What Markets Do Right and Wrong
On the price that carries knowledge, the price that carries the crowd, and the bandwagon that turns one into the other
Keywords: markets; prices; Hayek; Keynes beauty contest; bandwagon; information cascade; bubbles; efficient markets
A price is a small thing — a number — and it is asked to do one of the most extraordinary jobs in the world. When you pay what you pay for a loaf of bread, the number you hand over has quietly gathered inside itself the state of a wheat harvest in Kansas, the price of diesel, the wage a baker will accept, the rent on the shop, the tastes of ten thousand strangers who did or did not want bread this week, and a hundred other facts no single human being knows or could know. No one computed that number. No committee set it. It condensed, on its own, out of the choices of people who never met and never coordinated, and it tells the baker how much to bake and you whether to buy — and it is very often, astonishingly, about right. That is what markets do right, and it is a genuine wonder, and any honest account has to begin by kneeling to it before it says a word against markets.
But the same number does a second job, and the second job is where the trouble lives, and the two jobs look identical from the outside. Sometimes a price stops carrying facts about the world and starts carrying only the crowd’s opinion of itself — rising because it is rising, bought because it is being bought — and at that point the miracle inverts into a mania. The number that was a signal becomes a mirror pointed at a mirror. And because it is the same number, wearing the same decimal point, almost no one can tell, in the moment, which job it is doing. That is the whole subject: one mechanism, two functions, indistinguishable at a glance, and the difference between a civilisation coordinating itself and a mob trampling itself.
What the price gets right: the knowledge no one has
Start with the miracle, stated precisely, because you cannot understand the failure without first understanding the thing that fails.
The deep problem of any economy is that the knowledge required to run it does not exist in any one place. It is scattered — in millions of heads, each holding a scrap: this farmer knows his field flooded, that engineer knows a cheaper alloy, this shopper knows she has gone off beef, that machinist knows his lathe is failing. No planner, however brilliant, however well-equipped with computers, can gather these scraps, because most of them are fleeting, local, tacit, and never written down; the man who holds the scrap often could not tell you he holds it. This is the permanent, insurmountable obstacle to running an economy from a central desk, and it is why every attempt to do so has produced shortage and waste on a heroic scale — not because the planners were stupid or wicked, though some were, but because they were blind, and no amount of good intention cures blindness about facts you cannot possibly collect.
The price is the answer to the unanswerable. It does not gather the knowledge; it summarises it, without anyone having to know it. When the flooded field and the failing lathe and the shopper’s change of heart all press on the market, the price moves, and the movement transmits the upshot of a million scraps to everyone who needs to act on them — and to no more than the upshot, which is all they need. The machinist in Ohio adjusts his order because tin got dearer, and he never learns that a mine in Bolivia collapsed, and he does not need to learn it; the price told him the only thing his decision required. This is the market’s true genius: it is a device for acting rightly on knowledge you do not have, a computer made of strangers, and nothing else humanity has built comes close to it as a way of coordinating people who cannot possibly coordinate themselves. When someone tells you markets are efficient, this — and only this — is the defensible thing they mean: that the price aggregates dispersed knowledge no central mind could hold.
Notice the shape of this healthy process, because it is the tell. When a price carries knowledge, it runs on negative feedback: a high price cools demand and calls forth supply, which pushes the price back down; a low price does the reverse. The system leans against its own movement. It wanders around the underlying value and keeps returning to it, self-correcting, stable, like a marble in a bowl. That leaning-against is the sound of a market doing its job. Hold the image, because the disease is precisely its inversion.
What the price gets wrong: the mirror pointed at a mirror
Now the second job, and here the great witness is not the market’s champion but its shrewdest observer, who understood exactly how the miracle curdles.
Imagine, said Keynes, a certain newspaper competition. Readers are shown a hundred photographs of faces and asked to pick the six they find prettiest — but the prize goes to the reader whose choices best match the average of all readers’ choices. What does a clever competitor do? He does not choose the faces he finds prettiest; that would be a fool’s move. He chooses the faces he thinks others will find prettiest. But the others are doing the same thing, so really he must choose the faces he thinks others think others will find prettiest — and so on, down a hall of mirrors with no floor. The competition has nothing to do with beauty. It is a pure exercise in guessing the guesses of people guessing your guesses. And a great deal of what happens in a market, Keynes observed, is exactly this: not an estimate of what a thing is worth, but an estimate of what other people will shortly think it is worth, which is an estimate of their estimates of yet others’ estimates, all the way down.
When a market slips into this mode, the price detaches from the world and locks onto the crowd, and the feedback flips from negative to positive. Now a rising price does not cool demand — it inflames it, because the rise is read as a signal that others know something, that the thing is going up, that one must get in before it goes higher. Buying begets buying. The rise causes the rise. The marble is no longer in a bowl; it is on a dome, and every movement away from the centre accelerates. This is the bandwagon, and it is not a metaphor but a mechanism — the same aggregation that carried knowledge now amplifying belief, the same wisdom of crowds inverted into the madness of them.
The economists have a colder name for one engine of this: the information cascade. Suppose each person has a little private information and can also see what others do. The first few people act on their own information. But once enough have acted the same way, the next person reasons — rationally — that the crowd’s collective information must outweigh his own small scrap, and so he ignores his own signal and follows. So does the next, and the next, each rationally discarding what he knows in deference to what everyone is doing. And now the terrible feature: once people stop acting on their private information and start merely copying, the crowd stops gathering information at all. It is no longer aggregating a million scraps; it is echoing one early accident, amplified. A cascade can be enormous, confident, unanimous — and completely wrong, because after the first few movers, no one added any knowledge to it. Everyone was looking at everyone else. The mirror had swallowed the world.
There is a cruel twist to this, and it is among the deepest things anyone has said about financial manias: the calm itself breeds the storm. When a market has been stable for a long stretch, the memory of the last crash fades, the cautious are gradually punished for their caution by the gains they miss, the reckless are rewarded, and the whole system quietly rewrites its sense of what is safe. Debt that would have terrified everyone a decade earlier comes to seem prudent; leverage that once looked insane becomes ordinary, then almost mandatory, because the firm that refuses it is out-competed by the firm that embraces it. Stability, in other words, is not the opposite of the bubble. It is the incubator of it. The longer the good weather lasts, the more completely the crowd forgets that weather changes, and the larger the position it builds on the assumption that the sun is a fixed feature of the sky — so that the eventual reversal, when it comes, finds everyone leaning the same way at once, and the stampede for the exit is exactly as unanimous, and exactly as blind, as the stampede in. The market does not bubble because people are stupid. It bubbles because people are adaptive, and the thing they adapt to is the absence of the very danger that is, all the while, accumulating underneath them.
This is how you get tulip bulbs worth houses, and companies worth billions for having no revenue and a clever name, and mortgages that could not possibly be repaid bundled into securities that could not possibly fail, and, always, a chorus of intelligent people explaining why this time the dome is a bowl. The bandwagon does not feel like madness from inside. It feels like knowledge — because everyone around you agrees, and their agreement is real, and you mistake the fact of their agreement for evidence about the world. But their agreement is a fact about them. A stock is not made more valuable by a stampede toward it, exactly as snow is not made whiter by a show of hands. The price is going up because it is going up, and that sentence, when it is true, is the whole of the disease.
The steelman: markets fail better than anything else
I have to stop and make the market’s defence at full strength, because the account so far could be mistaken for a case against markets, and it is not. It is a case for understanding them, which is a different and more demanding thing, and the strongest reply to the bandwagon is genuinely strong.
The reply is this: yes, markets bubble, panic, cascade, and stampede — and they are still the least-bad mechanism we have, because their errors are visible, dated, and self-correcting, while the errors of the alternative are hidden, permanent, and enforced at gunpoint. A bubble pops. A fraud is eventually exposed when reality fails to arrive. A mania corrects, often brutally, but it corrects, and the correction restores the price to something like the truth. Compare the central planner, who makes exactly the same kinds of errors — mispricing, misallocation, confident wrongness — but whose errors do not self-correct, because there is no price to contradict him and no competitor to prove him wrong; his mistake simply becomes the five-year plan, and the shortage it causes is not a signal to be heeded but a failure to be denied, blamed on saboteurs, and repeated. The market’s manias are loud and embarrassing and temporary. The plan’s errors are quiet and dignified and last a decade, and people starve inside them. A system that corrects itself through visible crashes is categorically safer than one that cannot correct itself at all, and the man who points at the crash as proof that markets fail has forgotten to ask what the alternative’s failures look like — which is famine, and no crash to end it, because there is no market to crash.
This is correct, and it is decisive against the fantasy of replacing markets with command. Grant it completely. Then notice exactly what it does and does not establish — because it is a defence of the market’s self-correction, and self-correction has limits that the defence quietly steps over.
Where self-correction stops
Self-correction is real, but it is neither free nor fast, and there is a whole class of failures it does not touch at all.
It is not free: a bubble destroys real capital and real lives on its way to correcting. The factories built for demand that was never real, the people who bought at the top and are ruined, the years of investment poured into the mania and set on fire when it collapses — these are not erased by the correction. “The market recovered” is cold comfort to the generation wrecked in the interval, and “in the long run it self-corrects” earns the reply that in the long run we are all dead. That the marble eventually rolls back does not un-break the things it broke on the way.
And there is a deeper limit, which is the important one: the market corrects its belief errors far better than its structural ones. A bubble is a belief error — the crowd is wrong about value, reality eventually asserts itself, the price falls. But some failures are not the crowd being wrong; they are the mechanism itself pointing in the wrong direction, and no amount of self-correction fixes a mechanism that is working exactly as built. The market never prices what escapes the transaction: the factory that dumps its filth in the river has no line on its ledger for the poisoned town downstream, because the town was not a party to the sale, and so the price — the honest, efficient, self-correcting price — is systematically wrong wherever costs land on people who are not at the table. The market tends toward monopoly, because winning firms buy or crush rivals, and a monopoly kills the very competition that made the price meaningful, and it does not self-correct because the monopolist’s whole project is to prevent the correction. Fraud pays, often, and often is not caught, because a lie about quality can be more profitable than quality, and the market rewards whatever sells until the truth arrives, which it sometimes never does. These are not the crowd being temporarily wrong. These are the mechanism doing precisely what it is built to do — pricing what is traded, rewarding what wins, ignoring what is not on the ticket — and producing, reliably, results no one would choose. Self-correction cannot reach them, because there is nothing malfunctioning to correct. The machine is working. That is the problem.
And there is a last category the price cannot reach, the mirror image of the externality: not the cost the market ignores but the benefit it cannot capture. Some of the most valuable things a society can have are goods that, once made, everyone can enjoy and no one can be charged for — clean air, the basic research whose fruits anyone may use, the lighthouse that shines on every ship including the ones that never paid. Because no seller can fence these off and bill for them, no market will produce them in anything like the quantity they are worth; each person, reasoning perfectly, waits for someone else to bear the cost, and so no one bears it, and the good goes unmade though everyone would gladly have had it. The same logic runs in reverse across a shared resource that anyone may deplete — the common pasture, the fishery, the aquifer — where each user takes the whole of what he takes and shares only a sliver of the cost of the taking, so that the rational course for every individual is to consume it toward destruction, and the resource that could have fed all of them indefinitely is exhausted by all of them at once. Neither of these is a belief error the crowd will eventually correct. They are structural, permanent, and invisible to a mechanism built to price only what can be owned and sold — which is exactly why they, too, wait on a rule from outside the market to supply what the market cannot see.
The teeth: the price is not an oracle, and it is not a fraud
Which brings us to the two errors that bracket this whole subject, mirror images, both lazy, both false.
The first is the market-worshipper’s, and it reads the miracle as an oracle: the price is always right, whatever is profitable is good, the market’s verdict is the truth of value, and to question a price is to imagine you know better than the wisdom of millions. This mistakes what the price is. The price carries knowledge and crowd, mixed, and the mixture is invisible, so that a number swollen by a stampede wears the identical mask as a number condensed from a million honest facts. “The market rewarded it” is not the same sentence as “it was correct,” any more than “he won the vote” is the same as “he was right” — in both cases a real aggregation of real people has occurred, and in both cases the aggregation may be carrying wisdom or carrying a mob, and nothing on the surface tells you which. A rising price is evidence of buying. It is not evidence of worth. Treat it as an oracle and the oracle will lead you, confidently and with the roar of the crowd behind it, straight over the edge of the dome.
The second is the market-hater’s, and it is the same error inverted: because markets bubble and cascade and fail, they are worthless, a casino, a fraud, a thing to be abolished and replaced by the rule of clever people with a plan. This throws out the knowledge-aggregation miracle — the one genuinely irreplaceable thing, the computer made of strangers — because the same mechanism that performs the miracle also, sometimes, herds. It is the reasoning of a man who burns his house down because the wiring occasionally sparks. The failures are real; they are not the whole story, and they are not an argument for blindness, which is what abolishing the price actually buys you.
Both errors make the identical mistake: both treat the market as a single thing with a single verdict — an oracle to obey or a fraud to abolish — when it is a mechanism that does one job superbly and another job disastrously, using the same number for both. The entire discipline of thinking clearly about markets is the discipline of asking, of any given price, the question the worshipper refuses and the hater never gets to: what is this number made of — knowledge, or the crowd? You cannot always answer it. But the man who is even asking it has already escaped both the cult and the mob, and he is the only one in the room who might see the dome for what it is while everyone around him is calling it a bowl.
So markets do right the thing nothing else can do: they let strangers coordinate on knowledge no one holds. And markets do wrong the thing crowds always do: they mistake their own momentum for a message from the world. The price is a miracle and a mirror, and it never tells you which one it is being. That is the first thing to know about markets, and it turns out to be the first thing to know about democracies too — but that is the next essay.
Ask not whether the price is rising. Ask what the rising is made of.
Is it so?
State. Classify. Done.
The scaffolding is standard and checkable: Hayek’s account of the price system as a mechanism for using knowledge that exists in no single mind; Keynes’s newspaper beauty-contest, in which one guesses others’ guesses rather than the underlying value; the theory of rational information cascades, in which people optimally ignore their own information and follow the herd, so that a large confident consensus may carry almost no aggregated knowledge; and the standard treatment of externalities, monopoly, and fraud as failures the price mechanism does not self-correct because it is functioning as designed. Nothing here rests on an unnamed authority; every claim is meant to answer to the world it describes.