Was it easier to deport migrants to France before Brexit?

This BBC podcast is supported by ads outside the UK.

Live in the Bay Area long enough, and you know that this region is made up of many communities, each with its own people, stories, and local realities.

I'm Erica Cruz-Guevara, host of KQED's podcast, The Bay.

I sit down with reporters and the people who know this place best to connect the dots on why these stories matter to all of us.

Listen to The Bay, new episodes every Monday, Wednesday, and Friday, wherever you get your podcasts.

Hello and welcome to More or Less, the show that takes the M and the B and the S out of numbers.

Leaving us with

still workshopping this, we will get there eventually.

Send in your suggestions to more or less at bbc.co.uk.

I'm Tim Harford, by the way, and this week we follow a tangled trail of confusion about the UK's ageing population.

That trail begins in Jackson Hole and ends, well, no spoilers yet.

We ask, where's the beef?

And why is it so expensive these days?

And while ChatGPT isn't writing our scripts just yet, some people think it is being used to write the speeches of parliamentarians.

Is it?

Can the numbers tip us off?

But first, thousands of illegal migrants are arriving on the British coast in small boats.

On some days, more than a thousand arrive.

On some days, there are none at all.

The government would like to send more of them back to where they came from and have done a deal with France to that effect.

The one-in, one-out policy.

On the latest reported figures, this seems to have been taken almost literally.

Three people have so far left the UK on the scheme.

Why is the UK finding it so hard to deport illegal migrants to France?

This was Lib Dem leader leader Sir Ed Davies' response on GB News.

So is this true?

In a sense.

Peter Walsh is a senior researcher at the Migration Observatory at the University of Oxford.

What he's referring to is the Dublin system of which we were a part before Brexit, when we were a part of the EU.

And that allowed EU member states to transfer asylum seekers between each other.

So EU member states would, if they believe that an individual had arrived in their country from another member state, could ask that initial member state, that first country of arrival, to take the asylum seeker back.

So that would be a Dublin transfer-out out request.

And then it would actually be up to that country, country, for example, like Italy, like Greece, like Spain, to decide actually whether they wanted to take the individual back.

And they did have the option to do so.

So Sir Ed was not wrong that the government had this tool at their disposal.

But was the Dublin system a powerful tool?

Were large numbers of illegal migrants sent back to the EU countries they first arrived in?

The data data that we have on Dublin transfers goes back to 2008.

So there are broadly speaking two phases.

The first from 2008 to 2015 inclusive where the UK was a net sender of people under Dublin so it's transferred out more than it transferred in.

In 2008 there were 1,200 transfers out of the UK and 403 transfers in.

So the UK transferred out 814 more people than it received.

But in the last five years that the UK was a part of Dublin, so that's from 2016 to 2020 inclusive, the UK received more asylum seekers under Dublin than it transferred out.

And the numbers were in fact quite small.

From 2016 to 2020 then, under these Dublin arrangements, the UK took in more extra migrants than it returned to other countries.

And while Ed Davie said that the Dublin arrangements applied no questions asked, that doesn't seem to be how the system actually worked.

The reason that the UK was a net recipient in the last five years of Dublin membership is because a declining share of its requests to transfer individuals out, so requests to send asylum seekers back to other EU member states, did not result in an actual transfer.

And there's a big shift here.

So in 2008, the UK made around 2,250 requests to transfer people out, and it actually transferred out around 1,200.

So it's a 54%

conversion rate.

But that fell substantially after 2013 to a low point in 2020 where the UK was really hyperactive in making transfer out requests.

It made 8,500 in 2020, but it only transferred out about 100 people, so 1%.

Why were the percentages of successful return requests so low?

Well, the short answer is, I don't know.

But there are two explanations that have been floated.

And the first is related to poor management at the Home Office.

And lawyers would report that the third country unit was what it was called that was responsible for Dublin transfers.

They would say that actually it was a fairly dysfunctional unit.

Another explanation is that it might be related to better access to justice in the UK than in other countries, meaning that asylum seekers were better able to challenge their transfer elsewhere.

Right.

So either this is a story about the UK rigorously upholding human rights or it's a story about UK bureaucrats being a bit sleepy and incompetent.

Sleepy and incompetent or just first countries, you know, countries that already have very high numbers of asylum seekers refusing to take anymore and that option was always open to them.

Yeah.

So, I mean the numbers that you were describing, they don't sound very big compared to the total volume of asylum seeking.

Certainly not big compared to the total amount of immigration.

Certainly, they're not large numbers.

If you look at the whole data set from 2008 to 2020, so in that 13-year period, that's all of the years for which we have data.

The UK actually

sent out about 2,000 more individuals than it received.

So it was a transfer.

But in the grand scheme of things, that's not very many.

For a bit of context, in 2024, 108,000 people claimed asylum in the UK.

That was the largest number since records began in 1979.

The previous peak was 103,000 back in 2002.

The numbers were lower in between, mostly between 20 and 40,000 per year.

But however you cut it, the number of returns under under the Dublin arrangements would constitute a tiny fraction of this group of migrants.

You know, when we speak with asylum seekers and ask them why the UK or why other countries, Dublin was never mentioned in the pre-Brexit period, which is understandable because the numbers are so low.

And you could imagine as well that, you know, the handlers, the smugglers themselves, they have a vested interest in downplaying the chance of being returned.

And in fact, that chance of being returned was very small.

So there's not much evidence that the Dublin system acted as a deterrent.

However, Peter says that interviews with potential illegal migrants who want to come to the UK now do suggest that not being in Dublin and not being part of the EU's migrant fingerprint database has become one of the many reasons they want to come here.

A recurring theme has been when they are asked why the UK,

they will say to escape the Dublin system, and they understand that they can claim asylum in the EU, potentially even be refused, and that they would get another bite at the cherry if they could reach the UK, because in all likelihood, we wouldn't know that they've claimed asylum elsewhere and we would end up processing that claim.

Thanks to Peter Walsh from the Migration Observatory.

It's not every day that more or less spots a factual error from the governor of the Bank of England.

I'd really hope they'd check their facts pretty thoroughly.

But there, apparently, it was, reported in black and white on the Guardian website.

In their reporting of a speech by the Governor Andrew Bailey at an event for policymakers at somewhere with the unlikely name of Jackson Hole in the US, Bailey was quoted describing the productivity challenge the UK is facing, followed by this line in the story.

By 2040, 40% of the UK population will be older than 64, he added.

Wow.

Just wow.

I know all you loyal listeners will have the ONS population projections memorised, so you will know that this is an absolute howler.

40% by 2040?

Come on.

But just in case you're not absolutely confident in the correct numbers, we found a proper demographer to tell us what was what.

My name is Jennifer Dowd and I'm a professor of demography and population health at the University of Oxford.

So the claim 40% of the UK population is going to be 65 or over by 2040, true or false?

False.

Okay.

What's true then?

Well, it's important to say these are always projections when we're thinking about 2040.

We don't have a crystal ball, but the ONS estimates are that 22.6% of the UK population will be 65 or over in the year 2040.

So 22.6, that's, I mean, I don't want to get super technical, that's a lot less than 40%.

So we just over

2 in 10 rather than 4 in 10.

Yes, it's quite a big difference.

And it's also not that different from the 18.9%

that we are around now.

But now, the bombshell.

It turns out that this line in the Guardian story is not a report of what Andrew Bailey said.

Thankfully, the governor of the Bank of England is not massively wrong on basic facts about the country whose economy he is governing.

Phew.

Instead, we have a somewhat less lurid tale of journalists making mistakes.

That 40% figure was reported in the Financial Times, and it is at this moment I should remind you that my day job is working for the FT.

The FT journalist told us that this was a simple mistake based on eyeballing a PowerPoint slide that the Governor had up on screen.

The FT corrected the story to say that approaching 30% of the adult population would be 65 or over by 2040, which is fine.

Yes, so the ONS projections estimate that 28% of the adult population will be 65 or over by 2040.

But how did the Guardian mysteriously make the same mistake?

Simple.

They copied the FT and it looks like like they picked the wrong FT story to copy.

The Guardian also corrected their story, but they took a different angle.

Andrew Bailey's chart ran all the way to 2121, where the percentage does actually reach 40%.

So, in their correction, the Guardian changed the date in the story, so that the UK population will indeed have 40% of people over the age of 65, as long as you wait until the year 2121, 96 years hence.

It is quite an odd correction, but it is at least a correct correction.

So that claim is true.

The projection from the ONS for the year 2121 is indeed that the proportion of the adult population over age 65 will be 40 percent.

But we should also say that 2121 is a very long time away, so these projections are subject to a lot of uncertainty because we do not have a demographic crystal ball and generally can't predict how many babies humans will be having 80 years from now or more.

So mysteries solved, everything is safely back in the correct place, and yet the fundamental problem remains, namely our aging society, demonstrated by that slowly increasing proportion of the population aged 65 or over seen in Andrew Bailey's graph.

Can we zoom out for a second?

We hear all the time about this aging society.

The graph doesn't look great in as much as it does suggest that more and more people are going to be elderly.

There's obviously nothing wrong with being old, beats the alternative, but it creates certain pressures.

So why is our society aging?

What is the fundamental driver of that?

Yeah, it's interesting because I think a lot of people immediately think about increases in life expectancy that we're all living a lot longer when they think about why the population is aging.

But what's really true demographically is that population aging really reflects changes in how many babies we're having.

And it's that kind of demographic echo of changes in fertility in the past that really caused population aging.

So, for example, we know that we had a baby boom after World War II, and in the UK, there was also Volota births around 1960.

And, you know, those generations are now becoming the 65 and overs, and they're a relatively larger generation compared to the people born after them.

For example, there were nearly 900,000 babies born in the UK in 1961, compared to under 700,000 in 2021.

A big difference, especially given the population over that time period also significantly increased.

And so this is kind of baked in population aging that as those really larger birth cohorts get older, the share of the population at older ages increases.

But what's really interesting about population aging, I think, as a demographer is that eventually things kind of even out, especially if younger generations are more similar in size over time, then we would expect this population aging to eventually even out, so it can't really go on forever.

Some people talk about these really large cohorts like the baby boom being a pig and a python.

So once that pig works its way through, the issue of the aging society will become far less acute.

It just happens to be the case that the bulge is hitting about now.

Thanks to Professor Jennifer Dowd.

You're listening to more or less.

Stop settling for weak sound.

It's time to level up your game and bring the boom.

Hit the town town with the ultra-durable LG X-Boom portable speaker and enjoy vibrant sound wherever you go.

Elevate your listening experience to new heights because let's be real, your music deserves it.

The future of sound is now with LG XBoom.

And for a limited time, save 25% at lg.com with code Fall25.

Bring the boom.

X-Boom.

GiveMeTheVin.com is the best place to sell your classic or collectible, whether it's one car or an entire collection.

GiveMeTheVin want your muscle car, hot rod, or your old square body pickup, and they love buying resto mods.

Go to givemeTheVin.com today to sell your classic or collectible and avoid sellers' commissions, tire kickers, and low ballers on the internet.

It's as simple as going to givemethevin.com and entering your car's VIN number or license plate number.

GiveMeTheVin.com is America's best car buyer.

In a world so full of confusing modernities, TikTok, chat GPT, those laboo boo things, it's good to have a story so old school that it would prompt John Bull himself to pick up a quill pen and write in to complain.

And we have one, namely, what in the name of St.

George is going on with the price of beef?

Over the last 12 months, the price of beef in the shops in the UK has gone up by 25%.

Of all the foods and non-alcoholic beverages that make up the standard basket of foods that are used to calculate inflation, beef has seen the biggest increase in price.

25% seems like a pretty big number, especially since the most popular broad-based measure of inflation, the Consumer Prices Index, is up by less than 4% over the past year.

But why is it happening?

Who will help us explain these meaty numbers?

If it's it's numbers you want, I'm your man.

Perfect.

If you're going to do it, do it right.

This is David Swales, head of economics and analysis for livestock at the Agriculture and Horticulture Development Boards, the AHDB.

This 25% inflation figure does seem stark.

When I look at the cattle prices which farmers are getting paid, it actually seems like it's quite a low figure to some of the numbers I look up.

So, John, if I look at deadweight cattle prices in January of this year, farmers were getting around about £5

a kilo for sort of prime cattle.

And that was a relatively high price.

John, our prices tend to be anywhere between £3 and £5 a kilo on average.

So in January, those prices were quite high.

Now, those prices actually went up to about £7

a kilo in May.

Yeah.

So that's up 40% from what was already

a fairly high level.

Exactly.

To be be honest, John in AHDB, we are analysing this sort of data, looking at it all the time and reporting on this information to farmers so they understand what's going on.

And we were having to redraw all of our graphs because the line had gone off the top.

It was that severe.

So beef prices in the shops are up because the supermarkets and meat processing companies that buy beef from farmers are having to pay more for it.

Why are they having to do that?

Basically, it's a lot of it's to do with economics.

It's about supply and demand.

Of course, the loyalist of loyal listeners will remember that over 15 years ago, a drug dealer gave me the same answer.

In this case, there's not that much going on on the demand side.

We're not seeing a run on beef.

What we've got is a supply problem.

At the minute, what we see happening in the UK is our supplies of cattle have fallen and are continuing to fall.

And what you sometimes see is when our domestic situation has a bit of a shortage, we can kind of go outside the UK and import a little bit more beef from abroad but actually lots of the places we import our beef from are short as well so there aren't those ready available supplies for us to suck in so the whole market is a little bit short of beef and therefore the price has started to rise up let's talk about domestic supply first why have farmers been producing less beef.

Okay, so our domestic supplies in the first half of this year so this is sort of january up until the end of july are down about four percent this year so that's quite a drop we don't normally see drops of that order and i think probably there's a number of factors here it has been a bit of a challenging year weather-wise so we've not had as much rain as we would normally have and maybe grass growth has been probably less good than farmers are used to which maybe has slowed down animals coming to market but i think probably the the big factor here is over the last couple of years are suckler farmers so the people focused on producing beef lots of those farmers have have been struggling to make decent profits from their business david says that one of the chief causes here is the gradual phasing out of the subsidies that farmers used to receive when the uk was in the eu under the common agricultural policy But since Brexit, we've had control of our own agricultural policy.

Indeed, there are different policies in different parts of the UK.

So there's an England policy, a Scotland, and a Wales agricultural policy.

And certainly in England, that policy is probably a little bit less generous to farmers.

There's less support payments.

And indeed, the sort of basic support payments which farmers get are being phased out entirely.

So by 2027, those basic payments farmers receive will no longer exist.

And those payments have historically helped support incomes in some of these farms.

And maybe if a farm might have been lost making in a particular year, the support payments would have helped the farm make money in total.

So domestic supply is down and imports can't pick up the demand because supply in the countries we buy from is also down.

For example, in the biggest beef importer to the UK.

Supplies are down in Ireland, so it's around sort of four or five percent down.

Now, economics being economics, the high price of beef is likely to stimulate beef production.

But don't get too excited.

From an animal giving birth to a steak appearing on our shelf is typically about a two-year period.

So when I look at the data, we've got some really good data.

If you go to that age of cattle under 12 months old, there is a bit of an increase there.

But the bad news is, as consumers, we probably have to wait a little while before that's going to hit our shelves.

Thanks to David Swales.

This is more or less.

And our email is more or less at bbc.co.uk.

Now, if you've ever had the good fortune, nay, the privilege, to listen to a session of the House of Commons, you'll be used to it sounding a bit like this.

But recently, something strange has been occurring.

New phrases are cropping up.

Phrases more at home in US political dramas.

Thank you, Madam Deputy Speaker.

And I rise to speak.

Thank you very much, Madam Deputy Speaker, and I rise to speak.

Thank you, Madam Deputy Speaker.

I rise to speak in this debate today.

I rise to speak in support of the amendment in my name.

This might not seem newsworthy, but it has sent some members of parliament into a bit of a tizzy.

I rise to speak.

I rise to speak.

I rise to speak.

Chat GPT knows you're there, but that is an Americanism that we don't use.

But still, keep using it, because it makes it clear that this place has become absurd.

This isn't just a complaint that British MPs are sounding a bit too American and talking about elevators and sidewalks all of a sudden.

The implication is that MPs are using ChatGPT and other large language models to write their speeches, rather than doing it the old-fashioned way and getting an intern to do it.

Let us, as ChatGPT would say, do a deep delve into this question with our large language model correspondent, Lizzie McNeill.

Hello, Lizzy.

Hi, Tim.

So there has been some controversy about people using large language models, aka AI, to write essays or speeches for quite a while.

And now this suspicion has spread to the House of Commons?

It has.

You see, AI has certain tells that can show its metaphorical hand.

For example, using Americanisms such as IRIS.

Right, but why would a large language model use Americanisms for a British speech?

Well, it's all to do with how these language learning models learn.

So we spoke to Anthony Cohn, professor of automated reasoning at the University of Leeds, about the inner workings of AI.

Humans tend to overuse certain words and phrases too.

In both cases, it's a result really of the language being exposed to.

In our case as humans, it's our upbringing, friends, family, the podcasts we listen to, the TV and the radio shows we listen to.

In large language models, it's the same, but their training data comes from the vast swathes of text that's been given to them from the internet and so forth.

The point being, ChatGPT and the others learn from the internet.

A lot of stuff on the internet is comes from America just because it's a much larger country.

But are Americanisms cropping up more in parliament?

They do seem to be.

So we spoke to Zoe Craufer, a journalist for Politics Home, who has been using Hansard, the government's archive of all speeches made in the Commons and the House of Lords, to see whether the term IRISE has indeed increased.

It was used 601 times so far this year.

I think I did it up to the end of August, compared to 131 times in the same period last year.

It was also used only 227 times in the same period the year before.

Hmm.

I mean, I rise to speak is weird.

It's hardly a smoking gun.

Maybe they were watching US TV or maybe one MP used it and the others all started copying them without realising.

We're social beings after all.

Although 601 times versus 131 times a year ago, that is a very sudden change in linguistic habits.

Yes, and it's not just I rise to speak.

So things like, it's not merely underscores, bustling, just kind of chat GBT buzzwords that the software likes to use.

So I think combined with both the statistical analysis and the anecdotal evidence, I think it's clear that there clearly has been a spike that's at least in some part influenced by AI.

And these these terms have all seen spikes on Hansard.

It still doesn't feel conclusive.

Humans spit out clichés all the time.

Oh, thinking outside the box, Tim.

I like it.

Yes, you're right.

And also, the increase in people using AI may lead to a general increase in people using these terms independently because they've heard other people using them.

But as you say, the speed with which these phrases have appeared is suspicious, and it's not just in Parliament.

So, one paper which analysed the usage of words in scientific papers and looked at their frequency per million words of scientific paper outputs, they looked at four words in particular, pivotal, intricate, showcasing, and realm.

And in all these cases, we can see the frequency of their use up until the release of ChatGPT and shortly after was basically bubbling along at about the same level, somewhere between 20 per 1 million words.

And then suddenly after that, they really start shooting up to become much, much more likely, like 70 times more likely.

So AI being used to write speeches in Westminster is pretty consistent with what's going on in the wider world.

Yeah, it wouldn't be that surprising if politicians were using AI to write speeches.

I mean, they've outsourced that job for decades.

Quite often they'll deny direct use of it themselves, but...

in a way that's clearly suggesting that other people must be using it.

But if everyone's saying that everyone else is using it,

then, you know, anecdotally, I think it is probably pretty widespread, particularly probably among parliamentary staff who often also have a really big workload.

Anthony says that it's almost impossible to conclusively determine whether a piece of text was produced by AI.

But then in the case of MPs.

I think in the MP case, it probably is.

But, you know, would I want to swear in a court of law that it is?

No, I wouldn't.

It seems like it could be a case of overworked and exhausted junior staff drafting speeches with the help of ChatGPT.

What do you think, Lizzy?

Yes.

Is there anything else you'd like me to delve into?

The Honourable Member will withdraw her remarks immediately.

And she has a week to think about it, because that is all we have time for.

Please do keep your questions and comments coming in to moreorless at bbc.co.uk.

We'll be back next week.

And until then, goodbye.

More or less was presented by me, Tim Harford.

The producer was Tom Coles, with Nicholas Barrett, Nathan Gower, and Lizzie McNeill.

And thanks to Josh McMinn, who tipped us off for that last story.

The programme was recorded and mixed by Gareth Jones.

The production coordinator was Maria Ogondole.

And our editor is Richard Varden.

In the future, will your taxi fly?

I'm Greg Foot, host of the BBC Radio 4 show and podcast Slice Bread.

And now, Doe.

In Doe, we explore future wonder products that might rise to success and redefine our lives.

Might delivery drones make popping to the shops a thing of the past.

On-demand drone delivery could be absolutely huge.

Will we really let our cars do the driving for us?

If you say this whole driving thing, it's a thing that's only ever meant for humans.

That's obviously for the birds.

Each episode, I sit down with entrepreneurs and experts to discuss what today's everyday technology may look like tomorrow.

Like tomorrow.

Find out on Doe.

Listen first on BBC Sounds.

GiveMeTheVin.com is the best place to sell your classic or collectible, whether it's one car or an entire collection.

GiveMeTheVin wants your muscle car, hot rod, or your old square body pickup, and they love buying resto mods.

Go to givemeTheVin.com today to sell your classic or collectible and avoid sellers' commissions, tire kickers, and low ballers on the internet.

It's as simple as going to givemethevin.com and entering your car's VIN number or license plate number.

GimmeTheVin.com is America's best car buyer.