What would make this page better?
[ THE ULTIMATE GUIDE ]
Eighty-three arguments, for and against, side by side — and every one of them opens. Underneath: the greatest improvement in human welfare ever recorded, in five charts; thirty-three cases ranked best to worst, including the ones that are indefensible; and a system designer that tells you which real country you just built. Start here, though: almost nobody in this fight is arguing about the same thing.
[ 01 — WHAT IS ACTUALLY BEING ARGUED ]
Most capitalism arguments are two people using one word for two different systems and getting angrier that the other person won't concede. One of them is defending a Danish pharmacy and the other is attacking a Russian oligarch. This section is the part you have to read to make the rest of the page mean anything. It is not a warm-up. It is the argument.
Goods and services the world economy has produced since you opened this page. World output runs about $120 trillion a year — roughly $3.8 million every second, and about fifteen times as much per person as in 1820. Nobody planned any of it. Nobody could. The single most interesting fact about the modern world is that this number exists at all, and it is the thing any argument about capitalism has to explain.
THE ACTUAL AXIS // who the rules serve — not how big the government is
Notice what is not on this axis: tax rates, welfare, healthcare, unions, or how much you like billionaires. Those vary enormously within the first two boxes. Denmark taxes far more than the United States and sits one step to the left of it — because it is easier to start a firm there, easier to close one, and harder to buy a rule. The enemy of capitalism has never been socialists. It has always been incumbents.
Capitalism is private ownership of the means of production, with goods and capital allocated by prices arrived at through voluntary exchange. That is the definitional core, and it is the mirror image of the socialism definition: who owns the machine that makes the thing, and who decides what it costs.
But ownership and prices alone describe a system that can still rot, so the working definition needs two more legs. Free entry — anyone may start a competing firm — and hard budget constraints, meaning a firm that loses money actually dies. Strip either one and you still have private owners and nominal prices, and you get Russia in 1996 or American finance in 2008: capital with none of the discipline that made capital worth defending.
All four rest on a fifth thing that is not economic at all: rule of law. A contract you cannot enforce is a favour. A title you can be stripped of is a licence. Every one of the catastrophes in Section 04's bottom tier is a case where private ownership existed and one of the other four legs was missing — and that is the most useful thing on this page for telling a working market from a costume.
Private ownership, market prices, open entry, hard budgets, a modest state that mostly enforces contracts and provides public goods. Switzerland, New Zealand, Singapore, the United States on its better days. This is the thing defenders of capitalism are usually describing.
All of the above, plus a very large tax-and-transfer state on top. Denmark, Sweden, Norway, Finland, the Netherlands. Note carefully: these countries are more capitalist than the United States on business freedom, property rights, trade openness and ease of firm entry and exit. They tax the results harder and leave the machine alone. This is what most Americans who tell pollsters they like socialism are actually describing, which is one of the great ironies of the whole argument.
Private firms and market prices in most of the consumer economy, with the state owning the banks, the land, the energy majors, and picking national champions. China, the Gulf states, Korea in the 1970s, Japan in the 1960s. It has produced extraordinary catch-up growth and no frontier innovation to speak of, and the reason is in Section 02.
Private ownership where the rules are for sale: licences, tariffs, concessions, bailouts, occupational licensing, zoning, and regulatory moats written by the firms they protect. Russia, Hungary, Mexico's telecoms, and large slices of the United States. This is not a deviation from capitalism that critics invented — it is capitalism's own failure mode, and defending it is the fastest way to lose the argument you should be winning.
This deserves to be conceded loudly and early, because it is true and because pretending otherwise makes the pro-capitalist case look dishonest. "Real capitalism has never been tried" is exactly as bad an argument coming from this side as it is coming from the other one.
The United States runs roughly $11 trillion a year of combined federal, state and local spending — about 37% of GDP, within a few points of Canada and the United Kingdom. It socializes about 48% of health financing, runs the largest centrally planned organization on earth in the Department of Defense, grants twenty-year monopolies through the patent system, guarantees mortgages, insures deposits, subsidizes corn, restricts coastal shipping to American-built ships, and requires an occupational licence for something like one in five jobs.
It also, in 2008 and again in 2023, demonstrated that sufficiently large financial institutions do not face a hard budget constraint. By this page's own definition, that is a hole in the system, not a smear against it.
So the honest American question was never "capitalism, yes or no." It is which parts of this machine are load-bearing, which parts are barnacles, and what the rule is for telling them apart. The rest of this page is an attempt to find that rule.
Gallup's 2025 reading: 54% of Americans view capitalism positively, the lowest since they began asking, against 39% for socialism. Among adults 18–34 the two have crossed — 49% positive on socialism against 43% on capitalism. Among Democrats, socialism now outpolls capitalism 66% to 42%.
And then 81% of the same country views "free enterprise" positively, and "small business" polls near-universally. Those are not different systems. They are different words for the same system, and the gap between them is the whole political problem in one number.
The reasonable reading is not that voters are stupid. It is that "capitalism" has come to name the parts people can see — insurance denials, private equity in the local hospital, a landlord raising rent by an algorithm, a bailout — while "free enterprise" names the parts they like. If defenders of markets want the word back, the way to get it is not to insist the visible parts are fine. It is to attack the visible parts harder than the critics do, because most of them are cronyism, and cronyism is the enemy.
Forget the labels. If you want to predict how an economy will actually behave, there are four questions, and they do almost all the work:
1. Who owns the productive assets? Private owners, worker collectives, or the state. This determines who captures the gains and who eats the losses.
2. Who sets prices? Buyers and sellers, or an administrative body. This determines whether the economy has a functioning information system — which Section 02 argues is the single most important variable on this page.
3. Who is allowed to start a competing firm? Anyone, or whoever has a licence. This is the most underrated variable in the entire debate. Entry is how an economy corrects a mistake nobody at the top has noticed yet, and it is the first thing an incumbent buys off.
4. What happens to an organization that loses money? It dies, or it gets covered. János Kornai called the second one the soft budget constraint, and it is the quiet cause of more failure — socialist and capitalist — than any ideology.
Denmark answers: private, market, anyone, it dies. The USSR answered: state, administrative, nobody, it gets covered. Russia in 1996 answered: private, market, whoever the president likes, it gets covered — which is why it produced neither growth nor equality, and why "private ownership" on its own guarantees nothing at all.
"Socialism" was named by people who wanted it. "Capitalism" was named by people who didn't — the term enters wide circulation through nineteenth-century critics, Marx among them, and it was always a diagnosis rather than a manifesto. There was never a capitalist founding document, a capitalist party, or a capitalist five-year plan, because the thing is not a programme. It is what happens when a reasonably lawful society stops preventing people from trading.
That asymmetry explains a lot about the argument's shape. Socialism gets judged against its stated intentions; capitalism gets judged against its worst outcomes. Every socialist catastrophe can be called a betrayal of the idea, because the idea was written down first. Every capitalist catastrophe gets called an expression of the idea, because nobody wrote the idea down at all.
The fair standard is the one this page uses in both directions: judge each system by what it repeatedly produces under a wide range of leaders, cultures and starting conditions. On that standard, the results are in Sections 02 and 04, and they are not close — which is exactly why the parts where markets genuinely fail deserve to be taken more seriously, not less.
[ 02 — THE MECHANISM ]
Leonard Read's point in 1958 was not a cute one. No single human knows how to make a pencil — not the cedar logger, not the graphite miner in Sri Lanka, not the chemist who makes the lacquer, not the man who runs the ship. There is no pencil committee. Nobody is in charge. And a pencil costs a dime. The mechanism that does that is the single most underrated invention in human history, and it is an information system, not a moral philosophy.
THE GREAT ENRICHMENT // pick a measure of human welfare — any of them
Line chart of a selected global human-welfare measure over the last two centuries. Use left and right arrow keys to step through years. Full figures are in the data table below the chart.
// For roughly 100,000 years, human material life did not measurably improve. A peasant in 1700 lived about as well as a peasant in 1000 BC. Then, in a couple of centuries, in the places that developed secure property, open markets and free entry first, it went up by a factor of fifteen or thirty and never came back down — and every other place that later adopted the same institutions did the same thing, in a fifth of the time. Any theory of capitalism has to account for that graph. Most critiques of capitalism never mention it.
WHO IS ACTUALLY CAPITALIST // three different measurements — they do not agree
Horizontal bar chart comparing selected countries on the chosen measure. Full figures are in the data table below the chart.
// Flip between those three metrics and watch the ranking scramble. On government spending, Denmark looks like the least capitalist country on the list. On property rights and ease of doing business — the things that actually make a market work — Denmark beats the United States, and has for twenty years. On state ownership of firms, Singapore is above both of them. The metric you pick decides who you think is capitalist, and the metric that predicts outcomes is not the one either side argues about.
If you read nothing else on this page, read these seven. Everything in the ledger and everything in the historical record is downstream of them — including the last two, which cut against markets and which any serious defence has to carry rather than dodge.
Hayek's 1945 paper asked a question that sounds trivial and isn't: how does a tin manufacturer in Ohio find out that a mine collapsed in Bolivia? Nobody tells him. The price of tin goes up, and he uses less of it. He does not need to know why. He does not need to know Bolivia exists.
The knowledge an economy runs on is not statistical. It is dispersed, local, tacit and constantly changing — which loading dock is slow this week, which supplier cuts corners, which customer will actually pay. No agency can collect it, because most of it cannot be written down and all of it is stale by Tuesday. A price is a single number that summarizes the entire world's competing claims on one thing, updates continuously, and requires no one to understand it.
This is why the calculation debate matters even now. Mises argued in 1920 that without market prices for capital goods, a planner literally cannot compute whether a factory creates or destroys value, because there is no non-arbitrary way to compare a ton of steel against a thousand hours of labour. Sixty years of attempts produced enormous evidence and no counterexample. Gosplan balanced about 2,000 aggregated commodity groups in an economy containing on the order of 12 million distinguishable products.
Everyone defends profit and almost nobody defends loss, which is backwards. Profit says keep going. Loss says stop, you are destroying value, those resources are worth more somewhere else. A system with profit and no loss is not capitalism running hot — it is capitalism with the sensor removed.
Kornai's soft budget constraint is the name for what happens then. When firms know they will be covered, managers stop economizing and start bargaining — for inputs, for allocations, for another year. Capital never gets released from what isn't working. Cost figures stop meaning anything, because costs are only real if someone eats them.
This is the test that catches Soviet steel mills and Wall Street banks with the same instrument, and it is why this page docks the United States for 2008 in exactly the way it docks Hungary in 1985. Roughly one in ten American businesses closes each year. That is not a symptom of a sick economy. It is the immune system.
Schumpeter's insight was that capitalism's defining feature is not efficiency at a moment in time — it is the constant destruction of the existing arrangement from within. Kodak invented the digital camera and died of it. Blockbuster was offered Netflix for $50 million. Nokia had half the global handset market in 2007. Sears was Amazon for a century.
No committee killed any of them, and no committee could have. Each was destroyed by a customer choosing something else, one at a time, millions of times, in a process nobody had to authorize. That mechanism is the difference between an economy that has televisions and an economy that has better televisions every year.
And the cost is real and lands on specific people. The word "destruction" is in the phrase on purpose. The honest pro-capitalist position is not that this is painless — it is that the pain should be met with income support and retraining rather than by protecting the incumbent, because protecting the incumbent preserves the job and cancels the gain, which is precisely the deal Britain's coal industry, America's steel tariffs, and India's licence raj each took.
The number that predicts whether an economy corrects its mistakes is not the number of competitors today. It is whether a new competitor can appear tomorrow. An industry with three firms and free entry behaves better than an industry with thirty firms and a licensing board.
This is why India's licence raj is the cleanest natural experiment in the file: same people, same markets, same private ownership — but from 1951 to 1991 you needed government permission to expand output, change product, or import a machine. Growth ran around 3.5% a year. After 1991 it ran 6–7%+, and extreme poverty fell from roughly 45% to roughly 10%. Nothing changed about Indians. What changed was who was allowed to try things.
And this is where the American system is genuinely weakest. Occupational licensing covers about a fifth of the workforce. Zoning makes it illegal to build the obvious thing on most urban land. Certificate-of-need laws let existing hospitals veto new ones. None of that is socialism. All of it is incumbents using the state to close the door behind them, and it is a betrayal of capitalism by people who describe themselves as capitalists.
An owner is the person who gets whatever is left after everyone else is paid — the residual claimant. That position is what makes someone care about the thousand small decisions no contract can specify and no supervisor can observe: whether to check the weld, whether to chase the invoice, whether to fix it now or in spring.
When nobody is the residual claimant, those decisions get made by people whose actual incentive is to satisfy a metric. That is the principal-agent problem, and it is not unique to state ownership — it is the standing disease of large public companies too, which is exactly why executive pay is tied to equity and why founder-run firms behave differently from professionally managed ones.
The strongest evidence is agricultural and it is enormous. China de-collectivized farming after 1978 — the same land, the same people, the same tools, plus the right to keep the surplus — and food output rose more than 40% in six years. Nothing technological happened. Somebody just became the person who got what was left.
A defence of markets that cannot name their failures is advertising. There are four well-defined ones, they are in every economics textbook including the most promarket ones, and every serious argument the other column makes is a specific instance of one of them.
1. Externalities. A price reflects what the two parties to a sale care about. It is structurally blind to everyone else — the downwind town, the fishery, people who do not exist yet. Climate change is the largest externality in the history of pricing, and left alone the market's correct level of abatement is zero forever.
2. Public goods. Non-excludable and non-rival: national defence, basic research, mosquito control, a legal system. If you cannot charge for it and my use doesn't diminish yours, no firm will supply enough of it. This is why DARPA and the NIH exist, and no serious market economist argues otherwise.
3. Information asymmetry. When the seller knows what the buyer cannot evaluate, competition rewards the wrong thing. Akerlof's lemons, Arrow on medical care — you cannot comparison-shop an ambulance — and the entire architecture of food safety, drug approval and securities disclosure.
4. Market power. Textbook markets assume nobody sets the price. Real ones contain natural monopolies, network effects that tip to one winner, one-hospital towns, and non-compete agreements. A monopolist is not running a market; it is running a private planning system with better branding.
Hold both of these at once. Prices are the best allocation mechanism anyone has ever found, and there are four defined conditions under which they demonstrably fail. That is not a contradiction — it is the actual state of the evidence, and the verdict at the end of this page is built on both halves.
One extra percentage point of annual growth, sustained for a century, is the difference between your great-grandchildren being 2.7 times richer than you and 7.2 times richer. There is no redistribution policy in the history of government that moves anything by a factor of seven.
This is the argument that ought to end the pretence that growth is a right-wing hobbyhorse. Everything the left wants is easier in a richer country: pensions, hospitals, clean air, shorter hours, better wages, absorbing refugees. A 2% economy funds a welfare state that a 1% economy has to keep cutting, which is the actual reason European social models look strained and Nordic ones do not.
And it runs the other way with equal force. A system that redistributes brilliantly and grows at zero is dividing a fixed pie forever — which is precisely the trap the stagnation tier of Section 04 sits in, and why "the pie has to be baked before it can be divided" is not a slogan but arithmetic.
[ 03 — THE LEDGER ]
Eighty-three arguments, matched left and right. Every card opens into the actual numbers plus the steelman — the strongest response the other side has to that specific point, written by someone who takes it seriously. Filter by topic, search the text, and mark which arguments actually move you. The scale keeps score.
Mark arguments as MATTERS or DECISIVE as you read and the scale will tip. Nothing weighed yet — the beam is level.
▲ THE CASE FOR
▼ THE CASE AGAINST
[ 04 — THE RECORD ]
Nobody has to speculate about this either. Market economies have been built dozens of times across three centuries, on every continent, by decent people and monstrous ones, in rich countries and desperately poor ones. Here is the file — ordered by what it did to the people living in it, best first — with what each one actually did, why it produced what it produced, and the strongest honest counterpoint. The last tier is the one defenders of capitalism should read hardest: it is the tier where private ownership and market prices did the damage themselves, and no amount of "that wasn't real capitalism" gets you out of it.
THE NATURAL EXPERIMENTS // same people, same land, same starting point, different rules
Line chart of GDP per capita over time for the selected pair of economies, with a marked year where their systems diverged. Use left and right arrow keys to step through years. Full figures are in the data table below the chart.
[ 05 — THE AMERICAN QUESTION ]
"Capitalism: yes or no" is not a question anyone can answer, because it is seven questions wearing a suit. So separate them. Move the controls, and the panel on the right will tell you what the system you just specified does — and, at the bottom, which real country in history you have most closely rebuilt. Two of these dials have nothing to do with left and right, and they are the two that decide almost everything.
Can you enforce a contract, keep what you built, and lose a case against the government? The precondition for all of it. Denmark ≈95, the US ≈80, Russia ≈22.
The most underrated variable in the entire debate. Entry is how an economy corrects a mistake nobody at the top has noticed yet — and it is the first thing an incumbent buys off.
Access to the rest of the world's goods, ideas, money and workers. Every catch-up miracle in Section 04 turned this dial up first.
The floor under people. This is the welfare-state dial, it is not the same lever as ownership, and turning it up does not turn the others down.
// A teaching model, not a forecast. Weights and formulas are in sources & methodology.
Whatever you conclude about theory, these four numbers are the scale of the thing being argued about. They get quoted wrongly by both sides constantly, so here they are with the sourcing attached.
Share of humanity living in extreme poverty — 1820 against today. In 1990 it was still 37.9%. Roughly 1.1 billion people have left extreme poverty since then, about three quarters of them in China and India, in the exact decades those two countries opened to prices, entry and trade. This is the largest improvement in human material welfare ever recorded, and it happened inside one lifetime.
Fall in the labour price of light between 1800 and the 1990s, per Nordhaus. An hour of work bought roughly ten minutes of reading light in 1800; it now buys years of it. Pick almost any physical good — a shirt, a chicken, a mile travelled, a phone call, a watt of solar — and the same graph appears. Wage statistics understate this badly, because they price a 2026 basket in 1970 categories.
Global life expectancy at birth, 1800 against today. Child mortality fell from roughly 43% of all children dying before age five to about 3.7%. Literacy went from about 12% of adults to about 87%. Every one of those curves bends at the same time and in the same places, and none of them has anything to do with who you voted for.
Share of US GDP that is government spending — about $11 trillion a year across federal, state and local. Add licensing for a fifth of jobs, twenty-year patent monopolies, a mortgage guarantee, and a demonstrated bailout for large banks. Whatever this system is, "unregulated free-market capitalism" is not a description of it, and arguments that assume it is are aimed at nothing.
Not hypothetical either. Americans are running experiments right now, and it is worth watching them with the analytic framework above rather than a team jersey.
THE MOST IMPORTANT DOMESTIC PRICE IN THE COUNTRY
Housing costs are the single strongest emotional fuel for anti-capitalism in America, and the binding constraint is government restriction of supply — zoning, parking minimums, height limits, discretionary review, and neighbours with a veto. Where entry was allowed, prices behaved: Austin, Minneapolis and Tokyo all built and all saw rents flatten or fall while comparable cities that didn't build did not. This is a market failure only in the sense that the market was made illegal.
2005–PRESENT
Utility-scale solar costs fell roughly 90% since 2010 and lithium-ion battery packs by a similar order; American natural gas went from a projected import crisis to the world's largest exporter in a decade. Each involved public research money at the front end and thousands of competing private firms grinding cost curves down at the back end. It is the exact division of labour Section 02 argues for, and the fastest decarbonization mechanism yet found is a falling price.
SINCE 2008, ARGUABLY SINCE 1984
The United States runs hard budget constraints for restaurants and soft ones for large financial institutions. Profits are private and catastrophic losses are socialized. This is a genuine, unanswered indictment of the American system on exactly the grounds this page uses in its favour — and any honest reader of Section 02 has to hold it against the market column rather than wave it away.
2010–PRESENT
Voluntary exchange is the moral engine of the whole system, and it assumes the buyer knows what they are buying. A product optimized by machine learning against a teenager's dopamine response is a genuinely new object, and "nobody made you use it" is a weaker answer than defenders of markets want it to be. It is the information-asymmetry failure mode running at a scale Akerlof never imagined, and markets have not solved it on their own.
[ 06 — THE HONEST VERDICT ]
If you read all eighty-three arguments and the whole record, a pattern emerges that is sharper than "capitalism good." Nearly every serious anti-capitalist diagnosis that survives contact with evidence is about distribution and externality — who gets what, who is hurt by transactions they were not party to, and who has power over whom. Nearly every anti-capitalist prescription that survives contact with evidence leaves the price system standing. And nearly every capitalist catastrophe in Section 04's bottom tier is a case where one of the four legs was missing. Six rules fall out of that, and they are what this page is actually for.
A price tells you what is scarce. It says nothing whatever about what anyone deserves, and it was never designed to. Treating a market outcome as a moral judgment is the mistake that discredits capitalism; treating a moral objection as a reason to switch off the measurement is the mistake that destroyed Venezuela.
Loss is the half of the feedback loop everyone wants to skip. A bailed-out bank is a state enterprise with better bonuses, and the model on this page docks it the same way. Bankruptcy is not capitalism failing; it is capitalism working, and the twenty largest firms should be no more exempt than the corner shop.
Occupational licensing, zoning, certificates of need, tariffs, non-competes, and regulation written by the regulated. Every one of these is a private actor using the state to close a door behind them, and every one is more damaging to actual capitalism than any socialist has ever managed. Open entry is the whole ballgame.
Carbon prices, liability, disclosure, food safety, drug approval. An externality is not an argument against markets — it is an argument for a market that includes the missing party. Price the thing. Do not ban the transaction, and do not pretend the harm isn't there.
Not because it is a concession to the other side, but because the system is a machine for reallocating people away from what no longer works, and a society will not tolerate that machine without a landing. Denmark's flexicurity is the point: easiest firing rules in Europe, most generous unemployment support in the world, and both are load-bearing.
Not tax rates. Not the size of the state. Capture. Russia had private ownership and market prices in 1996 and produced an oligarchy; Denmark has punishing taxes and produces the freest business environment in Europe. The corruption dial in the designer above is the one that moves the grade furthest, and it is the one nobody campaigns on.
A market price is the only device anyone has ever built that tells eight billion strangers what is scarce — and it will never once tell you what anybody deserves.
Both halves of that sentence are true, and almost everybody in this argument only believes one of them. If you only believe the first half you end up defending outcomes no decent person should defend, on the grounds that a market produced them — and you end up defending cronies, bailouts and rent-seekers as though they were entrepreneurs. If you only believe the second half you end up doing what Venezuela did: seizing the machine that generates the information, in the sincere belief that the information was the problem.
So here is the case, stated as plainly as it can be. For roughly a hundred thousand years, nearly every human being who ever lived was poor, illiterate, and likely to bury half their children. Then a handful of societies stumbled into a specific combination — secure property, prices set by exchange, open entry, real failure, and law that applied to the powerful — and within two centuries the share of humanity in extreme poverty fell from three quarters to under a tenth, life expectancy went from 29 to 73, and a person of ordinary means acquired conveniences that emperors could not buy. Every country that later adopted that combination did the same thing faster. Every country that abolished it produced the same shortages in the same order. That is not an ideological claim. It is the most replicated result in the social sciences.
And none of that makes the grievances fake. The externalities are real, the capture is real, the insecurity is real, and the opioid epidemic, the leaded gasoline, and the plantation were all run for profit by private owners responding to prices. A defence of capitalism that cannot say those sentences is not a defence, it is a sales pitch. The correct response to each of them is a specific institutional fix, and in every case the fix keeps the price system and changes what the price is required to include.
Be relentlessly pro-market about production and relentlessly ambitious about distribution at the same time. Those were never opposites. Treating them as opposites is how a country ends up with neither the growth nor the floor — and how the best machine anyone ever built for lifting people out of poverty ends up defended mainly by the people who have already been lifted.
On bias: this page has a thesis, stated plainly rather than smuggled in — that the combination of private ownership, market prices, open entry, hard budget constraints and rule of law is the most effective arrangement anyone has found for producing material welfare, and that its documented failures are real, specific, and fixable without dismantling it. The arguments on the other side are written to be as strong as I can honestly make them, several of them are ones I think are simply correct, and the historical record here includes the atrocities as well as the miracles. Where the evidence is contested — pre-industrial income reconstructions, colonial death tolls, the productivity-pay gap, whether decoupling is real — it is presented as contested. Weights you set in the ledger are stored in your own browser and go nowhere. If you think an argument is missing or wrong, the button in the corner goes straight to me.
// jeremymsparks.com — last reviewed August 2026