The Luxembourg Institute of Socio-Economic Research (LISER)

is recruiting a

Research Associate (Post-doc, f/m) in Economics

• Ref: 24-32
• 12 months fixed-term working contract at LISER
• Full-time, 40 hours/week 
• Department: Living Conditions
• Work location: Belval (Luxembourg)
• Deadline for applications : February 14th, 2025
• Foreseen start date: September 1st, 2025

Project description 

The rise of Artificial Intelligence (AI) systems has revolutionized the capabilities of digital machines, increasing their precision and efficiency in performing tasks. However, numerous case studies highlight the potential risks to individuals and society, including widespread surveillance, manipulation of citizens' behavior and exacerbation of discrimination against marginalized communities. Recognizing these risks, policymakers, particularly in the EU, have begun to address these concerns through analysis and regulation, laying the groundwork for understanding and mitigating the negative impacts of AI.

Luxembourg's labor market is at risk of discrimination, particularly along ethnic, linguistic and gender lines. Our project examines statistical discrimination, which not only exacerbates this problem, but also converts linguistic or educational differences into significant disadvantages in the labor market, thereby perpetuating inequalities. We aim to shed light on the role of AI in addressing this subtle threat to social cohesion. More specifically, this project aims to develop a comprehensive framework for modelling the impact of AI technologies on discrimination, with a particular focus on discrimination in hiring practices. Divided into three main work packages, the project aims to address theoretical, observational, and experimental aspects of this issue.

The first work package involves theoretical modeling to distinguish between outcome inequalities driven by individual effort and those caused by discriminatory circumstances. This involves revisiting classical models such as Phelps (1972) and integrating modern theoretical insights from Equality of Opportunity theory (see Roemer and Trannoy 2015 and Bourguignon et al. 2007) to refine appropriate algorithms that assess occupational data, ensuring they can distinguish between bias and merit-based discrepancies.

In the second work package, the focus shifts to an observational study using data from the Social Security in Luxembourg (IGSS) and various job vacancy databases. Over a 20-year period, we first assess how technological changes have affected the content of job advertisement and the persistence of statistical discrimination. Using advanced Natural Language Processing (NLP) techniques, the project will develop metrics to detect and quantify biases in job listings of a large data set (the Lightcast database), including those related to gender and nationality. This process will be powered by MeluXina, one of Europe's leading high-performance computing facilities, ensuring robust data analysis capabilities. In a second step, we will link the aforementioned job advertisement data and those provided by the Public Employment Agency (ADEM) with IGSS records to study the recruitment outcomes for minority groups. This comprehensive data analysis aims to show how biases in job advertisements correlate with actual hiring disparities, thereby informing strategies to enhance diversity in the workplace.

Finally, the third work package involves conducting empirical AI experiments to directly measure how different AI-driven hiring interventions affect discrimination. Participants to an online survey will be acting as recruiters, choosing candidates from lists generated by different fairness-sensitive AI algorithms, specifically designed and trained to either ignore discriminatory factors ("reductionist") or compensate for them ("correcting"). The collaboration with ADEM will help refine these algorithms to ensure that they effectively promote equity. By integrating theoretical modeling, data-driven observations, and empirical testing, this project aims to provide both challenging and innovative research perspectives and valuable insights into the discriminatory impact of AI technologies, useful for developing actionable strategies to promote more equitable hiring practices in Luxembourg.

Your role

In this context, the successful candidate will have the following missions:

• participate in the TechnoD project in collaboration with local and international partners;
• produce excellent research outputs, which can be published in high-ranked, peer-reviewed journals;
• take part in the day-to-day department-related activities (workshops, lecture series, etc.), and participate in public outreach activities like policy briefs or social media communication related to the TechnoD project.

Your profile

• Ph.D. in economics, applied econometrics, computational social sciences or other relevant discipline;
• Excellent quantitative skills and skills in programming (Stata, R, Python), machine learning, data science;
• Hands-on experience working with large-scale survey or administrative data;
• Excellent writing skills, as evidenced by selected publications (or working papers);
• Excellent communication skills and ability to work within a team;
• Fluency in English (both spoken and written); fluency in French is considered as an asset.

LISER particularly encourages female researchers to apply

What we offer

• A dynamic and international research environment with attractive working conditions and a stimulating work environment;
• Flexibility in the organisation of the working hours and the possibility of teleworking;
• Competitive remuneration according to the Collective Labour Agreement in force (half of a 13th month salary, meal vouchers, etc.);
• Investment in career support and development (trainings, seminars, participation to international meetings and conferences);
• 32.5 days of annual leave for a full-time contract.

How to apply

Please submit your complete application in English via https://jobs.liser.lu/jobs by including the following documents before February 14th, 2025 :

• Curriculum vitae;
• Motivation letter;
• Recent piece of research;
• Copy of your PhD diploma;
• Two reference letters (to be uploaded with the application documents or sent separately to recruitment@liser.lu).

Deadline to submit applications: February 14th, 2025

Would you have any work-related question, please contact Prof. Eugenio PELUSO at eugenio.peluso@liser.lu. 

For administrative matters, please contact Mrs. Vanya KIROVA at recruitment@liser.lu.

Why LISER? 

LISER is a publicly funded research institute located in Luxembourg and dedicated to applied empirical research in economic, social and spatial sciences. The Institute attracts top researchers from all over the world and high-level student training is a vibrant part of the Institute’s activities. LISER staff consists of approx. 200 employees, about 60% of the staff being researchers; mainly from the fields of economics, geography and sociology. The vision of the institute is to be a socio-economic research institute internationally recognised, focused on scientific excellence and societal impact, able to contribute through multi-disciplinary and intersectoral research, in an active and inclusive way to a sustainable and inclusive society at national and international level. 

The institute is located on the new Belval campus in the south of Luxembourg (Cité des Sciences, Luxembourg), which hosts the University of Luxembourg and a substantial part of the country’s publicly funded research facilities, i.e. LISER, the Luxembourg Income Study (LIS) cross-national data centre, the Luxembourg institutes of Health (LIH) and of Science and Technology (LIST). Information on research in Luxembourg is accessible via the national EURAXESS platform. 

LISER is an Equal Opportunity Employer