Should Socrates have drunk the hemlock or was Antigone right to claim that one may morally disobey political authority? What is the state for? The economic concept of “public good” is central to answering these questions in economics and political philosophy. Here’s a short introduction—with some further questions.
A public good (sometimes called “collective good”) is something for which all members of a group are willing to pay some price, something that they will all be able to consume simultaneously, and for which it is impossible (or too costly) to charge a price to consumers. Think about fireworks over a city or national defense. Whether or not the state is required for the production of public goods is what makes James Buchanan a limited-government (classical) liberal and Anthony de Jasay an anarchist (even if other considerations such as the effects of political competition play a role).
The theory of public goods as today’s economists understand it was elaborated by Paul Samuelson in two articles of the Review of Economics and Statistics: a mathematical exposition in “The Pure Theory of Public Expenditure” (Novembre 1954), and a graphical and less technical presentation in “Diagrammatic Exposition of a Theory of Public Expenditure” (November 1955).
A public good can also be viewed as something that transmits positive externalities to all individuals in a group. An Econlib article by Bryan Caplan provides a good and short explanation of the concept of externalities. Although externalities and public goods are often thought as distinct concepts, their overlap has been noticed in the economic literature—from identifying “public good externalities” as a category of externalities (Francis M. Bator, “The Simple Analytics of Welfare Maximization,” American Economic Review [March 1957]), to noting that, between the two concepts, “there are no purely formal differences as far as necessary and sufficient conditions for top-level optimality are concerned” (E.J. Mishan, Introduction to Normative Economics, Oxford University Press, 1981). A public good exists, then, when its production for the benefit of paying customers automatically creates positive externalities for all those who do not pay.
Externalities are a more complex concept than many imagine. As noted by Mishan, they extend to the mere “an awareness of what is happening to others.” You transmit a negative externality to me if you privately do (or even just think about) something I don’t like being done (or thought about). This follows from the definition of externalities combined with the subjectivity of individual preferences. In a sense, public goods are a simpler concept than externalities because, by definition, they are unanimously liked. They are also more difficult to find.
Yet, there is little doubt that public goods exist, but the question is whether they exist for a group defined over a country, that is, for everybody in the group occupying a large territory dominated by a central state. Most if not all public goods are “public” only for a specific subgroup of society.
Are there public goods that justify the domination of a (central) state over a (wide) territory? De Jasay answers negatively. For Buchanan the answer is positive: individuals contract (implicitely) to form a country because there are some public goods such as national defense that they want.
Many questions remain. Buchanan uses a less technical or a priori concept of public goods that makes their existence depend on the definition of property rights in the social contract. Does not this open a Pandora box? What have the participants in the implicit social contract really “signed”? How can the state, if it is necessary, be limited? On de Jasay’s side, how can anarchy maintain something resembling a free society with formal individual rights and the rule of law?
Of course, other anarchist and liberal theorists exist. I am focussion on de Jasay and Buchanan because their theories are especially interesting and because they show the central place of public goods.
Can we avoid choosing between de Jasay and Buchanan by rejecting the relevance of public goods in social theory and framing the social problem in terms of coordination by evolved rules? This is the approach of David Hume, the Scottish Enlightenment, and Friedrich Hayek. But it is not sure that this third way provides an escape. De Jasay claims to be following Hume. And Buchanan suggests that “evolutionary and contractarian explanations can be complements rather than substitutes” (Economics and Philosophy 4 [1988]).
De Jasay and Buchanan share many ideas. They are both methodological individualists, that is, they study society starting from the positive positive observation of individual preferences and choices. Normatively, they believe that no economic or moral value exists outside the preferences of individuals.
In a sense, both de Jasay and Buchanan believe that anarchy is the ideal. De Jasay is not sure that an anarchic society could avoid being taken over by a foreign state. Buchanan sees the liberal contractarian state as protecting “ordered anarchy.” He echoes French philosopher Raymond Ruyer, who, in his 1969 book Éloge de la société de consommation (“In Praise of Consumer Society”), wrote that “real anarchism, feasible and realized … is simply the [classical] liberal economy.” In the same vein, another French classical liberal, Émile Faguet, wrote that “an anarchist is an uncompromising liberal” (Politiques et moralistes du dix-neuvième siècle [“Nineteenth-Century Politicians and Moralists”], 1891).
This suggests another approach to avoiding a definitive choice between de Jasay and Buchanan: the closer we can get to anarchy, the better; if we can’t get the real thing and the state is necessary, it is to produce or finance the few public goods (narrowly defined) that would go unproduced or seriously underproduced under anarchy. In this perspective, paradoxically, the humble role of the state is to maintain and protect as much anarchy as possible.
READER COMMENTS
Dylan
Dec 6 2019 at 10:28am
Deja vu all over again? Wasn’t this posted last week?
Pierre Lemieux
Dec 6 2019 at 10:43am
You have a keen and early eye, Dylan… up to a point. A draft was mistakenly posted for a few hours last week, due to a scheduling mistake of mine. I rescheduled it when, at breakfast, I discovered my mistake. If your keen eye rereads and compares the current version, it will discover a large number of changes (including substantial changes) and a whole new section at the end.
Jon Murphy
Dec 6 2019 at 1:40pm
I’m glad you’re thinking about this. I am doing some similar work from a Buchanan/Adam Smith framework and focusing on pecuniary externalities. It’s a tough question to wrestle with.
Pierre Lemieux
Dec 6 2019 at 2:43pm
Jon: The way I see it (and I think Buchanan and de Jasay would have agreed), pecuniary externalities are nothing but transfers; so they are not externalities. Do tell us more about your research, and how it could affect (if it does) the importance of public goods in the debate between classical liberalism and anarcho-capitalism. And what would Smith say about this?
Jon Murphy
Dec 7 2019 at 2:34pm
I agree that pecuniary externalities are transfers, but I do not agree that fact renders them not externalities. An externality occurs when one person’s behavior has an effect on another person. More precisely, using Scitovsky’s (1954) definition, if a firm i produces output q_i using inputs (x_1, y_1…), Firm 1 has a production function of:
q_1=f(x_1, y_1,…;q_2, x_2, y_2…)
then a technical externality exists when the portion of the production function to the right of the semicolon is non-zero. In other words, a technical externality exists when the actions taken by Firm 2 affects the physical output of Firm 1, holding constant Firm 1’s level of input usage.
Continuing on with Scitovsky’s definition, given that a firm’s profits are also a function of inputs and outputs, so Firm 1’s profits can be represented:
P_1=f(x_1, y_1,…;q_2, x_2, y_2…)
where P is profit (I typically represent profit as pi, but I don’t know how to type that here). Again, if any of the terms to the right of the semicolon have a non-zero effect on P_1, then an externality exists but if the externality has no effect on output, then it is pecuniary. Pigou and Viner also make a similar distinction. Note also how this definition has an accord with Buchanan’s definition with Stubblebine:
Furthermore, Alchian and Allen define externalities as: “effects on other people whom adequate – mutually agreeable- compensation is not made.” (Universal Economics, pg 36). They include pecuniary externalities in this area (pg. 38), but note that pecuniary externalities do not deserve compensation in the same way physical externalities do because no one has a right to market value.
So, I do not think that calling pecuniary externalities a transfer overrules the fact that they are external effects on third parties.
Which leads me to public goods. As you point out above, a strict interpretation of a public good implies it is “unanimously liked.” Uanimonity is sometimes cited as the ideal, and Buchanan points to his justification for using the unanimity rule as a starting point for constitutional analysis based on the fact he sees it as a market analogy (see The Demand and Supply of Public Goods, pgs 7-8 of the Liberty Fund edition. See also pg. 79). However, market transactions are never unanimous because of the existence of pecuniary externalities (see also Russ Sobel and Randy Holcombe’s 2001 Public Choice paper “The Unanimous Voting Rule is Not the Political Equivalent to Market Exchange”). Indeed, the point of a competitive economic system is to produce pecuniary externalities; that is, to reduce “extra-normal” profit. When a person chooses to buy from one seller and not another (or even chooses not to buy at all), there is some third party that may object and claim harm (this argument is very popular. Indeed, it’s used by the President to justify his tariffs). So, the unanimity criteria favored by Buchanan and the Social Contract Theorists can never be, even as a “state of nature” type of argument.
Anarcho-capitalism seems to rely on unanimity, at least to some degree, to me. They seem to fall into the same trap of “market exchange is purely voluntary, so we can just extend that to politics.” But any change from a status quo, either from one equilibrium to another or within a disequilibrium situation, will have some folks who face pecuniary externalities and thus will not consent to the change. Who, in an anarchist society, decides whose objections will be heard and whose will be not? What the “rules of the game” are? Many anarcho-capitalists seem to be working under the assumption of unanimously consented to institutions that are prior to the system; in other words, anarcho-capitalism seems to fall right back into SCT faults.
So, with unanimity out, the question of “what are the rules of the game” remains open. If politics and markets are not consensual, what remains? One option is coercive, and that does explain a lot of both institutional frameworks. But so does consent.* Why is it that firms and individuals can be compensated if their utility is reduced by another individual (Buchanan’s definition of externality, also Scitovsky’s), but not their profits? Why are property rights situated in that manner?
I argue, following the “third way” you discuss above, that convention is the answer. Markets and politics are conventional, not consensual. What constitutes a “public good” in practical political language is indeed conventual. As I wrote back in July (and I think no disagreement around these parts), what a public good is depends a lot on the analyst looking at the good; it is not as precise as strict theory might make it seem. Thus, we rely on some conventional roles of government (ie, national defense) to justify public good classifications.
Now, it is obvious to me that, for the next steps on this research, I need to address de Jasay’s arguments that he is following Hume (I think I have at least sketched out my disagreement with Buchanan viz contractarianism and evolution). Frankly, I do not know enough de Jasay to respond right now.
*To be clear, I am not saying consent plays no role in markets and politics. Rather, I am saying it is not the foundation of either.
Pierre Lemieux
Dec 8 2019 at 11:38pm
Interesting defense of the third way, Jon. This is an important topic. A few disorganized comments:
If you have not read it, Robert Sugden’s book The Economics of Rights, Cooperation and Welfare presents a similar defense. Moreover, I think your approach is consistent with the “Pandora box” opened by Buchanan who says that whether something is a public good or not depends on how property rights are defined.
Resurrecting Scitovsky’s “two concepts of external economies” is interesting too. I can’t immediately find his article on my bookshelves, but I think your equation 2 should include input and output prices. It is not obvious to me that Scitovsky’s pecuniary externalities are consistent with Buchanan and Stubblebine’s definition.
Distinguishing externalities from transfers appears to be a useful distinction. They allow viewing exchange and markets as consensual. Otherwise, you may have to introduce moral dei ex machina such as “no right to market value,” instead of deriving the principle from the benefits of exchange.
Consider that Scitovsky (to the extent that I remember his thesis and understand your exposé) transforms too many things into externalities, including competition itself. If it is useful to extend the concept, why not extend it instead to externalities in consumption? It then becomes obvious that property rights are a way to control externalities (if everybody had the same living room, externalities would be everywhere)—in a way perhaps less convoluted than identifying transfers to externalities.
We should not forget that Buchanan’s political unanimity is unanimity on rules, which rules are the rules of the game.
De Jasay and Sugden rely on conventions to establish “property rights”—although Buchanan would replace this concept by “liberties”: see his Social Justice and the Indian Rope Trick, and my Regulation review.
Jon Murphy
Dec 9 2019 at 7:12am
Thanks for the comments, Pierre. A lot of what you’ve said has been critiqued to me in the past and it is stuff I am working on/thinking on. The stuff I do not respond to here is because I don’t have a good response, yet.
Which I think does come into play given excludability depends partly on rights. But whether or not something is a public good depends also on the cost-influencing choice of the person providing that good. An outdoor concert hall, for example, can be a public good. But, for the right cost, the provider could make it a club good by building up walls and a dome. By not doing so, the provider chooses to make the good public. So, I just mention this as part of the discussion on public goods that defining one in real life is tricky.
You’re right. An error of transcription on my part.
I agree that for the purposes of describing markets as consensual, distinguishing transfers from externalities is useful. But it is precisely that description of markets I think is incorrect and such the distinction ends up being misleading. I actually think the concept of pecuniary externalities helps us avoid the moral dei ex machina, or at least helps us discuss in normative terms, the allocation of rights. But that’s a longer point I’m going to punt on for now.
I agree.
Yes, this is absolutely correct, but as I read his metaphor usage I cite above, he uses that criteria on rules because it is how the market works.
Alex
Dec 7 2019 at 10:57am
I have a problem with how this problem is often framed. A public good is valued enough to pay for its cost, but because of externalities gets underproduced. The essential feature here is non-excludability, not transaction cost. Transaction costs are a weak argument in the age of information technologies. The real problem is that goods with positive externalities cause everybody to think they can free ride by pretending not to value them, since everybody thinks they get them anyway.
Imagine a facial recognition system on your smartphone good enough to function as a reliable lie detector. In that case, we could just ask everybody if they want to pay for the good. The whole public good problem would disappear. If enough people agree to pay a price high enough to allow the public good to be produced, the case is closed. If you agreed, it’s a violation of contract no to pay. It’s not an aggression to force you to pay. If not enough people agree, they really don’t want the good. In that case, it isn’t a public good and nobody can argue otherwise.
The real problem with public goods is that we don’t have a reliable lie detector. We don’t trust each other.
Jon Murphy
Dec 7 2019 at 2:34pm
Excludability is a transaction cost issue. That’s more or less how it is defined.
Alex
Dec 8 2019 at 10:15am
Okay, I see. Excuse me, I’m neither a student of economics nor is English my first language.
But I think my point stands. Since the problem vanishes the moment we assume agents not capable of lying (using a reliable lie detector), this shows that the production of a public good could be done by markets, no matter its non-excludability. Public goods are goods that (in a libertarian society) should only produced for and paid by people who consent to their production and these people have reason to consent because by definition they unanimously want the good and they know they can’t lie about their preferences.
It all boils down to a problem of trust. Solvable by a reliable lie detector, which would eliminate the possibility of free riding or forced riding.
By the way, it’s not important that no public good exists that everybody values, all that matters in a libertarian society is that enough people exists who value a non-excludable good to pay for its cost of production and that nobody in this group can free ride (and nobody can be forced to ride if not in the group).
To make this operational, a producer just ask the people he/she wants to make pay if a non-excludable good shall be produced. Everybody has time to declare himself/herself (using the lie detector). No declaration counts as yes. If enough people opt out to make production unprofitable, it’s no public good.
Jon Murphy
Dec 8 2019 at 11:06am
I don’t see how lying matters at all. Everyone has an incentive to lie about their willingness to pay regardless of excludability or not. As you rightfully point out, all that matters is that costs are covered. Thus, we do have many public goods provided by non-governments (eg, the lighthouse, outdoor concert halls, etc).
Alex
Dec 9 2019 at 5:54am
I can’t follow. In case of an excludable good (that I value), lying would be irrational. If Disney+ ask me if I want a streaming subscription and I judge it worth the price, but nevertheless pretend not to want it, I just don’t get it. End of story.
In case of a non-excludable good pretending not to want it (not signing a contract), make sense, since if it gets produced (if enough other people are honest) I can free ride. Now, if everybody tries to free ride, it won’t get produced in the first place which is precisely our problem. This allows the state to come in and force people to pay.
Jon Murphy
Dec 9 2019 at 6:54am
Very true. And yet, there are countless so-called public goods which are produced, and at amounts to satisfy market demand, privately. Ronald Coase’s paper “The Lighthouse in Economics” (Journal of Law & Economics, 1974) highlights one such case. As does countless outdoor amphitheaters scattered throughout this country. My friend Rosolino Candela and his co-author Vincent Geloso also identify another example in their paper “The Lightship in Economics” (Public Choice, 2018).
So, then, we are faced with a problem: why do so many supposed public goods exist absent the state? Do people not compulsively lie about their preferences despite the incentives? Possible; that is likely part of it. But could something more be at play?
Mark Brady
Dec 8 2019 at 3:55pm
A question for Pierre. In your post you used neither the word “nonexcludable” or the word “nonrival.” Was this a deliberate choice on your part? And if it were a conscious choice, why did you avoid either term?
Pierre Lemieux
Dec 8 2019 at 6:47pm
Mark: I was conscious of that, but it seemed to me that my definition amounts to the same and is perhaps easier to grasp for non-economists, don’t you think? (I am willing to admit that this pedagogical goal took a beating later on!)
Pierre Lemieux
Dec 8 2019 at 11:45pm
Moreover, I was trying to make some room for Buchanan’s expansive definition. For example, if people unanimously agree with Locke’s proviso, all non-privatized public land becomes a public good (in Buchanan’s sense).
Michael Pettengill
Dec 10 2019 at 6:25am
Externalities are things economists believe god created in finite supply, like clean air, clean water, roads, water and sewer lines, power distribution, communication, thus being in infinite supply, the price can’t be higher than $0.
What is considered an externality has changed over the past few decades to include built capital paid for by taxes before most people were born, thus seemingly created by god.
For example, building detached single family homes was much cheaper 25 to 50 years ago because government was taxing and spending on roads, water and sewer, utilities on rural land, so buying a tenth of an acre meant it was easy to build an 800 sq-ft home. The taxes were paid by existing homeowners, thus “transferring wealth” to farmers, land developers, in the current economic view. Current homeowners see the infrastructure that allows them to have the house built before they were born (25 years ago) as god given, and external to their economy. This in turn means that 97% of the land in the US can be used for housing, its external to the economy because god did not provide roads, water and sewer, ….
By not building roads, water and sewer, etc, the amounted of land for a house must be large, or the capital cost for recycling a small quantity of water in the house from waste to drinking, and generating electricity to be consumed, etc, is very much higher than for 99% of existing homes. This violates the economic concept of substitutes, ie, a 2000 sq-ft house built in 2020 should cost the same as building the equivalent in 1970.
And you can’t increase the supply of the detached single family homes by building 20 unit housing buildings.
Jon Murphy
Dec 10 2019 at 8:15am
No. Externalities are when the action of one person imposes costs or benefits on another person not originally involved in the transaction. It has nothing to do with what God created or anything about finite supply.
No. There may be spillover effects from that capital, but the capital itself is not an externality.
Your following example doesn’t make any sense given these issues with your conception of externalities here.
I mean, you keep saying “current economic view” and “economists believe.” Write in the active voice: what economists “believe”? What “current economic view”? Who are saying these things?
Comments are closed.