Study on Time-Varying Macroeconomic Announcement Risk in the Journal of Econometrics
A study by research fellow Norman Seeger (Vrije Universiteit Amsterdam) and co-authors Michael Johannes (Columbia University, United States) and Jonathan R. Stroud (Georgetown University, United States), published in the Journal of Econometrics, shows that the volatility spike around major economic announcements — such as FOMC decisions, employment reports, and inventory data — is not constant but varies by as much as a factor of 10 over time.
The paper demonstrates that announcement-driven market volatility is not fixed but highly time-varying and predictable, a finding that challenges the common practice in macroeconomic event studies of treating raw announcement returns as clean policy shocks and calls for a fundamental rethinking of how researchers and practitioners measure the market impact of economic news.
Using 13 years of 5-minute crude oil futures data and a flexible multifactor stochastic volatility model, the authors disentangle announcement risk from other volatility sources and find that the variation is largely driven by the prevailing level of slower-moving, interday market volatility. This time variation is highly predictable: ex-ante model-based volatility forecasts are strongly correlated with subsequently realized announcement volatility, meaning the same type of news release can produce vastly different market reactions depending on the broader volatility environment. The findings imply that raw announcement returns — widely used as policy "shocks" in event studies and risk management — can be misleading, and should be standardized by predictive volatility to properly assess their true impact.
Article citation
Michael Johannes, Norman J. Seeger, Jonathan R. Stroud, Time-varying macroeconomic announcement risk, Journal of Econometrics, Volume 254, Part B, March 2026. doi.org/10.1016/j.jeconom.2026.106194.