National Statistical

Building momentum to deliver further improvements to the quality of our core statistics is the big focus for the Office for National Statistics as we enter the new year. ONS Director General of Surveys and Economic Statistics James Benford looks ahead to some of the planned changes that are due to be rolled out this year to improve the quality of key economic statistics.  

Last year was a pivotal one for the ONS. Under new senior leadership, we reprioritised urgently to focus our resources on core economic statistics as part of a wider strategic recovery plan.  We have been working in the open by publishing our plans for economic statistics and surveys, and delivering quarterly progress reports on them.  We have reset the way we lead, with a new leadership commitment to build trust between teams, encourage challenge and take decisive actions on priorities.  We have initiated a business planning process that will produce a multi-year plan for the Office that is realistic and sequenced to manage dependencies and deliver what matters most. 

We also made important improvements to our statistics. House price statistics are now less prone to revision and producer price statistics have been reinstated with improved weights. We have changed our presentation of monthly GDP statistics to focus on the less-volatile 3-month on 3-month change, and through last year’s Blue Book and Pink Book incorporated better methods for research and development statistics and to account for the UK output and trade of multinationals. We restored the Labour Force Survey (LFS) achieved sample sizes to similar levels to those seen pre-pandemic and made important steps to progress the Transformed Labour Force Survey (TLFS).  

Across this year, we will continue to build momentum and I want to share some of the key milestones for 2026 which will help shape the future of economic statistics and surveys.  

Improving measurement of consumer prices  

In March, we plan to introduce supermarket scanner data into our headline inflation statistics. For 50% of the groceries market, instead of collecting 25,000 prices per month directly from shops by price collectors, we will use approximately 300 million price points derived from sales of over a billion units of products per month, collected directly from supermarket scanners at the checkouts or online. This will give us a more comprehensive measure of grocery prices that will also be able to take into account broader factors affecting prices like store card discounting.   

Improving labour market statistics and linking data for better insights 

A major milestone in the third quarter of last year was restoring the LFS achieved sample sizes similar to levels seen prior to the pandemic.  We know, however, that questions remain about how variations in that sample size and the underlying response rates have affected the results from the survey. Early this year, we will publish initial findings from our work linking HM Revenue & Customs tax data with LFS responses. This project will help identify potential variation in non-response biases to the survey, informing our work to improve the robustness of our labour market statistics. Leveraging administrative data to enhance the insights from our survey data is a key step to mitigate the global changes we are seeing to how households are responding to surveys. We are also launching a new project, in collaboration with the Economic Statistics Centre of Excellence to explore how we can use a range of methods to combine at an aggregate level different labour market indicators.   

Preparing for TLFS transition 

Following the successful launch of a shorter core survey last summer and of assisted completion in the autumn – making participation quicker and easier for households – we continue to make important progress towards the transition to the TLFS.  We are working to introduce in April a ‘data rotation’ element that will mean households do not have to repeat answers they have given in earlier waves of the survey.  This will make it easier and faster to complete and improve data quality and so support response rates in the later waves of the survey.  In July 2026, we will conduct a readiness assessment with key users. If we and our users are ready, we aim to transition in November 2026, though this may extend into 2027 if they are not.  

Harnessing AI for smarter surveys` 

In 2025 we built ‘ClassifAI’ into the Annual Survey of Hours and Earnings Survey. ClassifAI automates the classification of employment by occupation and sector. This has saved hundreds of hours of work, allowing us to devote more resources to quality checks and clearing survey data faster.  Throughout 2026, we will continue rolling out AI technology across our survey operations to increase the time saved to 1000s of hours. We will deploy ClassifAI to the Business Register Employment Survey and are investigating its use to code products on other surveys. These innovations will help us deliver data that is not only more accurate but also more responsive to the needs of policymakers, businesses, and the public. 

Launching a new statistical business register 

We are developing the Statistical Business Register (SBR) to replace our existing Inter‑Departmental Business Register via a phased transition starting in the autumn. SBR broadens coverage of smaller business, including full Companies House data and additional administrative sources, increasing the sampling frame we use for our business surveys from 3 million to 6 million. It’s built on a modern platform and will underpin adoption of SIC2026, the official way the UK classifies companies, and future international standards in economic statistics. Key programme milestones include a maturity checkpoint by March 2026, SIC2026 incorporation by March 2027, and an aim for all surveys to use SBR frames by March 2028 ahead of implementing core elements of the internationally agreed System of National Accounts (SNA25) by 2030. 

Improving statistical processes 

More broadly we are working hard to modernise the end-to-end processes, all the way from data collation from surveys and wider sources to the processing and analysis of that data that leads to our statistics. As part of that we are introducing a new process excellence standard to streamline and de-risk our processes, including through automation using reproducible analytical pipelines. We are piloting that on the retail sales indicator, learning from the error on how we seasonally adjusted the data last year, and will look at how we roll it out more broadly as part of our business planning work.   

We are also working with external partners to look at processes across the system that feed into our statistics. An important focus for us this year will be work to strengthen system-wide processes underpinning our public sector finances release. We are working closely with HMRC who are implementing improvements to quality assurance processes following an error with VAT data in 2025. This year we will work with them on their reviews of HMRC statistics to minimise the risk of errors in their figures, which are inputs to core ONS outputs. The Office for Statistics Regulation will also provide an independent perspective on this. We have also formed a joint Local Government Financial Information task force with HM Treasury, the Ministry of Housing, Communities and Local Government, and the Office for Budget Responsibility to improve the quality and timeliness of expenditure data by local authorities that is also a key input to our public sector finance statistics. This began in late 2025 and we will report on progress through our regular economic statistic updates. 

Prioritising quality over quantity 

Underpinning all this, we will also be putting in place actions to narrow the portfolio of the ONS’s work. To create much needed space, we are well underway with a three-year business planning exercise and are listening to feedback from our users about where we should be focusing our efforts. 

One aim is to reduce the number of outputs that the ONS produces annually by around 10% this year. Alongside a new website, we will improve our publications, focusing on what audiences need most and delivering streamlined, clearer information. We have been engaging with various users on our prioritisation areas and will update on our decisions in the coming months. Business planning will also be critical to making sure our ambitions to improve our statistics are realistic, given available resource, and carefully sequenced. It won’t be possible to improve all areas of statistics at the same time, and we need to work through how we will deliver the new international standards for macroeconomic statistics alongside work for Census 2031. 

Working in the open 

Our approach throughout the year will be to make all these changes with the utmost transparency so that we can be held accountable and receive feedback to help us focus on what matters most for users.  We are also looking to expand the channels with which we engage with various users, following a very useful seminar with academics and professional economists hosted by the National Institute of Economic and Social Research before Christmas.  I would welcome suggestions on fora that we might attend to seek further feedback. 

Building trust, supporting our people and places 

To embed our commitment to build trust between teams and foster greater challenge, listening and decisive action, we are rolling out 360-degree feedback and performance review processes this year.  

We remain fully committed to our headquarters in Newport; our presence there dates back to our predecessor bodies in the 1970s and it has been at the heart of our organisation for two decades. At the same time, we are proud to employ talented colleagues across our other bases in London, Darlington, Manchester, and Hampshire. This distributed model helps us draw on a wide range of skills and perspectives, strengthening the work we do for the country. 

This will be a year of transformation and further progress. We still have a lot of work to do in 2026, but we will continue to improve the quality of surveys and statistics in 2025 and will build on that momentum. By embracing innovation, strengthening partnerships, and maintaining our commitment to quality, we will ensure that our statistics continue to serve the nation effectively – today and in the years to come.