UK statistics agency to offer bigger cash rewards to tackle data flaws

Unlock the Editor’s Digest for free

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

The UK’s statistics agency plans to offer bigger cash rewards for filling out questionnaires as it rushes to fill holes in official figures on the labour market and pursue broader reforms of data collection.  

Evidence that £20 payments led to a much higher response rate when the Office for National Statistics ran its labour force survey in Northern Ireland has led officials to look at offering more generous financial inducements across the UK as they seek ways to boost responses and resume the publication of essential jobs data.

“We are quickly looking at other incentives we can deploy in the most cost effective way,” said Darren Morgan, ONS director for economic statistics production. “You need a sweet spot for effectiveness and value for money . . . We haven’t settled on an amount,” he added.

The ONS has come under intense scrutiny after a series of embarrassing revisions and qualifications of core data.

In September, it published revised GDP figures showing the UK’s recovery from the pandemic had been much stronger than previously thought — painting a radically different picture of how some sectors had performed. Bank of England officials have cast doubt on its estimates of wage growth, which are used by ministers to uprate pensions and other benefits.

But the most serious problem is the ONS’s current inability to publish basic information on the state of the labour market. Last month, it gave only bare-bones estimates of unemployment and employment because the survey on which its usual detailed figures are based looked too unreliable to use.

The ONS’s handling of these glitches has raised concerns among senior officials about whether the agency is transparent enough in communicating uncertainties around its data. It also suggests strains on capacity as the ONS pursues ambitious plans to make greater use of digital data and government records, alongside traditional survey-based methods.

“I doubt that there is a systemic issue but I am nevertheless a little concerned,” said Sir Charlie Bean, a former Bank of England deputy governor who led a review of UK economic statistics in 2016.

He gave the ONS credit for a “huge improvement in culture and effectiveness” since then, making it “less defensive, more self-critical and more outward-looking” and turning it from an “also ran” into a statistics agency recognised as an international leader.

But Bean said “somewhat cloth-eared” communication — in particular, the failure to realise how politically sensitive the GDP revisions would be — came on top of other worries, including an apparent lapse in high-level stakeholder meetings and some complaints of researchers finding it harder to access ONS micro data. The combination “makes one worried that they have either taken their eye off the ball or else don’t have the resources necessary to deliver everything that is expected of them”, he said.

On Tuesday, when the next interim estimates on employment and unemployment are due, the independent Office for Statistics Regulation will publish a review of what went wrong with the labour force survey, along with recommendations to the ONS on how to fix it.

It is likely to focus on how the agency should communicate uncertainty around its data, in particular when headline figures are broadly reliable, but breakdowns by local area, age group or other groupings are not robust enough for the uses researchers and policymakers want to put them to.

It will also examine whether the ONS allocated enough resource to keeping the labour force survey in operation, while it devoted efforts to launching a “transformed” survey that is running in parallel, and will replace it from next spring.

Morgan said running the two surveys in tandem had strained the team concerned, in the context of a rapid expansion of the ONS’s output and constraints on funding.

The latest spending review committed the UK Statistics Authority to efficiencies that included savings worth 10 per cent of its baseline budget in 2024-25.  

The overhaul of the LFS could therefore become a testing ground for the ONS’s ability to keep its existing surveys running smoothly, while shifting its focus to new methods of data collection.

The pandemic accelerated this shift: when lockdowns made it impossible for the ONS to knock on doors, the agency began publishing a much bigger range of experimental indicators based on real-time, digital data and drawing on administrative data held by other parts of government.

Tax records, redundancy notifications and jobless benefits claims are at least partly filling the current gaps in labour market data.

The ONS now tracks online job ads, spending on debit and credit cards and traffic camera activity, among other signals of economic activity. From next year, it will start incorporating data from online second hand car sales and supermarket scanners into its estimates of consumer price inflation. Next month, it will begin publishing much more granular figures on rental prices — split by property type and size for each local area.

None of this will remove the need for resource-intensive surveys, which remain the only way to answer some key questions — including the extent of self-employment and of precarious work, and the reasons why some people are outside the workforce.   

Meanwhile, a separate review under way, as part of broader scrutiny of all public bodies, will report to the Cabinet Office by the end of this year on the quality of the ONS’s output, its governance arrangements and ability to deliver efficiency plans.

The ONS itself has a further fix in mind for the LFS, given growing public apathy for surveys — to make the new version shorter. “It’s a treasure trove of information,” Morgan said, but added: “We’ve cut out some non-essentials. It’s a useful time to take stock.”