Labour market transformation – update on progress and plans: November 2025

Transformed Labour Force Survey

• The short Core Survey (the streamlined, longitudinal, labour-market-focused questionnaire) successfully launched in July 2025, and provisional data indicate that it is reducing the number of partial household responses and improving representativeness by increasing the proportion of larger households responding.
• Supported Completion, which enables interviewers to assist respondents within their own homes to complete the TLFS online, went live in October 2025 and is expected to improve response rates among harder-to-reach groups.
• The quarterly sample size for the TLFS Plus Survey (the separate, cross-sectional survey to meet needs for additional labour market, and wider household, socioeconomic and local data) increased from 50,000 to 90,000 households in October 2025 and now matches the Core Survey sample size.
• Data rotation (not re-asking the full questionnaire for each wave) and improvements to pay and earnings questions are now being aimed for implementation in April 2026 because the scale and complexity of testing is too great to complete by January 2026.
• The timing of transition to the TLFS remains an evidence-led decision, with the first readiness assessment conducted in collaboration with our main users scheduled for July 2026.
• We aim to transition to the TLFS for our published headline labour market statistics in November 2026 if we and our users are ready, though if our (or users’) assessment of quality requires more data to be collected and assessed, the transition may extend into 2027. 
• Additional complexities in the implementation of the remaining design changes and methods experienced since our last update may affect the decision as to whether more data will be required.
• Moving the implementation of data rotation to April means that by July 2026, only one calendar quarter (three-month period) of data reflecting the change will have been collected, allowing for only limited quality assessment of its impact.
• We will keep the impact of this under review with our main users and provide further updates as part of this regular series of labour market transformation articles. 

Labour Force Survey

• The Labour Force Survey (LFS) remains the lead measure for data on the supply of labour and continues to run in parallel with the TLFS while further development takes place.
• As updated previously, as a result of the ongoing improvements to address quality concerns with the LFS, response levels and rates have shown clear improvement, with Wave 1 response levels now very close to their pre-coronavirus (COVID-19) pandemic level.
• 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 takes place.
• Recent increases to interviewer capacity for Waves 2 to 5 will further improve response levels and representativeness, while a full reweighting of the LFS and Annual Population Survey will further improve the coherence of the estimates.

Design implementation

In our July 2025 update article we announced the launch of the short Transformed Labour Force Survey (TLFS) Core Survey (a streamlined, longitudinal, labour-market-focused questionnaire with a Wave 1 sample size of 90,000 households per quarter across Great Britain).

The Core Survey has now been live for four months and provisional data indicate that it is reducing the number of partial household responses and improving representativeness by increasing the proportion of larger households responding. However, we have not yet seen any improvement in overall response rates.

Following additional quality assurance of our labour market eXperimental (LMX) analysis published in April 2025, we have revised our expectations for response rates for the Core Survey. More information can be found in the updated technical design review analysis. We are continuing to explore options to improve response, and a fuller update on the Core Survey performance will be included in the next labour market transformation article in early 2026.

Further improvements have been implemented including Supported Completion, which enables interviewers to assist respondents within their own homes to complete the TLFS online, which went live in October 2025. This will ensure that digitally excluded or less confident respondents can still participate and aims to improve response rates among harder-to-reach groups.

The main outstanding design changes necessary to meet critical user needs are data rotation, pay and earnings, and complex variables (Standard Industrial Classification (SIC) and Standard Occupational Classification (SOC)). Testing and implementation of these improvements has been more challenging than initially expected.

New pay and earnings questions for proxy responders aim to improve coverage of earnings data for individuals who cannot respond directly, while data rotation has been highlighted by stakeholders and methods experts as an important improvement for enhancing data quality. It aims to improve response and data quality across the longitudinal waves of the Core Survey. By rotating previously collected data between waves, respondents are either not asked questions they have previously answered that will not change between waves, or are presented with their previously submitted information to check that it still applies.

While progress has been made on testing of both data rotation and pay and earnings questions, the original estimate of January 2026 for implementation of these improvements will not allow sufficient time for testing. Therefore, we are now aiming for April 2026 to align with the April to June quarter of data collection.

Ensuring that data on complex variables, SIC and SOC, are collected to a sufficient quality continues to be a priority for the TLFS Programme. Progress has been made but further time is needed to iterate testing and development of innovative solutions for these variables.

We are developing a search-as-you-type question and a respondent-friendly activity frame and are testing these with survey participants. These have shown potential, but the tests have identified that more refinement is needed before implementing them on the TLFS.

Alongside this, we are continuing to evolve both our longer-term survey assist (AI) approach and the post-collection processing solution. For the post-collection processing tool, we are assessing whether wider survey information (such as SOC) can help classify responses that we are not currently able to code. We will also explore probabilistic and heuristic matching to further improve coding. We are working on implementation plans for solutions and we will provide more detail on these in the next labour market transformation article in early 2026.

Methods development

Alongside the design changes, a suite of methods is being developed to maximise data quality. This is an iterative process as the methods will need to be tested for suitability and potentially revised once all the major survey design changes, such as data rotation, are implemented. The updated timeline for the introduction of data rotation therefore has an impact on the timeline for the methods development.

An important priority for the programme has been to develop a strategy for dealing with missingness in the TLFS data. As is expected with any self-completion survey, a small proportion of respondents skip questions or drop out before reaching the end of the survey, leading to partial responses. This drop out is strongly linked to the number of questions respondents are asked.

On TLFS, this means those from economically active groups are more likely to drop out and may be underrepresented in later questions. There are established methods for dealing with missingness that mitigate the risk of bias in responses. These include imputation and weighting.

Imputation uses all available data to fill in missing responses by borrowing a value from a statistically similar respondent, based on robust matching criteria. We can also use weights to deal with missing data; by only using full responses at different points in the survey, we can effectively scale up these responses to account for the missing ones.

The missingness strategy recommends using a combination of imputation and weighting and has been developed in collaboration with users and methods experts. This approach is regarded as the optimal method for maximising the quality of TLFS data. While introducing more methods is necessary to meet our quality standards, the approach is complex and will require careful consideration, and time, to develop and implement effectively.

A crossover survey to understand and interpret differences in response between the TLFS and the Labour Force Survey (LFS) has launched, with around 6,000 addresses that have previously completed the LFS being asked to complete the Core Survey during October and November 2025. This research will help us to understand and interpret potential differences between the two surveys. We will report on the findings as part of this regular series of labour market transformation articles.

Plus Survey and household, socioeconomic and local data

In October 2025, the quarterly sample size for the TLFS Plus Survey – the separate, cross-sectional survey to meet needs for additional labour market, and wider household, socioeconomic and local data – increased from 50,000 to 90,000 households and now matches the Core Survey sample size.

Significant activity is under way in collaboration with users to ensure both Core and Plus Surveys are meeting the need for additional labour market, and wider household, socioeconomic and local (HSL) data.

In parallel, we have begun our analysis of data collected on the Plus Survey. Our work in this area aims to assess the quality of the data collected and support the research and development of survey content. This includes redesigning survey topics collected on the LFS so that they are suitable for the online mode of the Plus Survey.

Early conversations with stakeholders have emphasised the importance of quality and achieved sample size for the Plus Survey. We have been capturing measures of success as part of our stakeholder engagement, which will inform our ongoing work to develop acceptance criteria for transition to the TLFS for HSL statistics.

Transition timeline

As previously announced, the timing of transition to TLFS will be evidence-led and decisions will be taken in collaboration with our main users at a series of readiness assessments starting in July 2026 and if needed, continuing throughout 2027.

To support this decision-making, measures of success (a comprehensive set of quality criteria covering different aspects of survey performance, including achieved sample size, precision, stability and timeliness) have been developed in collaboration with stakeholders. The measures of success refer to the minimum acceptable level for transition, and there will be further work required after transition to continue to mature the other success measures. They will be tested further over the coming months to ensure they can provide sufficient evidence to support effective decision making as we conduct the data quality assessment and make decisions following the first readiness assessment in July 2026.

We aim to transition to the TLFS for our published headline labour market statistics in November 2026 if we and our users are ready. If our (or users’) assessment of quality requires more data to be collected and assessed, the transition may extend into 2027. 

Additional complexities in the implementation of the remaining design changes and methods experienced since our last update in July 2025 may affect the decision as to whether more data will be required. Moving the implementation of data rotation to April means that by July 2026, only one calendar quarter (three-month period) of data reflecting the change will have been collected, allowing for only limited quality assessment of its impact. We will keep the impact of this under review with main users and provide further updates as part of this regular series of labour market transformation articles. 

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 (NISRA)’s Labour Market Survey (LMS). The Office for National Statistics (ONS) will then combine the LMS data with the corresponding data from the Great Britain LFS (and, in time, the TLFS), which will be seasonally adjusted to enable the production of NI and UK results.

The ONS and NISRA are working through potential solutions to overcome challenges to NI and UK production timelines and maintain aggregate UK labour market statistics as far as possible.

The latest updates on NISRA’s plans can be found on their Transformation of Labour Market Statistics in Northern Ireland webpage.

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 transition to the TLFS.

Quality

As set out in our previous articles, we have taken several actions to address quality concerns with the LFS. These include reinstating the sample boost, returning to face-to-face interviewing, increasing incentives and the ongoing recruitment of additional interviewers.

Following discussion with our main users, a further update on LFS quality was published alongside the labour market statistics release on 16 September 2025, as two full calendar quarters (three-month periods) of LFS data, which include the main interventions made to address quality concerns in all five waves, became available.

The main findings were:

• response levels and rates have shown clear improvement because of these interventions, with Wave 1 response levels now very close to their pre-coronavirus (COVID-19) pandemic level, although response levels for Waves 2 to 5 are still below their pre- coronavirus level
• the composition of respondents according to different characteristics has also shown some improvement since late 2023; however, representativeness is lower than before the pandemic and has not improved for all characteristics

We continue to closely monitor LFS data quality and update users. We have increased interviewer capacity for Waves 2 to 5 to further improve response levels to their pre-coronavirus levels, and a full reweighting of the LFS and Annual Population Survey planned for 2026 will further improve the coherence of our estimates.

The next LFS quality article is planned to follow the workforce jobs and labour market releases in December 2025 and will include more on coherence of employment data and further analysis on the quality of the LFS.

The improvements to the quality of LFS data continue to be welcomed by users, including the Stakeholder Advisory Panel on Labour Market Statistics, as they increase confidence that the LFS can be fit for purpose until transition takes place.

This article has primarily focused on the Labour Force Survey (LFS) and Transformed Labour Force Survey (TLFS), however, significant work is ongoing to improve our wider portfolio of labour market statistics. These are the main updates.

Increasing the role of administrative data

We continue to progress our work on administrative data, including linking HM Revenue and Customs (HMRC)’s Pay As You Earn (PAYE) Real Time Information (RTI) to TLFS and LFS with a view to improving our understanding of possible bias in our surveys.

Since our last update, we have strengthened our internal governance around this work, bringing it into a formal project and have progressed a proof-of-concept analysis using the 2021 Census. Findings from this analysis will inform our next steps where the priority is on extending the analysis beyond the census period in 2021 to understand how our analysis varies over time. In parallel, work is under way to assess linkage quality between the LFS and Demographic Index, an essential component in interpreting the findings of any wider analysis of linked data.

Alongside this, we have continued to progress our work on a linked employer-employee dataset (LEED). A LEED provides information on the interaction between employers and employees in the labour market, including over time. Since the last update, work on LEED has focused on developing a LEED prototype, linking our Longitudinal Business Database to HMRC’s PAYE RTI data.

We also continue to prioritise our areas of focus on administrative labour market data with our main stakeholders through a range of external engagement, including the Stakeholder Advisory Panel on Labour Market Statistics.

Annual Survey of Hours and Earnings

Significant work is under way to modernise and improve our Annual Survey of Hours and Earnings (ASHE) statistics, given their importance in informing critical decisions. We recently published a blog on how the ONS is improving its annual earnings survey, which outlined the main areas of progress and future plans, including the modernisation of the ASHE production systems.

As a first step, we have introduced a new tool, which uses artificial intelligence to code occupations, with further process improvements expected for ASHE 2026 and 2027.

Over the last year we have also launched our new quarterly Earnings User Group, attended by around 20 different government departments and other public sector bodies, as well as co-hosting an Earnings Symposium to engage with wider users. We are committed to continue working closely with main users and experts as we develop and implement any changes.

Economic Statistics Plan and Survey Improvement and Enhancement Plan

Work to advance a programme of improvements to our labour market statistics continues to be our highest priority. The importance of, and our commitment to, this work was highlighted in our plans for Economic Statistics and Survey Improvement and Enhancement, published in June 2025. An update on progress with both plans will be published on 4 December 2025.

Annual Population Survey

In order to support the delivery of our Economic Statistics and Survey Improvement and Enhancement and Improvement plans, we need to narrow the scope of our statistical portfolio. We published details of this in a letter from the ONS Permanent Secretary to Penny Young, Deputy Chair of the UK Statistics Authority on 12 November 2025, including a review of the Labour Force Survey (LFS) boost, which enables the Annual Population Survey (APS).

ONS Economic Forum

The latest updates on labour market transformation progress and plans outlined in this article will be presented at the next virtual Office for National Statistics (ONS) Economic Forum on 24 November 2025 alongside a presentation on the Annual Survey of Hours and Earnings development plans.

These forums are an opportunity for users to hear directly from the ONS on the latest economic and social statistics developments and ask questions. Further information on the Economic Forums, including how to register, is available.

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, including:

• assessing the Transformed Labour Force Survey (TLFS) data against main quality criteria, comparing estimates with the current Labour Force Survey (LFS) and administrative sources such as HM Revenue and Customs (HMRC) Pay As You Earn (PAYE) Real Time Information (RTI) data; this comparison helps us to understand how TLFS responses align with other labour market measures and identify potential improvements
• formally establishing the Household, Socioeconomic and Local Project within TLFS, ensuring requirements beyond labour market statistics are captured and delivered
• setting up a new project to explore data linkage opportunities, initially focusing on linking household surveys with census and PAYE RTI data to investigate potential sources of bias
• external peer review and working groups continue to support the reliability and accuracy of TLFS data, including through targeted deep dives into specific methods and estimates