Boris Johnson has repeatedly made “incorrect” claims about the level of child poverty under the Tories, the country’s statistics watchdog has found.

The UK Statistics Authority’s Office for Statistics Regulation (OSR) has concluded that the prime minister’s main suggestion – that the number of children in poverty had fallen by 400,000 since 2010 – was one of several inaccurate statements on the issue since he was elected in December.

It has now written to No.10 to draw attention to the correct statistics, advising the PM’s briefing team that they should pay regard to its conclusions.

The OSR was responding privately to a complaint from the End Child Poverty Coalition over what it called three separate “misleading” statements by Johnson in recent months.

The watchdog agreed that the PM’s remarks to the Andrew Marr Show on December 1, 2019, that there “are 400,000 fewer children in poverty than there were in 2010” was incorrect.

It also found that his line in Prime Minister’s Question Time last month – that “absolute poverty and relative poverty have both declined under this government” and “there are hundreds of thousands, I think 400,000, fewer families living in poverty now than there were in 2010” – was similarly untrue.

The following week, when Johnson was challenged by Keir Starmer to correct the record, he made a further incorrect statement.

The PM’s new claim – that “there are 100,000 fewer children in absolute poverty and 500,000 children falling below thresholds of low income and material deprivation” – was also wrong, the OSR said.

The watchdog found that the number in absolute poverty had fallen by 100,000, not 400,000, since 2010. And the number below the poverty line is not 500,000 – it is 1.5m.

In addition, since 2010/11 the number of children in relative poverty, by the DWP’s own measure, has actually gone up by 500,000.

Relative poverty refers to households with income below 60% of the UK’s median household income, while absolute poverty means people with inflation-adjusted income below 60% of the median income in a particular base year, in this case 2010/11.

The OSR has blogged before on how poverty is measured in different ways.

Shadow education secretary Kate Green said: “It is shameful that the prime minister is unable to tell the truth about the hardship faced by so many families struggling to make ends meet.

“Children and families in such difficult circumstances deserve better than this shabby treatment from an out-of-touch prime minister who has repeatedly failed to be honest about the challenges they face.

“The prime minister must now correct the record, both publicly and in Parliament, and ensure that when he next raises his government’s damning record on child poverty, he comes clean about what the stats are saying.”

No.10 pointed to a written parliamentary answer last month where the PM appeared to change his definition once more.

He said at the time: “Children living in workless households are around four times more likely to be in absolute poverty (after housing costs) than those where all adults work.

“As of December, the number of workless households has fallen by 1 million since 2010, meaning there are over 740,000 fewer children living in a household where no one works.”

Anna Feuchtwang, chair of the coalition, had written to complain to the watchdog that it could not be right that such figures are used “selectively”.

“While it is expected – and right – that child poverty should be the subject of robust political debate,” she wrote, “it cannot be right that official figures on something as fundamental as how many children are in poverty continue to be used selectively, inaccurately and, ultimately, misleadingly.”

Responding to her letter, OSR director general for regulation Ed Humpherson replied: “Our team has investigated the statements which you highlight and has reached the same conclusion that these statements are incorrect.”

Feuchtwang said she welcomed the conclusion from the OSR that the prime minister had used child poverty statistics incorrectly and said the watchdog would not have made its judgement lightly given its unwillingness to referee political disputes.

But she criticised the OSR for not publishing its verdict and only putting it in a private letter.

“It is deeply insulting to the children and families swept into poverty when data about them is used selectively and misleadingly at the whim of politicians,” she said.

“The simple fact is that by any measures child poverty is rising but instead of tackling the problem the government risks obscuring the issue and misinforming the public.

“The lives of real people are at stake and we need consistent use of information and urgent action.”

Imran Hussain, director of policy and campaigns at Action for Children, said: “This isn’t about the Punch and Judy of PMQs. Admitting that rising numbers of ordinary families are struggling to keep their children clothed and well fed matters to good policy making.

“You can’t ‘level up’ the country if you’re sweeping under the carpet the big rises in [relative] child poverty clearly shown by the official figures. The longer we’re in denial about the scale of the problem, the harder it will be to fix it.”

Child Poverty Action Group chief executive Alison Garnham added: “The hard truth is that child poverty is growing in the UK but the Government is in denial on this – that has to shift.

“If we are to make progress, the problem must be confronted, not circumvented. If the will and the focus are there, a strategy can be agreed and action taken to prevent more children from being damaged by poverty.”

A spokesperson for the OSR told HuffPost UK: “Measuring poverty is complicated, and different measures tell different parts of the story. In this instance, the Office for Statistics Regulation published an article on its website which sets out the landscape of poverty statistics and suggests how they can be used. 

“We also wrote privately to the End Child Poverty Coalition, and spoke with officials in No.10 about how best to use poverty statistics. Where a response remains private, we nevertheless record it on our published issues log which is updated periodically and summarises anonymously all the issues brought to our attention.“