Daniel sent us this one — what's the actual minimum a government should do, according to libertarian theory? Just courts, cops, and defense? And then the real-world gut check: which Western democracies come closest to that maximum privatization, and which ones lean hard the other way into state control? It's basically a three-part question hiding inside one prompt. Theory, then scoreboard, then the opposite end of the spectrum.
It lands at a moment when the real world is, as usual, making a mess of the theory. The UK's water utilities are drowning in sixty billion pounds of debt and sewage spills are up fifty percent since twenty nineteen. France fully renationalized its electricity giant EDF in a nine point seven billion euro buyout. So we're not debating this in the abstract. These are live, expensive, sometimes smelly experiments.
Smelly is a good adjective for a privatization discussion. Of course it is. So let's start with the theory, because I think most people hear "libertarian minimum state" and imagine something between a Ron Paul rally and a survivalist compound. What's the actual philosophical architecture here?
The canonical text is Robert Nozick's "Anarchy, State, and Utopia" from nineteen seventy-four. He articulates what he calls the night-watchman state. The government's only legitimate functions are protecting citizens against force, theft, and fraud, and enforcing contracts. That translates to national defense, police, and a court system. That's it. No public education, no public healthcare, no state pensions, no publicly owned utilities, no transportation infrastructure beyond maybe — maybe — some minimal public goods like lighthouses where you genuinely can't exclude non-payers.
Lighthouses are always the libertarian carve-out. It's the one thing they'll grant you.
It's the classic example. But here's the thing — even within libertarianism, Nozick was the moderate. Murray Rothbard, in "The Ethics of Liberty," argued that even the night-watchman state is illegitimate because taxation itself is theft. He wanted everything — police, courts, defense — provided voluntarily through the market. So the libertarian camp has its own internal spectrum from "tiny state" to "no state at all.
Which makes the phrase "libertarian minimum state" already a compromise position. Like the moderate wing of the absolutists.
And I think that's important because when we go looking for real-world examples, we're not going to find Rothbard's anarcho-capitalist utopia anywhere. We're looking for the Nozick end — countries that have shrunk the state's ownership role as far as any functioning democracy has managed.
How do we actually measure that? Because "privatization" is one of those words that gets thrown around to mean about six different things.
We need to distinguish between two things. One is ownership privatization — the government literally selling state-owned enterprises, transferring assets to private hands. The other is service delivery privatization — contracting out public functions to private providers while the government still funds and regulates them. Think private prisons, charter schools, or a municipality hiring a private company to run its water system. The state still pays, but it doesn't operate.
Those are very different animals. Selling off an airline is one thing. Keeping the funding tap on but handing the wrench to a contractor is another.
That second category is where a lot of the mess lives. But to answer the question of which countries are most privatized, the best single yardstick is the Fraser Institute's Economic Freedom of the World index. The twenty twenty-five edition ranks countries on multiple dimensions, but the two most relevant sub-indices are Size of Government — which measures government consumption, transfers, and subsidies as a share of the economy — and Legal System and Property Rights.
The Fraser Institute is a Canadian free-market think tank, right? So we should probably acknowledge the ideological tilt.
It's produced by economists sympathetic to limited government, and their methodology weights things like low marginal tax rates and absence of state ownership. But it's still the most comprehensive cross-country dataset we have for this question. And the rankings are revealing, even if you disagree with the normative framework.
So who tops the list?
Singapore is number one. Switzerland number two, New Zealand number three, Ireland number four, the United States number five. And that top five immediately tells you something fascinating.
The place with Temasek Holdings, the gigantic state-owned investment firm?
This is the Singapore paradox, and it's the most important thing to understand about how these rankings work. Singapore has massive state-owned enterprises. Temasek alone manages a portfolio worth over two hundred billion US dollars, with controlling stakes in Singapore Airlines, Singtel, DBS Bank, and dozens of other companies. The government also owns essentially all the land through the Land Authority. And yet it ranks first in economic freedom.
How does that square?
Because the Fraser index measures government consumption, transfers, subsidies, and the regulatory burden — not just ownership. Singapore's government spending is only about eighteen percent of GDP. Its top marginal income tax rate is twenty-two percent. There's no capital gains tax. The regulatory environment for business is extraordinarily light. So what you have is a government that owns a lot of things but runs them on strictly commercial terms and doesn't tax or spend heavily. The state-owned enterprises compete in open markets and are expected to turn a profit.
It's state capitalism with a libertarian tax code. That's not really the minimum state in Nozick's sense — the government's fingers are in everything. It just has a very light touch on the wallet.
And this is why a single number never captures the whole story. Singapore is the freest economy by the Fraser metrics, but nobody would describe it as a libertarian society. It's an authoritarian city-state with pervasive state involvement in housing, land, and industry. The state just happens to be a very efficient, commercially minded owner.
Which brings us to the actual closest contender among Western democracies. You said New Zealand is number three.
New Zealand is the poster child for the democratic privatization experiment. Between nineteen eighty-seven and nineteen ninety-nine, the country underwent what's known as Rogernomics — named after Finance Minister Roger Douglas. They privatized sixteen major state-owned enterprises. Telecom New Zealand, now Spark. Air New Zealand. The Bank of New Zealand. New Zealand Rail, later rebranded as Tranz Rail. The forestry service. Even the postal service was corporatized and partially sold.
They basically looked at the crown jewels and put a price tag on all of them.
The motive was pretty stark. By the early nineteen eighties, New Zealand had one of the most regulated, protected economies in the OECD. Per capita income had fallen from third in the world in the nineteen fifties to twentieth. The country was facing a balance of payments crisis. Douglas and Prime Minister David Lange — both from the Labor Party, notably — decided the only way out was to shrink the state dramatically.
A left-wing party doing the privatization. That's the kind of detail that upends the standard political narrative.
It's one of the great ironies. The Labour Party dismantled the welfare state it had built. By the end of the reforms, the New Zealand government had exited airlines, banking, telecom, rail, shipping, forestry, and hotels. What remained in state hands was essentially defense, police, courts, the core social welfare safety net, and public healthcare and education.
Even New Zealand, the most aggressive privatizer in the democratic world, kept healthcare and education public.
That's the key finding. No Western democracy has ever come close to the Nozickian night-watchman state. Even at peak Rogernomics, New Zealand retained a universal public health system and public schools. The political reality is that voters in every democracy demand these things, and no party that proposed eliminating them would survive an election. The libertarian minimum state is a philosophical benchmark, not a political destination.
Which is itself a data point worth sitting with. The closest approach in the real world still keeps the two biggest civilian functions in state hands. So the theory hits a democratic wall.
That wall has a name: the median voter. The median voter in every advanced democracy wants public healthcare and public schools. So let's talk about the United States, because it's number five on the list and there's a huge misconception here.
The American self-image is that of the rugged, small-government outlier.
The reality is that US public spending was thirty-seven percent of GDP in twenty twenty-four. Social Security alone is the largest government program in human history. Medicare covers sixty-five million Americans. The Veterans Health Administration is the largest integrated healthcare system in the country. Public schools educate fifty million children. The US Postal Service employs over half a million people.
The US is not a small-government country. It's a large-government country that's very good at pretending otherwise.
It's a large-government country with a cultural allergy to admitting it. The Fraser index ranks the US highly because American taxes are relatively low by OECD standards and the regulatory environment in many states is light. But in terms of absolute state involvement, the US is a massive mixed economy. The libertarian rhetoric is vastly out of step with the fiscal reality.
There's something almost theatrical about it. The performance of small government without the substance.
That theater has real consequences because it shapes what gets debated. Americans argue about whether government should do things that it's already doing on an enormous scale. It's like arguing about whether to buy a car while you're already driving one.
Alright, let's flip the lens. We've looked at the lean end of the spectrum. What does the other end look like? Which democracies maintain the most state control over industry and services?
France is the archetype. According to INSEE data, state-owned enterprises accounted for twelve percent of French GDP in twenty twenty-three. That compares to about two percent in the United States. The French state owns or controls EDF, the electricity giant that runs the country's fifty-six nuclear reactors. It owns SNCF, the national railway. La Poste, the postal service. It holds majority stakes in Orange, the telecom operator, and Air France-KLM. And in twenty twenty-two, it took EDF fully back into public ownership in a nine point seven billion euro buyout.
The renationalization is the part that catches my attention. The trend in most countries over the last forty years has been toward privatization. France is running the film in reverse.
The EDF case is instructive about why. The company was partially privatized in two thousand five, with the state retaining about eighty-four percent. But EDF was saddled with building a new generation of nuclear reactors — the Flamanville project — that went spectacularly over budget. Originally estimated at three point three billion euros, the final cost is projected at over thirteen billion. Meanwhile, the government was capping electricity prices to shield consumers from inflation. So you had a publicly controlled company with private minority shareholders, forced to sell electricity below cost while financing a construction disaster.
The worst of both worlds. Public obligations, private expectations.
The minority shareholders sued, arguing the price caps were effectively expropriation. The full nationalization was partly a solution to that legal mess. The state bought out the remaining sixteen percent and now owns EDF outright. It's a case study in how partial privatization can create governance nightmares.
What about the Nordic countries? Because they're often cited as proof that you can have a huge welfare state without massive state ownership of industry.
This is the Nordic hybrid model, and it's distinctive. Sweden and Denmark have government spending around forty-five to fifty percent of GDP — among the highest in the world. But they have relatively little direct state ownership of commercial enterprises. Sweden privatized most of its state-owned companies in the nineteen nineties and early two thousands. The government sold off its stakes in banking, telecom, and manufacturing.
Where does all that spending go?
Transfers and services. Pensions, healthcare, education, childcare, unemployment insurance. The Nordic model is high-tax, high-spend, but the state is a payer and a regulator, not an owner. It redistributes income and provides public services directly, but it doesn't run airlines or steel mills.
Norway is the exception though, right? Because of the oil.
Norway is the fascinating asterisk. The government owns sixty-seven percent of Equinor, the oil and gas company formerly known as Statoil. It was partially privatized and listed on the Oslo and New York stock exchanges in two thousand one, but the state kept its majority stake. And then there's the Government Pension Fund Global — the oil fund — which now holds one point five percent of every publicly listed company on earth, worth roughly one point seven trillion dollars as of twenty twenty-five.
One point five percent of everything. That's an extraordinary number.
It means the Norwegian state is a minority shareholder in about nine thousand companies worldwide. It's not controlling these firms, but it has a seat at the table. The fund has an ethics council that excludes companies involved in things like coal mining, tobacco, and certain weapons. So Norway is simultaneously a model of market-oriented social democracy and, through the oil fund, one of the largest state investors in global capitalism.
It's the sovereign wealth fund as quiet activist. The state doesn't run the companies, but it nudges them.
It raises a question about what "state control" even means. If you own one and a half percent of Apple and Microsoft and Nestlé, are you controlling those companies? Are you exerting influence? The Norwegian model is ownership without operational control, which is a different thing from the French model of direct state management.
What about other European examples? You mentioned Austria and Italy.
Austria has something called ÖBAG — the Austrian state holding company — which manages the government's stakes in several major firms. It owns just over thirty-one percent of OMV, the oil and gas company. It holds a significant stake in Telekom Austria. It controls Post AG, the postal service. The state is a major shareholder but typically not the sole owner. It's a portfolio approach.
Italy is one of Europe's heavyweights in state ownership. The Ministry of Economy and Finance directly controls Eni, the oil and gas giant, and Enel, the electricity utility, along with Poste Italiane. These are not small companies — Eni and Enel are multinational energy corporations. The Italian state also owns Ferrovie dello Stato, the rail operator, and holds a controlling stake in Leonardo, the aerospace and defense contractor.
If you're flying on an Italian military helicopter, the state built it, fueled it with state-owned oil, and the electricity powering the hangar came from a state-owned utility.
That's the Italian model. And it's worth noting that these companies are not basket cases — many of them are profitable, well-managed enterprises that compete internationally. Eni operates in dozens of countries. The question isn't whether state-owned enterprises can be well-run. The evidence shows they can be. The question is under what conditions.
Competition and commercial discipline. The state-owned enterprises that work — Singapore's Temasek portfolio, Norway's Equinor — tend to operate in competitive markets with professional management and hard budget constraints. The ones that fail tend to be monopolies with soft budgets and political interference in operational decisions.
Which brings us to the UK water situation, which is basically a masterclass in how to get privatization wrong.
This is the cautionary tale. In nineteen eighty-nine, Margaret Thatcher's government privatized ten regional water authorities in England and Wales. These became private monopolies — not competitive markets, but regulated monopolies with a captive customer base. The idea was that private management would be more efficient and private capital markets would fund infrastructure investment.
How did that work out?
By twenty twenty-five, the combined debt of the English water companies reached sixty billion pounds. Meanwhile, sewage spills into rivers and coastal waters increased by fifty percent since twenty nineteen. Several companies — Thames Water in particular — have been on the brink of insolvency. The regulator, Ofwat, has been criticized for being too weak to enforce investment requirements.
The private owners loaded the companies with debt, paid themselves dividends, and let the infrastructure rot.
That's the critique, and the data supports a lot of it. Between twenty ten and twenty twenty-two, English water companies paid out over seventy billion pounds in dividends while accumulating the sixty billion in debt. The financial engineering worked beautifully for shareholders. The pipes and treatment plants, less so.
This is the privatized natural monopoly problem. If you can't have competition — nobody's installing a second set of water pipes to your house — then you've just swapped a public monopoly for a private one, and the private one has a profit motive to underinvest.
The theory of privatization assumes competition drives efficiency. But for natural monopolies like water networks, rail infrastructure, or electricity grids, competition is structurally impossible. You're left with regulation as the only check on monopoly power, and regulation is a political process that can be captured or weakened.
The UK rail story has a similar arc, doesn't it?
British Rail was privatized between nineteen ninety-four and nineteen ninety-seven, split into infrastructure — Railtrack — and about twenty-five train operating companies. The infrastructure company collapsed in two thousand one after a series of fatal accidents and was replaced by Network Rail, a publicly owned entity. The operating franchises limped along with declining service quality until the government effectively renationalized the system in twenty twenty-one, bringing most operations under a new public body called Great British Railways.
A thirty-year experiment that ended roughly where it started, with a publicly controlled railway.
Several billion pounds of transaction costs and consultant fees along the way. The lesson isn't that privatization always fails. The lesson is that the structure of the market matters enormously. If you privatize a natural monopoly, you're not creating a market. You're creating a rent-extraction machine that requires vigilant regulation to prevent abuse. And that regulation is expensive and politically difficult to sustain.
Let's step back to the broader question, because I think a listener trying to make sense of privatization debates in their own country needs a framework, not just examples. What's the mental model for thinking about what should be public and what should be private?
I'd suggest three questions. First, is competition possible? If you can have multiple providers competing — like airlines or telecom — privatization tends to work. Second, is the service a natural monopoly? If yes, you need either public ownership or extremely strong regulation. Third, are the outcomes easily measurable? Healthcare is hard to privatize well because quality is difficult to measure and cost-cutting can be invisible until someone dies.
The measurability point is underappreciated. If you're privatizing garbage collection, you know if the trash is piling up. If you're privatizing prison healthcare, you might not know about the preventable deaths until an inspector general's report comes out three years later.
That's the information asymmetry problem. The private provider always knows more about its own operations than the government contract manager does. If the outcomes are hard to observe, the contractor can cut corners in ways that won't be detected until a crisis.
There's also the question of what you might call the "social cohesion" functions. Public schools aren't just about teaching math. They're about creating a shared civic culture. Public parks aren't just about grass. They're about common spaces where people of different incomes encounter each other.
This is the argument that some goods have intrinsic public value that markets don't price. A private park might be more efficiently maintained, but it can exclude people. A private school might have better test scores, but it doesn't create the same shared experience across class lines. These are philosophical questions about what kind of society you want, not just technical questions about efficiency.
The libertarian framework doesn't really have a vocabulary for that. Nozick's night-watchman state is built on individual rights, not collective goods.
That's its philosophical strength and its political weakness. It's internally consistent. It's just that almost nobody actually wants to live in it.
Let's talk about where this is all heading, because the ground is shifting under our feet. You mentioned AI and automation in your notes.
Here's the forward-looking question. The libertarian minimum state assumes most people can support themselves through market labor. But if AI and automation systematically reduce the demand for human labor — and I'm not saying this is certain, but it's a plausible scenario — then you have a structural problem. Millions of people who can't earn a living in the market, in a system where the state provides no income support beyond a residual safety net.
The political pressure for public provision would become overwhelming.
You might see even historically libertarian-leaning countries move toward something like a universal basic income or expanded public services. The minimum state is viable only when the labor market clears at a living wage for most people. If that condition breaks, the political consensus breaks with it.
Which suggests that the size of the state is partly a function of underlying economic conditions, not just ideology. When the market delivers broadly shared prosperity, the small-government argument is politically viable. When it doesn't, people demand public solutions regardless of what the philosophers say.
That's exactly the pattern we see historically. The welfare state expanded dramatically after the Great Depression because markets had visibly failed. The privatization wave of the nineteen eighties and nineties happened during a period of broad economic growth. The financial crisis of two thousand eight brought a new wave of state intervention. The pandemic brought another.
We're not actually arguing about principles. We're responding to shocks and then rationalizing after the fact.
I think that's too cynical, but only slightly. The principles are real — people disagree about the proper role of the state. But the direction of travel is heavily shaped by events. And right now, the UK water crisis and the French energy renationalization are events that are shifting the Overton window back toward state involvement in certain sectors.
Before we wrap up, what should a listener who wants to evaluate their own country's position actually do? What's the practical takeaway?
The Fraser Institute's Economic Freedom of the World data is publicly available and updated annually. You can look up your country's score on the Size of Government sub-index and compare it to others. But don't stop at the headline number. Dig into the components — government consumption, transfers and subsidies, government investment, and the top marginal tax rate. That tells you whether your country is high-spending because it employs lots of civil servants or because it writes lots of checks.
The ownership question is separate. A country can have low government spending but extensive state ownership, like Singapore, or high spending and low ownership, like Sweden. They're different dimensions.
The OECD also publishes data on state-owned enterprise share of GDP, which is the direct measure of public ownership. Between those two sources — Fraser for spending and regulation, OECD for ownership — you can triangulate where any country sits on the public-private spectrum.
The answer to the prompt's three-part question, in summary: the theoretical minimum state is defense, police, courts, and maybe a lighthouse. The closest democratic approach is New Zealand after Rogernomics, and even they kept healthcare and education public. At the other end, France, Italy, and Austria maintain significant state ownership of industry, with France fully renationalizing its electricity sector just a few years ago. And the whole spectrum is being reshaped in real time by infrastructure failures and technological disruption.
The one thing I'd add is that the real debate isn't state versus no state. It's which institutional forms work for which functions under which conditions. The evidence doesn't support the ideological extremes. It supports a case-by-case analysis that takes market structure, measurability, and social values seriously.
That's a less exciting bumper sticker than "government is the problem" or "privatize everything," but it has the advantage of being true.
Now: Hilbert's daily fun fact.
Hilbert: In the eighteen sixties, French chemist Louis Pasteur observed that fermenting beet juice rotated polarized light to the left, while fermenting cane sugar rotated it to the right — the first demonstration that fermentation chemistry produces optically active isomers specific to the sugar source.
My sauerkraut is doing geometry.
I'll be thinking about that all day.
Thanks to Hilbert Flumingtop for producing. This has been My Weird Prompts. If you want to dig deeper into any of this, the Fraser Institute data is free online, and we'll link the OECD state ownership reports at myweirdprompts.Leave us a review if you enjoyed the episode — it helps other people find the show. We're back next week.
