Matt Goodwin came up with a number to scare people. It’s already wrong.

Before June last year, 2063 was just a year. It was the date the water companies promised our rivers would finally be clean. It was a line on a chart of NHS spending that nobody read. It turned up in an argument about South Africa, in a piece about Zimbabwean industry, and in one man’s timeline for the arrival of the machines. It was a long way off and nothing much lived there.

Then Matt Goodwin, failed Reform candidate and enthusiastic ChatGPT user, gave it a new meaning in a report he published through a centre that employs him.

The next morning the Telegraph put it up, and the number stopped belonging to its author.

I looked into exactly how he came up with that number. Georgina Sturge, a statistician formerly of the House of Commons Library, did some excellent work on this here. Matt uses a bunch of assumptions, and most of them are already wrong. I wanted to look into that, but also into how this date has been used and spread around, not only on X but in more mainstream media.

Every post below is real and every figure is X’s own. The chart stays with you as you scroll; it marks where each one sits in the run.

You can clearly see three peaks of popularity. The first was the initial tweet and the Telegraph top line. The second was a Reform and Restore spat. The third was Rupert Lowe. So with the date established in parts of the political landscape, how did our friend Matt get to this point?

A projection is not a finding. It is not a fact. It is a set of assumptions with arithmetic laid over the top, and it is worth exactly what those assumptions are worth. Matt’s paper is fifty-four pages and the important parts are in a table and an appendix at the back. Do they stand up to serious scrutiny? Again with thanks to Georgina, here they are.

On 21 May the ONS published the latest migration statistics and they had fallen again. Goodwin posted about it at 13:54 that afternoon, at some length, and the post was seen 123,273 times and liked 5,929 times. It is worth reading properly because it is the moment the argument stops being about the model at all.

Read those three lines again, because they are his, not mine. Arrivals falling from 1.3 million to 948,000 to 813,000. Non-European arrivals falling from 1.1 million to 766,000 to 627,000. Every series he quotes is going down, and he is presenting them as evidence that it is getting worse.

His actual argument in the post is about churn: that a lot of people arrive and a lot of Britons leave, and that half the Britons leaving are aged 16 to 34. That is a real thing and worth arguing about.

But it is not the argument the model makes. The model does not run on churn. It runs on net migration, calibrated to an official projection of 340,000 a year, and net migration is the exact figure he is telling his readers to disregard.

The same post also contains this, which I include because it tells you what the demography is actually for.

Matt knows exactly what he is doing here, and the goat herders line is one he has repeated since. It is deliberately putting a whole section of people in a box, and not a very nice box. Because he is not a very nice person.

The report counts people who ticked one box on a census form. That is not much of a rallying cry, so it gets swapped for something bigger. Here is what the 762 posts actually claim is disappearing, and how far each version travelled. Click any line for a real example.

The date travels just as badly. Three of the years in circulation are actually in the report, and they mean three different things: 2063 for white British people, 2079 for the foreign-born and their children, 2100 for the projected Muslim share. The posts also carry 2050, which is the report’s under-40s crossover promoted to the headline. And 2066, which belongs to an older projection by the Oxford demographer David Coleman. And then 2060, 2045 and 2030, which belong to nobody at all. All of this lets anyone using these “facts” say pretty much whatever they want. Make the date scarier, make the effect broader, throw in an “indigenous” here and there.

Back to that spike in February this year. What happened is that online accounts aligned with Rupert Lowe’s new party went and dug out a clip that was eight months old and started passing it round as a test of loyalty.

The clip is the one from June 2025. Steven Edginton, an investigative journalist and documentary maker then working for GB News, interviewed Richard Tice for the channel and put the new projection to him. Tice said it was a long way off and he would be long gone, and when Edginton kept at him on deportations and demography he told him he was “obsessed with this stuff”. GB News wrote up the exchange itself and posted the clip, which is where the two million views come from. Two days later Tice was back on the channel calling for a baby boom: “We need more British born people, and we need to be less reliant on immigration.”

That answer sat there quietly for eight months. Then Restore Britain needed a stick, and it turned out to be an extremely good one.

Worth a word on where that comes from. Basil the Great is an anonymous account, opened in November 2023, with 355,000 followers and a profile that reads, in its entirety, “Restore Britain”. It reports nothing. It is a conveyor belt of right-wing red meat: clips, screenshots, outrage, many times a day, submissions welcome by direct message. Nobody knows who runs it. On this story it consistently outperforms the politicians it promotes.

By June it was leaflet copy: Restore’s Makerfield leaflet offered voters a straight choice between agreeing with Tice that a minority white British Britain was fine because he would be “long gone”, and agreeing with Restore that it must be reversed. Lowe published the leaflet himself when a journalist asked him about it, and went a good deal further, promising mass deportations of “many millions”.

Worth noting what the quotation had become by then. Tice said “long gone”. Within two days it was being reported as “dead by then”. By February it had acquired a “So what” that he never said, and a question about white British people had turned into a question about “our people”. One account eventually had him saying he wouldn’t be dead by then, which is the opposite of the charge, and used it to make exactly the same charge.

A projection about census responses had become the loyalty test between two right-wing parties. Which is quite a journey for a table in an appendix.

The most-seen version of this claim anywhere does not belong to Goodwin, or Lowe, or any British politician. It belongs to Ian Miles Cheong, and it is worth knowing who that is.

Cheong is a Malaysian commentator based in the United Arab Emirates. He started out as a video games journalist, editing the site Gameranx, before moving into American and then British political commentary. He has 1.24 million followers and is among the accounts Elon Musk replies to and amplifies most often. He does not live in Britain and, as far as I can establish, never has.

His record with facts is a matter of record. In September 2020 he wrongly implicated a Black resident of Compton in the shooting of two Los Angeles sheriff’s deputies. More recently he posted an edited video claiming a United States congresswoman was running seminars coaching undocumented Somalis to evade deportation, which Musk then pushed to his own followers.

On 1 September last year he posted this. It was seen 5,596,117 times, which makes it the single most-seen thing in this entire story, larger than the Telegraph, larger than GB News, larger than anything the author of the report has ever posted.

There are no official projections. Asked directly in November 2024 whether it produces them, the Office for National Statistics replied: “We do not produce projections by ethnic group.” It explained why, too. The births and deaths data you would need are not broken down by ethnicity in the first place.

So what we have is one man’s model, published without peer review by a centre at which he is a visiting professor. The census tick box has been swapped for “native Britons”, which nobody counts. And the single word carrying the whole sentence, the word that turns a contested projection into a settled fact, is “official”.

It is also worth saying that the Leeds demographers who have done this properly, across three different scenarios, put white groups at around 70 per cent of the population in 2061. That projection has a fraction of the readership. It does not fit on a leaflet. It does not scare people, and it does not help populists offer simple solutions to complex problems.

Fifteen posts since publication, 2.07 million views between them. Read in order, you can watch the caveats fall off.

Goodwin was professor of politics at the University of Kent from 2015 until he took voluntary severance in July 2024. He is now a GB News presenter, honorary president of Reform UK’s student wing, and stood as Reform’s candidate in the Gorton and Denton by-election this January, where he came second with 28.7 per cent to the Green candidate’s 40.6. He publishes through a Substack and through the Centre for Heterodox Social Science at Buckingham, where he is a visiting professor. The 2063 paper was not peer reviewed.

He has form with confident numbers. In May 2017 he said Labour would not reach 38 per cent at the general election and offered to eat his own book if they did. They got 40, and he chewed a page of it live on Sky News, which I thought was a decent way to take it.

In March this year his self-published book Suicide of a Nation was examined by the journalist Andy Twelves and by Byline Times, who reported quotations attributed to Cicero, Friedrich Hayek, Roger Scruton and Christopher Lasch that cannot be found in anything those men wrote. Two of the book’s twelve footnotes carried ChatGPT addresses. Goodwin’s answer was that he had used AI only to obtain datasets, which he had “cross-checked with the original source”. He has not, as far as I can find, addressed the quotations themselves. The nickname that stuck was MattGPT.

I give this flavour and background not to be unkind about a man’s career but because it is relevant. This is the standard of evidence behind a date that now gets spouted by right-wingers all over the media.

None of which settles whether Britain’s population is changing, because it plainly is, or whether a person is entitled to mind about that. Those are real arguments and people should have them.

It settles something narrower and, I think, more useful. The specific number is being used to make people angry, to stoke up fears and worries. There are plenty of legitimate arguments to have about immigration, about government policy on both sides, and about the country we want to live in. But why make up a number, know it is wrong, and keep pushing it, if you are so sure of your position?

And here is the worst bit. Matt, and the rest of them, know the number is wrong. They are not stupid. But it serves their purpose. And they think they can get away with the deception, because they think the people they are feeding it to are the stupid ones.