The Measurement of Revision Effects on Business Cycle Turning Points.
Posted: 22 Oct 2019
Date Written: October 11, 2017
This paper examines the effects of revisions on quarterly GDP estimates in highlighting effects on business cycle turning points. We integrate a Markow-switching process to a dynamic factor model of monthly GDP. The quarterly averages of our monthly GDP estimates are, by construction, exactly equal to quarterly GDP estimates.
This study finds that the model used in incorporating the preliminary data is a flexible framework in estimating of the 'true' quarterly GDP. Data precision estimates are improved when the revision errors get smaller and unimproved on the other hand. When the model is estimated using data available immediately after 2001, it identifies a recession in 2001-02. But as revised data are added to the information set, the model assigns progressively lower probabilities to the recession state in 2001-02.
Keywords: Revision effects and turning points,'True' GDP estimates, Preliminary GDP, Dynamic factor model, exact constraint
JEL Classification: C01, C11, C13, C15
Suggested Citation: Suggested Citation