What Is a Public Good? Why Some Benefits Are Hard to Sell

Non-rivalry means that one person's use does not significantly prevent another person from enjoying the same benefit. If a lighthouse signal helps one ship navigate, that does not use up the signal for another ship. Non-excludability means it is difficult or costly to prevent people from receiving the benefit once it is provided.

These two properties form the core answer to what a public good is. National defence is another standard textbook example because protecting a territory can benefit residents broadly, and one person's protection does not normally subtract the same amount of protection from someone else. Real goods can fit these categories imperfectly, so economists often discuss degrees rather than perfectly pure cases.

Imagine a service that benefits everyone in a neighbourhood whether or not each person pays for it. An individual may reason that the service will probably be provided by other people's contributions, so they can keep their money and still receive the benefit. If many people make the same choice, too little money may be collected to provide the service at all.

This is the free-rider problem, and it helps explain why a public good matters in economics. A private seller may struggle to charge everyone who benefits because non-payers cannot easily be excluded. The market can therefore provide less of the good than people collectively would prefer, even when the total benefit is substantial.

The economic definition is based on rivalry and excludability, not on who owns or pays for something. A government can provide goods that are rival or excludable, while private organisations can sometimes provide goods with public-good characteristics. A public park, for example, may be widely accessible but can become crowded, making it partly rival when use is very high.

When learning about public goods, compare them with private goods. A sandwich is rival because if you eat it, someone else cannot eat the same sandwich, and it is excludable because a seller can refuse to hand it over without payment. Thinking through those two questions for each example is more reliable than memorising a list of government programs.

Economists have argued about lighthouses for a century. They were the textbook public good until Ronald Coase pointed out that British lighthouses were in fact built and run privately for centuries, paid for by fees collected from ships at nearby ports, which shows that even a textbook case can sometimes be financed by a workaround. A related category is the common-pool resource: fisheries, groundwater and grazing land are non-excludable like public goods but rival, because one boat's catch is gone for the next.

Garrett Hardin called the resulting overuse the tragedy of the commons, and Elinor Ostrom won the Nobel Prize for showing that communities often manage such resources well through their own rules, without either privatising them or handing them to government. The two-by-two grid of rival or not and excludable or not is the framework behind all of this, and it is worth drawing out and filling in with examples before an exam.

The two-by-two grid has a fourth box that is easy to forget: goods that are excludable but non-rival, often called club goods. A streaming service is a good example. One more viewer does not use up the film for anyone else, yet the company can block people who have not paid. Toll roads when they are not congested, private parks and many software licences work similarly. Club goods can be sold in ordinary markets because exclusion is possible, even though sharing them costs very little.

Where a good sits is not fixed by nature. Broadcast television was close to non-excludable, since anyone with an aerial could watch, so it relied on adverts or public funding. Encryption and subscription boxes turned much of it into a club good. Knowledge shows the same tension. A new scientific idea is non-rival, because your using it does not stop me, and once published it is hard to exclude people. Patents create temporary, legal exclusion, while governments fund basic research directly because markets would tend to supply too little of it.

Picture a street of ten households that would all like a new streetlight. It costs 200 pounds, and each household values the extra safety at 30 pounds, so the total benefit of 300 pounds clearly exceeds the cost. A neighbour collects voluntary donations. Each household reasons that its own 20-pound share makes little difference to whether the light is built, and that if the others pay, it will enjoy the light anyway. If enough households think this way, the collection falls short and a project worth 300 pounds to the street never happens.

There are several ways round this. A local council can pay for the light through compulsory taxes, which removes the option of free riding. Communities can also use pledges that only count if a target is reached: each household promises 20 pounds, but nobody pays unless the full amount is pledged. Some crowdfunding platforms use exactly this all-or-nothing rule, so a pledge is never wasted on a project that fails. Neither method is perfect, because people can still understate how much they value a public good when asked to contribute.

Because nobody pays for a public good at a till, there is no market price to show how much people value it. Yet a government deciding whether to build a flood barrier or protect a wetland needs some estimate to compare with the cost. Economists first look for revealed preferences, clues hidden in choices people already make. Hedonic pricing compares the prices of similar houses near to and far from a park or a quiet street, so the premium buyers pay reveals part of the value. The travel cost method counts the time and money visitors spend reaching a national park.

Where no such clues exist, researchers use stated preferences, asking people in surveys how much they would pay. This method, called contingent valuation, was used to estimate the damage after the Exxon Valdez oil spill in Alaska in 1989, and it remains controversial. Answers are hypothetical, so people may give generous figures to signal support, or low ones if they fear a bill. Answers can also ignore scale. In one well-known survey, people offered roughly the same amount to save 2,000 migrating birds as to save 200,000. Careful survey design reduces these problems but cannot remove them entirely.

A public good is a good that is non-rival and non-excludable in the pure economic definition. Those properties can create free riding because people may benefit without paying, which can lead markets to provide too little. Test any example by asking whether use reduces availability and whether non-payers can realistically be excluded.