The factors behind output gap revisions

Estimates of potential output and the output gap are widely used in monetary, fiscal and structural policy analysis, but they are unobservable and subject to substantial real-time uncertainty. This paper examines revisions to output gap and potential growth estimates produced by the European Commission, the Eurosystem, the International Monetary Fund (IMF) and the Organisation for Economic Co-operation and Development (OECD) for the euro area and its member countries. Using annual forecast vintages covering up to 20 euro area countries from 2002 to 2025 and a balanced panel of 11 countries from 2007 to 2025, the paper compares the reliability of estimates across institutions, decomposes output gap revisions into their main sources and estimates the impact of real GDP data revisions and forecast errors on potential growth revisions. The results show sizeable differences across institutions: European Commission estimates revise least in the balanced country sample, OECD estimates revise most, and Eurosystem estimates have become markedly more stable over the past decade. Output gap revisions are driven mainly by revisions to real GDP data and potential growth, while nowcast errors play a smaller role. Panel regressions show that both real GDP data revisions and medium-term forecast errors are significantly associated with revisions to potential growth, implying that revised or unexpected GDP developments are to some extent absorbed into estimates of potential growth. The findings highlight the uncertainty surrounding real-time output gap estimates and support the use of complementary indicators of economic slack. JEL Classification: E23, E27, E32, C53, C82