Market Failure, Information & Behaviour

Economics & Law - NIT Northern Institute of Technology / TUHH, Hamburg · part of my Technology Management MBA · study notes for revision.

---

Earlier chapters made a big promise: leave a competitive market alone and, as if by magic, it settles at the point where demand meets supply - the outcome that squeezes the most total value out of society’s scarce resources. Adam Smith called it the invisible hand: everyone chasing their own profit somehow lands on the socially good result. If that were always true, governments would have no business meddling in markets. And yet they meddle constantly. This chapter is about why - the cases where the invisible hand slips, and what a government can (and can’t) do about it.

Tip

The one-line idea. A market failure is any situation where a free market, left to itself, does not reach the efficient outcome. When that happens, there may be a case for the government to step in - but every fix has side effects, so “there’s a failure” never automatically means “so intervene”. We’ll spend the chapter naming the failures, then the tools, then the catch on both.

1 · What “market failure” actually means

First, let’s be precise, because the word “failure” is misleading. A market failure is not a market that crashes or a firm that goes bust. It’s a market that works perfectly well by its own logic and still lands on the wrong outcome for society.

Recall the benchmark from earlier chapters: a market is efficient when it maximises total surplus - the combined value enjoyed by consumers (consumer surplus, the gap between what they’d have paid and what they did pay) and producers (producer surplus, the gap between the price and their cost). Kaldor-Hicks efficiency is reached at the competitive equilibrium where demand equals supply. A market failure is when the free-market equilibrium misses that point - trading too little, too much, or the wrong things.

Caution

A failure is a reason to look, not a verdict to act. The whole rest of the chapter rests on this. Spotting a market failure only tells you the private outcome isn’t the socially best one. Whether a government can do better is a separate question - because governments have their own failures (bad information, lobbying, blunt tools). Keep both columns in mind.

Economists group the reasons a market fails into four families. We’ll take them one at a time.

Market powermonopoly

Externalitiesthird-party effects

Public goodsfree-riding

Asymmetric infoone side knows more

The four classic sources of market failure. A fifth - behavioural failure, where the “rational” assumption itself breaks down - gets its own section at the end.

2 · Source one - market power (monopoly)

We met this one already, so this is a quick recap. In perfect competition, firms are price-takers: each is so small it simply accepts the going price, and that price ends up equal to the marginal cost (the cost of making one more unit). Price equals marginal cost is exactly the efficient condition - the last unit made is worth as much to a buyer as it costs to produce.

A monopoly (a single seller) breaks that. It faces the whole downward-sloping demand curve, so to sell one more unit it must drop the price on all units. That makes its marginal revenue (the extra money from one more sale) fall below the price. The monopolist maximises profit where marginal revenue equals marginal cost - which means price ends up above marginal cost.

Note

The real reason monopoly is inefficient - and it’s a common exam trap. It is tempting to say “monopolies are bad because the price is too high”. That’s incomplete. The high price is a symptom. The actual inefficiency is that the monopolist restricts output: it produces less than the efficient quantity, so some trades that would have made both buyer and seller better off simply never happen. That lost value is the deadweight loss. Higher price transfers money from consumers to the firm (a distribution effect); restricted output destroys value outright (an efficiency effect).

There’s a subtlety worth remembering: monopoly is inefficient once it exists (the ex-post view), but the promise of temporary monopoly profit is often what motivates the costly R&D in the first place (the ex-ante view). That’s the whole logic of patents - we knowingly grant a monopoly for a while to reward invention. It’s a classic economics trade-off: you often can’t get ex-ante and ex-post efficiency at the same time.

3 · Source two - externalities

An externality is a cost or benefit that lands on a third party - someone who is not the buyer or the seller and had no say in the deal. Because the market price only captures the private costs and benefits of the two parties trading, it quietly ignores the effect on everyone else. So the private optimum (what the market picks) drifts away from the social optimum (what’s best for society once the third-party effect is counted).

Negative externality too much made

• A factory makes steel; its smoke harms nearby residents’ health.
• The firm pays for labour and materials, but not for the pollution.
• So social cost > private cost → the good is over-produced.

Positive externality too little made

• You get a vaccine (or wear a mask); others are less likely to catch the disease from you.
• You pay the full cost but pocket only your own benefit.
• So social benefit > private benefit → the good is under-produced.

The pattern is symmetric. With a negative externality (pollution, noise, congestion) the market makes too much, because the polluter doesn’t feel the full cost. With a positive externality (vaccination, education, a well-kept garden) the market makes too little, because the do-gooder isn’t paid for the spillover benefit. In both cases the price is “wrong” - it fails to price in the third party.

The Coase theorem - sometimes you don’t need the government at all

Here’s a genuinely surprising result from Ronald Coase. The instinct is that externalities require government to fix. But Coase showed that if two conditions hold, the private parties can just bargain their way to the efficient outcome by themselves - and, astonishingly, it doesn’t matter who the law sides with.

The classic story is a rancher whose cattle stray onto a neighbouring farmer’s land and eat the crops.

1. Set the numbers. A cattle invasion costs the farmer 100 dollars in lost crops. A fence costs 50 dollars if the farmer builds it, 75 dollars if the rancher does. Efficiency says: avoid the 100-dollar damage as cheaply as possible → the farmer should fence, at 50.

2. Give the farmer the right (“rancher must keep cattle in”). The rancher is liable. But the rancher would rather pay the farmer something between 50 and 75 to build the cheaper fence than fence at 75 himself. Deal struck; farmer fences. Efficient.

3. Give the rancher the right (“farmer must keep cattle out”). Now the farmer bears his own losses - so he simply builds the 50-dollar fence himself, because that beats losing 100. Efficient again.

4. The punchline. Either way, the fence gets built by the cheapest builder. The legal right decides who pays (the distribution), but not whether the efficient thing happens. That’s the Coase theorem.

Coase theorem When transaction costs are zero, private bargaining reaches the efficient use of resources regardless of how the law assigns the property right.

Condition 1 Property rights are clear and transferable - everyone knows who is liable.

Condition 2 Transaction costs are low (ideally zero) - bargaining, drafting, enforcing is cheap.

Coase’s insight: with clear rights and cheap bargaining, the market fixes its own externality - the law only reshuffles who pays.

Caution

Where Coase runs out of road. The theorem needs those two conditions, and real life often breaks them. If transaction costs are high - hiring lawyers, writing contracts, monitoring compliance - a deal that should happen may not be worth the hassle. And crucially, if there are many parties (think a factory polluting a whole city of thousands), getting everyone to the negotiating table is hopeless, and free-riding creeps in. That’s exactly when we fall back on government tools instead. The Law-and-Economics view of the law’s job: lubricate transactions so parties can bargain, and when that’s too costly, replicate the outcome bargaining would have reached.

4 · Source three - public goods

Goods differ along two dimensions, and those two dimensions decide whether a market can supply them at all.

• Rival? Does my consuming it leave less for you? An apple is rival (I eat it, it’s gone); a radio broadcast is non-rival (my listening doesn’t reduce yours).
• Excludable? Can a seller stop non-payers from using it? A cinema seat is excludable (ticket at the door); the light from a lighthouse is not (every passing ship sees it, pay or not).

Cross the two and you get four kinds of good:

Public goods non-rival, non-excludable

• National defence, street lighting, a lighthouse, a stable climate.
• Can’t charge, can’t ration → the market under-provides.

Common goods rival, non-excludable

• Ocean fish stocks, a shared pasture, a public road.
• Can’t exclude but use runs it down → the tragedy of the commons (overuse).

Club goods non-rival, excludable

• Netflix, a gym, a toll motorway at quiet hours.
• Can charge and one more member costs nothing → markets do fine.

Private goods rival, excludable

• Food, clothes, an apartment.
• The textbook case - markets handle these well.

The rival × excludable grid. Private and club goods sell fine; public and common goods are where markets stumble.

The trouble child is the top-left box. A public good is both non-rival (one more user costs nothing extra) and non-excludable (you can’t stop non-payers enjoying it). Put those together and you get the free-rider problem: since I’ll get the benefit whether or not I chip in, my selfish move is to let others pay and enjoy it for free. But everyone reasons the same way, so nobody pays, and the good is under-provided or not provided at all - even though everyone would be better off if it existed.

Note

Free-riding in one tiny example. Three students each value a shared exam-prep book at almost 2 dollars; the book costs 2 dollars, so they agree each will donate 1 dollar. If I think the other two will pay, I’m better off not donating - I still get the book for free. If everyone reasons that way, no one donates and the book is never bought. That’s the free-rider logic that sinks voluntary funding of public goods, and it’s why they’re usually paid for through taxes, not donations.

This scales all the way up. Climate protection is a global public good: if one country cuts emissions, every country enjoys the more stable climate (non-excludable), and one country’s benefit doesn’t shrink another’s (non-rival). So every country is tempted to free-ride on everyone else’s costly cuts - which is exactly why binding international climate agreements are so hard to reach and hold together.

5 · Source four - asymmetric information

The fourth failure is the subtlest and, honestly, my favourite. Asymmetric information means one side of a deal knows something the other side doesn’t. It splits neatly into two problems that are easy to mix up, so pin down the timing: does the hidden thing exist before the deal, or does it happen after?

Adverse selection hidden info, BEFORE the deal

• The hidden thing is a fixed type / quality the other side can’t see.
• ”The wrong people enter the market” - bad types crowd out good ones.
• Model: Akerlof’s market for lemons.

Moral hazard hidden action, AFTER the deal

• The hidden thing is a behaviour the other side can’t monitor.
• ”People act differently once protected” - they take more risk.
• Model: insurance & principal-agent problems.

Adverse selection and the “market for lemons”

This is George Akerlof’s Nobel-winning idea (a “lemon” is US slang for a dud used car). Suppose sellers know whether their car is good or bad, but buyers can’t tell them apart. Watch the market eat itself:

1. Buyers can only guess the average. Say half the cars are good (worth 10,000 to a buyer) and half are lemons (worth 4,000). Unable to see quality, a buyer will only offer the expected value: 0.5 × 10,000 + 0.5 × 4,000 = 7,000.

2. Good sellers walk away. An owner of a genuinely good car values it near 8,000 and won’t sell for 7,000. A lemon owner (reservation price ~3,000) happily sells. So the good cars quietly leave the market.

3. Buyers realise, and re-price down. Now only lemons are left. Buyers work this out and refuse to pay more than the lemon value, ~4,000.

4. The market unravels. Bad quality has driven out good. In the worst case - if even the lemon owners have a positive selling cost - nobody trades and the market collapses entirely. Trades that should happen (good cars changing hands) never do. That’s the inefficiency.

Note

The same trap, three markets. Adverse selection is everywhere information is one-sided.

• Health insurance: sick people (who know their risk) buy the most cover; healthy people opt out; the risk pool worsens; premiums rise; more healthy people leave - the classic “insurance death spiral”.
• Credit: raise loan interest rates and your safe borrowers (who’d repay) drop out because it’s now too pricey, while risky borrowers - who may default anyway - happily stay. The bank’s borrower pool gets riskier. This is why a bank may refuse to raise rates even when loans are in demand (the Stiglitz-Weiss credit-rationing logic).
• Labour: employers can’t see worker ability before hiring, so they offer an average wage - which the best workers may reject.

Moral hazard - the hidden action

Flip the timing. Moral hazard is about a hidden action taken after the contract is signed, by someone who no longer bears the full consequences of their choices. The information problem isn’t a hidden type, it’s a hidden behaviour.

The purest case is insurance. Once my car is fully insured, I have less reason to lock it or drive carefully - the insurer pays if anything goes wrong. The insurer can’t watch me, so it can’t tell a careful driver from a careless one after the fact. Same shape appears in banking (a borrower takes wild risks with the loan - keeps the upside, dumps the downside on the bank) and in any principal-agent relationship, like an employer who can’t perfectly monitor an employee’s effort.

How markets (and laws) fight information problems

The good news: markets aren’t helpless. Several mechanisms exist to shrink the information gap, and they map neatly onto who acts.

Signallingthe INFORMED side acts - takes a costly action bad types can’t afford to fake

↓

Screeningthe UNINFORMED side acts - designs choices that make types reveal themselves

↓

Warranties & disclosurecontract or law forces quality info into the open

Three families of fix. The trick behind all of them: make honesty cheap for good types and expensive for bad types.

Fix	Who acts	How it works	Example
Signalling	The informed party (seller / worker)	Takes an action that is cheap for good types, painful for bad types, so faking it isn’t worth it	A degree signals ability (Spence); a long warranty signals a reliable product
Screening	The uninformed party (buyer / insurer)	Offers a menu so people sort themselves and reveal their type	Insurance: low-premium-high-deductible vs high-premium-low-deductible - safe drivers pick the first
Warranties / guarantees	The seller	Promise to repair or refund if the product fails	A 5-year car warranty a lemon-maker couldn’t afford
Mandatory disclosure	The law	Forces the informed side to reveal facts	Car-history reports, financial disclosure, food-safety labels

Tip

Why a warranty is such a clean signal. A warranty is a promise to pay if the product breaks. For a high-quality seller whose product rarely fails, that promise is cheap to make. For a lemon-seller it would be ruinously expensive - they’d be paying out constantly. So only good sellers can afford strong warranties, which is exactly why buyers read a long warranty as a credible sign of quality. It does triple duty: it insures the buyer against defects, it gives the seller an incentive to build quality (they bear the repair bill), and it signals quality. Same object, three jobs.

6 · The government’s toolkit - and the catch on every tool

So a market has failed and bargaining can’t fix it. Now the government reaches for its tools. The single most important lesson here: every remedy has side effects. A tool that helps one failure often creates a deadweight loss of its own - especially if aimed at a market that was actually working fine.

Taxes and subsidies (including the Pigouvian tax)

The natural fix for an externality is to put the missing third-party effect back into the price.

• For a negative externality, add a Pigouvian tax equal to the external damage. If a factory’s pollution costs society 2 dollars a unit, tax it 2 dollars a unit. Now the firm feels the full social cost, cuts output to the efficient level, and the externality is “internalised”. Carbon taxes and cigarette taxes are exactly this.
• For a positive externality, mirror it with a subsidy - pay people to do the thing that spills benefits onto others (vaccination, education, solar panels).

Caution

But taxes and subsidies distort healthy markets. Here’s the catch. In a competitive market with no externality, a per-unit tax is pure harm: it wedges apart the price buyers pay and the price sellers receive, cuts the quantity traded below the efficient level, and creates a deadweight loss. The government’s tax revenue does not fully make up for the surplus lost - some value simply vanishes. A per-unit subsidy is not a free gift either: it pushes quantity above the efficient level (over-consumption), and the government spends more than the extra surplus created - again a deadweight loss. (One exception to file away: a lump-sum tax - a fixed fee that doesn’t change with output - doesn’t distort the production decision, so no deadweight loss, though a big enough one can scare firms out of the market entirely.)

The lesson: a Pigouvian tax is only a fix because there was a distortion to cancel. Point the same tax at a well-functioning market and you create the distortion.

Price regulation - ceilings and floors

Instead of taxing, the government can dictate the price directly. Two flavours, mirror images of each other:

Price ceiling a legal MAXIMUM

• Example: rent control (a “Mietbremse”) capping rent below the market rate.
• Below equilibrium, demand exceeds supply → a shortage.
• Rationing by queues, connections, worse upkeep, side-payments (“buy the sofa separately”).

Price floor a legal MINIMUM

• Example: a minimum wage set above the market wage.
• Above equilibrium, supply exceeds demand → a surplus.
• In the labour market a surplus of workers is unemployment, hitting low-skilled workers hardest.

A binding ceiling (set below the market price) creates a shortage: more people want the good than there is to go round, so it gets rationed by waiting lists, personal connections, or quiet side-payments, and landlords under-maintain because the rent no longer pays for it. A binding floor (set above the market price) creates a surplus: with a minimum wage, more people want to work than employers want to hire, and the leftover supply of labour shows up as unemployment. Both carry a deadweight loss in a competitive market - the same story as taxes.

Note

The twist: the same tools can help against a monopoly. All that “intervention destroys surplus” talk assumes the market was competitive to begin with. Against a monopoly - which is already producing too little - a cleverly set price ceiling can actually push output up toward the efficient level and raise welfare. So the honest rule isn’t “regulation good” or “regulation bad”; it’s match the tool to the failure. A tax on a monopoly, by contrast, usually makes things worse, since the monopoly already under-produces and the tax makes it under-produce even more.

Mandated guarantees and disclosure

For information failures, the government can make the market’s own fixes compulsory. The EU’s consumer-sales rules, for instance, set a legal guarantee: a mandatory minimum warranty period (two years for new goods) that every seller must honour, on top of which sellers may voluntarily offer more. Mandatory disclosure laws (car-history reports, food labelling, financial statements) force the informed side to reveal what it knows. These attack the root - the information gap - rather than the price.

Market failure

• The market misses the efficient outcome: monopoly, externalities, public goods, bad information.
• A reason to consider intervention.

Government failure

• The fix misses too: wrong-sized taxes, shortages from price caps, lobbying capturing the rules, blunt one-size policy.
• Why “a failure exists” ≠ “so intervene”.

7 · A behavioural reality check on “homo economicus”

Everything above still assumed people are the cool, calculating homo economicus of Chapter 1 - perfectly rational, self-interested, maximising utility. Behavioural economics is the field that takes that assumption into the lab, and finds it… only sometimes true. People run on bounded rationality: limited attention, limited computing power, and a head full of mental shortcuts (heuristics) that usually work but sometimes misfire in predictable, systematic ways. Systematic is the key word - these aren’t random slip-ups you can average away; they lean the same direction every time, so they’re a genuine, patterned departure from the model. That makes them a fifth kind of market failure: a behavioural one.

Loss aversion & the endowment effect

• Losses hurt roughly twice as much as equal gains feel good (Kahneman & Tversky’s prospect theory).
• So we over-value what we already own: willingness-to-accept > willingness-to-pay. Two neighbours can both refuse to swap houses, each sure theirs is worth more.

Status-quo / default bias

• We stick with whatever the default is - out of inertia and procrastination.
• Auto-enrol people in a pension and most stay; make them opt in and many never do.

Anchoring & framing

• Anchoring: the first number sticks - a prosecutor’s “20 years” drags the final sentence toward it.
• Framing: “90% fat-free” beats “10% fat”; “200 saved” beats “400 will die” - same facts, different choice.

Over-confidence & present bias

• Over-confidence / optimism: “accidents won’t happen to me” - why seat-belt laws are needed.
• Present bias: we over-weight today over tomorrow - under-saving, over-spending, putting off the gym.

Four families of systematic bias that break the rational-actor model. There are more (availability, representativeness, base-rate neglect), but these carry most of the policy weight.

Note

Why this matters for law. If people don’t respond to incentives the way the rational model predicts, then laws built on that model can misfire - and laws can be designed around the real psychology instead. Small text health warnings did little; large graphic ones on cigarette packs, aimed at the emotional System-1 brain, did more. This is behavioural law and economics: take human quirks seriously when you write the rules.

Nudges - steering without shoving

If defaults and framing move behaviour, a government can exploit that gently. A nudge (Thaler and Sunstein) changes the choice architecture - how options are presented - to steer people toward better choices without removing any option or changing the payoffs. It’s the lightest-touch tool in the box.

Default: auto-enrol pensionsDefault: opt-out organ donationSalience: fruit at eye levelSimplify: shorter benefit formsFraming: personal, relevant messages

Switching organ donation from opt-in to opt-out, for instance, leaves everyone completely free to refuse - yet donation rates jump, purely because most people stick with the default. Putting fruit at eye level in a canteen restricts nobody’s choice but nudges healthier eating. Crucially, a nudge preserves freedom of choice - that’s what separates it from a tax or a ban.

Caution

Nudges are no magic wand - three honest criticisms.

• They can be weak. A gentle nudge may barely move behaviour against a strong habit, where a price change (a Pigouvian tax) would bite harder. Nudges and taxes can be substitutes or complements - sometimes you need both.
• They can backfire or mis-target. A message saying “75% of people already watch YouTube - you’re not like them, right?” accidentally advertises a social norm of doing the thing, and may increase it. A personal framing (“that’s time you could spend with family”) tends to work better. And nudges bite different people differently - children fall for supermarket placement more than adults do.
• They can manipulate. Steering people below their conscious radar raises an ethics question, which is why transparency matters. And the evil twin of a nudge is a sludge - deliberate friction (endless forms, tortuous cancellation flows) that makes the good choice hard, usually to benefit the firm rather than the person.

Revision summary

Tip

The one-minute recap

• A market failure is when a free market misses the efficient outcome. It’s a reason to look at intervention - never an automatic verdict, because government failure is real too.
• Four classic sources: market power (monopoly restricts output → deadweight loss, not just “high price”); externalities (third-party costs/benefits - negative → over-production, positive → under-production); public goods (non-rival + non-excludable → free-riding → under-provision); asymmetric information.
• The Coase theorem: with clear property rights and low transaction costs, private parties bargain to efficiency regardless of who holds the right - but it fails with high transaction costs or many parties.
• Asymmetric information splits by timing: adverse selection (hidden type, before the deal - Akerlof’s market for lemons, where bad drives out good and the market can unravel) vs moral hazard (hidden action, after the deal - the insured take more risk). Fixes: signalling, screening, warranties/guarantees, mandatory disclosure.
• The government’s toolkit: Pigouvian taxes (negative externalities) and subsidies (positive); price ceilings (→ shortages) and price floors (→ surpluses); mandated guarantees & disclosure (information). Every tool has side effects - in a competitive market these mostly create deadweight loss; against a monopoly some can help. Match the tool to the failure.
• Behavioural economics replaces perfect rationality with bounded rationality and systematic biases - loss aversion, status-quo/default bias, anchoring, framing, over-confidence, present bias. Nudges steer via choice architecture without removing choice, but can be weak, mis-targeted, or manipulative (and their evil twin is the sludge).

Next: International Trade → - why countries trade, and who wins and loses.