukacka

Stephen Sacht

University of Kiel ( email )

Olshausenstr. 40
D-24118 Kiel, Schleswig-Holstein 24118
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
171
Abstract Views
668
Rank
322,888
PlumX Metrics