Policymakers have been “left in the dark,” by official jobs figures since the pandemic, which may have “lost” almost a million workers according to the thinktank Resolution Foundation.

In a report, the thinktank said the regular snapshot from the Office for National Statistics may have painted an “overly pessimistic” picture of the UK labour market since the pandemic.

Principal economist, Adam Corlett, says in the report that response rates to the key Labour Force Survey (LFS) have collapsed, from 39% in 2019 to just 13% last year.

There are concerns that workers may be less likely to respond to the survey than people who are economically inactive – potentially skewing the results.

“Official statistics have misrepresented what has happened in the UK labour market since the pandemic, and left policymakers in the dark by painting an overly pessimistic picture of our labour market,” said Corlett. “The ONS Labour Force Survey appears to have ‘lost’ almost a million workers over the past few years compared to better sources.”

Ministers are expected to publish a Back to Work white paper within weeks, aimed at helping people back into the workplace – including by improving the service provided by Jobcentres, and joining up work and health support.

The official jobs data has shown employment in the UK failing to recover to pre-Covid levels – a different pattern from other major economies. In particular, the number of people out of work because of health conditions has risen sharply.

Resolution has constructed an alternative assessment, using tax and population data. This tracks the official figures closely until 2020, but then diverges sharply. It suggests the ONS may be underestimating the number of people in jobs by as much as 930,000.

The analysis suggests that the working age employment rate may be back to the pre-pandemic level of 76%, instead of the 75% currently estimated by the ONS.

Resolution Foundation urged the ONS to act swiftly to reconcile the official figures with alternative estimates.

“The government faces a significant challenge in aiming to raise employment, even if the rate is higher than previously thought. But crafting good policy is made harder still if the UK does not have reliable employment statistics,” said Corlett.

The Bank of England governor, Andrew Bailey, reiterated his concerns about the labour market data on Tuesday, after raising the issue at last week’s Mansion House dinner.

“The sample is now so small that it is not a statistically representative sample, that’s the problem – so we don’t know how much faith we can put into it,” Bailey told MPs on the cross-party Treasury select committee.

“It’s an important issue: it’s very important, and it’s not just important for us, it’s important for other aspects of public policy as well.”

The Bank’s understanding of how much slack there is in the labour market is a key factor in setting interest rate policy.

A spokesperson for the ONS said it is aware that other measures of employment may be giving “a more accurate picture” than the LFS, but insisted it is working to improve the figures.

“Our ongoing work to improve the LFS estimates by increasing the sample, reintroducing face-to-face interviews, increasing incentives for those taking part and re-weighting the data using the latest population information will all help to improve the quality of the survey,” they said.

The spokesperson added that the ONS is cooperating with outside experts, to see if more action needs to be taken.