Labour market transformation – update on progress and plans: August 2026

Transformed Labour Force Survey

• Transitioning to the online-first Transformed Labour Force Survey (TLFS) in 2027 is an important strategic priority for the Office for National Statistics (ONS) as the sustainable solution to declining participation and quality issues affecting traditional social surveys.

• Early data following the upgrades to the TLFS in April 2026, indicate improved performance, including higher return rates, fewer partial returns and shorter completion times, with median household completion times of 17 minutes for Wave 1 and between 9 and 12 minutes for Waves 2 to 5 following the introduction of data rotation between survey waves; this compares with around 40 to 45 minutes for the Labour Force Survey (LFS).

• Our first readiness assessment in July 2026, conducted with our main users, reconfirmed that now is not the right time to transition to the TLFS for our headline labour market statistics and as the effect of design and methods improvements are continuing to feed through into the data, we will not publish research outputs at this stage.

• For the remainder of this year, our focus will be on monitoring the effect of survey improvements, maturing priority datasets for transition, improving industry data quality and progressing analysis on Household, Socioeconomic and Local (HSL) data, leading us to a critical progress check in January 2027 where we will prepare a detailed transition plan that we will publish in the first quarter of the year.

• July 2027 will be an important decision point when, if we and our users are ready, we will start the process for a November 2027 transition to the TLFS; this remains the most likely path and balances a data-led approach with the increasing costs and quality implications of continuing the dual run of the TLFS and LFS.

• The quality of the TLFS will improve further after transition, when additional resource and data will be available, enabling further methodological upgrades and survey improvements to take place.

Labour Force Survey

• The LFS remains the lead source of labour market data and continues to run alongside the TLFS while development continues.

• Response levels for the LFS have shown clear improvement since 2023 because of interventions, including increases to survey incentives and interviewer numbers, with response levels for Waves 2 to 5 slightly above their pre-coronavirus (COVID-19) pandemic level in January to March 2026.

• These improvements continue to be welcomed by users, including the Stakeholder Advisory Panel on Labour Market Statistics, as they increase confidence in the LFS data until the transition to TLFS completes.

• An operational issue occurred in May 2026, which led to temporary under-resourcing in LFS telephone collection operations; a number of immediate actions were taken in response, accompanied by a comprehensive lessons learned review, and our analysis indicates that the issue had a minimal effect on the headline estimates.

• Research findings to help understand potential employment-related variation in non-response bias by linking LFS survey data with HM Revenue and Customs (HMRC) Pay As You Earn (PAYE) Real Time Information (RTI) will be published in autumn 2026.

In response to user feedback, we have expanded the coverage of information on the Transformed Labour Force Survey (TLFS) in this article. We welcome further feedback from users by email at labour.market.transformation@ons.gov.uk.

Improved survey design and performance

As detailed in the previous update, a series of survey improvements were implemented in April 2026. These were the addition of data rotation, enhancements to pay and earnings questions, and an increase in the conditional incentive value from £10 to £20.

Data rotation, where some questions relating to household composition and education are not re-asked in full in subsequent waves, was particularly highlighted by stakeholders and methods experts as an important improvement for enhancing data quality. We are starting to see the effect of these changes in initial survey response metrics based on the first nine weeks of available data. These early data indicate improved performance, including higher response rates, fewer partial returns and shorter completion times.  

For the April to June 2026 quarter, the TLFS Core Wave 1 return rate was 39.9%; this compares with a return rate of 37.1% for the January to March 2026 quarter before the survey improvements were made. The longitudinal return rate for Waves 1 to 5 for April to June 2026 was 36.7%, a 3.7 percentage point improvement on the January to March 2026 response rate of 33.0%.

The longitudinal response rate calculations for the TLFS and Labour Force Survey (LFS) differ and cannot be directly compared; this is because of the TLFS subsampling a defined proportion of the Wave 1 sample for longitudinal follow up, whereas the LFS follows up all eligible Wave 1 cases.

Early analysis for the April to June 2026 quarter shows some notable improvements on the TLFS compared with the LFS. The TLFS Core Wave 1 is significantly quicker for respondents to complete than the LFS. Early data indicate that completion time is reducing further for Waves 2 to 5 following the introduction of data rotation, with median household completion times for those waves reducing to between 9 and 12 minutes.

The median household completion time for Wave 1 is approximately 17 minutes, compared with the LFS median completion time of 40 to 45 minutes. TLFS completion times across all waves represent a significant reduction in burden for respondents, and an improvement to their overall experience. Furthermore, to achieve its response rate, the TLFS also requires less interviewer time than the LFS because of its "online-first" approach and offers lower financial incentives for respondents.

We will be closely monitoring performance in this area in the coming months. Once we have two full quarters of data collected on the new basis (in November 2026), we will be able to assess more fully the effect of the design changes implemented in April 2026. Future articles will include more detailed performance metrics, including comparisons with the LFS, once the required data are available and have been analysed.

Crossover survey

To better understand differences between estimates from the TLFS and LFS, we conducted a crossover survey in 2025. The study invited respondents who had recently completed the interviewer-led LFS to also complete the online TLFS questionnaire.

By comparing responses from the same individuals across both surveys, the crossover survey controls for differences in sample composition and non-response, allowing us to better understand the extent to which differences between LFS and TLFS estimates may be influenced by survey mode and questionnaire design. More than 5,000 respondents provided usable data for analysis.

Analysis of the crossover survey helps us to understand how mode and questionnaire changes affect the measurement of respondents' economic activity. Overall, the crossover survey found high levels of agreement between LFS and TLFS responses for measures such as aggregated economic activity (whether someone is employed, unemployed, or inactive) and employment type (employee, self-employed and unpaid family worker). Responses were broadly consistent across the two surveys for many core labour market measures.

The largest differences, and lower levels of agreement, were seen for types of inactivity. For example, fewer respondents were classified as long-term sick on the TLFS, with some reporting instead that they were retired, looking after family or home, or economically inactive for other reasons.

These differences could reflect how some questions are asked on the TLFS, with changes to both the wording of some response options and to the order in which they appear to respondents. These changes were implemented, in part, to make this question less burdensome to respondents.

For example, the ordering of response options now better reflects the prevalence of different reasons for not having sought work (that is, reason for economic inactivity). Within the LFS questionnaire, the relevant question presents "retired" as the ninth response option, while for TLFS it is presented first as this is the most common reason given by respondents.

Findings from the crossover survey are strengthening our understanding of the differences between LFS and TLFS estimates and informing ongoing survey development and evaluation. Further analysis of reported disability is under way, alongside work to better understand the reasons for differences seen in economic inactivity classifications. Further updates will be provided as part of this regular series of articles.

Industry and occupation data

Collecting high-quality data on the complex Standard Industrial Classification (SIC) and Standard Occupational Classification (SOC) variables remains a priority and an important challenge for the TLFS. Progress continues in the development of multiple parallel solutions to meet two-digit SIC and four-digit SOC requirements.

The search-as-you-type question for the SIC, which helps respondents to describe their industry to capture the correct code, has performed well in stand-alone testing. A larger-scale test is running from July to September 2026 to confirm there is no major respondent drop-off or bias. Analysis and a decision on implementation will follow later in the year. Users will be kept updated as part of this regular series of articles.

Alongside this, newer iterations of the Classification Index Matching Service (CIMS) tool are being developed to classify responses after-processing for both the SIC and SOC. Recent analysis has shown an improved uplift in codability for the latest versions, however, work continues to improve the algorithms before implementation on live data.

Initial feasibility work is also under way into the use of administrative data to provide a contingency for the SIC if search-as-you-type testing does not provide the required accuracy. Work to link survey data to HM Revenue and Customs (HMRC) Pay As You Earn (PAYE) data offers an alternative route to infer an employer's industry classification that could be used to populate missing data, or validate responses, for employees.

UK data coherence

As outlined previously, in the future, labour market information for Northern Ireland (NI) will be collected by the Northern Ireland Statistics and Research Agency's (NISRA's) Labour Market Survey (LMS). The Office for National Statistics (ONS) will then combine the LMS data, including seasonal adjustments, with the corresponding data from Great Britain to enable the production of results covering the UK. Initially the Great Britain data will be from the LFS then, in time, from the TLFS.

Significant joint progress has been made since our last update. This includes testing of the integration of NI LMS and Great Britain LFS data, which has increased confidence that UK-level labour market outputs will continue to meet the current publication timeline.

Our Labour Market Technical Group and Stakeholder Advisory Panel have informed the prioritisation of the variables and outputs that will be fully consistent at the UK level, and those that will move from UK- to Great Britain-level following NISRA's transition. Once these have been finalised, we will update users as part of this regular series of articles.

The first UK-level person-based quarterly dataset combining LMS and TLFS data has been produced and is undergoing initial quality assurance ahead of the supply of a longer time series, which will undergo fuller quality assurance. The ONS and NISRA have established a joint Steering Group, comprising LMS and TLFS lead officials, to strengthen governance of the ongoing UK data integration work. Our joint focus over the coming months is to develop integrated transition plans, taking into account decisions the ONS has taken as part of its TLFS July 2026 readiness assessment, and NISRA's ongoing analysis and quality assurance that will underpin decisions around the timing of their transition to the LMS.

The latest updates on the main developments and future plans with NISRA's LMS can be found on their Transformation of Labour Market Statistics in Northern Ireland webpage.

Household, Socioeconomic and Local statistics

In June 2025, the Household, Socioeconomic and Local (HSL) project was initiated to strengthen understanding of user needs and our ability to meet them, beyond labour market and productivity statistics. The main domains include health, education, demographics and local-level insights.

Our aim is to identify priority topics and assess data quality, so that the TLFS can deliver high-value data, aligned with HSL stakeholder needs. This work is critical for supporting decisions on survey design and the decommissioning of the current LFS and Annual Population Survey (APS). 

We have continued to make progress in our assessment of HSL data from the TLFS. The data include a wide range of socioeconomic themes such as education, health, travel, and wider household characteristics to support local, granular analysis across the survey. We have progressed assessing quarterly data in the interim to provide stakeholders early insight on quality while we develop an annual dataset and supporting weights.

User engagement on the APS has reconfirmed our understanding of the need for high-quality annual HSL datasets as a replacement. This has been explored further through an engagement exercise with users, which aimed to seek detail on critical user needs for these data, and has been used to update the assumptions around HSL data within our programme plans. Our engagement work will continue over summer 2026 to gather feedback specifically from local stakeholders and fill gaps in our evidence base, providing a clearer understanding of what is essential to our users and continuing to inform the prioritisation of the TLFS development.

We have been assessing the quality of APS data, including breakdowns by main demographic and geographic characteristics, and analysis of longer-term trends in the main statistical outputs sourced from the APS.

In September 2026, we will publish an APS quality article to coincide with the publication of APS data in our Labour market in the regions of the UK: September 2026 release. This will set out what the previously announced reduction in the survey boost means for the reliability and use of APS outputs, and include some analysis to make an early assessment of the effect on the quality on publications, including for local geographies. This will be a practical illustration of how changes may affect selected APS outputs.

We will also outline how the ONS will support users in the future, both while quality improvements are under way and where longer-term alternative data sources would be needed.

We provide regular updates to the HSL Technical Group but are keen to engage a wider range of HSL users. Those with an interest in TLFS HSL data who are not yet involved are encouraged to email TLFSPlusTransformation@ons.gov.uk.

July 2026 readiness assessment

The first readiness assessment was conducted in collaboration with our main users in July 2026. It was an important milestone providing critical insights to inform our next steps, including final design decisions, future publication plans and the timing of the next readiness review. Crucially, it informed a credible forward plan needed to support a safe transition in 2027, which remains the most likely path.

As part of the assessment, we engaged extensively with our main users, including the Stakeholder Advisory Panel on Labour Market Statistics. The assessment was informed by our measures of success for transition, which we continue to review and refine in collaboration with users. The measures are centred around the three pillars:

• Statistical and Data Quality

• Operational Readiness

• User Confidence and Readiness

Further detail on each pillar was provided in Section 6 of our Transformed Labour Force Survey measures of success April 2026 article.

We had already announced that we would not be ready to begin transition in July 2026 as only one calendar quarter of data reflecting the latest set of design changes has been collected, resulting in limited quality assessment of their impact.

We are doing everything possible to enable a safe transition in 2027, including mitigating risks. For example, monitoring measures of change precision improvements closely, and establishing an administrative data workstream for the Standard Industrial Classification (SIC). We will continue to discuss remaining risks with users alongside any required mitigating actions, recognising that further trade-offs may be needed.

For the remainder of this year, our focus will be on maturing priority datasets for transition, improving industry data quality, progressing analysis on HSL data, and monitoring the effect of survey improvements in preparation for a progress check in January 2027. This check will not be a full readiness assessment, instead it will review progress and confirm the programme trajectory on the main areas of uncertainty. Following this, we will prepare a detailed transition plan that we will publish in the first quarter of the year.

Publication plans

In line with expectations, the readiness assessment also confirmed that TLFS research outputs should not be published at this stage, as imminent design and methods changes will shortly lead to revisions. Publishing outputs once these changes have been incorporated should be more helpful to users as they make their own preparations for transition.

As set out in our previous update in April 2026, this decision balances the need to make potentially useful data available to a wider range of users and the benefits of enabling broader feedback with the importance of sufficient data quality, user confidence, and minimising confusion in the labour market narrative. This approach is supported by the Labour Market Technical Group and the Stakeholder Advisory Panel on Labour Market Statistics.

TLFS data will be published before transition to allow users to understand and prepare for the impact of the change. The decision on when and how to publish TLFS data will be revisited at the review point in January 2027.

In the meantime, TLFS user guidance is available to provide users with information on the survey design, variables and methods, and how it compares with the current LFS and APS, to support their understanding and readiness ahead of transition.

Transition approach and timeline

Transitioning to the TLFS for our published headline labour market statistics is an important strategic priority for the ONS. The timing of transition to the TLFS remains a data-led decision that is being balanced with the cost of continuing the dual run with the Labour Force Survey (LFS), which is increasingly constraining data quality and limiting the ONS's ability to invest in wider statistics improvements.

Given the complexities and resource involved, neither the LFS nor the TLFS will reach their quality potential while we are running the two surveys in parallel. Our approach is therefore to get the TLFS data quality to an acceptable level for us and our main users to have confidence in turning off the LFS and APS. This will then allow the quality of the TLFS to improve further after transition, where additional resource will be available.

The transition to TLFS will be the start of a managed improvement process, not the point at which all development ends. Further methodological improvements are likely after transition, particularly as the SIC and SOC, and HSL data quality improve. Upgrades will also follow transition because additional data, based on the improved design, are required for optimisation.

Our approach is to manage the remaining uncertainty, mitigate the main risks and maintain a credible path to enable a safe transition to TLFS in 2027. We are focusing on improving data quality and working with users to agree any trade-offs needed to transition. For example, we are exploring where lower-priority activity can be de-scoped until after transition, and where some specific data quality limitations may be acceptable at transition.

Once any trade-offs have been agreed, we will provide an update as part of this regular series of articles, setting out what improvements users can expect to see before and after transition. We will also publish a detailed transition plan following the progress check in January 2027.

Looking ahead to July 2027, transition readiness is projected to improve reaching the minimum or target levels, across a number of the measures of success for transition. July 2027 will be an important decision point when, if we and our users are ready, we will start the process for a November 2027 transition to the TLFS.

The main activities and waypoints we are working towards over the next 18 months are now outlined.

Quarter 3 (July to Sept) to Quarter 4 (Oct to Dec) 2026

Goal: Analyse additional data and reduce uncertainty

• First two quarters with data rotation and incentives included.

• First annual datasets available for analysis.

• Outcomes of Standard Industrial Classification (SIC) search-as-you-type testing.

Quarter 1 (Jan to Mar) 2027

Goal: Check progress and publish detailed transition plan

• January progress check and transition plan published.

• Early knock-to-nudge evidence reviewed.

• Priority questionnaire changes made.

Quarter 2 (Apr to June) 2027

Goal: Build user readiness

• Priority UK data ready for headline labour market estimates.

• Potential publication of research outputs.

• Phase 2 delivery plan agreed.

Quarter 3 2027

Goal: Assess readiness and make decision on transition

• Decision on transition to TLFS.

• UK two-quarter longitudinal datasets ready to produce estimates of labour market flows.

• Potential publication of further research outputs.

Quarter 4 2027

Goal: If ready, transition to the TLFS as main headline estimate

• Transition of headline labour market data (subject to readiness).

• UK household quarterly datasets ready for the main estimates of households in the labour market.

From Quarter 1 2028 onwards

Goal: Ongoing transition and improvement

• Phase 2 delivery plan implemented.

• Transition of further labour market and Household, Socioeconomic and Local (HSL) data (including annual data).

The Labour Force Survey (LFS) remains the lead measure for data on the supply of labour while further development of the Transformed Labour Force Survey (TLFS) takes place. The LFS continues to run in parallel with the TLFS, at least until we complete the transition to the TLFS as the headline measure.

As set out previously, we have taken several actions to address quality concerns with the LFS, which are now fully embedded in the survey. These include:

• returning to face-to-face interviewing

• increasing incentives

• continuing to recruit additional interviewers

Our latest update on LFS quality was published alongside the labour market statistics release on 21 July 2026. The main conclusions from the article were that:

• response levels for the LFS have shown clear improvement since 2023 because of interventions, including increases to survey incentives and interviewer numbers, to improve quality, with response levels for Waves 2 to 5 slightly above their pre-coronavirus (COVID-19) pandemic level in January to March 2026

• coherence between the LFS and other labour market data sources, although improving, remains a challenge, and we continue to monitor this closely; you can find detail on the coherence of the latest estimates in our Labour market overview: July 2026 bulletin

• while single month and wave-based estimates can provide valuable additional context, they are best used alongside the headline three-month estimates; comparisons over three-month intervals and analyses using the wave structure can offer further insight into underlying trends

As the improvements continue to feed through to the published data, we recommend that the LFS is used as part of our suite of labour market indicators alongside workforce jobs, Claimant Count, and Pay As You Earn Real Time Information (PAYE RTI) estimates.

These improvements continue to be welcomed by users, including the Stakeholder Advisory Panel on Labour Market Statistics, as they increase confidence in the LFS data until the transition to TLFS completes. Although LFS data quality has improved, the online-first TLFS remains the sustainable solution to declining participation and quality issues affecting traditional social surveys.

Lessons learned from operational issue

An operational issue occurred in May 2026, which led to temporary under-resourcing in LFS telephone collection operations. This resulted in fewer achieved household interviews than expected during this period. Our analysis indicates that this had a minimal effect on the headline estimates, although it does suggest there is a small but noticeable effect on the estimates of average hours and therefore the total actual hours worked.

A subsequent lessons-learned review, supported by independent external input, found that the issue did not arise from a single decision or error. Instead, it resulted from a combination of operational changes and pressures within a complex dual-run environment. Collectively, these created a mismatch between operational demand and available interviewer capacity, affecting delivery of the LFS as planned. The review also identified scope to strengthen the early identification and escalation of any effects on interview volumes.

Based on the review's findings, immediate actions have been implemented to strengthen operational oversight, governance and performance monitoring. We are also undertaking a wider assessment of social survey operations to assess whether the same lessons learned apply in other areas to ensure a sustainable operating model across the survey portfolio. Further updates will be provided in our next Surveys and Economic Statistics Improvement Plan quarterly update.

Research is under way to understand whether linking HM Revenue and Customs (HMRC) Pay As You Earn (PAYE) data to the Labour Force Survey (LFS) and Transformed Labour Force Survey (TLFS) can help identify and address employment-related non-response bias. The work has progressed resulting in improved methods, updated data, and ongoing engagement with external experts.

Initial findings from the LFS analysis, which reconfirm that the LFS recovery activity has reduced survey bias, will be published in autumn 2026.

Alongside publication, the work will continue to develop our understanding of the presence, size and direction of any employment-related non-response bias. It will support ongoing monitoring of survey quality and help understand differences between the LFS and TLFS to smooth the transition.

While the research is promising, significant further development, testing and assurance are required before any recommendation on incorporating PAYE data into survey methods can be made. Consistent with external expert advice, the priority remains to allow TLFS methods to stabilise through transition. Subject to evidence of quality improvements, operational feasibility and readiness, any future implementation would therefore be considered as a potential post-transition methodological enhancement.

Linked Employer–Employee Dataset

Further progress has been made on the Linked Employer–Employee Dataset (LEED), which links workers to their employers and enables longitudinal tracking.

Work is progressing in assessing the feasibility of improving quality assurance for existing labour market statistics. There has been a coherence assessment between LEED and two business surveys – the Short-Term Employment Survey (STES) used in the private sector employee jobs component of workforce jobs estimates and the Monthly Wages and Salaries Survey (MWSS) used for monthly Average Weekly Earnings (AWE) estimates.

This assessment has shown strong alignment in employment and earnings at the aggregate level. Further micro-level investigation will be undertaken to better understand the findings and consider how LEED may be used to inform recommendations within existing statistical processes for these business surveys.

We plan to publish the first set of experimental labour market flows statistics in late 2026, which would capture movements in and out of payrolled employment, including job-to-job flows between and within industries. These will provide new insights into how people move through the labour market, including hiring, job changes, and separations from paid employment, helping to improve our understanding of how the labour market is evolving.

Alongside this publication, a methodological article setting out underlying data and methods will also be published.

We continue to gather feedback from a wide range of stakeholders in government, academia, and the wider research community, as part of our efforts to prioritise how best to maximise the impact of LEED, including ensuring the outputs remain relevant for our users.

Surveys and Economic Statistics Improvement Plan

In July 2026, we published our third quarterly update on the improvement work under way across ONS surveys and economic statistics. The update reflects refined priorities, reset milestones and clarity about where activity must wait until constraints are resolved. The update reinforced the importance of progressing the TLFS to become our headline measure for labour market statistics in 2027.

In an accompanying blog, firmer footing: our latest update on our Surveys and Economic Statistics Improvement Plan, we explain that while there is still much to do, putting quality ahead of quantity, and credibility ahead of volume remains the guiding principle driving the ongoing recovery in economic statistics.

The update also recognised a major milestone in that we have completed the move of the Annual Survey of Hours and Earnings to digital-first collection, reducing operational cost and risk, and speeding up data processing.

UK Labour Accounts

Beyond completing the projects outlined in April 2026, we are also investigating both existing and new data sources that could contribute to a future UK Labour Accounts framework. A main priority is the systematic assessment of the quality, strengths and limitations of these data sources, ensuring that they are suitable for use within a coherent framework aligned with UK National Accounts concepts and definitions. This quality assessment is a critical step in the development process, as it will provide the evidence needed to integrate data sources consistently and support the construction of the four Labour Accounts quadrants.

We also continue to engage with international partners and national statistical institutes to share knowledge, discuss methodological approaches, and identify best practice in the development of Labour Accounts.

Through this work, we aim to build a robust and transparent framework that delivers a more comprehensive and coherent view of labour market activity across the UK economy.

We continue to make progress with actions in response to the Office for Statistics Regulation (OSR)'s Systemic Review of Economic Statistics and their ongoing Regulatory Review of Labour Market Statistics. We regularly engage with the OSR on the significant work under way and our progress in important areas.

We consider the designation of statistics as an important signal of their trustworthiness, quality and value for users and have set out our approach in a letter to the OSR for outputs derived from the Labour Force Survey (LFS) and the Annual Population Survey (APS).

Reflecting the significant measures we have taken in recent years that have improved response levels from their lowest point to close to pre-coronavirus (COVID-19) pandemic levels, supported wider improvements in quality, and been recognised by important users, including the Bank of England and the Office for Budget Responsibility, LFS and APS-derived outputs that are currently official statistics in development will move to official statistics. We will not be seeking reaccreditation for these outputs, instead, our priority is for outputs based on the TLFS, as the online-first sustainable replacement for the LFS and APS, to achieve the standards required to become accredited official statistics and for the OSR to consider their accreditation at the earliest opportunity. We will provide an update on designation and an explanation of our approach within the next release of each affected output.

More widely, our Average Weekly Earnings statistics were recently reviewed by the OSR. The review found the statistics to be robust, credible and supported by methodological and system improvements. The OSR concluded that the statistics should retain their accredited official statistics designation.