Alternative Education Outside the State System

Three locks on the schoolhouse door

The first lock is compulsory attendance. German law requires attendance at a school rather than an education as such, and the courts have upheld enforcement to a considerable degree: in Wunderlich v. Germany in 2019 the European Court of Human Rights found no violation of family life where parents lost part of their parental authority and their four children were taken into care for three weeks after being withdrawn from school. Sweden’s Education Act of 2010 permits home education only in extraordinary circumstances, granted a year at a time. France replaced its declaration regime with prior authorisation on four defined grounds from the 2022 school year, and registered home instruction fell from roughly 72,000 children to roughly 59,000 in a single year.

The second lock is quieter and binds more widely. In most European systems a private school still teaches the national curriculum and still employs state-licensed teachers. A family can therefore leave the public system without reaching a different education: what the law offers is a choice of provider, not a choice of method. Countries that permit home education often apply the same logic to it, assessing the child against the curriculum they left.

The third lock sits at the far end, and it is the one that makes the other two heavier than they look. A school-leaving certificate is worth what universities and employers will accept it for. Whatever a family does when a child is fifteen has to be legible to an admissions office when the child is eighteen, which quietly rules out anything that cannot be translated back into the standard qualification.

When the rules do not bend, families move

The clearest illustration is a place nobody chose as an experiment. Åland is an autonomous, Swedish-speaking province of Finland a few hours by boat from the Swedish mainland, with its own education legislation and a Finnish constitutional guarantee of the parental role. In 2012 it registered eleven home-educated children. By 2024 the figure was 132, most of them from families who had left Sweden because they could not teach their children at home there.

This is jurisdictional competition in education happening by itself, without anyone designing it: a rule written in Stockholm, and a ferry. It is also a lesson in how little the mechanism guarantees. Åland tightened its own rules in 2021, allowing authorities to end home education where a child’s progress was judged insufficient, and has considered going further still. Asked about families who find the local law too strict, one supervisory teacher’s answer was that they are free to move away. Exit is worth exactly what the next jurisdiction is willing to allow, and nothing in the arrangement keeps the next one open.

Why a city can try what a country cannot

National caution here is not simple obstinacy. The subjects are children, the feedback loop runs a decade or more, and a reform that fails is discovered long after the cohort that absorbed it has left school. At the scale of a country that risk is close to unbearable, and the rational response is to change very little, very slowly, and to make deviation difficult in the meantime.

A jurisdiction of a few thousand people changes the arithmetic without changing the argument. A school of forty pupils is a real trial of a real method; a failure costs a few dozen families rather than a national cohort, is visible within a few years rather than a generation, and can be ended. That is the general case for jurisdictional competition, and education is where it bites hardest, because it is the field in which a national system is least able to run a trial on itself. It is also why the interesting work tends to appear at the edges: AmityAge Academy, a Bitcoin education centre being built on Roatán with Próspera’s support, or the Prague institutions built on Václav Benda’s idea of a parallel polis, where the response to an education you could not obtain was to build the institution rather than petition for one.

One case shows both the promise and its limit. My First Bitcoin began teaching a ten-week course to 38 El Salvadoran public school students in 2022, entirely outside the state system, and grew to tens of thousands taught. Its open-source workbook was then adopted by the country’s education ministry as the basis of a national curriculum, and the state programme ended in April 2025, after which the organisation closed its local chapter. A parallel structure successful enough to be adopted has handed its survival to whoever adopted it.

The contrary case shows what happens when a lock is simply released. Arizona opened its education savings accounts to every school-age child in 2022; enrolment went from around 12,000 to over 100,000 within four years, and a network of microschools grew up to meet it, because a dozen families with a few thousand dollars each is enough to pay a teacher and rent a room. Arizona is a US state rather than a Free City, and the point is the mechanism rather than the place: where the money follows the child and the provision is not prescribed, provision appears quickly.

The constraint that does not go away

A new jurisdiction can write its own school law. It cannot make a university in another country accept the certificate that law produces. This is why schools in special jurisdictions overwhelmingly route through an established international qualification, the International Baccalaureate or Cambridge International, rather than inventing one: it solves recognition by borrowing someone else’s, at the price of taking their curriculum with it. The genuine frontier is whether a small jurisdiction can build a credential of its own that anyone outside honours, and no one has done it yet.

That is a fair statement of where the field stands. The regulatory freedom exists in principle in several of the projects covered on this site, and very little has yet been built on it. What the Free Cities argument offers education is not a finished alternative but the one thing the national systems structurally cannot provide, which is somewhere to try, at a scale small enough that being wrong is survivable.