In today’s digital era, many types of markets—including labor markets—are moving online, transforming the way supply and demand are matched. This trend offers new opportunities to leverage internet data for social science research, particularly in labor economics, which is central to the mission of IDSC, IZA’s research data center.
When markets operate online, digital technology becomes a powerful tool for optimizing the matching of supply and demand—one of the core challenges of any market. The digital nature of these transactions allows researchers to efficiently generate, collect, and analyze data, essentially enabling them to “rewind and replay” market activities for better understanding. As technologies like AI-assisted matching continue to evolve, they offer promising new solutions to traditional labor market challenges while also presenting fresh complexities and research questions.
The 7th IDSC Workshop, co-organized by Nikos Askitas, Peter J. Kuhn, and Christina Gathmann (Labor Market Department Director at LISER), focused on these emerging dynamics of the virtual labor market. Hosted at LISER’s Belval campus in Luxembourg, this was the first face-to-face meeting since the COVID-19 crisis. The venue—a mix of modern facilities and historical industrial landmarks on the grounds of an old blast furnace—provided an inspiring environment for the workshop.
This year’s event was supported by the Luxembourg National Research Fund and LISER, with organizational help from IZA’s events team. Keynote speakers included Amanda Agan (Cornell University), Thomas Le Barbanchon (Bocconi University), and Barbara Petrongolo (University of Oxford). Alongside these keynote presentations, the workshop featured a roundtable with local stakeholders, a poster session, and a diverse set of papers representing research from 11 countries.
The workshop addressed a variety of relevant themes, including:
- Algorithms and AI in the Matching Process
- Networks, Social Media, and Mindfulness
- Helping Workers Find Jobs
- Crafting Job Ads: Ad Content and Applicant Behavior
Selected Presentations
Automating Automaticity (Amanda Agan, Cornell University)
Amanda Agan explored how algorithms that customize user content can unintentionally reinforce biases. Through a combination of field and lab experiments, she demonstrated that algorithms trained on automatic user behaviors—like quickly scrolling past a social media post—tend to encode in-group biases. In contrast, slower, more deliberate user actions, such as adding a new contact, exhibit fewer biases. These findings suggest that job boards and other platforms could minimize the spread of unconscious biases by training algorithms with more conscious, higher-stakes user decisions.
Traditional vs Machine Learning Methods (Sabrina Mühlbauer, IAB)
Sabrina Mühlbauer and her co-authors used comprehensive administrative data from Germany to predict job matching quality in terms of job stability and wages, using both traditional econometric techniques and machine learning (ML) methods. The study found that ML outperformed traditional methods in pattern recognition, data handling, and reducing prediction errors. By combining ML insights with algorithms that provide ranked job recommendations tailored to individual characteristics, the research shows significant promise in supporting job seekers and caseworkers to refine job search strategies.
Wage Information and Applicant Selection (Marc Witte, VU Amsterdam and IZA)
In a field experiment in Addis Ababa, Ethiopia, Marc Witte and co-authors examined how wage transparency in job advertisements affects workers’ application decisions. The experiment highlighted the role of recruiters’ choices in job ad content in shaping recruitment outcomes. Unlike the vast body of field experiments focused on employers’ responses to resume content, this study offers a rare insight into how job seekers respond to vacancy details—highlighting an essential part of the recruitment process that influences where individuals choose to apply.
With these and other presentations, the 7th IDSC Workshop offered critical insights into how technological advances are reshaping labor markets. From AI-driven matching processes to understanding the effects of wage transparency, the discussions underscored the dual nature of these advances: while they provide new tools for improving labor market efficiency, they also introduce new challenges that require careful exploration. This event highlighted the value of continued research in guiding policy and improving practices in the evolving landscape of labor economics.