9.11 — How Markets Actually Work

In 1958 the economist Leonard Read wrote a short essay in the voice of a pencil, which begins by observing that not a single person on the face of this earth knows how to make it.

The wood is cedar from Oregon, cut with saws made of steel from iron ore, by men who drink coffee grown in Brazil. The graphite is from Sri Lanka, mixed with clay from Mississippi. The eraser is a rubber-like product made from rapeseed oil from Indonesia and sulphur chloride. The brass ferrule needs zinc and copper from mines somewhere else. The lacquer, the paint, the glue and the machinery each have their own chains.

Millions of people, none of whom knows more than a fraction of the process, none of whom is directed by anyone, and most of whom would dislike each other's politics, cooperate to produce a pencil that costs a few rupees.

Nobody is in charge. This chapter is about the mechanism that makes that happen, and about where it fails.

Supply, demand and what a price is

A demand curve says how much people will buy at each price. It slopes down: cheaper means more buyers and more purchases per buyer.

A supply curve says how much producers will offer at each price. It slopes up: higher prices make it worth producing more, including by higher-cost producers who could not cover their costs before.

Where they cross is the market price, and the important thing about that point is what happens away from it. Above it, more is offered than is wanted, so unsold stock accumulates and sellers cut prices. Below it, more is wanted than is offered, so buyers compete and prices rise. The system self-corrects without anyone directing it.

And the price is not simply a number. It is a message carrying two instructions at once.

A rising price tells buyers: this is scarcer, use less, or find a substitute. It tells producers: this is valuable, make more, and it is now worth your while even if your costs are high.

Neither party needs to know why. If a drought in Brazil destroys the coffee crop, a shopkeeper in Nagpur does not need to hear about the drought. The price rises, some customers switch to tea, and some producers elsewhere plant more coffee. The information travels without the story travelling.

This is Hayek's argument (Chapter 9.6) and it is the strongest case against central planning: the knowledge required to run an economy is dispersed among millions of people, most of it local, tacit and constantly changing, and no committee can assemble it.

Which is also why price controls do specific damage. A price ceiling below the market price does not create supply; it creates a shortage, and the shortage is then rationed by queueing, by connections, or by a black market. Diocletian found this in 301 CE (Chapter 9.8) and every government since has rediscovered it.

Elasticity, and who actually pays a tax

Elasticity measures how much quantity responds to a price change.

Inelastic demand means people buy nearly the same amount whatever the price — insulin, salt, petrol in the short run, cigarettes for an addicted smoker.

Elastic demand means small price changes produce large quantity changes — one restaurant among many, one brand of biscuit, a holiday destination.

What determines it: whether substitutes exist, whether the item is a necessity, how much of your budget it takes, and how much time you have to adjust. Petrol demand is inelastic this week and elastic over ten years, because eventually people move house, change car or change job.

And the payoff is the tax incidence result, which is genuinely counter-intuitive.

Who legally pays a tax has nothing to do with who actually bears it. The burden falls on whichever side is less able to walk away.

Tax petrol, where demand is inelastic and consumers cannot easily stop buying, and consumers bear nearly all of it. Tax a luxury good with elastic demand and the producer bears most of it, because raising the price loses the customers.

The same logic applies to labour taxes. Employer and employee contributions to a provident fund are formally split, and the economic burden falls according to elasticity, not according to the form on the payslip.

This is why "we will make companies pay" is usually an incomplete sentence. A company is a legal fiction; the burden lands on its shareholders, its workers or its customers in proportions set by elasticities, and which one is an empirical question rather than a political choice.

Competition and its absence

Perfect competition — many buyers and sellers, identical products, free entry, full information — does not exist and is a benchmark. What matters is how close a real market is.

Monopoly. One seller. The monopolist maximises profit by restricting output and raising price, which is a transfer from consumers and, more importantly, a loss of transactions that would have benefited both parties. That lost value is called deadweight loss and it is the actual economic case against monopoly, distinct from the fairness case.

Natural monopoly is the case where competition genuinely does not work: where fixed costs are enormous and marginal costs tiny, one producer is cheapest. Water pipes, electricity grids, railway track. Building three competing water networks down one street is waste. The answer is regulation or public ownership, not competition, and every country does some version of it.

Oligopoly — a few large sellers — is the common real case, and it produces strategic behaviour: firms respond to each other's decisions, which is what game theory analyses, and which can produce price leadership, tacit coordination, or fierce competition depending on conditions.

Network effects create a modern version. A service becomes more valuable the more people use it — a phone network, a payments system, a social platform, a marketplace. This tends toward one dominant provider, and it is why competition policy for technology is so difficult.

And the case for competition is not that competitive markets are pleasant for producers. It is that competition is the only reliable mechanism for making producers serve customers rather than themselves. Where it is absent — a state monopoly, a protected incumbent, a cartel — quality falls and prices rise, consistently, regardless of who owns the firm. Chapter 6.28's Licence Raj is the Indian illustration.

Where markets fail

Chapter 9.6 listed them. Here they are with the mechanism.

Externalities. A cost or benefit falling on someone outside the transaction. A factory's pollution is a cost borne by people downstream who are not party to the sale, so it does not appear in the price, so too much is produced.

The remedies: a tax equal to the external cost, so the price tells the truth; a tradable permit system, capping the total and letting the market allocate; regulation; or, where the parties are few and rights are clear, negotiation between them — Ronald Coase's result, which holds when transaction costs are low and which is why it does not solve climate change.

Public goods. Non-excludable and non-rival — one person's use does not reduce another's, and non-payers cannot be excluded. National defence, clean air, basic research, a lighthouse. Everyone benefits and nobody has a reason to pay, so private provision fails and public funding is the answer.

Information asymmetry. Akerlof's used-car analysis: if buyers cannot tell good cars from bad, they offer an average price, so sellers of good cars withdraw, so the average quality falls, so the price falls further. The market for good cars can disappear entirely. The same logic explains why health insurance markets need regulation, why medicine is licensed, and why food safety is not left to reputation.

And the asymmetry generates two more problems. Adverse selection — the people most likely to claim are the most likely to buy insurance. Moral hazard — being insured changes behaviour (Chapter 9.2).

Coordination failures. Everyone would be better off if all switched together, and nobody moves first. Standards, electric vehicle charging networks, and Chapter 9.9's paradox of thrift.

And inequality is not a market failure at all, which is worth stating precisely. A market can be operating perfectly and produce an outcome most people find unacceptable, because it allocates by ability to pay rather than by need. That is a distributional question and it has a political answer, not a technical one.

Behaviour, and where the standard model breaks

Standard economics assumes people have consistent preferences and maximise expected utility. Real people do not, in documented and systematic ways.

Loss aversion. Losses feel roughly twice as bad as equivalent gains feel good, which explains why people hold falling investments, why sellers overprice their own houses, and why insurance sells.

Anchoring. The first number mentioned influences the judgement, even when it is obviously irrelevant.

Framing. A treatment with a 90 percent survival rate is chosen more often than one with a 10 percent mortality rate.

Present bias. People systematically over-weight the immediate, which is why saving for retirement requires automatic enrolment to work.

The practical consequence is that market design matters. Defaults, framing and the order of options change outcomes measurably. Chapter 13.8 covers where this becomes policy, and it is worth noting that it cuts both ways — the same findings that improve pension enrolment are used to design gambling apps and infinite feeds.

Where this shows up in your life

Every subsidy debate. A subsidy on a good with inelastic demand mostly raises the producer's margin; on one with elastic demand it mostly raises consumption. Whether India's fertiliser subsidy reaches farmers or manufacturers is exactly this question.

Surge pricing is the price mechanism doing its job visibly, which is why people hate it — it allocates scarce rides to those willing to pay most, and it brings more drivers out. Both effects are real and the second is why banning it produces longer waits rather than cheaper rides.

And the useful discipline for reading any economic claim is to ask three questions: what is the elasticity, what is the counterfactual, and who actually bears the cost? Almost every misleading argument in public economics fails at least one of them.

What the next page covers

If markets allocate within a country, trade allocates between them, and the argument about it is two centuries old and still live. Chapter 9.12 covers trade, tariffs and comparative advantage — what Ricardo's argument actually says and why it is genuinely surprising, what it leaves out, why every country that got rich used protection while telling others not to, what the evidence on trade and jobs shows, and how to think about the current turn back toward tariffs.