Compulsory taxation

Why must a service everyone wants be funded by a payment nobody is allowed to refuse?

Here is an odd pairing. Almost everyone says they want street lighting, coastal defences, clean air, courts. Almost nobody would send the money in voluntarily. If the want is real and near-universal, why does funding it require compulsion at all — and if compulsion is required, does that not suggest the want was never real in the first place?

Start by noticing what these particular goods have in common, because it is not that they are important. Bread is important, and nobody proposes a bread tax.

The relevant property is that you cannot keep a non-payer out. A sea wall protects the house of the man who refused to contribute exactly as well as it protects yours. Economists call this non-excludability, and it usually travels with a second property — one more person benefiting does not reduce what is left for you, which is non-rivalry. Bread has neither: I can refuse to hand you a loaf, and once you eat it, it is gone.

Now put yourself inside a market for a good you cannot be excluded from. The sea wall costs a large sum and there are ten thousand of you. What does your own contribution accomplish? If enough others pay, the wall is built and you are protected whether or not you contributed. If not enough others pay, the wall is not built, and your contribution is lost without buying anything. In both branches you are better off keeping the money. That is the uncomfortable part: contributing zero is not the choice of a selfish person, but the choice that follows for anyone reasoning about their own contribution alone, whatever everyone else does.

So the good goes unbuilt, or is built at a fraction of the scale everyone would have preferred — and nothing about the preference was false. Every one of the ten thousand may genuinely prefer everyone pays and the wall stands to nobody pays and the sea comes in, and still not pay. The failure is not in what people want; it is that the wanting cannot be converted into a payment by any means available to an individual.

What would fix it? Only something that makes your payment conditional on everyone else's, so that paying actually buys the wall rather than buying nothing. Compulsion does exactly that. Tax is not a way of extracting money from people who do not want the good; it is the device that makes "I will pay if the others do" enforceable. Paul Samuelson's treatment adds a twist: even a well-meaning planner cannot simply ask, because if your bill depends on your stated valuation you have reason to understate it, and if it does not, reason to overstate.

Then what becomes of consent? Nobody consents to a particular tax bill — that is what "compulsory" means. But you can consent to the rule, and James Buchanan and Gordon Tullock's point is that these are different questions asked at different levels. Whether to have a system of compulsory contribution is chosen in advance, without knowing which specific decisions will fall your way, and a person can rationally agree to be bound at that level precisely because the binding is what makes the good available. Consent to the constitution, not to the invoice.

The model above is clean, and it is also incomplete in ways worth stating plainly.

The starkest problem is that real people do not behave as it predicts. In public goods experiments — where participants privately decide how much of an endowment to put into a common pot that is then multiplied and shared — contributions typically start at around half the endowment, not at zero. They decay over repeated rounds as cooperators watch others free-ride, but toward a low level rather than to nothing, and where participants may pay a cost to punish free-riders, as in Ernst Fehr and Simon Gächter's work, cooperation is sustained instead. The better description is conditional cooperation: many people contribute if they believe others are contributing, which is a different animal from the pure free-rider.

That matters for what the argument establishes. It does not establish that people are selfish, nor that voluntary provision never works. Elinor Ostrom's fieldwork documented long-lived communities managing shared irrigation, forests, and fisheries without state compulsion, through local rules, mutual monitoring, and graduated sanctions. What those cases share is small numbers and visibility — contributions are observed and reputations exist. Scale and anonymity are what break voluntary provision, which is precisely the condition a national defence budget is in.

Two further notes. The argument justifies compulsion but says almost nothing about how much or for what: it tells you some enforcement is needed, not what the tax rate should be. And a great deal of what taxes actually fund is not a public good in this narrow sense at all — health care and pensions are both excludable and rival, and the reasons for funding them collectively, such as insurance against unpriceable risks and adverse selection in private markets, are separate arguments this one does not supply.

• Why the same logic produces international climate agreements that are hard to enforce and easy to leave.
• How Lindahl pricing tries to charge each person their own valuation, and why it founders on preference revelation.

Every term the collection defines is gathered in the glossary.