HookThe 5p that beat seven billion plastic bags
In October 2015, England introduced a 5p charge on single-use carrier bags at large retailers. The year before, the main supermarkets had handed out roughly 7.6 billion of them. Within a few years, single-use bag numbers at those retailers had fallen by well over 90% — one of the cheapest, fastest behaviour changes in British policy history, later extended to all retailers at 10p in 2021. Five pence. Less than the rounding error on a weekly shop.
Why did such a trivial price work? Because the bag had always been free but never costless — the litter, the marine plastic, the landfill were real costs paid by everyone else. The market price (zero) was a lie, and correcting the lie by even 5p forced shoppers to notice a cost they had been ignoring. That is this whole section: markets fail society when prices lie — about pollution, about the true value of education and vaccines, about goods no one can be made to pay for at all — and governments own a toolkit of taxes, subsidies, permits, rules and information to make prices tell the truth.
ModelWhere markets fail society
Three families of failure recur across the economy. Merit goods — education, healthcare, vaccination — are under-consumed if left to the market: people underestimate the long-term benefit to themselves (information failure) and ignore the benefit to others entirely. Demerit goods — tobacco, gambling, alcohol — are over-consumed for the mirror-image reasons: the harm is distant, misjudged, and partly dumped on other people, from passive smokers to the NHS budget.
Public goods fail hardest: the market does not under-provide them, it may not provide them at all. A public good is non-rival (my use doesn't reduce yours) and non-excludable (non-payers cannot be kept out) — street lighting, national defence, flood defences. Non-excludability creates the free-rider problem: since you get the flood defence whether or not you pay, nobody volunteers to pay, so no private firm can sell it — which is why flood defences are built by the Environment Agency out of taxes, not sold by the metre. Keep the definition surgical in the exam: the NHS is not a public good (a hospital bed is rival and excludable); it is a merit good the state chooses to provide. The test is the two properties, never who happens to provide it.
ModelExternalities — when prices lie
An externality is a cost or benefit landing on a third party who was never asked — someone outside the transaction. The framework: private costs are what the producer and consumer pay; external costs are what spill onto everyone else; social cost is the sum of the two. When a market prices only the private cost, the good is too cheap and over-produced. Air pollution is the canonical negative externality: UK government and academic estimates attribute somewhere in the region of 28,000–36,000 early deaths a year to air pollution, a colossal cost that appears on no polluter's invoice.
Positive externalities mirror the logic: vaccination protects the vaccinated and everyone they no longer infect, education raises the graduate's wages and the productivity of everyone who works with them. Because consumers weigh only their private benefit, such goods are under-consumed at market prices. The single most reliable analysis structure in this section: identify the third party by name, state whether the spillover is a cost or a benefit, and conclude over- or under-provision relative to the social optimum. Miss the third party and the answer collapses into vague talk about 'bad things' — the examiner is checking whether you can see the person outside the transaction.
MechanismThe policy toolkit — making prices tell the truth
Six instruments, each with a UK flagship. Indirect taxes put the external cost into the price: landfill tax started at £7 a tonne in 1996 and now stands at just over £100 a tonne — a price signal that helped transform UK waste from burying to recycling. Subsidies do the reverse for positive externalities: plug-in car grants and subsidised childcare push consumption toward the social optimum. Tradable pollution permits cap total emissions and let firms trade the right to pollute — the UK Emissions Trading Scheme (2021) shrinks the cap each year so the carbon price does the rationing. Regulation simply bans or limits: the 2007 indoor smoking ban did more for passive smokers than decades of duty rises; adult smoking has roughly halved since, from around 21% to about 12% by 2023 (duty, information and vaping all share the credit — say so, it is good evaluation). Information provision attacks the failure at its root: traffic-light food labels and calorie counts correct the misjudgement rather than the price. And state provision — the NHS, state schools — handles merit goods and public goods the market would starve.
Every tool carries a government-failure risk: taxes set at guessed values (the true external cost of a tonne of carbon is genuinely unknowable), subsidies captured by people who would have bought anyway, regulations that invite evasion. The strong answer picks the tool, shows the mechanism, then stress-tests it.
A factory emits 100,000 tonnes of CO2 a year under the UK ETS, with permits trading at roughly £40 a tonne (they have swung between about £30 and £75 since 2021). Doing nothing costs 100,000 × £40 = £4,000,000 a year in permits. An abatement technology can eliminate half its emissions at a cost of £25 per tonne: abating 50,000 tonnes costs 50,000 × £25 = £1,250,000, versus 50,000 × £40 = £2,000,000 for the permits those tonnes would have needed — a saving of £750,000, before selling any spare permits. That is the whole genius of the scheme: pollution now has a price, so every firm hunts for any clean-up cheaper than the permit — and the cheapest abatement in the economy happens first.
CaseULEZ — pricing the air over London
In August 2023 London's Ultra Low Emission Zone expanded to cover every borough: drive a non-compliant vehicle anywhere in Greater London and you pay £12.50 a day. It is a textbook externality charge — the driver of an older diesel finally pays something toward the health costs their exhaust imposes on other people's lungs. Transport for London reports that well over 90% of vehicles seen driving in the zone now meet the standards, and central London's roadside nitrogen dioxide is down substantially since the original 2019 zone — cleaner air achieved largely by people upgrading vehicles rather than paying the charge, which is exactly how a corrective price is supposed to work.
But the expansion also handed you the best evaluation paragraph in this section: distribution. The charge falls hardest on lower-income drivers — tradespeople, carers, shift workers — who own the oldest vehicles and can least afford to replace them, which is why a £2,000-per-vehicle scrappage scheme was bolted on and why the policy became politically explosive. Every corrective tax and charge has this property: efficient on average, regressive at the margin. Naming who bears the burden, and whether compensation reaches them, is what separates a level-3 from a level-4 answer on any intervention question.
VocabularyKey terms the mark scheme pays for
TrapsMisconceptions that cost marks
ExamWhat examiners want
Precision on definitions is where this section pays: external costs are costs to THIRD PARTIES not reflected in price, and public goods are non-rival AND non-excludable — examiners dock vague versions of both every session. In any externality answer, name the third party explicitly (residents near the runway, passive smokers, other drivers' lungs) before you conclude over- or under-provision.
On policy questions, the top-band recipe is mechanism → magnitude → government failure: show HOW the tax or permit changes the price signal, ask WHETHER the response will be big (that is a PED point — the bag charge versus fuel duty contrast is gold), then stress-test with a government-failure risk such as setting the tax at a guessed external cost. Finish with distribution: who actually bears the burden — ULEZ and every green tax question rewards the candidate who notices that corrective prices are often regressive, and that compensation schemes like scrappage decide the fairness verdict.