Monthly GDP, first introduced in 2018, is a closely followed statistic, showing changes in the size of the UK economy around 40 days after the end of a month. However, it has been questioned whether focusing on such short-term movements risks placing too much weight on a relatively noisy signal about the state of the economy. Liz McKeown sets out how, to help address this, we are changing our presentation of GDP to lead with changes across three-month periods.
When we first launched monthly GDP statistics in 2018, we explained that our lead indicator would be the three-month on three-month GDP movements, with the month-on-month picture also provided in the bulletin for completeness. This reflected that there can be volatility from month to month in the GDP figures and that the three-month on three-month figure smoothed out some of that volatility and presented a cleaner picture of the state of the economy.
However, with the onset of the Covid-19 pandemic – followed by significant events such as the war in Ukraine – our focus moved to highlighting the monthly figures to give policy makers and commentators the earliest read on how those events affected monthly output.
We’ve written in previous blogs about our ongoing efforts to continue improving how we communicate GDP . As part of that we have revisited the original analysis that underpinned our initial decisions on presentation of the monthly data and listened to the views of users.
As a result, from today, we will be reverting to leading on the three-month on three-month picture in our Monthly GDP bulletin and associated commentary. To further support the interpretation of trends within GDP we are also introducing a series of new charts into the bulletin that give more insights into the trends over time for GDP overall and for each of its components.
Reversion to leading on the three-month on three-month measures also means there is alignment in the headline measure for GDP growth in the monthly releases and those in our quarterly releases. This will make it easier for users to interpret those releases.
We have already begun to use this approach with last week’s retail sales release focusing much more on three-monthly changes and today’s bulletins on the Index of Services, Production and Construction also following suit. We intend to, in discussion with users, review whether it is appropriate for any other monthly publications to also adopt this approach.
In moving back to leading on the three-month on three-month picture we want to openly acknowledge that while such an approach brings considerable benefits there are some shortcomings with three-month on three-month measures that users should be aware of. In particular, they can be slow to pick up turning points, while they can also still be distorted by big, one-off, events, such as extra bank holidays. It is also worth noting that our bulletin will continue to provide coverage of the month-on-month picture including drawing out qualitatively – when they arise – themes from business returns from that month that cut across different industries.
In his blog last week, the new Director General James Benford highlighted how the ONS is focusing further on core economic statistics. In that context it is important that we reflect on the value of monthly GDP as a core economic output and the resource required to produce it. Monthly GDP corresponds to a weighted average of sub-indicators already produced and published (e.g. index of production, services and construction) – so there is limited additional work required to produce monthly GDP and the work that is undertaken supports the production and quality assurance of quarterly GDP. The monthly GDP estimates add up to give the quarterly GDP value so publishing these ahead of having data for the full quarter helps users by providing insight on the emerging outcome of the quarter.
The monthly data itself also continues to provide timely insights not readily available in the quarterly data. We saw that earlier this year where we could identify in the monthly data that some economic activity had been brought forward to February and March to avoid stamp duty and tariff changes. Where there is important insight within the month-on-month data we will emphasise it more. Where we do, we will explain the rationale for this reporting, such as in the case of any future energy price shocks or additional holidays.
We will also continue to provide users with information about the data content of the statistics in our monthly bulletin. Each month when we first publish our monthly GDP estimate, we have over 35,000 returns from businesses based on our comprehensive monthly business surveys and make wider use of administrative sources such as energy statistics from the Department for Energy Security and Net Zero and health statistics from NHS England. Our bulletins provide information on the response rates achieved in the survey data and highlight where users can go to find more information about how early estimates of GDP are revised.
We hope the changes I’ve highlighted in this blog will help improve the understanding and aid interpretation of changes in the UK economy each month. We welcome any feedback on them – please get in contact at gdp@ons.gov.uk
Liz McKeown is Director of Economic Statistics at the Office for National Statistics.