When Should the Government Provide Public Goods?

Cite this Article
Thomas A. Lambert, When Should the Government Provide Public Goods?, Truth on the Market (July 14, 2017), https://truthonthemarket.com/2017/07/14/when-should-the-government-provide-public-goods/

My new book, How to Regulate: A Guide for Policymakers, will be published in a few weeks.  A while back, I promised a series of posts on the book’s key chapters.  I posted an overview of the book and a description of the book’s chapter on externalities.  I then got busy on another writing project (on horizontal shareholdings—more on that later) and dropped the ball.  Today, I resume my book summary with some thoughts from the book’s chapter on public goods.

With most goods, the owner can keep others from enjoying what she owns, and, if one person enjoys the good, no one else can do so.  Consider your coat or your morning cup of Starbucks.  You can prevent me from wearing your coat or drinking your coffee, and if you choose to let me wear the coat or drink the coffee, it’s not available to anyone else.

There are some amenities, though, that are “non-excludable,” meaning that the owner can’t prevent others from enjoying them, and “non-rivalrous,” meaning that one person’s consumption of them doesn’t prevent others from enjoying them as well.  National defense and local flood control systems (levees, etc.) are like this.  So are more mundane things like public art projects and fireworks displays.  Amenities that are both non-excludable and non-rivalrous are “public goods.”

[NOTE:  Amenities that are either non-excludable or non-rivalrous, but not both, are “quasi-public goods.”  Such goods include excludable but non-rivalrous “club goods” (e.g., satellite radio programming) and non-excludable but rivalrous “commons goods” (e.g., public fisheries).  The public goods chapter of How to Regulate addresses both types of quasi-public goods, but I won’t discuss them here.]

The primary concern with public goods is that they will be underproduced.  That’s because the producer, who must bear all the cost of producing the good, cannot exclude benefit recipients who do not contribute to the good’s production and thus cannot capture many of the benefits of his productive efforts.

Suppose, for example, that a levee would cost $5 million to construct and would create $10 million of benefit by protecting 500 homeowners from expected losses of $20,000 each (i.e., the levee would eliminate a 10% chance of a big flood that would cause each homeowner a $200,000 loss).  To maximize social welfare, the levee should be built.  But no single homeowner has an incentive to build the levee.  At least 250 homeowners would need to combine their resources to make the levee project worthwhile for participants (250 * $20,000 in individual benefit = $5 million), but most homeowners would prefer to hold out and see if their neighbors will finance the levee project without their help.  The upshot is that the levee never gets built, even though its construction is value-enhancing.

Economists have often jumped from the observation that public goods are susceptible to underproduction to the conclusion that the government should tax people and use the revenues to provide public goods.  Consider, for example, this passage from a law school textbook by several renowned economists:

It is apparent that public goods will not be adequately supplied by the private sector. The reason is plain: because people can’t be excluded from using public goods, they can’t be charged money for using them, so a private supplier can’t make money from providing them. … Because public goods are generally not adequately supplied by the private sector, they have to be supplied by the public sector.

[Howell E. Jackson, Louis Kaplow, Steven Shavell, W. Kip Viscusi, & David Cope, Analytical Methods for Lawyers 362-63 (2003) (emphasis added).]

That last claim seems demonstrably false.  

Public goods are regularly supplied by private actors without government coercion. For example:

  1. Having one’s downtown be free of impoverished beggars is a benefit that is both non-rivalrous and non-excludable, yet privately funded homeless shelters and soup kitchens are common.
  2. A beautiful, well-kept garden provides a vista that multiple users can enjoy without depletion (non-rivalrous) and from which passersby cannot easily be barred (non-excludable), yet many homeowners in populated areas expend significant sums, not to mention hours of hard labor, tending their yards.
  3. A highly educated citizenry tends to make better political decisions and to generate a richer cultural environment—both benefits that are non-rivalrous and largely non-excludable—yet people routinely spend great sums educating their children.
  4. Private groups regularly clean up roadsides, even though the benefit they are creating is not depleted as more drivers use the road (non-rivalrous) and cannot be limited to people who contribute to the clean-up (non-excludable).
  5. Millions of people make donations on “crowd-funding” websites like Kickstarter to finance projects like community theatre spaces, thereby creating the non-excludable, non-rivalrous benefit of more cultured communities.

These five examples highlight several common situations in which private actors, lacking access to (or refraining from using) state power, create public goods. Homeless shelters and soup kitchens (example 1) are often established and operated privately by groups of individuals whose moral or religious convictions lead them to make charitable contributions. Private provision of garden vistas and youth education (examples 2 and 3) occur because the personal benefit the provider receives from creating the amenity at issue is greater than her cost of doing so; if, for example, you receive tremendous personal happiness from having a beautiful yard or a well-educated child, you will spend a great deal on yardwork or tuition—even though much of the benefit of your expenditure inures to others. Private groups sometimes clean up roadsides (example 4) not only as acts of charity but also because doing so entitles them to public recognition on “adopt a spot” road signs.

We can see from examples 1 through 4, then, at least three situations in which private provision of public goods regularly occurs: (1) when the amenity at issue so appeals to conscience that individuals with moral or religious convictions are likely to pony up funding; (2) when some individuals have personal preferences (i.e., high enough reservation prices for the amenity at issue) that justify their bearing all the cost of the amenity, despite the spillover of benefits onto others; and (3) when some organization has successfully tied participation in the provision of a public good to some sort of excludable benefit (e.g., favorable publicity, etc.).

Example 5 demonstrates how the forces that often lead to private provision of public goods can be combined and then amplified through the use of creative contracting.  A typical Kickstarter campaign involves the features discussed above: appealing to people’s charitable inclinations (This project makes the world better!), catering to idiosyncratically strong preferences for an amenity (You personally benefit from this project!), and the tying of excludable benefits to public good provision (If you contribute, we’ll give you something!).  But Kickstarter then adds, well, a kicker aimed at overcoming a problem that may inhibit private provision of public goods.

People who might otherwise be willing to contribute to public goods, either for charitable reasons or because of idiosyncratically strong preferences for the amenities at issue, may not give if they are concerned that the contemplated projects will fail and their contributions will be squandered.  A homeowner living near an unprotected portion of a river, for example, will worry that the organizers of a private levee project may ultimately end up with insufficient funding, resulting in either no levee or one that is too short to provide adequate flood protection.  Despite the substantial private benefit she would attain from levee construction, a homeowner may choose not to contribute to the effort because she’s worried that her donation will be effectively squandered.  If other homeowners reason similarly, the voluntary levee project is destined to fail.  Private providers of public goods thus need some way to assure potential donors that their contributions will not go to waste.

A mechanism often used by such providers—and a key to Kickstarter’s business model—is the “assurance contract.”  In an assurance contract, project organizers promise potential donors that if sufficient funds are not raised (or committed), the organizers will return (or refrain from collecting) the donors’ contributions.  On the Kickstarter platform, potential donors are assured that their credit cards will not be charged the amount pledged unless donors collectively pledged some amount—enough money to complete the project at issue—by a certain date.  That assurance protects donors from having their contributions squandered if the fundraising effort falls short.

A variation on the assurance contract may be even more helpful in facilitating the private production of public goods.  Even when a solicitor of funds has assured donors that their contributions will be returned if the financing effort doesn’t succeed, individuals with charitable inclinations or idiosyncratically strong preferences for the amenity at issue will still be reluctant to commit funds if they believe the fundraising effort will ultimately fail.  Why go through the rigmarole of making a donation and tie up the committed funds for some period of time if the project at issue is unlikely to make?  Of course, if too many potential donors think this way, failure of the fundraising effort becomes a self-fulfilling prophecy.  Accordingly, organizers of efforts to finance public goods need some way to get enough initial funds committed that the project appears likely to succeed and will appeal to less enthusiastic donors.

A “dominant” assurance contract can assist project organizers with this first-mover problem.  Under such a contract, an entrepreneur seeking to provide a public good promises potential donors that if the project fails for lack of sufficient funding, their funds will be returned and they will be paid some amount (or given some other tangible benefit).  The latter promise induces initial donors to pony up and get the ball rolling. As Alex Tabarrok has explained, “Pledging is now a no-lose proposition—if enough people pledge you get the public good and if not enough pledge you get the prize. A contract like this makes it a dominant strategy to pledge and so the public good is funded.”

It seems, then, that the aforementioned claim that public goods “have to be supplied by the public sector” is just wrong.  But perhaps I’m being unfair to the quoted economists. After all, they didn’t say the public sector must provide public goods if they are to be provided at all; rather, they asserted that such goods “are generally not adequately supplied” by private actors and therefore “have to be supplied by the public sector.” The implication is that because private provision of public goods tends to be inadequate, public provision of such goods must be the optimal approach.

Assessing that claim requires some thought about how to identify the optimal approach to securing public goods. One might think that the optimal approach would be the one under which every unit whose marginal benefit exceeds its marginal cost, but no unit beyond that point, is produced—in other words, the approach that generates the ideal, or theoretically welfare-maximizing, level of production. But that is the old Nirvana Fallacy! Theoretical perfection is impossible in our fallen world, so policymakers should strive not for ideal production (no approach can attain that) but for the best of achievable outcome. They should catalog available policies (when it comes to public goods, they are chiefly government provision of the amenity at issue or reliance on private sector provision), assess outcomes under each, and select the approach that generates the highest level of welfare. That is the optimal approach. And there are certainly reasons to doubt that government provision of non-rivalrous, non-excludable amenities is always optimal. In particular, both the Hayekian knowledge problem and public choice concerns rear their ugly heads.

When the government provides a public good, officials must decide, among other things, how much to produce. With some public goods—say, a lighthouse—there may not be many choices: one or zero. But most public goods—national defense, fire protection, flood control systems, public art—can be produced in different amounts (and even a lighthouse can be built to different heights and with lights that reach different ranges). Should our town have three fire trucks or ten? Should we install two giant sculptures downtown or six? Should the levee protect against only fifty-year floods, or should we build it high enough to protect us from the hundred-year deluge? Private providers of public goods confront similar questions, of course, but because they rely on voluntary participation by contributors, they face a constraint that will prevent overprovision: contributors’ willingness to support the project at issue. Government planners with access to the public fisc lack a similar mechanism to tell them enough is enough.

That is particularly problematic given that the funds the planners are spending are coerced from individuals who may not support the particular public good at all or, at a minimum, may want less of it. Pacifists, for example, are appalled at current levels of military spending; conservative Christians, at public funding of art projects like the 1987 photograph “Piss Christ” (for which the artist received $15,000 from the National Endowment for the Arts); many secularists, at abstinence-only sex education (for which federal funding is available). In light of inevitably divergent preferences among taxpayers and the absence of any feedback mechanism other than the ballot box, government planners are likely to make significant mistakes in deciding how much of a public good to produce.

But what about that ballot box? Won’t it punish (and ultimately prevent) mistakes about how much of a public good should be provided? Not very well. For one thing, an election assesses everything the candidates have done and promised and therefore sends only a muted signal about any particular expenditure of public funds. What’s more, government officials’ decisions to spend tax revenues on public goods exhibit a feature that makes them particularly impervious to political correction. While many or most members of society receive some benefit from government provision of public goods (e.g., protection from fire or flooding, a more aesthetically pleasing downtown), some individuals and groups enjoy special benefits that are not generally available. Firemen and levee builders receive business opportunities from government provision of fire and flood protection; artists, from enhanced expenditures on public art projects. While these special benefits are concentrated on small and discrete groups, their costs are borne by taxpayers as a whole. As public choice scholars have demonstrated, government programs involving concentrated benefits and diffuse costs tend to persist even when they reduce total welfare by imposing marginal costs in excess of marginal benefits. Such programs are resilient because small, easily organized interest groups have much to lose if the program is curtailed and will fight curtailment accordingly, whereas each individual beneficiary of curtailment—each taxpayer whose money is being wasted—receives only a tiny personal benefit from cutting back on the program and is thus unlikely to expend resources to secure that end.

In light of the knowledge problem and the potential for manipulation by special interests, government provision of public goods creates a need for careful monitoring. And therein lies the irony: The monitoring of government provision of public goods is itself a public good!  It is non-rivalrous, for when you enjoy my monitoring of public officials, your neighbor’s ability to do so remains undiminished. It is non-excludable, for if I monitor officials to protect my own interest, I can’t prevent you from benefiting from my efforts. This suggests that monitoring of government efforts to provide public goods—an activity that is essential to ensuring that government production approaches ideal production—will be systematically underproduced. Government provision of public goods, then, is no silver bullet.

Of course, none of this implies that government provision is never the optimal policy response to a particular public good. The point is simply that both government and private provision of public goods are likely to diverge from the ideal production level. The optimal approach is the one that is likely to diverge less. With respect to national defense, government provision is almost certainly optimal. Free-rider problems would likely hold private provision of national defense well below the ideal level, and while the knowledge problem and public choice concerns could generate some overproduction of government-provided defense services, any inefficiencies there would be dwarfed by those resulting from reliance on the private sector. Most publicly viewable art projects, by contrast, are probably best provided by the private sector. Although such art projects are susceptible to free-riding, people’s charitable inclinations, idiosyncratically strong preferences, and desire for excludable benefits that are frequently tied to the project at issue (e.g., a plaque listing donors) often lead to significant sums for publicly viewable art. Moreover, conflicting preferences among taxpayers are likely to lead to an intractable knowledge problem (How many art projects should be funded? Which ones?), and the concentrated benefits/diffuse costs problem will make it difficult, as a political matter, to cut back on public art programs. Flood protection may be somewhat in the middle. Free-rider problems exist, but many communities—especially smaller ones with more communitarian social norms—have witnessed private levee projects.

A good starting point for policymakers confronted with a non-excludable, non-rivalrous good is to ask whether this type of public good is regularly provided by the private sector in this type of community. Policymakers should always remember, though, that there are pros and cons to both private and government provision of public goods, and they should opt for the latter only when they have sound reasons to believe that it promises greater net welfare.