Research | Pablo Fiorentini

Research

I study how firms, institutions, and information interact in markets and policy, with a focus on industrial policy, certification markets, and regulation. No journal publications yet — below is the honest state of the pipeline, from the project I am building my job market around, down to the graveyard.

Current direction

Where the agenda is heading

Industrial Policy · Certification Emerging JMP direction

Green Conditionality in Industrial Policy: Certified Energy Savings and Threshold Incentives in Italy’s Transizione 5.0

Green subsidies increasingly pay for certified numbers. When public money is allocated on an energy-saving threshold certified by a paid third party, is the firm’s response real investment, optimistic reporting, or certifier-supplied leniency?

This project studies whether the notched subsidies of Italy’s Transizione 5.0 program induce real energy-saving investment or strategic responses in certified savings. When eligibility depends on a number certified by a paid third party, the measured elasticity of savings to incentives is an equilibrium outcome of the verification market, not only a technological parameter of firms. The aim is to decompose firms’ responses to the thresholds into real re-optimization, firm-side optimistic reporting, and certifier-supplied leniency.

Status: design & data-access stage · Last updated: July 2026

Active work

In progress

Monetary Economics Active

A Monetary Model of Money Laundering

with Fabrizio Mattesini

We develop a theoretical model of money laundering, with the goal of disciplining it with moments drawn from the Financial Intelligence Unit datasets of the Bank of Italy. The aim is to say something precise about the incentives, frictions, and aggregate implications of illicit financial flows.

Status: model development · Last updated: July 2026

Earlier & paused work

On hold, not forgotten

Projects I still believe in, parked while the agenda above takes shape.

Political Economy Paused

Do Lobbyists Inform or Mislead? Evidence from EU Public Consultations

Can policymakers extract truth when every messenger has an agenda?

A large-scale test of multi-sender cheap-talk theory: a novel dataset links the EU Transparency Register, the Have Your Say portal, and EUR-Lex at sentence level, following claims from lobbying submissions into legislative text. Convergence across opposed interests should signal informative content.

Status: paused — data pipeline built, analysis on hold

Macroeconomics & Growth Paused

Automation, Population Growth, and the Future of Human Capital

What happens to the labour share when machines can learn the skills?

Master thesis written at CEMFI under the supervision of Pau Roldan-Blanco. A capital-accumulation model with human capital, studying how the labour share, output, and inputs respond to population growth, saving rates, and education investment when automation is possible.

Status: paused — under revision

Urban Economics Paused

Urban Mobility, with a Focus on the City of Rome

What does a city’s mobility look like when you actually measure it?

Combines large-scale web scraping and news monitoring to build an empirical basis for studying transport dynamics, urban frictions, and mobility policy in a large metropolitan area.

Status: paused — data-collection stage

Digital Governance Paused

Digital Democracy and Gender Equality

Should gender bias in AI systems be treated as a structural externality?

ENGAGE.EU Research Label collaboration. With my co-authors I drafted a critical framework for understanding gender bias in AI across data and model levels, arguing these biases are structural externalities for governance and regulation. Event page →

Status: paused — op-ed drafted

⚰️ The Graveyard

Ideas that did not make it

Every research agenda has a shadow: projects that seemed brilliant at 2 a.m. In loving memory. May they rest in referee report 2.

R.I.P.

The Instrument That Wasn’t

2024 – 2024

Cause of death: the first stage. F-statistic at time of passing: 1.7.

R.I.P.

The Great Scrape

2023 – 2024

Cause of death: a website redesign that broke 4,000 lines of scraping code overnight.

R.I.P.

The Dataset Merge of Legend

2025 – 2025

Cause of death: fuzzy matching. Survived by 12,000 unmatched observations.

A representative sample — the full census is still being compiled.