The power of narrative sentiment in economic forecasts

B-Tier
Journal: International Journal of Forecasting
Year: 2023
Volume: 39
Issue: 3
Pages: 1097-1121

Authors (3)

Sharpe, Steven A. (Federal Reserve Board (Board o...) Sinha, Nitish R. (not in RePEc) Hollrah, Christopher A. (not in RePEc)

Score contribution per author:

0.670 = (α=2.01 / 3 authors) × 1.0x B-tier

α: calibrated so average coauthorship-adjusted count equals average raw count

Abstract

The sentiment, or “Tonality”, extracted from the narratives that accompany Federal Reserve economic forecasts (in the Greenbook) is strongly correlated with future economic performance, positively with GDP, and negatively with unemployment and inflation. More notably, Tonality conveys substantial incremental information in that it predicts errors in Federal Reserve and even in private-sector point forecasts of unemployment and GDP up to four quarters ahead. More favorable sentiment predicts economic performance that exceeds point forecasts. Higher Tonality also predicts positive monetary policy (fed funds rate) surprises and higher stock returns up to four quarters ahead. Quantile regressions suggest that much of Tonality’s forecasting power arises from its signal of downside risks to both economic performance and stock returns. If observed in real time, tonality would have been most informative about economic prospects and stock returns when economic uncertainty was high or when point forecasts called for subpar GDP growth.

Technical Details

RePEc Handle
repec:eee:intfor:v:39:y:2023:i:3:p:1097-1121
Journal Field
Econometrics
Author Count
3
Added to Database
2026-01-29