From Tone to Trajectory: Continuous Sentiment and the Shape of Monetary Policy Communication

Martin Feldkircher; Márton Kardos; Kristoffer Laigaard Nielbo

cs.CLAug 24, 2026

From Tone to Trajectory: Continuous Sentiment and the Shape of Monetary Policy Communication

Authors: Martin Feldkircher, Márton Kardos, Kristoffer Laigaard Nielbo

Organizations: Vienna School of International Studies (DA) · Oesterreichische Nationalbank (OeNB) · Centre for Applied Macroeconomic Analysis (CAMA) · School of Culture and Society - Center for Humanities Computing (AU)

Abstract

Central bank press conferences are not merely information releases --- they are structured narratives. We study whether the shape of sentiment within a statement, not just its average tone, carries policy-relevant signals. Constructing sentiment arcs for ECB and Fed press conferences along three dimensions --- monetary stance, economic outlook, and uncertainty --- we assess their predictive content for policy rate changes, inflation expectations, and forecaster disagreement. Our findings show that arc shape robustly predicts rate decisions beyond lexicon-based benchmarks at both institutions --- it is not merely whether a statement sounds hawkish or economically optimistic on average, but how these sentiments are sequenced and emphasized across the statement, that carries the policy signal. Arc features also shape how professional forecasters update inflation expectations and how much they disagree, pointing to a receiver-side effect distinct from the direct policy signal. These findings suggest that communication design --- the sequencing and emphasis of policy language across a statement --- is a first-order feature of the policy signal, not a second-order refinement.

Figures & tables

Panel of Figure 1 — Figure 1: Distribution of scores for seed phrases

Panel of Figure 2 — Figure 2: Full-sample average sentiment arcs

Dimension	Pole	Date	Score	Sentence
Panel A: ECB
MP stance	Hawkish	2022-03	$+0.237$	Price rises have also become more broadly based.
MP stance	Dovish	2020-09	$-0.124$	We will also continue to provide ample liquidity through our refinancing operations.
Economic	Optimistic	2018-03	$+0.209$	The latest economic data and survey results indicate continued strong and broad-based growth momentum.
Economic	Pessimistic	2022-04	$-0.140$	Some sectors face growing difficulties in sourcing their inputs, which is disrupting production.
Uncertainty	Uncertain	2002-11	$+0.085$	The view has prevailed to keep interest rates unchanged.

Table 1: Highest and lowest scoring sentences by CVP dimension

Panel of Figure 3 — Figure 3: Sentiment arcs before rate changes

Panel of Figure 4 — Figure 4: Sentiment arcs during crisis and non-crisis periods

	ECB ( $N=261$ )		Fed ( $N=218$ )
	Full arc	Full arc $\times$ C	Full arc	Full arc $\times$ C
(Intercept)	$0.531^{*}$	$0.588^{*}$	$-0.590^{*}$	$-0.187$
	$(2.065)$	$(2.260)$	$(-2.363)$	$(-0.819)$
mp_level	$4.908^{***}$	$4.625^{***}$	$3.227^{***}$	$4.036^{***}$
	$(5.266)$	$(4.876)$	$(4.690)$	$(6.563)$
mp_slope	$3.099^{***}$	$2.495^{**}$	$0.831$	$1.795^{***}$

Table 4: Rate change regressions: crisis interactions

	Inflation expectations ( $\pi^{e}_{t+1}$ )				Forecaster disagreement ( $\sigma_{t+1}(\pi^{e})$ )
	ECB ( $N=225$ )		Fed ( $N=184$ )		ECB ( $N=225$ )		Fed ( $N=184$ )
	M1	Full arc	M1	Full arc	M1	Full arc	M1	Full arc
(Intercept)	$0.694^{***}$	$0.928^{.}$	$0.916^{***}$	$0.829$	$0.213^{***}$	$-0.297$	$0.316^{***}$	$0.510$
	$(17.961)$	$(1.724)$	$(15.180)$	$(1.191)$	$(7.496)$	$(-0.715)$	$(8.515)$	$(1.204)$
$\Delta p^{\mathrm{yoy}}_{t}$	$0.561^{***}$	$0.499^{***}$	$0.474^{***}$	$0.441^{***}$	$0.033^{***}$	$0.042^{*}$	$0.015$	$0.018$
	$(42.824)$	$(21.169)$	$(26.171)$	$(18.835)$	$(3.389)$	$(2.317)$	$(1.380)$	$(1.230)$

Table 5: Inflation expectations and forecaster disagreement regressions

	ECB ( $\Delta i_{t+1}$ , $N=261$ )			ECB ( $\Delta i_{t+2}$ , $N=260$ )
	Base	Base $+$ D fg	Base $+$ D new	Base	Base $+$ D fg	Base $+$ D new
(Intercept)	$0.531^{*}$	$0.548^{*}$	$0.521^{*}$	$0.999^{*}$	$1.097^{**}$	$1.042^{*}$
	$(2.065)$	$(2.116)$	$(2.010)$	$(2.398)$	$(2.637)$	$(2.478)$
mp_level	$4.908^{***}$	$4.763^{***}$	$5.108^{***}$	$8.972^{***}$	$8.152^{***}$	$8.201^{***}$
	$(5.266)$	$(4.953)$	$(4.631)$	$(5.953)$	$(5.283)$	$(4.569)$
mp_slope	$3.099^{***}$	$2.956^{***}$	$3.235^{***}$	$5.354^{***}$	$4.548^{**}$	$4.826^{**}$

Table 6: Rate change regressions: ECB with communication-regime dummies

	ECB (contextual)	ECB (sentence)	Fed (contextual)	Fed (sentence)
(Intercept)	$0.531^{{}^{*}}$	$0.519^{{}^{*}}$	$-0.590^{{}^{*}}$	$-0.930^{{}^{**}}$
	$(2.065)$	$(2.395)$	$(-2.363)$	$(-3.084)$
mp_level	$4.908^{{}^{***}}$	$5.069^{{}^{***}}$	$3.227^{{}^{***}}$	$4.503^{{}^{***}}$
	$(5.266)$	$(4.011)$	$(4.690)$	$(3.727)$
mp_slope	$3.099^{{}^{***}}$	$2.139^{{}^{**}}$	$0.831$	$2.750^{{}^{***}}$
	$(3.697)$	$(2.661)$	$(1.593)$	$(4.779)$

Table 7: Full arc model: contextual vs. sentence-level arcs

Appendix figures & tables6 assets

Supplementary material from the paper’s appendix.

Appendix

NS arc factors: monetary policy sentiment
	ECB ( $N=269$ )					Fed ( $N=219$ )
	Mean	SD	Min	Median	Max	Mean	SD	Min	Median	Max
mp_level	0.041	0.037	-0.065	0.036	0.148	-0.015	0.049	-0.183	-0.013	0.149
mp_slope	-0.078	0.043	-0.198	-0.072	0.069	-0.012	0.062	-0.217	-0.012	0.135
mp_curve	0.055	0.139	-0.345	0.071	0.517	0.135	0.154	-0.294	0.149	0.484
NS arc factors: economic sentiment

Appendix

Table 8: Descriptive statistics: arc features

Panel of Figure 5 — Figure 5: Seed phrase validation: pairwise CVP score scatter

Panel of Figure 6 — Figure 6: Seed phrase validation: PCA of seed embeddings

Hawkish monetary policy (11 seeds)
Inflation remains too high and requires further policy tightening.
Rising inflationary pressures call for higher interest rates.
Strong wage growth poses upside risks to inflation.
Financial conditions remain loose given current inflation dynamics.
Interest rates will need to increase further to ensure price stability.
Persistent inflation requires a firm and timely policy response.

Appendix

Table 10: Seed phrases: Monetary sentiment

Economic sentiment: positive (14 seeds)
The economy is expanding at a solid pace with strong job growth.
Consumer demand remains strong and supports continued growth.
Labor-market conditions are strong and improving.
Business investment continues to strengthen.
Economic momentum has increased across many sectors.
Job gains remain robust and unemployment is low.

Appendix

Table 11: Seed phrases: Economic sentiment

High uncertainty (15 seeds)
The outlook remains highly uncertain.
The recovery process is likely to be uneven and subject to high uncertainty.
Economic growth is expected to remain uneven in an environment of uncertainty.
The outlook is subject to particularly high uncertainty and intensified downside risks.
The near-term economic outlook remains clouded by uncertainty.
The speed and scale of the recovery remain highly uncertain.

Appendix

Table 12: Seed phrases: Uncertainty sentiment

Explore similar work

Forecasting Intraday USD/CAD Exchange Rate with News-Derived Monetary-Policy Signals
Sep 26, 2026Maya Kodeih, Aliaa Alnaggar, Mucahit CevikFinance Financial Forecasting
Department of Mechanical and Industrial Engineering, Toronto Metropolitan University, Toronto, ON, Canada
Reading Copom's Tone: A Weighted LLM Framework for Hawkish-Dovish Sentiment, Forward Guidance, and Uncertainty
Aug 7, 2026Gabriel de Macedo SantosSentiment Analysis Financial Sentiment Analysis
Instituto de Tecnologia e Liderança
Mind the Shift: Decoding Monetary Policy Stance from FOMC Statements with Large Language Models
Mar 15, 2026Yixuan Tang, Yi YangStance Detection Finance
The Hong Kong University of Science and Technology

Sep 26, 2026J/K move · Enter open · S save

Forecasting Intraday USD/CAD Exchange Rate with News-Derived Monetary-Policy Signals

Maya Kodeih, Aliaa Alnaggar, Mucahit Cevik

Department of Mechanical and Industrial Engineering, Toronto Metropolitan University, Toronto, ON, Canada

Finance Financial Forecasting Feature Attribution

Open paper AskPDF arXiv

Monetary-policy announcements and central-bank communications play a central role in foreign exchange markets, yet their qualitative, unstructured form makes their forecasting value difficult to quantify. While prior research has largely focused on sentiment extracted from financial news, comparatively little is known about the relative contribution of different dimensions of monetary-policy communication. Existing studies primarily evaluate whether textual information improves overall forecasting performance but provide limited insight into which communication channels drive such improvements. To address this gap, this paper introduces a statistical attribution methodology that decomposes monetary-policy communication into interpretable channels and quantifies their incremental forecasting contribution under false-discovery-rate control. Monetary-policy news is transformed into structured communication signals using large language models (LLMs) and temporal feature engineering. These signals are evaluated using rolling-window experiments with tree-based machine-learning models. The results show that monetary-policy communication contains measurable predictive information. Attribution analysis shows that predictive value is concentrated in a small subset of signals, with communication timing providing the strongest individual feature-level contribution, targeted communication-activity measures also contributing positively, and LLM-derived sentiment providing complementary information at the group level. The findings indicate that communication-based forecasting value extends beyond sentiment alone and that attribution, rather than aggregate accuracy alone, is central to evaluating news-derived signals.

Abstract

Explore similar work

Forecasting Intraday USD/CAD Exchange Rate with News-Derived Monetary-Policy Signals

Reading Copom's Tone: A Weighted LLM Framework for Hawkish-Dovish Sentiment, Forward Guidance, and Uncertainty

Mind the Shift: Decoding Monetary Policy Stance from FOMC Statements with Large Language Models