Early findings from the Office for National Statistics suggest that incorporating PAYE Real Time Information into survey weighting could shift estimated employment rates by up to 0.8 percentage points.
At the start of 2024, the proposed weighting method would have raised the estimated employment rate by 0.8 percentage points compared with the published Labour Force Survey series, according to research published by the Office for National Statistics on 14 September 2026. The analysis linked HMRC Pay As You Earn Real Time Information records with Labour Force Survey records to identify whether people in PAYE employment respond to the survey at different rates than those without it — and found evidence that this bias has shifted over time.
The practical effect cuts both ways. Applying the PAYE RTI weighting method would lower estimated employment rates during the peak of the coronavirus pandemic, but raise them between 2022 and 2024. ONS says those changes in response patterns may help explain periods when Labour Force Survey estimates diverged from administrative measures of the labour market.
The survey’s sample size gives some context for why this matters. The achieved Great Britain sample had fallen to an approximate low of 14,000 households at the start of 2024 before beginning to recover — a drop that followed the pandemic-era shift to wholly telephone-based interviewing, which ONS says increased non-response bias. The agency introduced housing-tenure-based weights in October 2020 to address some of that distortion.
The response-rate gap between PAYE and non-PAYE addresses isn’t new. ONS research from 2018 found that addresses with PAYE income had a response rate around 6 percentage points higher than those without, and estimated a potential bias of about 103,500 people for the tax year ending 2016.
But PAYE RTI and the Labour Force Survey don’t measure exactly the same thing. RTI covers employees paid through company payroll; the survey captures all forms of employment, including self-employment. ONS is explicit that RTI currently provides the most reliable measure of employees, while stopping short of presenting it as a direct replacement for the survey.
ONS advises data users to continue relying on routinely published labour-market statistics while the research develops. The findings are described as early and sit within a wider programme to improve the quality and representativeness of labour-market statistics.
Source: @ONS