UK gender pay gap measurement bias and policy risk

The revelation that the Office for National Statistics has understated the gender pay gap by roughly one percentage point for more than two decades is not a niche methodological quibble. It is an institutional signal about how wage evidence is constructed, how policy bodies interpret it, and where the distributional stress sits in the economy. The finding turns on weighting inside the Annual Survey of Hours and Earnings, which appears to have leaned too heavily toward large employers that pay more on average and exhibit smaller male-female gaps, while under-representing smaller private firms where women are less well represented and gaps tend to be wider. Call it what it is: a quiet tilt in the evidence base that shapes national pay conversations. That is the core of the UK gender pay gap measurement bias.

The timing matters because Ashe underpins work by the Low Pay Commission, the Office of Manpower Economics and pay review bodies, as well as a host of Whitehall and devolved analyses of sectoral conditions. If the underlying survey gave outsized weight to larger firms, the state’s reading of pay pressures in smaller enterprises will have looked softer than reality. That does not mean the minimum wage was mis-set in any particular year. It does mean the calibration of risk around female-heavy segments in small and mid-sized businesses may have been chronically low. Policy that is designed around an average built on big-company structure will tend to miss fragility where HR capacity is lean, progression ladders are short, and bargaining power is thin.

The ONS says its sampling and weighting are under review and that improvements since 2024 will address elements of the critique. That is necessary, but the institutional question is broader. Ashe is an employer-based survey that feeds directly into decisions with real cash consequences, from the National Living Wage trajectory to the framing of public sector comparators. A one-point understatement of the gender gap is not catastrophic. It is, however, enough to tilt narratives on whether progress is holding, stalling, or reversing in female pay outcomes within smaller firms. Narratives matter, because they set the room temperature for what review bodies believe is feasible in a given pay round.

There is also a governance point about data asymmetry. Large employers produce cleaner returns and respond more consistently, which makes them attractive to statistical systems under budget constraint. Over time, that incentive creates representation risk. The public sector and big private employers dominate the lens, while the texture of pay in dispersed services, care, and hospitality remains under-lit. The result is a subtle policy bias toward the parts of the labor market that look tidy on paper. Once baked in, that bias propagates through wage distributions, earnings deciles, and productivity narratives that rely on Ashe as a keystone input.

Internationally, wage statistics offices have wrestled with similar problems, which is why many pair establishment surveys with household-based instruments and pursue aggressive calibration to firm size, sector, and hours. The point is not that the UK lacks these tools. The point is that when a flagship series becomes the default anchor for wage discourse, small methodological drifts can compound into policy priors. In this case the drift flattened the apparent difference between men and women in the strata where policy is trying to move the needle most.

The practical implications will land in three places. First, distributional assessment. Review bodies evaluating awards for feminized sectors will need to revisit their sense of baseline fairness, since a one-point wider gap across many years reframes what “progress” means in the data. Second, enforcement and compliance. If the small-firm universe looks worse than previously measured, monitoring resources for equal pay and discrimination risk will need a sharper small-business focus, not only headline transparency for listed companies. Third, wage-floor trajectory. A mis-weighted understanding of low-pay dynamics can lead to under-powered adjustments in one cycle and over-correction in the next. Smoother paths require better representation.

For the ONS, the fix is not only technical. It is also communicative. When survey methods change, users need a clear bridge between old and new series, an articulation of what revisions imply for trend interpretation, and a statement about which downstream decisions should be considered with caution. That includes guidance for departments and devolved administrations that have relied on Ashe-based dashboards for local pay setting. Without that, the reputational damage extends beyond a single survey and bleeds into broader confidence in UK official statistics at a moment when the macro conversation needs credible anchors.

There is an uncomfortable inference here for macro watchers. If the weighting scheme compressed the measured gender gap by privileging larger enterprises, other distributional estimates inside the same survey may carry similar biases. Median versus mean decompositions by firm size, sectoral earnings ladders, and progression estimates for part-time workers could all reflect versions of the same tilt. That does not invalidate the series, but it does require more humility about how tight our confidence intervals really are when we translate Ashe lines into policy moves.

The corrective path is straightforward in principle. Re-weight to employer population frames that better capture small-firm reality. Increase the rotation and sample re-contact discipline for under-represented segments. Pair the employer survey with more aggressive linkage to administrative data where lawful and proportionate. Publish a revision study that quantifies how far historical gender gap estimates would shift under the new weights. Then guide users on how to interpret the back-series. These are operational tasks, but they are also political in the small-p sense, because they change the story the state tells itself about wages.

What this signals is simple and important. The UK is not facing a crisis of wage statistics. It is facing a credibility test in how it measures where pay pressure and pay inequity actually live. A one-point understatement across two decades sounds modest. In policy time, it is not. It shapes tone, calibrates caution, and influences whether pay settlements for female-heavy sectors are seen as catch-up or as stretch. Expect review bodies to welcome the ONS audit. Expect them to ask for more. The posture may appear technical. The signal is institutional.