Forecasting U.S. equity market volatility with attention and sentiment to the economy

Paper: arXiv 2503.19767

Authors: Martina Halousková, Štefan Lyócsa

Abstract

Macroeconomic variables are known to significantly impact equity markets, but their predictive power for price fluctuations has been underexplored due to challenges such as infrequency and variability in timing of announcements, changing market expectations, and the gradual pricing in of news. To address these concerns, we estimate the public’s attention and sentiment towards ten scheduled macroeconomic variables using social media, news articles, information consumption data, and a search engine. We use standard and machine-learning methods and show that we are able to improve volatility forecasts for almost all 404 major U.S. stocks in our sample. Models that use sentiment to macroeconomic announcements consistently improve volatility forecasts across all economic sectors, with the greatest improvement of 14.99% on average against the benchmark method - on days of extreme price variation. The magnitude of improvements varies with the data source used to estimate attention and sentiment, and is found within machine-learning models.

Complexity vs Empirical Score

• Math Complexity: 3.5/10
• Empirical Rigor: 8.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper provides a strong empirical framework combining natural language sentiment/attention data with machine learning techniques to forecast cross-sectional U.S. stock volatility around macroeconomic events. While the underlying mathematical modeling is straightforward econometric and ML benchmarking rather than novel theoretical derivations, the empirical execution across 404 stocks and multiple alternative data sources is exceptionally rigorous.

Research Flowchart

  flowchart TD
    A["Research Goal: Improve U.S. Stock Volatility Forecasts"] --> B["Data Inputs: Social Media, News, Search Engine & Information Consumption"]
    B --> C["Estimate Public Attention & Sentiment for 10 Macroeconomic Variables"]
    C --> D["Apply Standard & Machine Learning Models (404 U.S. Stocks)"]
    D --> E["Outcome: Outperforms Benchmarks Across All Economic Sectors"]
    E --> F["Key Result: Up to 14.99% Forecast Improvement on Extreme Price Days"]

Cited References (124)

1. Stock Prices and Economic News (1985)
2. Stock returns, expected returns, and real activity (1990)
3. Stock returns, real activity, inflation, and money (1981)
4. Dynamic factor models, factor-augmented vector autoregressions, and structural vector autoregressions in macroeconomics (2016)
5. Forecasting with economic news (2023) — link
6. Time-varying causality between stock prices and macroeconomic fundamentals: connection or disconnection? (2022)
7. Words speak as loudly as actions: Central bank communication and the response of equity prices to macroeconomic announcements (2022)
8. Stock returns and real activity: A century of evidence (1990)

0 of 124 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.