Gross domestic product (GDP) is one of the most widely used economic indicators, informing decisions by policymakers, businesses and the public. One of the Office for National Statistics (ONS)’s monthly publications, GDP usually leads the broadcast news agenda and generates significant media coverage. It also gives citizens and business a view of the UK’s overall economic picture.
However, GDP can be difficult to interpret due to its complexity, the frequent revisions to underlying data sources and the inherent uncertainty in economic measurement. Revisions to GDP are inevitable, as statistics producers must respond to policymakers’ need for timely estimates on the economy. The challenge for producers, therefore, is how to best communicate this uncertainty, given that the process of producing timely estimates of GDP is not always understood.
Effectively communicating GDP statistics is essential to ensure that users can accurately understand and use the data, but it presents several challenges:
- Complexity: GDP is compiled using three different approaches and hundreds of data sources, which are reconciled to produce a single central estimate. This means that traditional approaches to measuring and communicating uncertainty, such as confidence intervals, cannot be used.
- Timeliness vs. Accuracy: In the UK, initial GDP estimates are produced approximately 40 days after the end of the quarter. However, producing estimates this quickly means that many data sources are not yet available, as it usually takes two to three years before all the information is available and a fully reconciled estimate is possible. Of course, waiting this long for estimates would reduce the usefulness and value of the statistics, particularly for policymakers, who need a much earlier view when assessing shocks to the economy (for example during COVID-19), setting taxes, government spending and interest rates.
- Mixing up revisions and errors: When people hear the word “revisions”, they often assume that these are changes made to correct errors, rather than changes due to factors such as the availability of improved data or the introduction of new and improved methodologies. This misunderstanding leads to misplaced scepticism about economic data. Communicating this distinction therefore is therefore crucial to conveying the inherent uncertainty in early estimates without undermining trust in economic statistics.
By effectively communicating uncertainty, statistical producers can enhance trust and credibility. And using improved methods of presenting uncertainty can support the use of the data and in turn lead to better-informed decision-making.
OSR reviews of GDP revisions
In September 2023, ONS invited OSR to review its practices around revisions to estimates of GDP, resulting in several requirements setting out how ONS can further enhance its communication of uncertainty. ONS continues to work on these recommendations.
The approach that ONS has used to communicate uncertainty in GDP estimates can be applied to other statistical domains to improve the overall communication of uncertainty in public data.
UK gross domestic product (GDP) is estimated to have fallen by 0.3% in Quarter 4 (Oct to Dec) 2023, following an unrevised fall of 0.1% in the previous quarter.
GDP monthly estimate, UK: January 2025
Historically, the absolute average revision between the initial quarterly GDP estimate and the estimate three years later is 0.2 percentage points, when more detailed information is available through the comprehensive annual supply and use balancing process.
GDP first quarterly estimate, UK: October to December 2023
This release includes revisions to Quarter 1 (Jan to Mar) to Quarter 3 (July to Sept) 2024; growth in real GDP in Quarter 1 2024 has been revised up 0.1 percentage points, while growths in Quarter 2 (Apr to June) and Quarter 3 2024 are unrevised.
GDP first quarterly estimate, UK: October to December 2024
GDP estimates for Quarter 1 2020 and Quarter 2 2020 are subject to more uncertainty than usual as a result of the challenges we faced in collecting the data under government imposed public health restrictions.
GDP first quarterly estimate, UK: April to June 2020
The mean absolute revision (MAR) to quarterly GDP from Quarter 1 (Jan to Mar) 2020 to Quarter 2 (Apr to June) 2024 in Blue Book 2024 was slightly smaller than in recent Blue Books.
There is some evidence that revisions are marginally statistically significant when looking at the final quarterly estimate of GDP across the entire time span from Quarter 2 1961 to Quarter 4 (Oct to Dec) 2021 compared with first estimates.
GDP Revisions in Blue Book: 2024
- Real-time databases record and track the history of data releases and revisions over time. They allow users to analyse how initial estimates are revised.
- Headline revision triangles are a helpful visual tool showing how estimates of a particular economic indicator change over successive publications.
- Component revisions triangles allow users to understand which components are revised more or less than others.
Data related to GDP first quarterly estimate, UK – Office for National Statistics
Impact
ONS told us that its approach has helped economic commentators and policymakers to make more-informed judgements on how and when they use GDP data because they now better understand the revisions process and the revisions properties of the data.
The impact goes beyond the outputs, with the production team more focussed on uncertainty in their internal ‘curiosity’ sessions on data quality.
There were initially some reservations from ONS that having the uncertainty narrative front and centre may impact confidence and trust in the headline numbers – but neither OSR or ONS has seen any evidence that this has been the case.
Related work
The Economic Statistics Centre of Excellence (ESCOE) has conducted a range of research on approaches for communicating uncertainty and how it impacts trust. This work found that:
The decision to communicate uncertainty information to the public matters. The way that this is communicated is also important. For instance, we show that communicating uncertainty information alongside the GDP point estimate or alongside estimates comparing productivity between the UK and other G7 countries improves public understanding of data uncertainty and does not reduce trust in the statistical office. Overall, the results call for greater communication of data uncertainty.