← All articles
Society

What explains the gender pay gap?

· first posted on LinkedIn (opens in a new tab)

Some factors widen the gender pay gap. Others narrow it. That is why a single headline number tells only part of the story.

Waterfall chart for the European Union in 2022 that breaks the 12.2% hourly gender pay gap into contributions from industry, full-time or part-time work, occupation, education, other adjustments and an unexplained component. Industry and working time widen the gap; occupation, education and other adjustments narrow it. (page 1 of 2)
Waterfall chart for the United States in 2015–2016 that breaks the 19.1% geometric-mean pay gap into contributions from industry, occupation, work history, education, weekly hours, other characteristics and an unexplained component. Education and weekly hours narrow the gap; the other measured factors widen it. (page 2 of 2)
Contributions to the gender pay gap, in percentage points: European Union, 2022 (first chart) and United States, 2015–2016 (second chart).

These two waterfalls unpack the contributions estimated in separate EU and US studies:

  • Industry matters: women and men work in different sectors with different pay levels.
  • Education offsets part of the gap in both studies.
  • Occupational profiles narrow the gap in the EU analysis but widen it in the US study.

What “unexplained” means

The remaining “unexplained” component is what the models cannot account for using their measured characteristics. It may include discrimination, unmeasured differences and data limitations.

And “explained” does not mean fair: access to occupations, promotions and career opportunities can itself be unequal.

About the data

Scope: EU data from 2022; US data from 2015–16. Different populations and methods mean these are separate illustrations, not a like-for-like comparison. US contributions are proportionally allocated from the study’s log-wage decomposition.

Countries in this post

Discuss on LinkedIn

Which part of the breakdown surprised you most?

More from The Data Atlas