Public services are good, but they're not "public goods" by Matthew Yglesias

Public services are good, but they're not "public goods"

But true public goods are also good and deserve more support

A recent David Zipper column in Bloomberg noted that mass transit is not a public good in the technical economics sense and that this is relevant to arguments about whether cities should invest in $0 fares.

Some of the discussion generated by the piece focused on whether (and how much) it matters that people are using a piece of economics jargon correctly. I have mixed feelings about this. Transit is certainly an example of a public service — something like parks and roads and fire services that are provided by the state in most places. These services normally exist in part because people think they are good, so if folks want to use “public good” to mean something that is publicly provided and that they think is good, then I’m not sure that’s the worst thing in the world.

The economics terminology, after all, is confusing. Most of what the public sector does is not a public good, and lots of public goods are provided by private nonprofits and sometimes even just by individuals. Matt Grossman, for example, runs a Twitter feed that’s light on commentary and long on neutrally-presented links to recent political science research. That’s a kind of public good even though it has nothing to do with the public sector.

Separate from the linguistic issue, though, I do think the concept of a public good as something distinct from “stuff the government does” or “stuff that I think is good” is useful. Because the point of the concept is that some things that are important are likely to be underprovided by a market economy, and we should do things to increase their provision. And we should!

Public goods are non-rivalrous and non-excludable

The economic definition of a public good requires it to be non-rivalrous and non-excludable.

This is perhaps easiest to understand by thinking about the opposite of a public good, something like a house. My house is my house, and your house is your house. Kate and I share our house with each other, but we do not share it with you. The house has a door that locks and we both have keys, but you do not have a key. We also have a little piece of paper that says we own the house. If you came into the house and started living here, that would be a big problem for us — a house is rivalrous. But also we would not let you do that. We have a locking door and if push comes to shove, the police will remove you from our house — it’s excludable.

The classic textbook public good is a lighthouse.

If a community gets together to put up a lighthouse to warn ships away from the rocks, they can’t stop other people who sail by from seeing the light — it’s non-excludable. At the same time, other people seeing the light and avoiding the rocks doesn’t degrade the community’s use of the lighthouse at all — it’s non-rivalrous.

A lot of stuff the public sector does falls in a kind of mushy middle. A playground is absolutely excludable, but in practice, most playgrounds do not exclude people. And a playground is also rivalrous — it can become so crowded as to degrade the experience of other playground users. But again, in practice, that’s not usually the case. A normal playground in a normal city park functions a lot like a lighthouse. It’s free to the public, all are welcome, and all users are pretty happy that it’s there.

But then you have something like a highway. Most highways are functionally non-rivalrous most of the time, but we’re also all very familiar with the existence of traffic jams. Lots of things are like this — not genuinely non-rivalrous public goods, but also amenable to a decent degree of sharing. These are “club goods,” and oftentimes government, especially local government, acts as a club.

In the club

A good example is something like a public library. These are very common in the United States, though libraries also exist outside the public library system — most notably the libraries that are run by colleges and universities. And in a world without government-run libraries, we’d likely see more private ones.

The reasoning behind libraries is not that “books are public goods” but that shared ownership of books can be highly efficient. Readers generally value having access to large quantities of books but only want to read a small number of books in any given week. Having each member of a town maintain their own personal stockpile of books at home is highly inefficient. The vast majority of people’s personal libraries would be quite small, limiting access to books, with the vast majority of the books sitting unused most days. Everyone who reads and has friends who read shares and swaps books for this reason. And it might make sense to formalize this arrangement and create book-sharing clubs, where you pay a regular membership fee to someone who maintains a big lending library of books.

A standard public library is essentially a town- or county-wide book-lending club with mandatory participation. Why structure it this way? Formally, organizing public libraries creates cross-subsidies from non-readers to readers and from rich to poor. It also in practice ensures that modern-day libraries are involved in the provision of public services beyond bookpooling.

In my younger and less Burkean days, I sometimes argued that this is a bad setup and we should eliminate libraries in favor of explicit cash transfers to poor people. Nowadays, I don’t see any real reason to subject longstanding community institutions to this kind of scrutiny. It’s just genuinely the case that local governments perform a lot of club-like functions, and this introduces a certain amount of arbitrariness. Almost every town has one or more parks. But what’s actually in the parks varies considerably from place to place, and I don’t think there’s any right answer derivable from the abstract principles of justice that will tell you whether maintaining public tennis courts is important. If we were talking about an HOA rather than a town government, we’d just say “well, it’s their condo/gated community/whatever, so they can maintain a pool or not just according to what they want.” And even though town governments are public rather than private, the distinction can get pretty fuzzy, which is one reason right-wing people can sometimes come to convince themselves that heavy-handed zoning rules are consistent with their free market principles.

I think understanding the services provided by local government — including, really, schools and public safety — as club goods helps explain a lot about how this stuff functions.

In particular, small towns really function as proper clubs. Everyone in Blue Hill, Maine either uses the town library or else chooses not to use it because he doesn’t find it useful or valuable. In Washington, D.C. most people who do use the D.C. Public Library system don’t actually use the majority of libraries — they just use the library by their house. Most MPD officers don’t work in my neighborhood. I go to city parks all the time, but the parks I frequently use are a small minority of the citywide park system. The larger club is able to perform more specialized functions and redistributes across a broader swathe of citizens, but it’s also much harder for users to monitor and allows for more waste.

Natural monopolies

Another attribute of club goods is that they tend to arise in situations where the idea of having lots of overlapping competing entities isn’t very appealing. You could imagine several different library clubs that compete with each other in a big city, but that would be pretty wasteful. In smaller communities, the addressable market just isn’t big enough. You could have competing chains of exclusionary parks (indeed, entities like SkyZone are a version of this), but it would ruin some of the spontaneity and fun of a stroll through the park if you had to stop and pay a fee.

Perhaps most salient of all are roads.

Roads can absolutely have gated access via tolls, and there are plenty of examples of private landowners operating private roads for private use, typically on big stretches of rural land. But while a heavily used highway certainly can be in private hands and tolled for profit, there are a few problems with this. The main one is geographical: it’s not logistically feasible to maintain a highly competitive market with many buyers and many sellers and free entry into and exit from the highway market. There just isn’t space for that. And when you think about city streets and suburban arterials, the problem is even worse. You can have various degrees of privatization, but you’re always going to be talking about low-competition, regulated-utility markets.

And the same is true of mass transit. There are different levels of privatization at work in different countries around the world (an odd fact about hyper-capitalist America is that our transit systems are unusually privatization-averse), but you don’t really see competition between rival transit networks. It just doesn’t make sense as a way to organize an urban rail system.

Advanced economies have lots of different models for handling natural monopolies. In the United States, we normally have direct state ownership of municipal water systems but private ownership of electrical utilities. Around the world, though, you can find places where water systems are privatized — even in the United States we have some public electrical utilities. But under either system, every place I can think of charges you for water or electricity because they’re concerned about over-use. By the same token, even though parks are mostly free, if you want to go see a popular national park like Acadia you need to pay. There are non-fee ways of avoiding park crowding, but the fees work and they help raise funds for the whole system.

I think one reason real transit-heads like Zipper and me and Jerusalem Demsas don’t like the $0 fare fad is that its feasibility depends on ridership being low. The LA Metro is never crowded, so it makes superficial sense to think of it as non-rivalrous. But if you care about transit, the sensible thing to do with extra money is to use it to make the product better (rather than, in LA’s case, a venue for drug abuse). The problem is that the urban progressive coalition’s natural inclination is to turn everything into a make-work jobs program rather than care about service quality.

We should have more public goods

Real public goods are a thing, though, and while the American government absolutely does invest in their production, there’s plenty of scope to do more.

The true public goods are things like information and research. The fact that I know what the National Weather Service forecasts for tomorrow doesn’t take anything away from you or anyone else who might also want this information. It’s of course possible to create paywalled sources of information to generate private profits. But this involves taking something that is by its nature non-rivalrous and creating artificial scarcity. Sometimes that’s the best available option — I run a subscription newsletter, so I’m not going to tell you otherwise — but scarcity is bad and information abundance is good for the world.

This is why something like the lack of timely national data on homicide incidence bothers me so much.

“How many people were murdered in February 2023, and roughly where did those murders happen?” is the kind of question you would expect there to be a good, accessible, public-sector source of information on. And lots of individual cities do in fact do this (typically jurisdictions where at some point crime has been a hot local political issue), but most of the country doesn’t, and there’s no standard reporting framework. And that’s probably the most straightforward criminal justice question we can ask. It would be nice to have full information about police stops, arrests, prosecutions, and the whole deal. It sometimes feels like facts and data play no role in our public debates, but I do think even the most inane partisan arguments tend to be, to some extent, downstream from research and policy development, which in turn is downstream from data.

Unfortunately we are moving backward in some areas, as with the drive to make Census data worse and less accurate out of a misguided zeal for privacy. I was glad to learn recently that they’ve shelved plans to apply the same methods to the American Community Survey, but we’re still left with the fact that one of the crown jewels of true public goods production is being deliberately sabotaged.

On a loftier level, it’s important to think about the public sector’s role in scientific research. The American government both directly and indirectly funds a lot of research, which is good. But there are big questions hanging around this. The NIH Director job, for example, is lingering open in part because it only pays $200,000 per year, which is much less than qualified applicants make — it’s about half what the NIH Director’s direct reports get. This is obviously not just a question of money, but it’s weird to cheap out on the leadership of an agency that distributes so much money. More substantively, approving grant applications by committee tends to generate a lot of consensus science, which is antithetical to the goal of generating breakthroughs. We also have enormous barriers to actually doing research. The ethical strictures on medical research seem to me to wildly understate the value of letting people volunteer to participate in experiments. I’ve never heard a social science researcher say anything good about the IRB process.

This stuff tends to be neglected, but while it obviously takes money to do research, there’s not a one-to-one relationship between dollars spent and interesting discoveries. The work has to actually be done on a granular level, and making that difficult is very costly to society.

End of sermon! The point is that public goods are a valuable concept and something we should be more focused on creating. There’s also lots of other stuff the public sector does and some of it is good, but relatively little of it has the special character of a public good.