Two questions that sort every good
The last guide showed one way markets stumble: an externality, where a transaction spills costs or benefits onto bystanders. This guide opens a different crack in the same wall. Some goods are built in a way that the price system simply cannot get a grip on — not because of any spillover, but because of two awkward properties of the good itself. To see them, we ask just two yes-or-no questions about anything you might buy.
Question one: is the good rival? That is, does one person's use of it use it up, leaving less for others? A sandwich is rival — if I eat it, you cannot. Question two: is the good excludable? That is, can the seller easily fence it off and stop people who do not pay from enjoying it? A cinema seat is excludable: no ticket, no entry. Almost everything you have studied so far — apples, haircuts, factory output — has been quietly assumed to be *both* rival and excludable, which is exactly what lets a price do its job.
EXCLUDABLE? NON-EXCLUDABLE?
(can fence it off) (cannot fence it off)
+-----------------------+-------------------------+
RIVAL | Private good | Common-pool resource |
(use uses up) | sandwich, shoes, | ocean fish, grazing |
| a haircut | land, congested road |
+-----------------------+-------------------------+
NON-RIVAL | Club good | PUBLIC GOOD |
(your use does | Netflix, a toll | national defence, |
not use it up)| bridge, cinema | clean air, a lighthouse|
+-----------------------+-------------------------+The pure public good: non-rival and non-excludable
A public good sits in the awkward corner: it is *non-rival* and *non-excludable* at once. The textbook example is a lighthouse. Its beam is non-rival because a hundredth ship reading the light takes nothing away from the first ninety-nine — the light is just as bright for all. And it is non-excludable because once the lamp is lit, the keeper cannot somehow withhold the beam from a ship that did not pay. National defence is the same: an army that protects the country protects every resident at once, and cannot shield only the taxpayers while leaving the dodgers undefended. So is clean air, a mosquito-eradication programme, or a freely broadcast weather forecast.
Most real goods are not perfectly public; they sit somewhere on a spectrum. A city street is non-rival at 3 a.m. but very rival in rush hour, when one more car genuinely slows everyone down. A scrambled satellite TV signal is non-rival (your watching costs me nothing) yet *excludable* (a decoder card can be withheld) — that makes it a club good, sellable by subscription. The two questions are dials, not switches, and where a good lands on them decides whether a market can supply it.
Why markets under-supply them: the free rider
Now the trouble. Suppose a small street of ten neighbours could chip in for a streetlight that benefits them all. The light is non-excludable: once it shines, you cannot keep it from a neighbour who refused to pay. So each neighbour reasons coldly — exactly the rational-choice logic from the very first rung — "if the others fund it, I enjoy it for free; if they do not, my single contribution is too small to matter anyway." Either way, the privately best move is to keep your wallet shut. This temptation to enjoy a good without paying for it is the free-rider problem.
Notice the shape of this. If everyone reasons that way, nobody pays, and the streetlight that *all ten would gladly have bought together* never gets built. The good outcome (everyone contributes, everyone enjoys the light) is wrecked by each person's private incentive to defect — which is precisely the prisoner's dilemma from the game-theory rung, now played by ten people instead of two. The free-rider problem is a many-player prisoner's dilemma in disguise, and that is why it bites so hard: each individual is behaving sensibly, and the group ends up worse off.
Put numbers on it briefly. Say the streetlight costs $1,000 and each of the ten neighbours values it at $150 — so the total benefit is $1,500, comfortably above the cost. From society's view it *should* be built. But to any single neighbour, paying the full $1,000 for a personal benefit of $150 is a terrible deal, and even paying a $100 share feels risky if you suspect others will dodge. The market, asking each person to volunteer, collects too little. The result is under-provision: a good worth more than it costs is supplied in too small a quantity, or not at all — a textbook market failure.
Mirror image: the tragedy of the commons
Flip the public good around and you meet its evil twin. A common-pool resource is *rival* but *non-excludable* — the bottom-left-meets-top-right box in our grid: ocean fisheries, a shared pasture, groundwater, the atmosphere's capacity to absorb carbon. Here each user *can* deplete the resource (it is rival), yet *cannot* be kept out (it is non-excludable). Where the public good was under-provided, the common resource is over-used.
The classic picture is a village pasture open to all. Each herder who adds one more cow gets the *whole* gain from that cow, but the cost — slightly thinner grass for everyone — is spread across all herders. So each rationally keeps adding cows, the grass is grazed to dust, and every herder ends up worse off than if all had shown restraint. This is the tragedy of the commons, popularised by Garrett Hardin in 1968. Recognise the engine: it is the same defect-versus-cooperate logic as the free rider, the tragedy-of-the-commons game you met under game theory, now driving over-extraction rather than under-funding.
Why some things need collective provision
If voluntary payment fails for public goods, what works? The oldest answer is to make payment *non-voluntary*: a government provides the good and funds it through taxes, so nobody can free-ride by opting out. This is why national defence, basic research, street lighting, and public-health programmes are so often state-supplied — not from ideology, but because the free-rider problem makes private supply chronically too thin. For common-pool resources the parallel fix is to put a fence where nature left none: fishing quotas, grazing rights, or a cap on emissions, which restore the excludability the resource was missing.
But collective provision is no free lunch, and honesty demands the caveats. First, the government still has to guess how much people value the good — and since the public good is non-excludable, people have every reason to *understate* their willingness-to-pay to lower their tax bill, so the right quantity is genuinely hard to pin down. Second, public provision can over-shoot as easily as the market under-shoots: a politically driven project can build a bridge to nowhere, a problem economists call government failure. The market failure does not vanish; it is traded for a different set of imperfections.
And government is not the only fixer. The political scientist Elinor Ostrom won the 2009 Nobel for showing that real communities — Swiss alpine herders, Japanese forest villages, Maine lobster fishers — have for centuries governed their own commons sustainably through local rules, monitoring, and graduated penalties, with no state and no privatisation. Her work is a standing rebuke to the idea that the tragedy of the commons is inevitable: where users can talk, watch each other, and punish cheats, the many-player dilemma can sometimes be solved from the bottom up. The lesson of this guide is not "markets bad, government good," but something subtler: when a good is non-excludable, *some* form of collective arrangement — state, custom, or contract — has to step in, because the price tag alone cannot do the work.