Working paper

In the Land of AKM: Explaining the Dynamics of Wage Inequality in France

Published on the 25th of March 2025
Authors : Damien Babet, Olivier Godechot, Marco G. Palladino

Working Paper Series no. 987. We use a newly constructed and quasi-exhaustive matched employer-employee database to study the contribution of firms to wage inequality in France. We implement a simple and tractable correction for the limited mobility bias. Our analysis, covering the period 2002-2019, reveals an increase in between-firm inequality, mainly due to the growing clustering of workers with similar market value. These phenomena are associated with increasing occupational specialization at the firm level. Our results highlight the importance of bias-corrected AKM estimates of the Abowd, Kramarz et Margolis (1999) model –hereafter AKM- in capturing the dynamics of wage inequality, and show how both observable job types and unobservable individual characteristics contribute to these patterns.

Figure 1 - Evolution of wage inequality - France

Image WP987
This figure shows the evolution over time of the variance of log-wage, the between-firm variance of log-wage, and the within-firm variance of log-wage. All individuals employed for at least 360 days by the same firm during the year are included for a given year. Individuals and firms in public administration are not included. Source: Base Tous Salariés (2002-2019).

Using a comprehensive database of matched employer-employee information, this paper examines how wage and workplace inequality evolved in France between 2002 and 2019. While wage inequality increased in many developed countries during this period, France presents an interesting case where overall wage inequality remained relatively stable (Figure 1). Underlying this apparent stability, however, were important changes in the distribution of workers across firms. 

First, there is increasing 'segregation' - workers with similar earning potential are increasingly likely to work together in the same firms. This means that high earners are increasingly concentrated in certain firms, while low earners are concentrated in others. In fact, this trend has been going on in France since the early 1980s. Second, there is a modest increase in 'sorting' - a growing tendency for high potential earners to be employed in firms that pay higher wages across the board. 

The increasing segregation between workers at different levels of earnings is mainly due to changes in the way occupations are distributed across firms. Firms are becoming more specialised in the types of occupations they employ. For example, firms are increasingly concentrating on either high-skilled occupations, such as managers and engineers, or low-skilled occupations, rather than having a mix of both.

The study also looked at whether these changes could be explained by other factors. We found only a modest role for rising returns to skill (where skilled workers command increasingly higher wages). We also found little evidence that changes in the way firms share profits with workers explain the patterns.

Rather, these trends reflect wider changes in the way work is organised across firms. Technological changes, particularly in information technology, have made it easier to coordinate work across company boundaries. At the same time, financial pressures have led firms to focus on their 'core' activities and to simplify their structures. This has led to more outsourcing and specialisation, with different occupational groups that used to work together within the same company now being spread across different companies.

This increasing segregation of workers by productivity level and occupation can have important social implications. As firms become more homogeneous in terms of the types of workers they employ, there are fewer opportunities for interaction across social and economic boundaries within workplaces. This could potentially reduce social mobility and increase inequality of opportunity, even if overall wage inequality remains stable.

Keywords: Wage Inequality, Worker Segregation, Occupational Sorting, Employer-Employee Data
Codes JEL: JEL: 23, J24, J31, J62

Updated on the 25th of March 2025