DYMOLAMO applies dynamic discrete choice models to labor markets, UnStruct extends structural models to unstructured data, and MARKET POWER examines macroeconomic trends.
FUNDACIO PRIVADA BARCELONA GRADUATE SCHOOL OF ECONOMICS
Barcelona-based economics research institution combining structural econometric modeling with machine learning for labor, energy, and media market analysis.
Their core work
The Barcelona School of Economics (BSE) is a graduate research institution specializing in advanced economic modeling, econometrics, and applied microeconomics. Their researchers develop structural models to understand labor markets, migration dynamics, media economics, and macroeconomic trends. More recently, they have expanded into applying machine learning and Bayesian methods to economic problems — bridging data science with traditional economic theory. Their work directly informs policy decisions on labor mobility, energy transitions, and media regulation.
What they specialise in
DYMOLAMO directly studies labor market mobility and human capital accumulation; ADEMU addressed economic and monetary union dynamics.
MIRAGE (their only coordinated project) investigates media independence, advertiser influence, and the impact of internet search engines on news quality.
UnStruct applies probabilistic machine learning and Bayesian inference to economic data; ENECML builds a machine learning toolbox for energy transition analysis.
ENECML (their largest-funded project at EUR 1.46M) uses ML methods to understand energy market dynamics.
How they've shifted over time
In their early H2020 period (2015–2018), BSE focused squarely on traditional structural economics — labor market mobility, migration, discrete choice models, and media market dynamics. From 2020 onward, their work shifted decisively toward computational and data-driven methods: probabilistic machine learning, Bayesian inference, and applying these tools to unstructured data and energy markets. This represents a clear pivot from classical econometric theory toward the intersection of economics and modern data science.
BSE is moving from pure economic theory toward ML-augmented economic modeling, making them increasingly relevant for projects that need rigorous economic analysis combined with modern data science capabilities.
How they like to work
BSE operates overwhelmingly as a participant (5 of 6 projects), contributing specialist economic research within larger consortia rather than leading them. Their single coordination role (MIRAGE) was a smaller ERC grant. With only 10 unique partners across 7 countries, they maintain a focused network — this is typical of an academic economics department that joins projects where deep methodological expertise is needed, rather than building large collaborative ecosystems.
BSE has collaborated with 10 unique partners across 7 countries, indicating a moderately diverse but not expansive European network. Their partnerships span multiple EU member states, consistent with ERC-funded research groups that attract talent internationally.
What sets them apart
BSE sits at a rare intersection: they combine world-class economic theory (structural models, equilibrium analysis) with growing machine learning capabilities — a combination few economics institutions in Europe can match. Their faculty-driven ERC portfolio (STG, COG, and ADG grants) signals individual research excellence across career stages. For anyone building a consortium that needs rigorous economic modeling with modern computational methods, BSE offers a credibility and skill set that is hard to replicate.
Highlights from their portfolio
- ENECMLLargest single grant (EUR 1.46M) and signals BSE's strategic move into energy economics powered by machine learning — a high-demand combination.
- MIRAGEBSE's only coordinated project, studying media independence in the internet age — a politically and commercially relevant topic connecting economics with digital platform regulation.
- UnStructBridges traditional structural economics with probabilistic ML and Bayesian inference applied to unstructured text data — methodologically distinctive.