CE26 - Innovation, Travail

Technology and Polarization of Employment – TOPAZE

Empirical studies are based on original datasets. For example, the methodologies for several tasks rely on confidential data provided by INSEE and DGFiP, which requires obtaining access and working on the CASD (Centre d'accès sécurisé aux données). They are also extremely computationally intensive both in term of space and machine speed. In the end, several tasks will be able to explore multi-dimensional aspects of wage inequality, which will allow to produce original results in the literature.
Macroeconomic models will also be used to improve the understanding of job polarization and its impact across gender and across generations. The macroeconomic models allow to quantify the importance of the forces at work, using counterfactual exercises.

in phase 1 of the project, the vast ajority of tasks are under data preparation. This is consistent with the agenda mentioned in the final ANR proposal.

In phase 2 of the project, the other research outputs will deliver results as announced in the project schedule.

2 articles submitted / under revision

Submission summary

Over the last 30 years, employment growth has been fast, not only in high-paid jobs (abstract, cognitive tasks requiring creativity), but also in low-paid jobs (manual, non-routine job requiring human interaction). In contrast, employment has decreased significantly among middling jobs (routine, repetitive), and those involving tasks that can be replaced by machines. As robots can now beat humans in performing complex tasks, employment becomes polarized with the disappearance of middle-income workers (replaced by machines) and an expansion in top and bottom jobs (performing tasks that cannot be computerized). Will this job polarization lead to a polarized society? This project aims at shedding light on this worldwide phenomenon using theoretical and empirical works.

Part 1 documents new stylized facts on job polarization. First, little is known about the geography of job polarization and its impact on job location, spatial wage dispersion and residential choices. We will tackle this issue on the Grand Paris, using micro data in Ile de France. The project will investigate the role of job polarization in the rising share of non-standard work using French survey data. Job polarization, with the disappearance of middling jobs and the expansion of low-skill manual occupations, might widen the gap between stable and unstable employment. The loss of stable jobs raises the spectre of an increasingly unequal society.

Part 2 documents wage inequality. Since the 1990s, most developed countries have seen median wage growth lagging behind productivity growth, a phenomenon dubbed “wage decoupling”. During this time, labor markets experienced polarization of employment and, in some countries, also of wages, as well as a decline in the labor income share. The project will investigate the nexus of polarization and wage decoupling. We will also study the potential role of Information and Communication Technologies in the rise of between-firm wage inequality using a unique combination of French firm micro datasets. In the policy discussion on job polarization, concerns have been raised on intergenerational effects of technological change as the disappearance of routine jobs forces routine workers to change occupations. Older workers may be particularly vulnerable to polarization as the occupational change involve loss of human capital, which is more costly at the end of the working life. The research agenda will then include the study of inter and intra generational wage inequality in a context of fast technological changes.

Part 3 explores gender issues. Gender skill biased technological change is at work in US data. Interestingly, in the US, we also observe a decline of marriage rates, a rise of single-headed households and share of out-of-wedlock childbearing. The project aims at exploring how gender employment polarization relates to marriage rates and a deterioration of family structure. In addition, little is known about job polarization in emerging countries. This empirical project fills this gap by focusing on Brazilian firm data. The originality of this project also lies in the focus on gender wage gap across occupations, in particular in exporting firms as they are more productive.

Part 4 studies institutional aspect of job polarization by first measuring the electoral consequences of job polarization using French data. Are areas with polarized employment more prone to populism? We will quantify the contribution of job polarization to polarized electoral outcomes, including during the 2017 presidential election. Finally, job polarization involves labor reallocation from laid-off routine workers to abstract or manual jobs. The extent and speed of this reallocation depends on labor market institutions (employment protection laws, minimum wage, … ). The study quantifies the interaction between job polarization and labor market institutions in order to predict employment and participation outcomes in European regions and in France.

Project coordination

Thepthida Sopraseuth (Théorie économique, modélisation et applications)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.


THEMA Théorie économique, modélisation et applications
FNSP - Département d'économie Département d'Economie de Sciences Po

Help of the ANR 338,126 euros
Beginning and duration of the scientific project: September 2018 - 48 Months

Useful links

Explorez notre base de projets financés



ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter