The published article and its Internet Appendix, our COVID-19 medical announcements dataset, the replication package, and the citation.
Journal of Financial Economics, Volume 157, article 103850, July 2024.
The published supplementary material: data construction, the list of newspapers, examples of tweet tone, LDA topics, and additional robustness tables and figures.
The hand-collected dataset behind the announcements study: 15,839 announcements from 20 countries, January 2020 to February 2022, time-stamped to the minute (UTC).
One row per announcement (590 KB). Free to use; please cite the paper.
Python, R, and MATLAB code that reproduces the paper's figures and tables, with a README in each folder.
| Column | Description |
|---|---|
time | Minute-level timestamp of the announcement, in UTC. |
country | Country code: AR, AU, BR, CA, CH, CL, CN, CN-HK, CO, DE, ES, FR, IN, IT, JA, KR, MX, NZ, UK, US. |
category | cases (case counts released), live (live-streamed press conference), president speech (led by the President or Prime Minister), or other. |
peak | 1 if the announcement came at the country's usual briefing time (regularly scheduled), 0 otherwise. |
In the replication package, only the announcements file is actual data. Newspaper tweet text and proprietary financial series (e.g. Bloomberg, Thomson Reuters, EPFR) are replaced by pseudo data because of provider sharing policies.
If you use the paper or the announcements data, please cite:
@article{ArteagaGaravitoCroceFarroniWolfskeil2024,
title = {When the Markets Get {CO.V.I.D}: {CO}ntagion, {V}iruses, and {I}nformation {D}iffusion},
author = {Arteaga-Garavito, Maria Jose and Croce, Mariano M. and Farroni, Paolo and Wolfskeil, Isabella},
journal = {Journal of Financial Economics},
volume = {157},
pages = {103850},
year = {2024},
doi = {10.1016/j.jfineco.2024.103850}
}