Are Public Goods Necessarily Undersupplied?

In economics, public goods are goods which are non-rival (a person’s use of the good does not reduce the ability of another person to use the same good, eg listening to the radio) and non-excludable (people who do not pay cannot easily be prevented from using the good, eg when a burglar is arrested, everyone in a neighborhood benefits, not just those who paid for the security service). Because of this definition of public goods, we tend to teach undergrads that public goods will therefore necessarily be undersupplied, that in a free market the amount of the good produced is less than the socially optimal level of production. As such, government may be able to step in and, though use of taxation, correct this underproduction (see, for example, Page 369 of Modern Principles of Economics by Tyler Cowen and Alex Tabarrok).

But is it necessarily the case that public goods are necessarily undersupplied in a free market? It does not seem clear to me that it is.

The first question we need to ask is “as compared to what?” What is the free market outcome undersupplied compared to? It is compared to what would be the socially optimal level where everyone who benefits pays the cost (the intuition here is this: if an individual can earn more producing something, they will produce more of it, all else held equal. Supply curves slope upward).

Now we need to ask: is this an attainable alternative? In a free market setting, it does not appear to be so. After all, as we argued above, given the characteristics of a public good, they will tend to be undersupplied. Getting people to pay for their use is difficult. A more technical way of saying this is the transaction costs are high. The marginal benefit of receiving the payments exceeds the marginal cost of obtaining those payments. In a zero-transaction cost world, the socially optimal level would be easily obtained. There is some bargain that could be reached where those who enjoy the benefit without paying the cost (free rider problem) could be incentivized to pay the cost and production would increase. This is just an application of the Coase Theorem.

If, however, as posited by the public goods problem, the transaction costs of solving the free rider problem are too high, then the socially optimal level is not necessarily an attainable alternative. It’s a fantasy alternative. Thus, it is an irrelevant comparison. It’d like saying “I’d be better off with a fairy godmother who grants wishes than needing to work for my well-being.” Sure, but given faeries don’t exist, that’s a meaningless choice. The choice is between working and living well or not working and living poorly.

If the socially optimal outcome of the model is not a real alternative, then the situation is already at an optimal outcome. There is no undersupply. Thus, public goods are not necessarily undersupplied.

A note of caution: none of what I just wrote should be taken to mean that the free market outcome is necessarily the best outcome. There may be better alternatives. Government (or some other non-market force) may be able to achieve an alternative arrangement that is superior to the free market outcome. For example, better defining property rights can lead to less undersupply of public goods. But in movement from one alternative to the other, we need to consider the transaction costs. Do the benefits of moving from the market alternative to the non-market alternative outweigh the costs?

With this article, I reiterate a point made by Ronald Coase, Carl Dahlman, Harold Demsetz, and many others before me: transaction costs matter. We need to compare attainable alternatives and consider how institutions actually work as opposed to an idealized version of them. Comparing a market outcome to an idealized, but unobtainable, alternative does not provide any guidance to our thought.