A public good has two defining properties. It is non-rival: my use does not reduce what's available for you. And it is non-excludable: you cannot prevent non-payers from benefiting. Clean air, national defense, basic research, scientific knowledge — these are the canonical examples. Markets systematically underproduce them, not because markets are bad but because the mechanism that makes markets work — exclusion by price — is impossible by definition for goods that cannot be excluded.

The Samuelson Formulation

Paul Samuelson's 1954 paper The Pure Theory of Public Expenditure is the foundational treatment. Samuelson showed mathematically that for a public good, the efficient quantity is the one where the sum of all consumers' marginal willingness to pay equals the marginal cost — not, as in private goods, where any individual consumer's willingness equals it. This means the financing has to come from the entire set of beneficiaries collectively, which means governments or some equivalent collective-action mechanism, because markets cannot collect the sum.

The Free-Rider Problem

The mechanism by which markets fail at public goods is straightforward. If I cannot be excluded from the benefit, I have no reason to pay. If nobody pays, the good is not produced. If somebody else pays, I benefit without contributing. In aggregate, everyone has an incentive to free-ride, and the good is undersupplied or unsupplied. National defense is the case where this is most stark — no private firm could possibly provide territorial defense on a paying-customer basis, because the defense necessarily covers the non-paying neighbor too.

What Markets Can Partly Do

Some apparent public goods can be partially marketized through exclusion technology. Toll roads convert what was a public good into a club good. Encrypted broadcast converts a non-excludable signal into an excludable one. Patents convert non-rival knowledge into a temporarily excludable monopoly. Each of these mechanisms imposes a deadweight cost — some users who would benefit are priced out — in exchange for making private production feasible. The patent system is the cleanest example: society accepts twenty years of monopoly pricing on a new drug because the alternative (no R&D investment at all) is worse, even though for those twenty years the drug is underused relative to the optimum.

Knowledge as the Most Important Modern Public Good

Scientific knowledge, software, mathematics, and basic research results are the largest categories of public goods in the modern economy. They are non-rival (my reading of a research paper does not exclude yours) and non-excludable (once published, the knowledge is in the world). The market provision mechanism for these — patents, copyright, secrecy — is partial at best, and the major engines of modern knowledge production are public: universities, national labs, international research consortia, and the open-source software movement.

Mariana Mazzucato's work has emphasized that even much of what looks like private innovation has public-good foundations. The iPhone's component technologies — touchscreen, GPS, voice recognition, the internet — were all publicly funded basic research that no single firm would have invested in because the returns were too diffuse to capture privately.

Ostrom and the Commons

Elinor Ostrom's 2009 Nobel Prize recognized her empirical work on common-pool resources — goods that are rival but non-excludable (fisheries, irrigation systems, forests). Ostrom documented hundreds of cases where local communities had successfully managed common-pool resources for centuries without either privatization or top-down state control, using locally evolved rules, monitoring, and graduated sanctions. Her work complicates the standard "private vs. public" dichotomy: there is a third category, governed commons, that can work under specific institutional conditions.

The Public-Goods Politics

The hardest political fights about public goods are usually about how to finance them and how to define their boundaries. Universal healthcare, universal education, climate stability, network infrastructure, public transit — all involve public-good elements and all involve large distributional fights over who pays. The economics is generally uncontroversial — markets undersupply them, collective provision is the only mechanism that scales — and the politics is generally about distribution.

The other recurring fight is about what counts as a public good. Conservative arguments often try to push goods from the public-good category into the private-good category by emphasizing exclusion technology (vouchers, fee-for-service models, patents). Progressive arguments often try to push goods in the opposite direction by emphasizing the externality that makes private provision insufficient (public health, basic income, knowledge commons). Both sides are arguing about the same underlying classification problem.

The Honest Reading

Public goods are real, the market failure they represent is unavoidable, and collective provision is the only mechanism that scales to provide them. The political question is which goods belong in this category, what level of provision is optimal, and how the cost should be distributed. Markets undersupply public goods by definition; that does not mean any specific collective provision proposal is the right one. The empirical work matters, and the political settlement matters, and the underlying economics is settled enough that the conversation should be about implementation rather than about whether the category exists.

The Modern Knowledge Commons

The largest public-goods category in the modern economy is knowledge itself — scientific research, mathematical results, software, open data. The patent and copyright systems convert some of this knowledge into excludable monopolies for limited periods, but most of the foundational knowledge that drives the modern economy is non-excludable by design. Public funding of universities and basic research, open-source software development, and Wikipedia-style commons all operate outside the market framework because the market framework does not produce public goods at adequate scale. Whether the political coalition for sustained public funding can hold through cycles of fiscal pressure is the recurring institutional question for the modern knowledge commons.

The Ostrom Synthesis

Elinor Ostrom's 2009 Nobel Prize work showed that the simple public-vs-private dichotomy misses a third category: governed commons. Locally evolved rules, peer monitoring, and graduated sanctions can sustain provision of common-pool resources without either privatization or top-down state control. Hundreds of cases — Maine lobster fisheries, Swiss alpine pastures, Spanish irrigation systems, Japanese forests — illustrate the pattern. The implication for modern policy is that public goods do not require central provision in every case; for specific categories with specific characteristics, governed commons can work. The framework is not a universal answer but is a real third option that the public-vs-private framing missed.