Why We All Need The World's Best Schools, And How This Ranking Supports Them

Start with the biggest fact in the economics of education. For fifty years, researchers tried to link education to national wealth by counting years spent in school — and the answers came out weak and messy. The breakthrough, led by economists Eric Hanushek and Ludger Woessmann, was simple: stop counting years and start counting what children can actually do, using international maths and science tests that go back to 1964.^[1]

The result changed the field. Once you know a country's real skill level, the number of years its children sit in classrooms tells you almost nothing extra about its growth. Skills, on the other hand, predict growth powerfully and consistently. Across forty years of data, a country that lifts its test performance by one standard deviation — roughly the gap between an average pupil and a good one — grows about two percentage points faster per year, every year.^[2] Compounded over a generation, that is the difference between two entirely different countries. The authors call this "knowledge capital." It explains why East Asia caught up with the West while other regions, which expanded schooling just as fast, did not: enrolment rose everywhere, but learning did not.

Two more findings matter here. First, this is not just correlation. The relationship has been stress-tested — by tracking skill gains within countries over time, and by comparing the earnings of migrants educated under different systems — and it holds up.^[2]Second, and crucially: when the researchers split the skill distribution apart, the share of a country's top performers predicted growth on its own, separately from the share of children reaching basic literacy. Both matter. They work together. But the top end is not a bonus sitting on top of the average — it is a second engine.^[1]

How much work does the top of the class actually do for everyone else? In 2011, Heiner Rindermann and James Thompson gathered test data from 90 countries and asked a simple question: what predicts a country's wealth better — the ability of its average student, or the ability of its top 5%? The answer was the top 5%. That small group predicted GDP, patents, Nobel prizes and the number of leading scientists more strongly than the average did.^[5] In a knowledge economy, prosperity is produced out of proportion by the few people working at the frontier.

The individual-level data says the same thing. Raj Chetty and colleagues linked 1.2 million American patent-holders back to their childhood test scores and family tax records. Children in the top 5% of third-grade maths were many times more likely to grow up to be inventors than everyone below them — and the curve is shaped like a hockey stick, exactly what you would expect if the top end drives innovation.^[6] But their second finding is the one this essay turns on: talent alone did not convert. Children with the same high scores, but from poor families or from places with no inventors around them, went on to patent at a fraction of the rate. The researchers call them "Lost Einsteins" — and they estimate that if girls, minorities and poor children invented at the same rate as well-off boys who grew up around innovation, America would produce four times as many inventors.^[6]

The same pattern shows up at the very peak of ability. Ruchir Agarwal and Patrick Gaulé followed every International Mathematical Olympiad medalist for twenty years — the most objectively identified mathematical talent on Earth. Medalists go on to push the boundaries of mathematics at extraordinary rates. Yet a medalist born in a poor country — with a medal proving identical talent — produces far less over a career. Not for lack of ability. For lack of institutions that turn ability into output.^[7] Talent, in short, is born everywhere. Turning it into something is done by institutions.

If exceptional results were produced wherever exceptional children happen to sit, top performers would be scattered thinly across thousands of schools — roughly wherever bright children live. They are not. Glenn Ellison and Ashley Swanson studied the American Mathematics Competitions, a nationwide talent search taken by hundreds of thousands of students. The students reaching the highest scores came overwhelmingly from a tiny group of schools — perhaps twenty in the whole United States — while the typical American high school, which statistically contains plenty of very bright children, produced essentially none.^[8]

Our own data tells the same story at global scale. We matched the schools in this ranking against Wikidata's registry of notable people — scientists, founders, artists, heads of state — under a published methodology.^[16] The concentration is extreme:

Fifty schools — three percent of our database — account for over seventy percent of every documented notable life in the index. Even after discounting generously for fame and for the depth of old schools' record-keeping, one conclusion is unavoidable: the idea that "talent succeeds wherever it sits" does not survive contact with that chart.

Here is the strongest argument against everything above, and we state it at full strength because our readers deserve it. Elite schools choose their students. Of course their results look spectacular — they admitted spectacular children. Scatter the same children across ordinary schools, the objection goes, and the results would follow the children, not the buildings.

Economists have tested exactly this, using an elegant trick: compare the last student admitted to a selective school with the first student rejected. Statistically, they are the same child — one got in, one didn't. The most famous of these studies, pointedly titled "The Elite Illusion," looked at Boston's and New York's exam schools and found the student who just got in gained little or nothing in test scores compared with the student who just missed.^[9] Dobbie and Fryer found the same in New York.^[10] Studies at the UK grammar-school borderline report similarly modest short-run exam effects. We do not dispute these findings. They are sound — and any honest case for elite schools has to be built on top of them, not around them.

And when researchers measure the things the borderline studies cannot, the picture changes. Seth Zimmerman followed Chilean students on both sides of the elite admission line for decades: getting in sharply raised the odds of reaching top incomes and company leadership — an effect that ran through the friendships and networks formed at school.^[12] Damon Clark and Emilia Del Bono followed a British cohort for fifty years: attending the selective school produced real long-run gains in education and earnings, especially for women.^[11] In Romania, where schools are assigned centrally nationwide, getting into a better school produced genuine achievement gains — the cleanest large-scale evidence that schools themselves, not just their students, carry quality.^[13] And David Card and Laura Giuliano showed that even mild concentration — one gifted classroom inside an ordinary district — lifted the achievement of high-ability minority students, at almost no cost.^[14]

History has run the experiment for us. Between 1965 and 1980, England closed most of its selective grammar schools. In 1976 it also stopped funding the 170-odd "direct-grant" schools — schools that picked children by ability and charged most families nothing. If closing the best schools really did spread their quality around, this is where we would see it.

Here is what actually happened. Ordinary students did just as well as before: national averages barely moved. But the excellence was not shared out — it was moved out of reach. Almost every direct-grant school, Manchester Grammar among them, survived by going private. A ladder that had been free, and open to any child who could pass an exam, became a ladder you must pay to climb. The few grammar schools that remained became so scarce that places now go disproportionately to families who can pay for tutors or for a house near the school — exactly the privilege the reform set out to end. And the long-run evidence suggests the old ladder had been delivering real lifetime gains to the children who climbed it, especially women.^[11]

The same thing happens wherever it is tried. When San Francisco switched its flagship Lowell High School from merit admission to a lottery in 2021, failing grades among new students roughly tripled, and the public demanded merit admission back within eighteen months. Notice what none of these stories show: averages falling. What they show is simpler, and worse. The middle of the system holds steady. The brightest do worse. And the machine that built the top end is destroyed, priced up, or quietly rebuilt under another name — because ambitious families never stop demanding it. The only real choice a country has is whether that machine is public, visible and open to talent — or hidden behind fees and postcodes.

A fair follow-up question: selective state schools are one thing — what happens if a country goes further and abolishes private schools? Outright bans barely exist. The United States Supreme Court ruled one unconstitutional in 1925, and most democracies protect the right to run a school. So the evidence comes from the two countries that came closest — and they point in tellingly different directions.

Finland is the strongest case for the reformers. Between 1972 and 1977 it effectively nationalised its private grammar-school tier: schools passed into municipal ownership or survived only tuition-free, on the national curriculum, at public expense. Because the reform rolled out region by region, economists could measure it cleanly. Social mobility improved substantially — the link between a father's income and his son's fell by about a fifth. Average test scores barely moved, and the gains that did appear went to children of less-educated parents.^[17] But look closely at what Finland actually did: it kept the buildings, the teachers and the ethos, and removed only the price. That is nationalisation of access — not demolition of the machinery. And the ledger has a missing page: no study has measured what the reform did to Finland's top tail, and today its share of top performers is modest while its celebrated PISA average has been sliding since 2006.^[3]

South Korea is the cautionary case. Its Equalization Policy, begun in 1974, replaced selective high-school admission with residential lotteries; in 2019 the government ordered its elite autonomous private schools converted into ordinary ones. The results are well documented. Private tutoring spending — the thing the policy was meant to reduce — did not fall; it kept climbing into one of the world's largest shadow education industries. Where equalization was extended, achievement fell by roughly a tenth of a standard deviation, concentrated among the highest-performing students, with knock-on effects on university entry and later earnings.^[18]The middle was untouched; the top tail was harmed. And the demand for selection did not disappear — it moved into evening tutoring academies and into house prices near the better schools, a pattern economists have measured on three continents.^[19]The elite-school abolition itself was reversed in 2024. Fifty years on, the policy still cycles back and forth, because the pressure it tried to abolish is a constant.

There is also a third model, worth naming because it splits the difference: the Netherlands funds private schools fully from the public purse — pluralism kept, price removed. The honest synthesis across all three: a country can buy real mobility gains by taking private schools public, if, like Finland, it preserves the institutions and removes only the fee. What no country has managed is to abolish selection and make the demand vanish. Korea shows where it goes instead — into tutors and postcodes, with the bill paid by the top tail and by families too poor to pay twice.

Pull the threads together and you get a testable claim: countries with more excellent schools should — through the skills channel (Part I), the top-end channel (Part II) and the conversion channel (Part III) — show stronger innovation, higher productivity and deeper leadership pipelines. The international test data fits: in PISA 2022 mathematics, 41% of Singapore's students reached the top proficiency levels, against an OECD average of 9%^[3] — and Singapore's rise from poverty since 1970 needs no introduction. Hanushek and Woessmann put numbers on the prize: realistic improvements in what a nation's children learn are worth several times its entire current GDP over the long run.^[4]

We are careful with this claim, because it is the easiest one to overstate. Across countries it is a correlation, and rich countries can afford good schools. But the evidence assembled above supplies the mechanism that mere reverse-causation cannot: skills cause growth,^[2] the top end drives patents,^[5][6]and institutions convert talent into output.^[7] We know how the flywheel turns, link by published link. A country that lets its best schools decay is not trimming a luxury. It is taking apart one of the few machines known to compound.

This is why the essay lives on a ranking website. You cannot protect what you cannot see. Every argument above depends on one ability: being able to say, with evidence, which schools are actually excellent. Not the most expensive ones — our own Value 100 research shows price and quality are only loosely linked, and forty schools in this ranking charge nothing at all. Not the ones with the best marketing. The ones that demonstrably teach, stretch, and turn talent into results.

A great school takes a century to build and one decade of bad policy to destroy. England's direct-grant schools disappeared from public ownership with barely a dataset to mourn them — nobody had put their value on the record, so when the moment came, nobody could defend them with evidence. That is the job of an independent ranking: to document what these institutions produce, before they are at risk, in a form no one can wave away. It is not a consumer accessory bolted onto education. It is part of the protection. This ranking accepts no payment for placement, publishes its methodology, and re-verifies its data every edition — because an argument this important deserves evidence this clean.