Britain’s brain drain, measured in mathematical olympians

This is Post Haste, our home for swift thoughts, responsive takes, and first attempts at solving problems. Want to write for us? Email editor@britishprogress.org

By Julia Willemyns (@jujulemons)

You can’t escape brain drain anxiety: it’s there in dinner party laments about friends in Dubai or New York, in the reflexive “if X gets in power, I’m emigrating,” and in the inflammatory opinion pieces that ask whether Britain is losing its best people.

I’ve also been worried about this. Whenever I talk to people in high human-capital fields like AI, they report that the most talented people in their circles are leaving. I too seem to be losing friends. Anecdotes are not trustworthy, but we have few empirical ways to assess whether this diffuse anxiety is warranted.

Until November 2025, the ONS estimated British emigration using the International Passenger Survey – essentially, people questioned at airports about their travel intentions. Not very surprisingly, this turned out to systematically miss actual emigrants.

A new method, using the Department for Work and Pensions’s RAPID database, was rolled out which infers emigration from tax and benefits records to measure who has actually stopped being economically present in the UK. The new method showed that the level of British emigration has been roughly three times what the old survey said – for example, 257,000 in 2024 compared to the old estimate of 77,000.

We now know far more Britons than previously expected are leaving the country. But the ONS revision doesn’t tell us whether things are getting worse – it doesn’t give us a reliable trend. In fact, it turns out that it is extremely difficult to capture British brain drain.

But there is one group small enough to track individually, distinguished enough to have public career records, and consequential enough that their choices actually matter for the national interest. International Mathematical Olympiad medallists.

The International Mathematical Olympiad is the world’s most prestigious mathematics competition for under-20s. Six students represent the UK each year; to medal is to be identified, in adolescence, as operating in a different category entirely.

The conditional probability that an IMO gold medallist will go on to win a Fields Medal (the highest honour in mathematics) is two orders of magnitude higher than for a PhD graduate from a top-ten mathematics programme. About 3% of IMO winners later receive major scientific prizes, making them 60 times more likely than MIT alumni to produce major scientific discoveries. The UK’s cohort includes two Fields Medallists, the designer of the chip inside 99% of the world’s smartphones, and the mathematicians who invented public-key cryptography.

And that’s just the medallists. Demis Hassabis (who founded DeepMind and accepted the Nobel Prize in Chemistry last year) and Alexandr Wang (who founded Scale AI at 19) both represented their countries as students – although neither medalled. The competition has a habit of identifying people who go on to matter.

This is not a perfect proxy for British brain drain. IMO medallists are not representative of British talent, but they are trackable in a way that talented Britons broadly are not. And their choices matter because these are, disproportionately, the people who go on to build things, discover things, and found things. If there is a canary in the coal mine for high-end talent flight, it is probably them.

This analysis is based on three datasets. Career data for UK medallists was compiled by us from public sources (LinkedIn, university websites, IMO records, and occasional bursts of determined Googling). Cross-country retention figures come from Tolga Yuret’s work, covering IMO cohorts from 1985 to 2005. Data on foreign medallists passing through UK and US institutions comes from a Substack series by Neeloy Banerjee, covering gold medallists from 1997 to 2017. I am grateful to both for sharing their work.

All three datasets have gaps, especially for recent cohorts and for people who have declined to maintain a public profile. A lot of what follows is built on small numbers, so read it with that in mind. The starting population is also already tiny: we send six kids to the IMO each year and not everyone will win a medal. That dataset then gets smaller every time you slice it.

Some of the figures below rest on a few dozen people. That’s enough to see the shape of things (and the shape matters), but it’s not enough to lean hard on any single percentage. So treat what follows as a sketch.

The retention rate for British IMO winners has historically been relatively high – 71% stay in the country, putting us 8th in the league table.

If you are British and won an IMO medal, you probably went to Trinity College, Cambridge – 88% of medallists with a known undergraduate institution went to Cambridge. Before 2016, not a single UK medallist did their undergraduate degree abroad. Since then, two have. Both went to MIT, the university that attracts the most IMO winners globally.

Overall, the data seems to suggest that our retention trend has broadly stayed the same over time. But there could be a flicker of something worth watching: the most recent career-stage cohort has dipped to 60% retention. It is too early to know whether this is the start of a trend or a statistical blip.

Retention is only half the story. Historically, the UK has been the world’s second-largest importer of IMO mathematical talent, after the US. But the US has consistently attracted many more winners than we have.

There is some evidence that we could attract more of this talent. Survey data shows that roughly one in five IMO participants from developing countries would prefer the UK for their undergraduate degree, but fewer than one in ten end up here.

Other indicators point to the UK attracting exceptional mathematical talent from around the world. We have no systematic demographic data on IMO competitors, but surnames can serve as a rough proxy for ancestry. This analysis suggests that the composition of the British team has changed dramatically over time.

Before 1990, 87% of UK medallists had British-Isles surnames. In the 2021+ cohort, that figure is 32% – the first time British-Isles surnames have been a minority. East Asian surnames have grown from negligible to 50% of the most recent cohort and South Asian surnames have risen steadily too (mostly of Indian and Sri Lankan origin). The competition is increasingly won by children of immigrants.

Before 1990, 78% of British medallists with known careers went into academia. In the most recent career-stage cohort, that figure is much lower. The shift away from academia started with quant trading in the 1990s and 2000s. Jane Street, Citadel, and their London equivalents began hoovering up mathematically exceptional graduates a generation ago.

Recently, there has been a marked shift toward AI, with 22% of British medallists in the post-2000 cohort now working in the field. Various winners are working at Anthropic and Google DeepMind. Only one is working for a British startup (Blaise AI).

Even those who stay in Britain are increasingly working for American companies. If you are a British IMO medallist working in industry, you are most likely at Anthropic, OpenAI, or Jane Street – only 33% of British IMO winners in non-academic roles work for a British firm.

The headline numbers – 71% retention, 8th in the world – sound reassuring. The UK has been a successful attractor of mathematical talent from abroad, consistently second only to the United States.

But look closer and the picture is less comfortable. The most recent cohort of medallists old enough to have settled careers shows the lowest retention rate on record. Before 2016, not a single UK medallist had done their undergraduate degree abroad; since then, two have gone to MIT. Among those who go into industry rather than academia, the majority work for American companies whether or not they are physically in the UK.

Many of the people best placed to build the technologies that will define the next decade from around the world are largely being absorbed into an American ecosystem. That is the uncomfortable finding underneath a possibly reassuring headline.

Share

It is too early to know whether a structural shift is underway or whether we are reading noise into small numbers. The data is consistent with both stories. What is clear is that the competition for this kind of talent has intensified sharply and that the US has structural advantages Britain has not found a way to match.

The honest truth is that, for many of these people, going to America is simply the correct decision. Research tracking the careers of thousands of IMO alumni finds that mathematicians who migrated to the US were up to six times more productive (by publications and citations) than equally talented peers who stayed home. Those who came to the UK were two to three times more productive. Britain improves the output of the mathematical talent it attracts, but it doesn’t improve it as much as the US. The individual-optimal and nationally-optimal choice is not always the same thing. That is what makes this hard to fix.

The dinner party laments may be premature. But they are not obviously wrong.