Bryan Caplan  

Public Goods and Pedagogy

PRINT
Freddie, Fannie, and the Fragi... The Income Heritability Puzzle...

I'm increasingly convinced that the way economists teach undergrads about public goods and externalities is needlessly confusing. Here's my radical solution: Purge discussions of public goods and bads, and replace them with discussions of positive and negative externalities.

Is this a libertarian stealth strategy? Hardly. Just because you've stopped talking about public goods, doesn't mean you can't talk about big externalities.

But doesn't the concept of public goods have extra content? Yes - and that's why it's pedagogically confusing. According to the textbook definition, a public good is both "non-excludable" and "non-rival." But all that "non-rival" really means is that the marginal cost of production is zero. And there's already a chapter in the textbook where we discuss the problem of price exceeding marginal cost - namely, the chapter on monopoly. Blending externalities and monopoly problems together when students are struggling to separately understand them makes little sense.

Once economists clean up our undergraduate teaching, we can immediately clean up our internal communications. In my experience, the way the economists apply the concept of "public goods" to the real world is extremely sloppy. How many times have you heard an economist say that "education is a public good," when at most he means that "education has positive externalities"? Given this sloppiness, I suspect that my terminological switch wouldn't just be better for undergraduate instruction; it would also clarify the thinking of professional economists.

If you've ever lectured on the subject of public goods - or listened to such a lecture - I'd like to hear your thoughts. Would my approach have clarified the subject without significant loss of content?

P.S. Here's my current lecture on externalities and public goods. It doesn't actually drop all discussion of public goods, but it does reduce it to a special case of externalities.


Comments and Sharing





TRACKBACKS (1 to date)
TrackBack URL: http://econlog.econlib.org/mt/mt-tb.cgi/874
The author at Dean's World in a related article titled Public Goods writes:
    Bryan Caplan thinks we might be better off dropping “public goods” from undergrad econ classes, in favor of a broader discussion of externalities. And he’s got a point — most people do not understand what economists mean by R... [Tracked on July 25, 2008 1:29 PM]
COMMENTS (19 to date)
jsalvati writes:

I totally agree with this, but to some extent I think this already happens; I don't think we discussed "public goods" very much in my intermediate micro class, but we did discuss externalities.

Though, come to think of it, I think we did discuss "public goods" in my intro micro class, and it was confusing.

JH writes:

Completely agree.

1. It's in the common thought that "public goods" are "government-provided goods" and that's an impression that is impossible to fight. A lot of economic terms -- "rent-seeking" "marginal analysis" even, I'd argue, "demand" -- are technical terms, and yet ones that use common words. Tends only to muddy the waters.

2. Non-rival and non-excludable are imprecise. If one wants to be pedantic, do any goods fit this bill? What are the principles that guide where we draw the line?

3. Does calling something a "public good" clarify anything? Is it trying to define what a state should and shouldn't do? What is efficient for state action?

Oh, here's one textbook's introduction to its public goods section. "Even at the height of laissez-faire capitalism in the mid-nineteenth century, people recognized that the state had to take responsibility for providing some things." Eww.

Manjira Dasgupta writes:

I am not sure how non-rivalry in consumption, one customary feature of public goods would be accommodated.

MS writes:

I believe Arrow did just that in his 1970 “Political and Economic Evaluation of Social Effects and Externalities” in The Analysis of Public Output ed. J Margolis (NBER).

William H Rogers writes:

I like the approach but I see one difference between public goods and positive externalities: positive externalities are spillovers from some transaction that will presumably take place without the spillover, while public goods will not exist at all without collective action. For example, some amount of education will take place without collective action but without collective action no amount of national defense will.

The policy implication is that governments can subsidize buyers or seller to deal with externalities but with public goods governments must act as the sole buyer or seller.

Carl the EconGuy writes:

I don't agree at all. The reason you teach public goods is to show that the solution is to sum up Marshallian demand prices vertically to cover the cost of the fixed output, and then you lead into a nice discussion of Wicksell's wonderful piece on the theory of just taxation. This is classical non-general equilibrium economics, and it should still be taught.

Externality is another general-equilibrium concept that is completely useless. What you want to teach instead is Coasian transaction cost theory, which leads into a discussion of relative institutional efficiencies in solving transaction cost problems.

Transaction costs are the common denominator btw all the so-called "market failure" concepts, so that's what you want to convey. Again, that's a non-general equilibrim concept, of course.

The central issue you want to teach in welfare theory is the Arrowian impossibility theorem -- there is no unique collective ranking, and therefore general equilibrium welfare economics should be completely relegated to the scrap heap of history.

Only wish that were the case in our silly intermediate and advanced microtexts.

Patrick Walsh writes:

I absolutely agree, at least for Principles. In Principles I do a lecture on + and - externalities, with examples of both. Then I do a lecture on the Tragedy of the Commons as a special case of neg externality. For all the reasons Bryan points out, I don't use the term "public good".

In my public finance course, on the other hand, I do both the + and - externality discussion, and the "textbook" definition of a public good. It may take a few more iterations before these flow together as smoothly as I'd like.

liberty writes:

I can't say that I agree.

Neither positive externalities (my neighbor is also protected if I buy defense) nor negative externalities (I litter on the common area) can really get across the idea that people will have a zero wtp for these goods because they expect someone else to pay for them.

The term "public goods" makes a lot of sense to me because it conveys this idea well: I don't want to pay or exert effort, for something that the whole public can use and the whole public can clean up after.

Non-excludability and non-rivalry, although a bit confusing at first, get this across more technically. You can also show it on the s&d curves. You can link it in to positive and negative externalities too - but those are possible for things which people will provide privately, and don't compose the whole problem with public goods.

LB writes:

I think David D. Friedman suggested the same either in Hidden Order or in his Price Theory Text.

Ryan writes:

I agree. I am an economist in a policy school that caters mainly to Masters of Public Administration (MPA) students. The typical MPA student believes that a public good is (by definition) a good provided by the government. I have spent an unjustifiable amount of time trying to get across the traditional economic concept of a public good, usually with little success. I have switched to a teaching strategy based on externalities--students seem to be able to grasp the concept much more easily--perhaps a little too easily. Students seem to eagerly embrace the idea that if someone does something you don't like then government intervention is the appropriate course of action.

Sinclair Davidson writes:

Also think about positive externatility as a failure to create a barrier to entry and negative externality as a failure to create a barrier to exit.

Jim Glass writes:

"Public Good" is the worst name ever thought up by economists for anything.

*Everybody* thinks "public good" = "something good for the public" = "something the gov't thus must provide, like oh, public schools!"

Take it out and shoot it.

Take out the economist who thought up the name and shoot him.

Just take all the textbooks and replace the term with something else. "A bartfizzle is non-rival and non-excludable..."

The world will be a better place.

Nick Rowe writes:

A paper by Frances Woolley, agreeing with you: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=907381

Non-gated version: http://economics.ca/2006/papers/0901.pdf

Robin Hanson writes:

Makes sense to me.

I agree that we'd be better off without the sloppy use of the term. But the precise definition, because it distinguishes between non-rivalrous and non-excludable goods, is useful to learn and think about. I think it's true that any "social cost" or benefit can ultimately be analyzed in terms of externalities. I still like to think about the categories.

For example, using the precise definition, it was easier for me to see how the provision of public goods does not have natural monopoly characteristics -- a fact with important political and religious consequences.

Frances Woolley writes:

I've made these arguments in my paper "Why public goods are a pedagogical bad" that Nick Rowe refers to above - when I present it I get almost exactly the same reactions as people have given here.

What's slightly depressing is that economic journals aren't willing to have the kind of discussion about fundamental economic concepts that's happening in this blog, though it's quite possible to publish some trivial variation on well-known public goods experiments.

Will Wilkinson writes:

Hmmm.... This is a philosopher's answer, not an economists...

There is a class goods that has a set of features such that the good will be produced only if a certain kind of collective action problem is solved. Suppose those features are non-excludability and non-rivalrousness. What do we call this class of goods?

Externality doesn't seem in any sense relevant. If the point of producing Snickers bars is consumption by individuals, you wouldn't say that the individual's utility from consumption is a positive "external" effect of production or exchange in the good. That the consumer and producer internalize the benefit from exchange and consumption is the reason the good is produced. Likewise, the public's enjoyment of a public good isn't "external" to production of the good. The point of producing it is that its benefits are internalized by the public. The difference between Snickers bars and classic public goods has to do with the difference in the means necessary to coordinate production.

I may be missing something important, but it's mysterious to me what positive externalities has to do with it.

Jeff Hummel writes:

I agree wholeheartedly with your ciriticism of the way public goods are taught to undergraduates. My solution, however, is less radical than the one you propose in your blog and closer to what you do in your lecture. As you know, you are not the first to point out that non-rivalness is really part of the monopoly question. David Freidman and Steven Shmanske are among the many economists who have emphasized that point. So what I have been doing since I first taught the subject in 1989 is essentially what David does in his price theory text: define a public good purely as nonexcludable and tell students we are going to ignore non-rivalness, even though they've read about it in their principles text, because non-rivalness really belongs with monopoly. That allows me to present a spectrum, with pure private goods at one pole, pure public goods at the other pole, and excludable goods with some positive externalities in between. (I also mention that although nonexcludability and non-rivalness on occasion come together, there is no reason they have to. And I started doing this long before Mankiw put it into his principles text.)

My approach is not as clean and straight forward as your blog proposal. But it has two partially offsetting advantages: (1) Since the students, if they take more economics courses, are going to encounter the "official" concept of public goods (or may have encountered it already), I've prepared them and partially immunized them from confusion. (2) Public goods as only nonexculdable coincides with the popular understanding of that term among non-economists, and since the term has become imbedded in political discourse, I would rather accept the popular definition (as with the term "inflation") than jettison it entirely. I'm not sure we would ever be able to get economists to stop using the term "public good," much less the general public.

My order of presentation is slightly different than your lecture as well. I start with public goods as I've defined them. Then take up negative externalities. And finally bring up positive externalities as a refinement of our understanding of public goods. I'm not entirely satisfied with this approach, but it does seem to work with students.

Jeff Hummel writes:

I agree wholeheartedly with your criticism of the way public goods are taught to undergraduates. My solution, however, is less radical than the one you propose in your blog and closer to what you do in your lecture. As you know, you are not the first to point out that non-rivalness is really part of the monopoly question. David Freidman and Steven Shmanske are among the many economists who have emphasized that point. So what I have been doing since I first taught the subject in 1989 is essentially what David does in his price theory text: define a public good purely as nonexcludable and tell students we are going to ignore non-rivalness, even though they've read about it in their principles text, because non-rivalness really belongs with monopoly. That allows me to present a spectrum, with pure private goods at one pole, pure public goods at the other pole, and excludable goods with some positive externalities in between. (I also mention that although nonexcludability and non-rivalness on occasion come together, there is no reason they have to. And I started doing this long before Mankiw put it into his principles text.)

My approach is not as clean and straight forward as your blog proposal. But it has two partially offsetting advantages: (1) Since the students, if they take more economics courses, are going to encounter the "official" concept of public goods (or may have encountered it already), I've prepared them and partially immunized them from confusion. (2) Public goods as only nonexcludable coincides with the popular understanding of that term among non-economists, and since the term has become imbedded in political discourse, I would rather accept the popular definition (as with the term "inflation") than jettison it entirely. I'm not sure we would ever be able to get economists to stop using the term "public good," much less the general public.

My order of presentation is slightly different than your lecture as well. I start with public goods as I've defined them. Then take up negative externalities. And finally bring up positive externalities as a refinement of our understanding of public goods. I'm not entirely satisfied with this approach, but it does seem to work with students.

Comments for this entry have been closed
Return to top