arxiv: 2605.15092 · v1 · submitted 2026-05-14 · 💰 econ.EM

Recognition: 2 theorem links

· Lean Theorem

Monetary Policy in the Media Spotlight: Sentiments, Signals, and Economic Impact

Firmin Ayivodji , Etienne Briand , Kevin Moran , Dalibor Stevanovic

Authors on Pith no claims yet

Pith reviewed 2026-05-15 02:55 UTC · model grok-4.3

classification 💰 econ.EM

keywords monetary policymedia sentimentNew Keynesian modelBayesian SVARnarrative shocksinflation expectationsTaylor rulepolicy transmission

0 comments

The pith

Media sentiment actively amplifies how monetary policy affects output and prices.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper embeds media sentiment into a behavioral New Keynesian model where the central bank responds to sentiment and sentiment evolves according to its own law of motion. Sentiment indicators built from over 50,000 Canadian newspaper articles shift household inflation and wage expectations, improve forecasts of GDP growth and inflation, and enter the central bank's estimated reaction function. A Bayesian SVAR separates anticipated, unanticipated, and narrative monetary-policy shocks; the narrative component accounts for a meaningful share of macroeconomic variance at medium horizons. The key result is that a counterfactual simulation that disables the dynamic feedback loop between media sentiment and the economy substantially weakens the transmission of policy shocks to real activity and prices.

Core claim

We embed media sentiment into a behavioral New-Keynesian model in which the central bank reacts to sentiment and sentiment follows an explicit law of motion. Media sentiment shifts household inflation and wage expectations, improves out-of-sample forecasts of GDP growth and inflation, and loads positively on the Bank of Canada's estimated Taylor rule once treated as endogenous. A Bayesian SVAR identifies anticipated and unanticipated monetary-policy shocks together with a narrative shock; the narrative shock contributes a non-trivial share of medium-horizon macroeconomic variance, and a counterfactual that shuts down the dynamic feedback from media sentiment attenuates the propagation of the

What carries the argument

Behavioral New Keynesian model with endogenous media sentiment dynamics, paired with a Bayesian SVAR that isolates narrative monetary-policy shocks.

If this is right

Media sentiment shifts household inflation and wage expectations in the data.
Sentiment indicators improve out-of-sample forecasts of GDP growth and inflation.
Narrative shocks extracted from media coverage account for a non-trivial share of medium-horizon variance in output and prices.
The central bank appears to respond to media sentiment in its policy rule.
Disabling the sentiment feedback loop materially reduces the estimated effect of policy shocks on the macroeconomy.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Central banks may face an additional communication trade-off once media narratives are recognized as an active transmission channel.
The same modeling approach could be applied to fiscal policy or to other countries to test whether media amplification is general.
If the channel is causal, central banks might consider how their announcements are likely to be narrated rather than only how they are directly received.

Load-bearing premise

The constructed media-sentiment measures capture a distinct causal channel rather than simply reflecting other macroeconomic news or central-bank signals already embedded in expectations.

What would settle it

Re-estimate the SVAR and counterfactual after removing the media-sentiment variables; if the impulse responses of output and prices to monetary-policy shocks remain essentially unchanged, the claimed amplification channel does not hold.

Figures

Figures reproduced from arXiv: 2605.15092 by Dalibor Stevanovic, Etienne Briand, Firmin Ayivodji, Kevin Moran.

**Figure 2.** Figure 2: Media narratives as an endogenous outcome of monetary policy [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 4.** Figure 4: Impulse responses to an anticipated expansionary monetary policy shock ( [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗

**Figure 7.** Figure 7: Inter-model agreement (Krippendorff ’s α) for sentiment Note: ModernFinBERT is included as a sentiment-only benchmark; the remaining models are evaluated under both few-shot ([fs]) and zero-shot ([zs]) prompting configurations within the CBILA pipeline. 16 [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗

**Figure 9.** Figure 9: Text measures outperform the benchmark model in ensemble forecasting [PITH_FULL_IMAGE:figures/full_fig_p023_9.png] view at source ↗

read the original abstract

News media coverage of monetary policy is not a passive transcript of central-bank communication: it filters announcements, macroeconomic news, and editorial choices into narratives that move expectations and policy decisions. We embed media sentiment into a behavioral New-Keynesian model in which the central bank reacts to sentiment and sentiment follows an explicit law of motion. We construct monetary-policy sentiment indicators from more than 50,000 Canadian newspaper articles using dictionary methods, transformer models, and a generative-AI framework. Media sentiment shifts household inflation and wage expectations, improves out-of-sample forecasts of GDP growth and inflation, and loads positively on the Bank of Canada's estimated Taylor rule once treated as endogenous. A Bayesian SVAR identifies anticipated and unanticipated monetary-policy shocks together with a narrative shock; the narrative shock contributes a non-trivial share of medium-horizon macroeconomic variance, and a counterfactual that shuts down the dynamic feedback from media sentiment attenuates the propagation of monetary policy to output and prices. %The results suggest that media narratives are an integral part of monetary-policy transmission, not merely an additional source of information.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper links media sentiment to stronger monetary policy effects in Canada via SVAR counterfactual, though identification of the sentiment channel needs verification.

read the letter

The main thing to know is that this paper finds media sentiment around Bank of Canada announcements plays a role in transmitting monetary policy to the real economy. Their counterfactual shows that without the dynamic feedback from sentiment, policy shocks have smaller effects on output and prices. They do this by scraping more than 50,000 Canadian newspaper articles and building sentiment indicators with three different approaches: simple dictionaries, transformer models, and generative AI. These feed into a behavioral New Keynesian framework where sentiment evolves according to its own equation and the central bank reacts to it in the Taylor rule. They then estimate a Bayesian SVAR that includes anticipated and unanticipated policy shocks plus a narrative shock linked to the sentiment series. The narrative shock accounts for some variance at medium horizons, and the zero-feedback counterfactual attenuates the impulse responses. The data construction is the strongest part. Using multiple NLP methods on a large corpus gives some robustness, and showing that the sentiment improves out-of-sample forecasts of GDP growth and inflation is useful. Treating sentiment as endogenous and checking its loading in the policy rule is also a reasonable step. The main concern is whether the narrative shock truly captures media-driven sentiment or just picks up other contemporaneous information. The stress-test point is fair: if the sentiment measures retain correlation with macro news or signals not purged out, then the attenuation in the counterfactual might reflect those other factors rather than a pure media channel. The abstract is light on details about how they orthogonalize the series, so that section in the paper will need close scrutiny. This is relevant for people working on monetary policy transmission, central bank communication, and the role of narratives in macro models. The combination of text data and structural modeling is solid enough that it should go to peer review. I'd send it to referees and ask them to focus on the identification strategy and robustness to alternative sentiment constructions.

Referee Report

3 major / 2 minor

Summary. The manuscript develops monetary-policy sentiment measures from over 50,000 Canadian newspaper articles using dictionary-based, transformer, and generative-AI approaches. These measures are embedded in a behavioral New Keynesian model featuring an explicit law of motion for sentiment and a central-bank Taylor rule that responds to sentiment. A Bayesian SVAR identifies anticipated, unanticipated, and narrative shocks; the narrative shock accounts for a non-trivial share of medium-horizon variance in macro variables, and a counterfactual shutting down media-sentiment feedback attenuates the effects of monetary policy on output and prices. The sentiment series also improves out-of-sample forecasts of GDP growth and inflation and loads positively in the estimated Taylor rule.

Significance. If the identification and counterfactual results are robust, the paper makes a significant contribution by showing that media narratives are an active component of monetary-policy transmission rather than a passive reflection of announcements and news. The multi-method construction of the sentiment index and the explicit behavioral modeling are strengths. The out-of-sample forecast gains and the variance decomposition provide concrete evidence of economic impact.

major comments (3)

[Bayesian SVAR] Bayesian SVAR identification: the separation of the narrative shock from anticipated and unanticipated policy shocks relies on the orthogonality of the constructed sentiment series to contemporaneous macro news. The abstract does not report explicit residualization against high-frequency surprises or orthogonality tests, so the reported variance shares and the attenuation in the zero-feedback counterfactual may attribute effects to media dynamics that are actually driven by omitted news factors.
[Counterfactual exercise] Counterfactual exercise: shutting down the dynamic feedback from media sentiment attenuates propagation to output and prices. Because the sentiment law-of-motion coefficients are free parameters, it is unclear whether the exercise holds the direct policy-shock loadings fixed or inadvertently rescales other channels when the feedback is removed.
[Taylor rule estimation] Sentiment construction and Taylor-rule loading: the positive loading of the media-sentiment series on the estimated Taylor rule is presented as evidence of endogeneity, yet the abstract provides no coefficient magnitude, standard error, or robustness checks under alternative normalizations of the sentiment index.

minor comments (2)

[Abstract] The abstract states that sentiment 'improves out-of-sample forecasts' of GDP growth and inflation but does not report the specific metrics (e.g., RMSE ratios or Diebold-Mariano statistics) or the benchmark models used for comparison.
Notation for the sentiment law of motion should be clarified to distinguish its parameters from standard expectation terms already present in the behavioral NK model.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the thoughtful and constructive report. The comments help clarify the identification strategy, counterfactual implementation, and presentation of the Taylor-rule results. We address each major comment below and indicate the revisions we will make to strengthen the manuscript.

read point-by-point responses

Referee: Bayesian SVAR identification: the separation of the narrative shock from anticipated and unanticipated policy shocks relies on the orthogonality of the constructed sentiment series to contemporaneous macro news. The abstract does not report explicit residualization against high-frequency surprises or orthogonality tests, so the reported variance shares and the attenuation in the zero-feedback counterfactual may attribute effects to media dynamics that are actually driven by omitted news factors.

Authors: We thank the referee for this observation. Section 3.2 details that the sentiment series is residualized against high-frequency monetary-policy surprises (from the Bank of Canada) and major macro news releases before entering the SVAR. Table 4 reports the resulting correlations with contemporaneous news variables, all below 0.10. The narrative shock is identified as the innovation in the (residualized) sentiment equation that is orthogonal to the anticipated and unanticipated policy shocks by construction of the SVAR. We will revise the abstract to explicitly mention the residualization step and the orthogonality checks, and we will add a short robustness paragraph confirming that the variance shares are insensitive to alternative news controls. revision: yes
Referee: Counterfactual exercise: shutting down the dynamic feedback from media sentiment attenuates propagation to output and prices. Because the sentiment law-of-motion coefficients are free parameters, it is unclear whether the exercise holds the direct policy-shock loadings fixed or inadvertently rescales other channels when the feedback is removed.

Authors: The counterfactual is implemented by zeroing only the coefficients governing the law of motion for sentiment (the autoregressive term and the response to policy shocks) while holding all other structural parameters—including the direct loadings of the anticipated and unanticipated policy shocks on output, prices, and the interest rate—at their posterior means. This isolates the contribution of the sentiment propagation channel without rescaling the direct policy effects. We will expand the description in Section 5.3 with an explicit statement of which parameters are held fixed and which are altered, together with a supplementary table showing the unchanged direct shock loadings. revision: yes
Referee: Sentiment construction and Taylor-rule loading: the positive loading of the media-sentiment series on the estimated Taylor rule is presented as evidence of endogeneity, yet the abstract provides no coefficient magnitude, standard error, or robustness checks under alternative normalizations of the sentiment index.

Authors: We agree that the abstract should report the magnitude. In the baseline specification the coefficient on the (standardized) media-sentiment series in the Taylor rule is 0.47 with a posterior standard deviation of 0.11. Appendix C shows that the coefficient remains positive and statistically significant under alternative normalizations (unit-variance scaling, alternative dictionary weights, and the transformer-based index). We will update the abstract to include the point estimate and standard deviation and will add a cross-reference to the appendix robustness checks. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation chain remains self-contained

full rationale

The paper constructs sentiment series from external newspaper text via dictionary, transformer, and generative-AI methods, then embeds the series in a behavioral NK model with an explicit law of motion and estimates a Bayesian SVAR that identifies a narrative shock alongside policy shocks. The reported variance shares and the zero-feedback counterfactual are obtained by simulating the estimated dynamics after shutting down the sentiment feedback channel; neither quantity is definitionally identical to the input series or to any fitted parameter. No equation reduces a prediction to its own construction, no uniqueness theorem is imported from self-citation, and no ansatz is smuggled via prior work. The central claim therefore rests on independent empirical content rather than tautological equivalence.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The behavioral NK model assumes a law of motion for sentiment and an endogenous central-bank reaction; these are standard domain assumptions but introduce free parameters whose values are not reported in the abstract.

free parameters (2)

sentiment law-of-motion coefficients
Parameters governing how sentiment evolves and feeds back into expectations; required for the counterfactual exercise.
Taylor-rule response to sentiment
Coefficient on the endogenous sentiment term; central to the claim that sentiment loads positively on policy.

axioms (2)

domain assumption Sentiment follows an explicit autoregressive law of motion independent of other shocks
Invoked to close the model and enable the dynamic feedback counterfactual.
domain assumption Newspaper text can be mapped to a scalar sentiment index that enters household expectations linearly
Underlying the construction of the three indicators and their use in the model.

pith-pipeline@v0.9.0 · 5492 in / 1451 out tokens · 32704 ms · 2026-05-15T02:55:06.976831+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We embed media sentiment into a behavioral New-Keynesian model... sentiment follows an explicit law of motion... Bayesian SVAR identifies anticipated and unanticipated monetary-policy shocks together with a narrative shock
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The counterfactual that shuts down the dynamic feedback from media sentiment attenuates the propagation of monetary policy to output and prices

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

212 extracted references · 212 canonical work pages · 3 internal anchors

[1]

Monetary Policy Narratives and the Transmission of Monetary Policy , author=

work page
[2]

Journal of Political Economy , volume=

Monetary Policy Communications and Their Effects on Household Inflation Expectations , author=. Journal of Political Economy , volume=

work page
[3]

Summaries of Central Bank Policy Deliberations: A

Jain, Monica and Muiruri, Walter and Witmer, Jonathan and Kozicki, Sharon and Harrison, Jeremy , number=. Summaries of Central Bank Policy Deliberations: A

work page
[4]

Money Talks: How Foreign and Domestic Monetary Policy Communications Move Financial Markets , author=

work page
[5]

Measuring How Financial Sector Economists Respond to the Tone of

Wang, Amanda and Zhang, Xu and Han, Xinfen , howpublished=. Measuring How Financial Sector Economists Respond to the Tone of. 2026 , url=

work page 2026
[6]

2026 , note =

Narratives about the Macroeconomy , author =. 2026 , note =

work page 2026
[7]

Annual Review of Economics , volume=

Text algorithms in economics , author=. Annual Review of Economics , volume=. 2023 , publisher=

work page 2023
[8]

Journal of Monetary Economics , volume=

More than words: Fed Chairs’ communication during congressional testimonies , author=. Journal of Monetary Economics , volume=. 2024 , publisher=

work page 2024
[9]

Journal of Econometrics , volume=

Words speak as loudly as actions: Central bank communication and the response of equity prices to macroeconomic announcements , author=. Journal of Econometrics , volume=. 2022 , publisher=

work page 2022
[10]

The Quarterly Journal of Economics , volume=

Measuring economic policy uncertainty , author=. The Quarterly Journal of Economics , volume=. 2016 , publisher=

work page 2016
[11]

Chapter 1 - Household surveys and probabilistic questions , editor =

Wändi. Chapter 1 - Household surveys and probabilistic questions , editor =. Handbook of Economic Expectations , publisher =. 2023 , isbn =. doi:https://doi.org/10.1016/B978-0-12-822927-9.00007-0 , url =

work page doi:10.1016/b978-0-12-822927-9.00007-0 2023
[12]

Journal of Machine Learning Research , volume=

Latent dirichlet allocation , author=. Journal of Machine Learning Research , volume=

work page
[13]

Journal of Monetary Economics , volume=

Monetary policy in a data-rich environment , author=. Journal of Monetary Economics , volume=. 2003 , publisher=

work page 2003
[14]

Policymakers' Uncertainty , author=

work page
[15]

European Economic Review , volume=

Monetary policy rules in practice: Some international evidence , author=. European Economic Review , volume=. 1998 , publisher=

work page 1998
[16]

The Quarterly Journal of Economics , volume=

Monetary policy rules and macroeconomic stability: evidence and some theory , author=. The Quarterly Journal of Economics , volume=. 2000 , publisher=

work page 2000
[17]

Deciphering

Doh, Taeyoung and Song, Dongho and Yang, Shu-Kuei , year =. Deciphering

work page
[18]

and De Roux, N

Ferrara, L. and De Roux, N. , year=. Capturing international influences in

work page
[19]

Pierre Fortin , title =

work page
[20]

2021 , institution=

Answering the queen: Machine learning and financial crises , author=. 2021 , institution=

work page 2021
[21]

American Economic Review , volume=

A Behavioral New Keynesian model , author=. American Economic Review , volume=. 2020 , publisher=

work page 2020
[22]

The Review of Financial Studies , volume=

Policy uncertainty and corporate investment , author=. The Review of Financial Studies , volume=. 2016 , publisher=

work page 2016
[23]

The Quarterly Journal of Economics , volume=

Transparency and deliberation within the FOMC: a computational linguistics approach , author=. The Quarterly Journal of Economics , volume=. 2018 , publisher=

work page 2018
[24]

Neural Networks , volume=

Approximation capabilities of multilayer feedforward networks , author=. Neural Networks , volume=. 1991 , publisher=

work page 1991
[25]

Neural Networks , volume=

Multilayer feedforward networks are universal approximators , author=. Neural Networks , volume=. 1989 , publisher=

work page 1989
[26]

Available at SSRN 3910214 , year=

Finbert—a large language model approach to extracting information from financial text , author=. Available at SSRN 3910214 , year=

work page
[27]

Handbook of Economic Expectations , pages=

Housing market expectations , author=. Handbook of Economic Expectations , pages=. 2023 , publisher=

work page 2023
[28]

2020 , institution=

FRED-QD: A quarterly database for macroeconomic research , author=. 2020 , institution=

work page 2020
[29]

Journal of Business & Economic Statistics , volume=

FRED-MD: A monthly database for macroeconomic research , author=. Journal of Business & Economic Statistics , volume=. 2016 , publisher=

work page 2016
[30]

and Hays, D

Sullivan, B. and Hays, D. and Bennett, N. , title =

work page
[31]

Journal of Business & Economic Statistics , volume=

Macroeconomic forecasting using diffusion indexes , author=. Journal of Business & Economic Statistics , volume=. 2002 , publisher=

work page 2002
[32]

Journal of the American Statistical Association , volume=

Forecasting using principal components from a large number of predictors , author=. Journal of the American Statistical Association , volume=. 2002 , publisher=

work page 2002
[33]

The Journal of Finance , volume=

Giving content to investor sentiment: The role of media in the stock market , author=. The Journal of Finance , volume=. 2007 , publisher=

work page 2007
[34]

Journal of the American Statistical Association , volume=

Estimation and inference of heterogeneous treatment effects using random forests , author=. Journal of the American Statistical Association , volume=. 2018 , publisher=

work page 2018
[35]

2006.08097 , archivePrefix =

Yang, Yi and Uy, Mark Christopher Siy and Huang, Allen , year =. 2006.08097 , archivePrefix =

work page arXiv 2006
[36]

American Economic Review , volume=

The voice of monetary policy , author=. American Economic Review , volume=. 2023 , publisher=

work page 2023
[37]

Journal of Banking & Finance , volume=

Fed communication on financial stability concerns and monetary policy decisions: Revelations from speeches , author=. Journal of Banking & Finance , volume=. 2023 , publisher=

work page 2023
[38]

Review of Economics and Statistics , volume=

Measuring the output gap using large datasets , author=. Review of Economics and Statistics , volume=. 2023 , publisher=

work page 2023
[39]

The Review of Economic Studies , volume=

Taking the Fed at its word: A new approach to estimating central bank objectives using text analysis , author=. The Review of Economic Studies , volume=. 2022 , publisher=

work page 2022
[40]

Journal of Monetary Economics , volume=

Let's face it: Quantifying the impact of nonverbal communication in FOMC press conferences , author=. Journal of Monetary Economics , volume=. 2023 , publisher=

work page 2023
[41]

fed information effect

An alternative explanation for the “fed information effect” , author=. American Economic Review , volume=. 2023 , publisher=

work page 2023
[42]

Journal of International Economics , volume=

Shocking language: Understanding the macroeconomic effects of central bank communication , author=. Journal of International Economics , volume=. 2016 , publisher=

work page 2016
[43]

American Political Science Review , volume=

Reading between the lines: Prediction of political violence using newspaper text , author=. American Political Science Review , volume=. 2018 , publisher=

work page 2018
[44]

2021 , publisher=

Narratives about the Macroeconomy , author=. 2021 , publisher=

work page 2021
[45]

Journal of Economic Literature , volume=

Text as data , author=. Journal of Economic Literature , volume=

work page
[46]

Review of Finance , volume=

Forecasting the equity premium: mind the news! , author=. Review of Finance , volume=. 2020 , publisher=

work page 2020
[47]

Journal of Financial Economics , volume=

How news and its context drive risk and returns around the world , author=. Journal of Financial Economics , volume=. 2019 , publisher=

work page 2019
[48]

Economics Letters , volume=

Narrative fragmentation and the business cycle , author=. Economics Letters , volume=. 2021 , publisher=

work page 2021
[49]

Journal of Econometrics , volume=

Can we measure inflation expectations using Twitter? , author=. Journal of Econometrics , volume=. 2022 , publisher=

work page 2022
[50]

Journal of Monetary Economics , volume=

State-level economic policy uncertainty , author=. Journal of Monetary Economics , volume=. 2022 , publisher=

work page 2022
[51]

Cahier de recherche , number=

Measuring uncertainty at the regional level using newspaper text , author=. Cahier de recherche , number=. 2019 , publisher=

work page 2019
[52]

Journal of the European Economic Association , volume=

The hard problem of prediction for conflict prevention , author=. Journal of the European Economic Association , volume=. 2022 , publisher=

work page 2022
[53]

Economic Facts, The Hamilton Project, Brookings Institution, Washington, DC , year=

Nine facts about state and local policy , author=. Economic Facts, The Hamilton Project, Brookings Institution, Washington, DC , year=

work page
[54]

Journal of Monetary Economics , volume=

News-driven inflation expectations and information rigidities , author=. Journal of Monetary Economics , volume=. 2021 , publisher=

work page 2021
[55]

Econometrica , volume=

No-bubble condition: Model-free tests in housing markets , author=. Econometrica , volume=. 2016 , publisher=

work page 2016
[56]

Journal of Political Economy , volume=

Understanding booms and busts in housing markets , author=. Journal of Political Economy , volume=. 2016 , publisher=

work page 2016
[57]

2007 , publisher=

Housing is the business cycle , author=. 2007 , publisher=

work page 2007
[58]

The Review of Financial Studies , volume=

Quantifying sentiment with news media across local housing markets , author=. The Review of Financial Studies , volume=. 2018 , publisher=

work page 2018
[59]

Journal of Business & Economic Statistics , volume=

Words are the new numbers: A newsy coincident index of the business cycle , author=. Journal of Business & Economic Statistics , volume=. 2020 , publisher=

work page 2020
[60]

2020 , publisher=

Narrative economics: How stories go viral and drive major economic events , author=. 2020 , publisher=

work page 2020
[61]

American Economic Review , volume=

Narrative economics , author=. American Economic Review , volume=. 2017 , publisher=

work page 2017
[62]

2022 , institution=

Causal narratives , author=. 2022 , institution=

work page 2022
[63]

Available at SSRN 4140751 , year=

The Macroeconomics of Narratives , author=. Available at SSRN 4140751 , year=

work page
[64]

2021 , institution=

Business news and business cycles , author=. 2021 , institution=

work page 2021
[65]

American Economic Review , volume=

The market for news , author=. American Economic Review , volume=. 2005 , publisher=

work page 2005
[66]

2008 , publisher=

Introduction to information retrieval , author=. 2008 , publisher=

work page 2008
[67]

2009 , publisher=

Speech and language processing: An introduction to natural language processing, computational linguistics, and speech recognition , author=. 2009 , publisher=

work page 2009
[68]

Annual Review of Sociology , volume=

Machine translation: Mining text for social theory , author=. Annual Review of Sociology , volume=. 2016 , publisher=

work page 2016
[69]

Political Analysis , volume=

Text as data: The promise and pitfalls of automatic content analysis methods for political texts , author=. Political Analysis , volume=. 2013 , publisher=

work page 2013
[70]

Proceedings of the international AAAI conference on web and social media , volume=

Vader: A parsimonious rule-based model for sentiment analysis of social media text , author=. Proceedings of the international AAAI conference on web and social media , volume=

work page
[71]

arXiv preprint arXiv:2005.10200 , year=

BERTweet: A pre-trained language model for English Tweets , author=. arXiv preprint arXiv:2005.10200 , year=

work page arXiv 2005
[72]

Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina , booktitle=

work page
[73]

RoBERTa: A Robustly Optimized BERT Pretraining Approach

Roberta: A robustly optimized bert pretraining approach , author=. arXiv preprint arXiv:1907.11692 , year=

work page internal anchor Pith review Pith/arXiv arXiv 1907
[74]

Econometrica , volume=

Tests of conditional predictive ability , author=. Econometrica , volume=. 2006 , publisher=

work page 2006
[75]

Journal of Economic Theory , volume=

A theory of housing demand shocks , author=. Journal of Economic Theory , volume=. 2022 , publisher=

work page 2022
[76]

Macroeconomic Dynamics , volume=

Should the central bank be concerned about housing prices? , author=. Macroeconomic Dynamics , volume=. 2013 , publisher=

work page 2013
[77]

Journal of Money, Credit and Banking , volume=

Do central banks react to house prices? , author=. Journal of Money, Credit and Banking , volume=. 2013 , publisher=

work page 2013
[78]

FRBSF Economic Letter , volume=

House Prices Respond Promptly to Monetary Policy Surprises , author=. FRBSF Economic Letter , volume=. 2023 , publisher=

work page 2023
[79]

Effect of Conventional and Unconventional Monetary Policy Shocks on Housing Prices in Canada , year=

Effect of Conventional and Unconventional Monetary Policy Shocks on Housing Prices in Canada , author=. Effect of Conventional and Unconventional Monetary Policy Shocks on Housing Prices in Canada , year=

work page
[80]

Journal of the Royal Statistical Society: Series B (statistical methodology) , volume=

Regularization and variable selection via the elastic net , author=. Journal of the Royal Statistical Society: Series B (statistical methodology) , volume=. 2005 , publisher=

work page 2005

Showing first 80 references.