New blow to official UK data with fall in responses to GDP and inflation survey

Unlock the Editor’s Digest for free

Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.

Response rates to a survey underpinning estimates of UK GDP and inflation rates have fallen in a similar way to a decline in the jobs survey, prompting the Office for National Statistics to take emergency action in order to ensure the quality of its figures.

The statistics agency said it had put more interviewers into the field earlier this year to boost responses to its survey on living costs and food, in an effort to pre-empt data problems of the kind caused by the collapse of its labour force survey (LFS).

Problems with the LFS have left policymakers unable to gauge the true state of the jobs market — and the ONS says this crucial input for decisions on interest rates may not be fixed until 2027.

The living costs and food survey, which collects information on household spending patterns and living costs, is an essential input for economic measures.

The ONS uses it, alongside other data sources, to compile estimates of GDP and to weight the basket of goods and services underpinning its consumer price indices, in particular the retail price index. The latest GDP estimates will be published on Monday morning.

The survey is also the most comprehensive source of information for researchers and policymakers seeking to understand how the cost of living crisis has affected households at different income levels, and the effect of changes in taxes and benefits on living standards. 

But, as with the LFS, the response rate for the living costs and food survey has declined over time, sliding from 60 per cent at the start of the millennium to 40 per cent in 2019.

It fell abruptly when Covid lockdowns disrupted face-to-face interviews and hit a new record low of 22 per cent in the financial year to 2023.

The number of responses from “co-operating” households in 2022-23 was just 4,061, a sharp drop from the sample of more than 5,000 that underpinned results in the previous five years.

Adam Corlett, principal economist at the Resolution Foundation think-tank, said results based on the 2022-2023 data, published in August, had shown a drop in real-terms spending that looked “implausible” compared with national accounts data for the same period. 

Meanwhile, researchers are still waiting for microdata that would usually have been made available by the ONS in September, but has been delayed into next year by staff shortages. 

“The whole world changed in February 2022 —and we still don’t have data to see how that affected households,” said Peter Levell, senior research economist at the Institute for Fiscal Studies, adding that new electronic data sources were no substitute for the overarching view given by the survey. 

The ONS said it had begun boosting the sample for the survey in April, increasing the number of interviews conducted from about 900 in the first quarter of 2024 to roughly 1,240 in the third quarter. It also cross checks results against a range of other data. 

“Where the living costs and food survey feeds into headline household expenditure estimates, its findings are triangulated with other sources, such as business surveys and trade data, to create a robust picture of household spending patterns,” the agency said, adding that it would begin checking results against card-spending data early in 2025. 

Levell said the way the ONS used the LCF and other sources to construct the national accounts was “opaque” and that the lag in publishing the data was “increasingly an issue” because policymakers used the figures to assess the impact of tax changes on households. 

The LCF feeds directly into the annual weighting of the RPI. This no longer has the status of official statistics, but it is used to calculate some index-linked bond repayments and student loan repayments and to set annual increases in train fares and phone bills. 

The survey has always had a lower response rate than others the ONS conducts, because it is unusually burdensome. Respondents, including both adults and children, are asked to fill in a diary over a two-week period itemising their spending.

The Office for Statistics Regulation, which polices the quality of UK statistics, flagged “significant” concerns about the quality of the LCF in 2022 and urged the ONS to “invest time and resource” in fixing it.

It said in its assessment that a sample of 5,000 was too small for some users to draw “useful and robust conclusions”. The volatility of the data created “a risk of reputational damage” for the ONS because it could mean genuine errors were not identified, the regulator added, with “a large impact on the price indices which use the LCF data for their weighting”. 

The ONS took action at the time to address the most urgent problems, and the regulator said this gave enough assurance of the data’s quality. However, the ONS’s progress in “digitising” the diary process has been slow — a new tool to scan receipts automatically will be rolled out only from the end of next year. 

Funding constraints have also delayed a longer-term plan to merge the LCF with other surveys and create a single, streamlined source of data on income, spending and wealth.