Gender pay gap underestimated in official statistics

Researchers have revealed that the data has not been weighted to account for small, young, private sector firms.

woman presenting to group of men
The response rate among the largest firms was around 50 percentage points higher than among the smallest firms.

A new study reveals that the UK's gender pay gap is larger than official estimates because the data used to calculate it is not weighted to account for jobs in small, young, private sector firms.

Researchers at University College London (UCL), Bayes Business School, the University of the West of England and the University of Stirling reviewed the Office for National Statistics' (ONS) Annual Survey of Hours and Earnings (ASHE), which is used to calculate the UK gender pay gap.

Under-estimated for 20 years

The study found that, despite efforts to weight the sample to be representative of the breadth of the UK workforce, it had not accounted for a higher survey non-reponses rate from some types of employers.

Having developed and applied a more representative revised weighting scheme, the researchers re-estimated the size of the UK gender pay gap. They found the gap has been consistently under-estimated over the past 20 years by a small but noteworthy margin of around one percentage point.

This happened because the original weighting under-represented smaller private sector firms, especially for women, and over-represented larger and public sector employers, where pay is generally higher and the differences between men and women within jobs are generally smaller. This under-representation was due to smaller private firms being less likely to respond to the ONS survey.

man in glasses looking at camera
Dr Carl Singleton
Senior Lecturer in Economics
The implications of our findings are wide-reaching, as the information used in this annual survey affects everything from pay deals to the National Minimum Wage.

Despite being a mandatory requirement, on average only 63% of employers responded to the survey between 1997 and 2019, dropping to just 46% since 2020. The response rate among the largest firms was around 50 percentage points higher than among the smallest firms, and around 15 percentage points higher among public sector employers than private sector.

The study also used the revised weighting to review the mechanisms and policy used by governmnet to set the National Minimum Wage (NMW) and National Living Wage (NLW). 

The researchers found that the government had inadvertently achieved its target of setting a NLW of two-thirds of median hourly earnings one year ahead of schedule, in 2023. This was because non-response to the survey meant they had over-estimated median hourly earnings.

The researchers recommend a review of the ONS’s Annual Survey of Hours and Earnings, which collects the data, to improve the accuracy and reliability of its methodology. 

Co-author Professor Alex Bryson, of the UCL Social Research Institute, said: “When we saw the higher non-response rate among some firms, we questioned whether this could be skewing the sample and, therefore, distorting its findings.

“By adjusting the weighting, we accounted for this bias and saw that the UK gender pay gap has been consistently under-estimated for decades.”

Implications are far-reaching

Co-author Dr Carl Singleton, of the University of Stirling Business School, said: “The implications of our findings are wide-reaching, as the information used in this annual survey affects everything from pay deals to the National Minimum Wage.”

Lead author Professor John Forth, of Bayes Business School, part of City St George's, University of London, said: “We welcome ONS’s recent announcement that they will be conducting a wide-ranging review of methodology behind the Annual Survey of Hours and Earnings. As part of the review, we encourage them to look at how the data can be made more representative of jobs in all types of organisations, so that the issues we observed can be addressed.”

The Annual Survey of Hours and Earnings is used by the Low Pay Commission to monitor the impact of the minimum wage and by the Office for the Pay Review Bodies to inform public sector pay settlements. It is also used by the World Economic Forum for international comparisons.

The research forms part of the Wage and Employment Dynamics project and was funded by ADR UK (Administrative Data Research UK) and the Economic and Social Research Council.

The Representativeness of the Annual Survey of Hours and Earnings and Its Implications for UK Wage Policy is published in the British Journal of Industrial Relations.