Source: Baker, S. R., Bloom, N. & Davis, S. J. (2016) · Quarterly Journal of Economics 131(4), 1593–1636 · DOI: 10.1093/qje/qjw024
Problem it solves
Economic "policy uncertainty" is widely invoked (by the FOMC, IMF, the press) but hard to measure: it is not directly observed, and survey/option-based proxies (VIX, forecaster surveys) are not specific to policy. The paper builds a transparent, reproducible, real-time Economic Policy Uncertainty (EPU) index from newspaper text, validates it against human readings and economic outcomes, and shows policy uncertainty foreshadows weaker investment, output, and employment. A landmark "text as data" macro-finance measure.
The method
The headline US monthly EPU index (from January 1985) is a newspaper-coverage measure built from 10 leading US papers (USA Today, Miami Herald, Chicago Tribune, Washington Post, LA Times, Boston Globe, SF Chronicle, Dallas Morning News, NY Times, WSJ):
Count, per paper-month, articles satisfying a triple keyword condition — the article must contain a term from each of three sets: (1) uncertainty: "uncertain(ty)"; (2) economy: "economic"/"economy"; (3) policy: the audit-selected set "congress", "deficit", "Federal Reserve", "legislation", "regulation", "white house" (with variants).
Scale each raw count by the paper's total article volume that month (to remove volume drift), standardize each paper's series to unit SD over 1985–2009, average across the 10 papers, then normalize to mean 100 over 1985–2009.
A daily index and category-specific indices (monetary, fiscal, healthcare, national security, etc.) are built similarly; historical US/UK indices back to 1900 use six long-running papers and an expanded term set.
Note: an earlier-draft version combined the news measure with (a) the dollar value of scheduled federal tax-code expirations and (b) dispersion of professional forecasts of government purchases and CPI. The published headline index drops these and uses the newspaper component alone; the other components are reported separately at policyuncertainty.com.
Assumptions & inputs
Inputs: digital newspaper archives and a validated term set; firm/macro panels for the applications.
Validation (the paper's key credibility step): an audit study of ~12,000 randomly selected articles (coded by trained RAs against a 65-page manual) was used to select the policy term set and benchmark the automated count. Human- vs. computer-generated indices correlate 0.86 (quarterly, 1985–2012) and 0.93 (annual, 1900–2010), with discrepancies uncorrelated with GDP growth or the EPU level. The daily index correlates 0.85 with the monthly index.
How to use it
As a macro-finance regressor / risk overlay: relate asset returns, volatility, investment, and hiring to EPU innovations; the authors use firm-level panels and a panel VAR for 12 economies.
Country and category variants are publicly available, enabling cross-country and event-window studies.
Headline findings (the validation evidence): EPU spikes near tight presidential elections, Gulf Wars I & II, 9/11, the Lehman failure, and the 2011 debt-ceiling dispute; higher EPU is associated with greater stock-price volatility and reduced investment and employment in policy-sensitive sectors (defense, healthcare, finance, infrastructure); at the macro level, EPU innovations foreshadow declines in investment, output, and employment. The long-span index rose sharply from late 1931 and has drifted up since the 1960s.
Limitations & pitfalls
The measure depends on the chosen term sets and newspaper set; media coverage is a proxy, not true uncertainty, and can reflect editorial slant (the authors test and find slant doesn't seriously distort the index).
EPU is correlated with volatility and recessions, so isolating a causal policy-uncertainty channel is hard; the VAR/firm-heterogeneity designs mitigate but don't fully resolve this.
Term sets must be re-derived per language/country, limiting strict cross-country comparability.
Key references
Baker, S., Bloom, N. & Davis, S. (2016) — Measuring Economic Policy Uncertainty — Quarterly Journal of Economics
Bloom, N. (2009) — The Impact of Uncertainty Shocks — Econometrica
Pástor, Ľ. & Veronesi, P. (2012) — Uncertainty about Government Policy and Stock Prices — Journal of Finance
Provenance: verified/generated from the paper's full text.
