Deciphering Federal Reserve Communication via Text Analysis of Alternative FOMC Statements

Taeyoung Doh, Dongho Song and Shu-Kuei X. Yang

No RWP 20-14, Research Working Paper from Federal Reserve Bank of Kansas City

Abstract: We propose a text-based measure of monetary policy stance that models FOMC statements as convex combinations of dovish and hawkish alternatives, providing a tractable representation of the Committee's position along the policy spectrum. Leveraging staff-drafted alternative statements, we fine-tune a pre-trained language model to capture both quantitative precision and semantic tone. Stance is defined as the product of tone and novelty, and decomposed into expected and surprise components using high-frequency financial data. Surprises arise from shifts in tone relative to expectations or from statement novelty. Our framework enables counterfactuals showing how alternative communication could have moved markets.

Keywords: FOMC; Alternative FOMC statements; Counterfactual policy evaluation; Monetary policy stance; Text analysis; Natural language processing (search for similar items in EconPapers)
JEL-codes: E30 E40 E50 G12 (search for similar items in EconPapers)
Pages: 50
Date: 2020-10-06, Revised 2025-10-16
New Economics Papers: this item is included in nep-big, nep-cba, nep-cmp, nep-mac and nep-mon
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (10)

Downloads: (external link)
https://www.kansascityfed.org/research/research-wo ... n-via-text-analysis/ Full text (text/html)

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:fedkrw:88946

Ordering information: This working paper can be ordered from

DOI: 10.18651/RWP2020-14

Access Statistics for this paper

More papers in Research Working Paper from Federal Reserve Bank of Kansas City Contact information at EDIRC.
Bibliographic data for series maintained by Kira Lillard ().