Forecasting US Recessions in Real-Time Using Regional Economic Sentiment

Christian Garciga and James Mitchell

Economic Commentary, 2025, vol. 2025, issue 13, 12

Abstract: Measures of regional economic sentiment, extracted from the Beige Book using natural language processing methods, consistently delivered reliable real-time forecasts of US recessions from the mid-1980s through the COVID-19 pandemic recession. Since then, recession risk probabilities have been choppy, with several false alarms. We attribute this unreliability to a post-2021 disconnect between measures of economic activity and the sentiment of business and community leaders.

Date: 2025
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://doi.org/10.26509/frbc-ec-202513 Persistent link with full text (text/html)
https://www.clevelandfed.org/-/media/project/cleve ... ons-in-real-time.pdf Full text (application/pdf)
Our link check indicates that this URL is bad, the error code is: 404 Not Found

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:fip:fedcec:102077

Ordering information: This journal article can be ordered from

DOI: 10.26509/frbc-ec-202513

Access Statistics for this article

More articles in Economic Commentary from Federal Reserve Bank of Cleveland Contact information at EDIRC.
Bibliographic data for series maintained by 4D Library ().