Data Insight: New evidence on gender pay gaps in Great Britain

This Data Insight by the Wage and Employment Dynamics team presents new evidence on the size of the gender pay gap in Great Britain.

First, it presents revised estimates of the gender pay gap in hourly earnings for full-time employees on average and across the wage distribution from 2004 to 2023. The analysis uses new weights for the Annual Survey of Hours and Earnings (ASHE) that account for non-random missing employer responses to the survey. Second, for the tax year 2018-19, we show the differences in men’s  and women’s weekly, monthly and quarterly employee earnings – aggregating across all jobs – using tax records from the linked ASHE- HM Revenue and Customs Pay-As-You-Earn (HMRC PAYE) dataset.

Read publication

Summary

The Wage and Employment Dynamics (WED) team has developed revised cross-sectional sampling weights to address a limitation of the ASHE dataset. In the first half of this Data Insight, we use these weights to produce new estimates of the gender pay gap in Great Britain, comparing them to those that would be estimated using the Office for National Statistics (ONS) standard procedure to address non-random sampling of employees. In doing so, we find that UK national statistics have highly likely been underestimating the gender pay gap consistently over the past two decades. In the last few years (2021-2023), for the headline measure of the gender pay gap, namely comparing median hourly earnings for full-time employees, this underestimation is about one percentage point, giving a total gap of nine rather than eight percentage points.

In the second half of this Data Insight, we present estimates of gender pay gaps in weekly, monthly, and quarterly tax data for those with at least one employee job in the tax year 2018-2019. These estimates are new and are only available due to the linkage of ASHE to the HMRC’s PAYE data, which record all earnings for all employee jobs subject to taxation in a tax year. These gender pay gaps at the median are roughly three-times larger than the median hourly gaps calculated with ASHE and also rise markedly with age.

What we found

• The impacts of addressing employer non-response bias in ASHE on national gender pay gap statistics

• Revised sample weights result in larger gender pay gap estimates for Great Britain

• The gender pay gap for total weekly earnings across all an employee’s jobs is over 30%

• The gender pay gap for total weekly employee earnings is greater among older workers

Why it matters

These new analyses of the gender pay gap matter for a variety of reasons. First, the analyses of the ‘enhanced’ ASHE, with weights accounting for employer non-response sample bias, show that gap in median hourly earnings between men and women working full-time has most likely been under-estimated in national statistics. This is probably explained for by the achieved ASHE sample under-representing employers with a larger gender pay gap, such as small private sector firms. 

Second, the ASHE-PAYE analysis provide a new perspective on the gender pay gap, by estimating the gaps in total earnings across all the jobs that employees undertake in a given period. This indicates that earnings gaps between those with at least one employer in a single tax year are considerably larger than those estimated at a single point in time for full-time employees only.

However, for the simple average gender pay gap, earnings gaps estimates do not appear to be particularly sensitive to the length of the period used -  whether it is one week, one month, or a quarter. This suggests that hours of work within a week rather than weeks of work and multiple job holding are the principal factors explaining average earnings gaps, though a deeper analysis looking beyond averages may reveal different patterns. Even so, since household consumption and well-being are reliant on total earnings received in a given period, this new perspective is a useful complement to the traditional focus on gender differences in hourly earnings at a point in time.