The wage premium that disappears when history enters the model

The French evidence examines a different object from the U.S. job-transition study: not what happens after entering a remote-eligible job, but why employees who report working from home earn more at a given moment. The central question is whether that wage-level gap survives controls for where people work, who they are, which firms employ them, and what they earned before the pandemic. Following that control ladder shows why a visible work-from-home premium can be largely a selection pattern rather than a direct return to remote work.

This distinction is essential. A worker can enter a remote-eligible role and receive higher wage growth because the role is a better match, while people who work from home at a particular time can still have higher wages because higher-paid workers are more likely to obtain the arrangement. One result concerns a transition; the other concerns the composition of workers who hold an arrangement. Treating them as competing estimates of the same effect would make both harder to understand.

What the French wage comparison begins with

The study links France’s Labor Force Survey to firm-registry information and social-security records. The survey supplies work-from-home status, hourly wages, occupation, education, location, and other worker characteristics. Firm data add employment, age, productivity, and related employer attributes. Administrative employment and wage histories provide a view of workers’ pre-pandemic outcomes. The post-pandemic baseline covers the second quarter of 2022 through the fourth quarter of 2024, after pandemic restrictions were no longer the main explanation for who worked at home.

The baseline WFH measure is broad: an employee counts as working from home if they report doing so during the previous four weeks. It includes workers who do some, most, or all of their work at home. The study also checks a stricter contractual remote-work measure and separates different levels of intensity. That breadth is useful for measuring the prevalence of any home work, but it means the baseline category is not equivalent to full remote employment or to the employer-side eligibility measure used in the U.S. study.

With only year fixed effects, workers who report WFH have hourly wages 35.2% higher than workers who do not. That figure is descriptive, not a causal estimate. WFH is not distributed randomly across the labor market. Higher-paying occupations may be easier to perform from home, better-educated workers may have more bargaining power, and firms with different management practices may be more willing to offer flexibility. A raw comparison combines all of those differences with any effect of the arrangement itself.

The first control step narrows the comparison to workers in more similar labor-market positions. The study adds detailed occupation, industry, and commuting-zone controls, each interacted with year effects. The wage premium falls from 35.2% to 12.0%. This large reduction shows how much of the initial gap reflected where people worked and which occupations were represented among WFH employees. It does not mean that occupation and place are irrelevant to remote work; it shows that they are part of the sorting process that must be separated from the arrangement.