research | Verónica C. Pérez

2026

Working Paper
Regulatory Decentralization and Construction Productivity.

V.C. Pérez, and E. Stamm

2026

Abs Bib Slides

Residential construction is dominated by small, local firms and has seen little productivity growth in the last several decades, while the rules governing how buildings are built are written and enforced by thousands of local governments. We ask whether the decentralization of building standards, helps explain small firm size and low productivity in residential construction. Across U.S. metropolitan areas, jurisdictional fragmentation is associated with fewer housing units per construction worker and smaller establishments, conditional on land-use regulation. To study the mechanism, we build a new dataset of every local amendment to California’s statewide building standards code, extracted from the full text of all ordinances filed with the California Building Standards Commission. Neighboring cities rarely amend the same sections: among cities that amend, the median city shares about nine percent of its amended sections with its neighbors. Cities whose amendments overlap more with their neighbors’ build more housing units per construction worker, and large builders concentrate in jurisdictions with similar codes.
@techreport{PS_RegDecentralization, title = {Regulatory Decentralization and Construction Productivity.}, author = {Pérez, V.C. and Stamm, E.}, year = {2026}, }
Working Paper
Localization and Steering of Economic Uncertainty in Large Language Models.

R. Wang, V.C. Pérez, C. Schlesinger, and L. Sun

2026

Abs Bib PDF

Economic uncertainty shapes investment, hiring, and asset prices, making its measurement from text a central task in economics and finance. While LLMs accurately measure uncertainty in financial texts such as earnings call transcripts, it remains unclear whether they develop coherent internal representations or merely pattern-match on surface lexical cues. In this study, we discover that LLMs linearly represent economic uncertainty with a single direction in the residual stream. Using activation patching on two synthetic datasets of contrastive earnings-call statements with varying linguistic styles, we localize this direction and find that models aggregate the uncertainty signal at the final token regardless of lexical patterns. The extracted direction perfectly separates held-out high- and low-uncertainty statements both within- and cross-dataset. Causal interventions show that adding or subtracting this direction monotonically flips uncertainty predictions, including cross-dataset transfer from templated to naturalistic text. Finally, in a downstream portfolio allocation task using real earnings-call excerpts, steering along the uncertainty direction shifts model investment toward safe assets, consistent with economic theory. Together, our results establish that LLMs encode economic uncertainty as a structured, causally active, and transferable representation, offering a foundation for interpretability-based auditing and control of LLMs deployed in financial analysis.
@techreport{WPSS_EconUncertaintyLLM, title = {Localization and Steering of Economic Uncertainty in Large Language Models.}, author = {Wang, R. and Pérez, V.C. and Schlesinger, C. and Sun, L.}, year = {2026}, }

2025

Measuring Industrial Policy: A Text-Based Approach.

R. Juhász, N. Lane, E. Oehlsen, and V.C. Pérez

May 2025

Since the 18th century, policymakers have debated the merits of industrial policy (IP). Yet, economists lack basic facts about its use due to measurement challenges. We propose a new approach to IP measurement based on information contained in policy text. We show how off-the-shelf supervised machine learning tools can be used to categorize industrial policies at scale. Using this approach, we validate longstanding concerns with earlier approaches to measurement which conflate IP with other types of policy. We apply our methodology to a global database of commercial policy descriptions, and provide a first look at IP use at the country, industry, and year levels (2010-2022). The new data on IP suggest that i) IP is on the rise; ii) modern IP tends to use subsidies and export promotion measures as opposed to tariffs; iii) rich countries heavily dominate IP use; iv) IP tends to target sectors with an established comparative advantage, particularly in high-income countries.

@techreport{JLOP,
  title = {Measuring Industrial Policy: A Text-Based Approach.},
  author = {Juhász, R. and Lane, N. and Oehlsen, E. and Pérez, V.C.},
  year = {2025},
  month = may,
  url = {https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5262841},
  online_version = {https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5262841},
  coverage_show = {true},
  coverage_links = {<p><a href="https://www.wsj.com/articles/this-part-of-bidenomics-needs-more-economics-2cea1641"><strong>Wall Street Journal:</strong> This Part of Bidenomics Needs More Economics</a></p><p><a href="https://www.bloomberg.com/news/features/2023-07-25/global-subsidy-wars-force-us-allies-to-pay-up-for-chips-evs?in_source=embedded-checkout-banner"><strong> Bloomberg:</strong> Subsidy Wars Heat Up With US Allies Forced to Pay Up or Lose Out</a></p><p><a href="https://iap.unido.org/articles/trends-global-industrial-policy"><strong>UNIDO Insights:</strong> Trends in Global Industrial Policy</a></p><p><a href="https://www.unido.org/sites/default/files/files/2023-03/IID%20Policy%20Brief%201%20-%20Global%20Industrial%20Policy%20-%20Measurment%20and%20Results%20-%20FINAL.pdf"><strong>UNIDO Policy Brief Series:</strong> Insights on industrial development</a></p><p><a href="https://d1e00ek4ebabms.cloudfront.net/production/uploaded-files/JPM_The_Long_term_Strate_2023-01-27_4318021-91952e6d-55c7-4f46-97ad-3cbc444609c7.pdf"><strong>JP Morgan:</strong> Global Long-term Strategy 2023</a></p><p><a href="https://steg.cepr.org/podcasts/who-what-when-and-how-industrial-policy-nathan-lane"><strong>STEG Podcast:</strong> Conversations on Transformation Series</a></p><p><a href="https://res.org.uk/wp-content/uploads/2023/03/RES-Newsletter-201-Digital-edition.pdf"><strong>RES:</strong> When industry means hard work by Daron Acemoglu</a></p>}
}

2022

M.A. Thesis
The effect of industrialization on political dynasties: Evidence from Colombian local governments.

V.C. Pérez

Jul 2022

Abs Bib

In this paper I find that industrialization reduces the concentration of political power, measured by the share of elected officials that were part of political dynasties, in the context of 19 Colombian departments. I use newly digitized data from departmental production by industrial sector since 1969 and the international price of manufactures as an exogenous source of variation in the industrialization levels. Using an instrumental variable approach, I find that an increase in one paid manufacturing employee reduces the share of elected individuals from dynasties by 1.15 p.p. Dividing the results by type of office indicates that the impact is more prominent in the election of dynastic individuals in the legislative brand compared to those elected in the executive one. One additional paid employee decreases the average share of dynastic mayors by 0.75 p.p., a 7.8% decrease in the average. On the other hand, an increase in a paid employee decreases the percentage of representatives from dynastic families by 3.24 p.p., which represents a diminution of 13.4%. I propose multiple mechanisms that could be driving these results including the relevance of industrialization to the creation of labor unions, and in the redistribution of resources such as education and income.
@techreport{PerezThesis, title = {The effect of industrialization on political dynasties: Evidence from Colombian local governments.}, author = {Pérez, V.C.}, year = {2022}, month = jul, url = {https://repositorio.uniandes.edu.co/entities/publication/17c39bcd-48f3-4434-85f9-037edc665e48}, online_version = {https://repositorio.uniandes.edu.co/entities/publication/17c39bcd-48f3-4434-85f9-037edc665e48}, }