My role. I am not an author of this paper. I am acknowledged in it for research assistance, alongside Gedeon Gbedonou, for work carried out at the Chair in Macroeconomics and Forecasting (ESG UQAM).
What the paper does
News 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. The paper embeds media sentiment into a behavioural New Keynesian model in which the central bank reacts to sentiment, and sentiment itself follows an explicit law of motion.
Monetary-policy sentiment indicators are constructed from more than 50,000 Canadian newspaper articles, using dictionary methods, transformer models and a generative-AI framework.
Findings reported
- Media sentiment shifts household inflation and wage expectations.
- It improves out-of-sample forecasts of GDP growth and inflation.
- It 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; that narrative shock contributes a non-trivial share of medium-horizon macroeconomic variance.
- A counterfactual shutting down the dynamic feedback from media sentiment attenuates the propagation of monetary policy to output and prices.
Authors. Firmin Ayivodji (IMF), Etienne Briand (UQAM), Kevin Moran (Université Laval, CIRANO), Dalibor Stevanovic (UQAM, CIRANO).
JEL codes. E52, E58, E71, D84, C32, C55.