Markets don’t always work perfectly

In your textbook, markets sound like a beautiful, self-correcting machine. Supply meets demand, price adjusts, resources flow to their best use. Efficient. Elegant. Automatic.

But here’s the thing economists don’t always shout loudly enough: markets only work perfectly under a very specific set of conditions. When those conditions break down, we get market failure — situations where the free market produces outcomes that are genuinely bad for society.

This matters enormously for policy. Climate change, vaccines, air pollution, traffic jams, street lighting — all of these involve some kind of market failure. Understanding market failures is basically understanding when and why the government might need to step in.

Negative externalities — when businesses dump costs on everyone else

Imagine a factory that produces cheap steel. Great for the economy, right? But the factory also pumps toxic smoke into the air. Nearby residents get respiratory problems. Hospital bills rise. Crop yields fall. Property values drop.

The factory doesn’t pay for any of this. It’s not in their accounts. But it’s a real cost — just one that gets shunted onto people who had nothing to do with the transaction. This is a negative externality.

The result? The factory produces too much steel relative to what’s socially optimal. Because it only sees its private costs (labour, materials, energy) — not the full social costs. Society would be better off with less steel and less pollution.

The diagram — what’s actually happening

Here’s the key insight:

• The factory supplies based on its Private Marginal Cost (PMC) — what it actually pays.
• Society cares about the Social Marginal Cost (SMC) = PMC + the pollution damage.
• The market equilibrium (where PMC = demand) gives too high a quantity.
• The socially optimal quantity is where SMC = demand — lower, with higher prices.
• That gap between the market outcome and the social optimum is deadweight loss — welfare that’s simply wasted.

The fix — the Pigouvian Tax

Arthur Pigou’s (1920) solution is elegant: tax the polluter exactly the amount of damage they cause. If every tonne of CO₂ causes £80 of damage to society, add an £80 tax per tonne. Now the factory internalises the external cost — it acts as if the social cost is the private cost. Market output falls to the socially optimal level.

Worked example

A chemical plant produces with PMC = £20/unit. It causes £8 of pollution damage per unit. Demand: P = 50 − 0.04Q.

• Market equilibrium: PMC = P → 20 = 50 − 0.04Q → Q_M = 750 units
• Social optimum: SMC = 20 + 8 = £28 = P → 28 = 50 − 0.04Q → Q* = 550 units
• Pigouvian tax: £8/unit (closes the gap)
• Deadweight loss from overproduction: ½ × (750−550) × £8 = £800 worth of wasted welfare

Positive externalities — when markets don’t produce enough

Now flip it. Vaccines don’t just protect you — they protect everyone around you. When enough people are vaccinated, diseases can’t spread (herd immunity). So when you get vaccinated, you create a benefit for every person you would otherwise have infected. They didn’t pay for that benefit. You didn’t charge them for it.

This is a positive externality. And the problem is the opposite of pollution: because the person making the decision doesn’t get all the social benefits, they don’t do enough of the activity. People get vaccinated at lower rates than is socially optimal. The market underproduces.

The fix? A subsidy to make the activity cheaper and bring quantity up to the social optimum. Or just provide it publicly — which is exactly why governments run free vaccination programmes.

The Coase Theorem — maybe we don’t always need the government

In 1960, Ronald Coase made a provocative argument. Forget Pigouvian taxes, he said. If property rights are clear and it’s easy for people to negotiate, then private bargaining will fix externalities automatically — and government intervention isn’t needed.

Example: A train emits sparks that burn a farmer’s crops. Damage = £500. Installing spark-catchers costs the railway £200.

• If the farmer has the right to be spark-free: the railway pays the farmer anywhere between £200–£500 to accept the sparks. But wait — actually, for £200 they can just install spark-catchers. So they install spark-catchers. Efficient.
• If the railway has the right to emit sparks: the farmer pays between £200–£500 for the railway to install spark-catchers. Same efficient outcome.

The point: the efficient solution (install the spark-catchers) happens regardless of who has the legal right, as long as they can talk to each other.

Public goods — the free-rider problem

Fireworks on New Year’s Eve. National defence. Flood defences. Street lighting. What do these have in common?

They’re all public goods:

• Non-rival: Your enjoyment of the fireworks doesn’t stop me enjoying them too. The army protecting you also protects me. One person’s use doesn’t reduce availability for others.
• Non-excludable: You can’t make someone NOT benefit from national defence even if they don’t pay taxes. The street light outside your house also lights the way for your neighbour.

The problem? Because nobody can be excluded from benefiting, nobody has an incentive to pay voluntarily. Everyone waits for someone else to pay — hoping to free-ride on their contribution. If everyone free-rides, nobody provides the good — even though everyone wants it.

This is why national defence, street lighting, flood defences and basic scientific research are government-funded. Private markets won’t supply them adequately — or at all.

Information asymmetry — when one side knows more than the other

George Akerlof’s 1970 paper “The Market for Lemons” is one of the most clever economics papers ever written. It asks: what happens in the used car market when sellers know more about the quality of the car than buyers?

Buyers can’t tell a good car from a bad one (“lemon”). So they offer the average expected price. But sellers of good cars think: “My car is worth more than that.” So they pull out. Now the average quality in the market falls. Buyers figure this out and lower their offers. More good cars leave. Eventually only lemons remain — or the market collapses entirely.

This “adverse selection” problem shows up everywhere: health insurance (sick people are more eager to buy), credit markets (riskier borrowers are more eager for loans), hiring (bad workers are more eager to take any job).

📋 Case Study: Sugar Tax and Carbon Trading

✏️ Practice Questions

🎯 Summary

• Market failure happens when free markets produce outcomes that are inefficient for society — usually too much or too little of something.
• Negative externalities (like pollution) cause overproduction — the Pigouvian tax corrects this by forcing firms to pay for the damage they cause.
• Positive externalities (like vaccines) cause underproduction — subsidies or public provision bring output up to the social optimum.
• The Coase theorem says private bargaining can fix externalities when transaction costs are low and rights are clear — but fails for problems like climate change.
• Public goods (non-rival, non-excludable) create free-rider problems — private markets won’t provide them adequately, justifying government funding.
• Information asymmetry (Akerlof’s lemons) causes markets to unravel when sellers know more than buyers — requiring regulation, disclosure, or public provision.

📚 References & Further Reading

1. Pigou, A.C. (1920). The Economics of Welfare. Macmillan. — The original case for taxing externalities.
2. Coase, R.H. (1960). The problem of social cost. Journal of Law and Economics, 3. — A beautifully written classic. Short and worth reading.
3. Akerlof, G.A. (1970). The market for “lemons.” Quarterly Journal of Economics, 84(3). — One of the cleverest papers in economics. Very readable.
4. Scarborough, P. et al. (2020). Impact of the UK Soft Drinks Industry Levy. PLOS Medicine, 17(2).
5. Dechezleprêtre, A. et al. (2018). Carbon trading and clean innovation. NBER Working Paper 24194.
6. Rennert, K. et al. (2022). Comprehensive evidence implies a higher social cost of carbon. Nature, 610, 687–692.