The Office for National Statistics (ONS), with its number-crunchers and crack-of-dawn data dumps, is an unlikely backdrop for turmoil.
But in recent months the Newport-based institution has come under repeated attack over questionable jobs figures, which experts have said leave policymakers“flying blind”. On top of this, questions are now being asked about survey data used to formulate its GDP readings, which were revised down for the second and third quarters on Monday.
The debate over jobs statistics comes at a crucial moment for the UK labour market, with unemployment rising and business lobby groups claiming Rachel Reeves’s budget tax policies will lead to job losses.
The Bank of England governor, Andrew Bailey, felt so concerned about the quality of the labour force survey (LFS) – the official measure of employment in the UK – that he raised it at the annual City shindig at Mansion House, calling it a “substantial problem”.
“I do struggle to explain when my fellow governors ask me why the British are particularly bad at this,” he said.
Response rates to the LFS, which involves a letter sent to households and followed up with a phone or in-person interview, were already falling before 2020, then plunged during the pandemic and remained low afterwards – hitting a nadir of 17.4% last year. The Living Costs and Food Survey – used in the GDP readings – receives similar response rates increasing the chance of GDP revisions, Bloomberg reported.
Economists are concerned that the labour market survey may be more likely to pick up people who are at home – potentially overestimating economic inactivity. The Resolution Foundation thinktank has suggested the LFS may have “lost” up to 930,000 workers.
That matters, because one of the key worries about the economy since the pandemic has been that the employment rate – the share of the population in a job – has never returned to 2019 levels, making the UK an international outlier.
The ONS has been building up a new index, the “transformed labour Market Survey”, to replace the LFS. But repeated twists and tweaks mean it may not be ready to launch until 2027.
The chair of the Treasury select committee, Meg Hillier, called it a “major blow”.
“These delays will make some of the most consequential decisions taken by the Treasury and Bank of England challenging at best and misinformed at worst,” she said.
“The LFS is essentially useless,” says Xiaowei Xu, senior research economist at the Institute for Fiscal Studies. “It’s a huge problem.”
Xu says the gap between the LFS and other data sources – such as payroll data – is particularly large for the 18-24 and 50-64 age groups, and raises the fear that sharp increases in economic inactivity for these age groups “could just be a data error”.
The ONS also admitted that it had reweighted its Northern Ireland data because of an error.
The organisation’s internal review of the drawn-out process of building the new labour market survey found it has had, “a profoundly negative impact on morale and wellbeing”.
Published earlier this month, with names redacted, the review found the process had been driven by arbitrary deadlines based on funding, and managers appeared unwilling to accept the timeline set was unrealistic.
“Many participants reported a reluctance to listen to expert advice and evidence about whether it would be feasible to achieve stated objectives,” the review said.
ONS insiders say the background to this crisis of confidence in the national number-cruncher is one that is familiar from other cash-strapped public services: cost-cutting, low pay and rock bottom morale.
after newsletter promotion
Its latest annual report showed that the ONS had implemented “efficiencies and cost savings”, of £17.8m over the previous 12 months.
On the frontline of the task of gathering information on citizens’ working lives are about 300 low-paid workers.
Phone interviewers entrusted with carrying out the LFS and other surveys are paid as little as £12.19 an hour. Field interviewers, who knock on participants’ doors, get £14.79 an hour if they work in central London, £12.50 outside the capital. Unions claim annual turnover among these frontline staff can be as high as 25%.
Back in Newport, the ONS is locked in a dispute with the PCS union, over back-to-the-office policies the union claims have been imposed unreasonably. Workers are expected to be at their desks two days a week, having previously understood they could work from home as much as they wanted.
ONS management reject this account, insisting staff were never told after the pandemic they would not need to return to the office.
Fran Heathcote, the general secretary of the PCS union, says “many of ONS management’s current problems are of their own making. If the ONS wants to salvage its reputation, then its leadership needs to listen to its own workers and their trade unions.”
However, the ONS’s defenders insist it is doing all it can to fix the faulty data. Jonathan Portes, a labour market expert at King’s College London, chairs the stakeholder advisory panel set up by the ONS to monitor its approach to the challenge.
He argues that the organisation should have realised at an earlier stage that falling response rates were a problem – but has been working hard to resolve the issue.
“Coming out of the pandemic, I have no doubt, and I think the ONS would accept with hindsight, that they didn’t realise how big the underlying structural problem was – and therefore they didn’t act quickly enough and put enough resources into it,” he says.
But he rejects the idea that policymakers are potentially being misled – because other indicators, including benefits data, are pointing in the same direction as the LFS.
“We’re still left with the point that there’s a lot more Britons on disability and sickness benefits than there used to be, and some of those at least are people who are now out of the labour market and weren’t before,” he says.
“I’m less worried about the exact size of the problem – which we don’t know – than I am about the fact that we still don’t really understand what’s happened and what is driving it, which means that solving it is very, very difficult.”