« Back to Results

Using AI to Measure Economic Dynamics

Paper Session

Saturday, Jan. 3, 2026 8:00 AM - 10:00 AM (EST)

Philadelphia Marriott Downtown, Grand Ballroom Salon A
Hosted By: American Economic Association
  • Chair: Martin Uribe, Columbia University and NBER

An Anatomy of Firms’ Political Speech

Pablo Ottonello
,
University of Maryland and NBER
Wenting Song
,
University of California-Davis
Sebastian Sotelo
,
University of Michigan and NBER

Abstract

We study the distribution of political speech across U.S. firms, using large language models to measure political engagement in firms’ communications. Our analysis reveals five facts: (1) Political engagement is rare. (2) It is concentrated among large firms. (3) Firms specialize in specific topics and outlets. (4) Large firms engage in a broader set of topics and outlets. (5) The 2020 surge in political engagement was associated with increased engagement by medium-sized firms and a shift in political topics. These findings suggest fixed costs to political engagement and the dominance of large firms’ views in the political space.

Expanding the Landscape of Cross-Border Flow Restrictions: Modern Tools and Historical Perspectives

Katharina Bergant
,
International Monetary Fund and CEPR
Andrés Fernández
,
International Monetary Fund
Ken Teoh
,
International Monetary Fund
Martin Uribe
,
Columbia University and NBER

Abstract

Employing large language models to analyze official documents, we construct a comprehensive record of daily changes in de jure restrictions on cross-border flows worldwide since the 1950s. Our analysis uncovers the wide array of instruments used to regulate cross-border financial flows and documents their evolving prevalence over the past seven decades. The fine granularity of the new measures allows us to characterize cross-country and time-series variation across eight categories of restrictions, further distinguishing by flow, direction, instrument type, and overall policy stance. We exploit the high frequency nature of the new data to document novel patterns in the use of these restrictions, as well as their relationship to crises, and political economy determinants. We validate our measures against established indicators of capital account regulation and show that our LLM-based classifications both replicate and substantially extend these benchmarks along multiple dimensions. Finally, we examine policymakers’ stated motivations for adopting these restrictions and account for the intensive margin of these policy actions.

Worker Rights in Collective Bargaining

Benjamin Arold
,
University of Cambridge
Elliott Ash
,
ETH Zurich and CEPR
W. Bentley MacLeod
,
Yale University and NBER
Suresh Naidu
,
Columbia University and NBER

Abstract

Collective bargaining agreements (CBAs) specify the contractual rights of unionized workers, but their full legal content has not yet been analyzed by economists. This paper develops novel natural language methods to analyze the empirical determinants and economic value of these rights using a new collection of 30,000 CBAs from Canada in the period 1986-2015. We parse legally binding rights (e.g., “workers shall receive…”) and obligations (e.g., “the employer shall provide…”) from contract text, and validate our measures through evaluation of clause pairs and comparison to firm surveys on HR practices. Using time-varying province-level variation in labor income tax rates, we find that higher taxes increase the share of worker-rights clauses while reducing pre-tax wages in unionized firms, consistent with a substitution effect away from taxed wages toward untaxed rights. Further, an exogenous increase in the value of outside options (from a leave-one-out instrument for labor demand) increases the share of worker rights clauses in CBAs. Combining the regression estimates, we infer that a one-standard-deviation increase in worker rights is valued at about 5.7% of wages.

Reading the Fund: A Systematic Analysis of IMF Fiscal Advice using LLMs

Anton Korinek
,
University of Virginia, NBER and CEPR
Jeremie Cohen-Setton
,
International Monetary Fund
Jantsankhorloo Amgalan
,
International Monetary Fund

Abstract

This paper presents a systematic analysis of the International Monetary Fund’s (IMF) fiscal policy advice using large language models (LLMs). Leveraging recent advances in natural language processing, we construct a novel framework to extract, classify, and evaluate the content and evolution of IMF fiscal recommendations across a comprehensive corpus of Article IV reports and program documents. Our methodology employs LLMs that allow for the automated identification of key themes, shifts in policy emphasis, and the alignment of advice with country-specific macroeconomic conditions. The analysis reveals patterns in the Fund’s fiscal guidance and subsequent policy actions in the corresponding member countries. We demonstrate that LLMs, when systematically aligned with domain-specific knowledge and historical experience, can provide valuable insights into the consistency, pragmatism, and contextualization of IMF fiscal advice.

Discussant(s)
Aakash Kalyani
,
Federal Reserve Bank of St. Louis
Anton Korinek
,
University of Virginia, NBER and CEPR
Thomas Drechsel
,
University of Maryland
Paul E. Soto
,
Federal Reserve Board
JEL Classifications
  • A1 - General Economics
  • E6 - Macroeconomic Policy, Macroeconomic Aspects of Public Finance, and General Outlook