Update on the Office for Statistic Regulation's review into GDP

In November 2023, the Office for Statistics Regulation (OSR) published a report on Revisions of estimates of UK gross domestic product (GDP). The report outlined eight requirements for improvement in the way we produce and communicate estimates of GDP.

Requirements 1 to 4 focused on how we at the Office for National Statistics (ONS) should communicate uncertainty in GDP estimates, and how we should help users to understand revisions. Requirements 5 to 8 focused on improving the data content for the three measures of GDP in the short-term (monthly and quarterly) estimates.

We implemented a number of improvements to how we communicate GDP estimates, which are detailed in our update to the OSR on GDP estimates, published in January 2024. In October 2024, we published information on the areas of focus for improving GDP data quality, and how they aligned to the three approaches to measuring GDP, in our GDP revisions in Blue Book: 2024 article. Although many of these improvements will take time to research, test and analyse, we have committed to them as part of our Plan for economic statistics. 

We provided an update on requirements 1 to 4 in the 31 October 2025 GDP revisions in Blue Book 2025 article. This article provides a full update on the remaining requirements looking at the output, expenditure and income approaches to measuring GDP.

We previously recognised two opportunities for using HMRC Value Added Tax (VAT) administrative data to improve our approach to measuring short-term estimates of the output approach to measuring GDP (GDP(O)). Both of these work strands are recognised in the Plan for Economic Statistics, section C2.3.

The first opportunity is around expanding the current use of VAT for producing GDP estimates, either through producing a timelier, more robust iteration of VAT turnover estimates or through broadening the range of industries for which we use VAT turnover data. The second is about building on the findings from a 2021 research paper on using VAT expenditure data to estimate intermediate consumption (PDF 1,637KB). Preliminary work on both of these strands has begun, including on building capability in data linking, setting up a secure platform to conduct the analysis, and gathering insights and lessons learned from previous work completed.

For the measurement of intermediate consumption in short-term GDP(O) estimates, we view VAT as the most promising option and where we will see the biggest gain. Using other data sources to measure intermediate consumption will however remain on our backlog of things to explore in future.

To improve our early GDP(O) estimates, we have also implemented a continuing programme of targeted improvements. In 2025, we have focused our attention on deflators. We delivered improvements to the construction repair and maintenance deflator, which were implemented in Blue Book 2025, with further improvements being considered for Blue Book 2026. Service Producer Price Indices (SPPIs) are used as deflators for many of the industries within the services sector. SPPIs are measured quarterly and splined to months for use in our monthly GDP estimates. We identified and tested an improved method for forecasting the next quarterly data point until it becomes available, which was implemented in spring 2025.

From September 2025, we have commenced a rolling series of industry deep dives. These focus on a broader set of data feeding into GDP(O) estimates, including survey data, administrative sources, deflators, our methods, and any external economic evidence we can gather. These reviews allow us to assess the quality, representativeness and reliability of the data being used to produce our GDP(O) estimates, resulting in areas where we can make quick and effective improvements and longer-term recommendations.

Our initial area of focus for the expenditure approach to measuring GDP is about exploring the potential for using non-survey data sources to either directly measure or supplement the current measurement of household final consumption expenditure (HHFCE), which is the largest component of the expenditure measure. This has been recognised in Section C2.2 of our Plan for Economic Statistics. 

From spring 2025, we have started using the UK consumer card spending data to quality assure HHFCE estimates for some areas of expenditure. We are using the data as a tool to help us understand consumer behaviour, comparing it with our traditional sources to help us better inform our estimates of what is happening in the real world. We are continually looking to expand the range of areas where we utilise these data for quality assurance purposes, while monitoring the comparability, coherence and reliability of the card data. 

As these sources are still relatively new to the Office for National Statistics (ONS), we are using a tactical approach, using them in conjunction with our traditional survey sources to help us better understand consumer behaviour. Our longer-term aim remains to see whether these administrative data sources can be used to inform the level of household expenditure, rather than only the growth rates in the latest periods. However, as we have set out in Section C2.2 of our Plan for Economic Statistics, this will be a phased approach as we need to better understand the methodological challenges, such as:

• mapping the data to the relevant classifications

• analysing and adjusting for market share

• assessing the reliability and strength of the data

We are also exploring a range of other alternative data sources, and we will conduct feasibility studies on the potential for using these as they become available to us. 

For example, in summer 2025 we carried out a feasibility study into whether we can utilise data being used for the transformation of the Consumer Price Index (CPI) to help us better measure HHFCE. These alternative data sources include rail fare data and secondhand car data. The feasibility study has shown the potential for using these data sources to support the measurement of HHFCE, and we are scoping the next steps for this research with the aim to introduce some of these alternative sources into our HHFCE processes in 2026.

Recognised within Section C2.1 of our Plan for Economic Statistics are two strands of work aimed at improving the income approach to measuring GDP.

We previously noted that large scale transformation to the Office for National Statistics (ONS) Labour Force Survey, along with the realisation of benefits from access to the HM Revenue and Customs (HMRC) Pay As You Earn (PAYE) Real Time Information (RTI) data will lead to improvements of earnings and employment estimates used within the Compensation of Employees. Through 2025, the ONS has secured access to the HMRC PAYE RTI data feed and we have begun comparative analysis with our existing compensation of employees (CoE) data sources. 

Additionally, within the income approach, we are seeking opportunities to improve early estimates of private non-financial corporations (PNFC) gross operating surplus (GOS). In autumn 2025, we began a collaboration with ONS partners in Durham University, who are supporting our research into finding more timely data sources to measure PNFC GOS.

The Office for National Statistics (ONS) has set a timeline for delivery of most of these improvements in our Plan for Economic Statistics, and we have committed to providing our users with regular updates on the progress being made against each of the milestones set out in the plan. Other items set out in this section include rolling reviews, and we will communicate any improvements identified through these in our regular statistical bulletins and articles.

We rely on other government departments and private organisations sharing their data with us to move this work forward. We will work closely with them to secure the timely and sustainable deliveries that are vital to make this a success.

The improvements set out in this article will form the basis for feasibility work into whether it is possible to create fully supply-use balanced measures of GDP on a quarterly basis.