Someone is wrong on the internet again*

By Kathy G.

The "someone" in question being Megan McArdle. This time she screws up the Coase theorem. McArdle illustrates the theorem with an example of a dispute between two neighbors, one of whom plays loud music, to which the other objects. She writes:

I pointed out that as long as the transaction costs for side deals are relatively low, which they should be in the kind of leafy, large suburb we were discussing, it didn't really matter whether he had the right to play music, or I have the right to enjoy silence; the Coase Theorem dictates that we will end up with preference maximization.

What's wrong with this statement? Where do I begin? First, let's start with "preference maximization." Though I've heard a number of different formulations of the Coase Theorem, none that I am familiar with have ever said anything about "preference maximization." The most widely accepted version of the Coase Theorem that I am familiar with says that, in a dispute concerning an externality, if transaction costs are low (or nonexistent) and property rights are well-defined, the parties can bargain their way to an efficient outcome.

Note that the result is an efficient outcome -- this is an entirely different concept than "preference maximization" (which I confess is a term I haven't heard in any of the econ classes I've taken, though I assume it means something like "utility maximization"). And by "efficient," most economists mean "Pareto efficient" -- that is, the best deal that can be made that leads to no one being worse off and at least one person being better off.

The second thing wrong with that statement is that McArdle doesn't include a very important qualifier "well-defined property rights." That is always crucial. In her example, if it's unclear whether or not the offended party has the legal right to peace and quiet, on what basis, exactly, are they supposed to bargain? Because then it's not clear if the wants-loud-music person owes anything to the wants-quiet person, or vice versa. And in the absence of that kind of baseline rule, the resulting uncertainty will be a significant barrier to negotiating. Even if there is some kind of anti-noise statute on the books, unless it is very clear -- such as, no music after 9 pm, or something similar -- then it will be hard to determine if the music is so loud as to be violating the law. Also, the property right has to be enforceable. If the cops don't respond to complaints, then the offender has little incentive to bargain.

The other problem with the Coase theorem, of course is the assumption of minimal (Coase actually says "no") transaction costs. In the real world, transaction costs always exist. It is costly to overcome certain barriers to negotiation in the real world, such as information asymmetries costs, coordination problems, enforcement costs, and the like. Even in McArdle's simple example, it would cost something to find out what the relevant laws about noise said, and also to enforce the contract.

Beyond the question of efficiency, the Coase theorem says nothing at all about equity. Assuming well-defined property rights and no transaction costs, a solution can be arrived at that is efficient, in the sense that it's an improvement for at least one party. But it need not be equitable or just.

Libertarians tend to be fond of the Coase theorem, but that is because they often misunderstand it. They're likely to be enamored of the idea of autonomous actors in a free market bargaining their way to an efficient solution, with government staying out of the matter entirely. They especially prefer the Coase solution to the Pigouvian mechanism, in which the externality is internalized via a tax. But in the real world, a Coase solution to an externality is not necessarily going to be any more efficient than a Pigouvian one.

And contrary to libertarian misreadings, the Coase theorem is positive, not normative. Its conclusion (an efficient outcome) will always logically follow from its premises, but only if the assumptions -- well-defined property rights and zero transaction costs -- hold. Which of course they pretty much never do in real life. Empirical studies that try to determine whether Coase holds in the real world have had mixed results at best, and in any case are complicated by the fact that vanishingly few real world situations ever meet Coase's preconditions.

Even Coase himself said that the assumption of zero transaction costs was "very unrealistic." In 1981, he wrote:

[W]hile consideration[s] of what would happen in a world of zero transaction costs can give us valuable insights, these insights are, in my view, without value except as steps on the way to the analysis of the real world of positive transaction costs. We do not do well to devote ourselves to a detailed study of the world of zero transaction costs, like augers divining the future by the minute inspection of the entrails of a goose. 

But just because you can't apply Coase in a simple-minded and literalistic way to the real world, it doesn't mean the theorem is without value.On the contrary, it can be "good to think with" (to borrow the phrase of a theorist I can't recall at the moment). Coasean-style transaction costs can be used as an analytical tool to understand how, and under what conditions, markets work. For example, the Nobel-prize winning economist Douglass North has built on Coase's work in interesting and fruitful ways to look at how institutions affect economic performance. But to argue that the Coase theorem "dictates" anything in the real world, as per McArdle, is to grossly misunderstand it.

*If you don't get the reference, click here.

Comments

[reason]

Of course Megan McArdle just meant that the noisy neighbour could be paid to turn down the music or use radio headphones. Then the disturbed neighbour has paid a price that he considers worth it for peace. But there is a small equity problem here for the observant. And the reality is that often the solution will be for the disturbed to threaten the noisy neighbour, precisely because equity is important to real humans, as against glibertarians.

[Gabriel]

There's nothing wrong with "preference maximization". It just means that out of the set of feasible actions, the agent will choose the actions for which the outcome is at least as good, subjectively, as that of any other action.

"Preference maximization" is "utility maximization" whenever the preference relation can be represented by a twice-differentiable, continuous function. But in many circumstances, you don't need that extra structure, i.e. the first welfare theorem can be proved without these assumptions.

The link between "Pareto efficiency" and "preference maximization" should be obvious. (No one can do better without someone doing worse?)

Also, it is bad sport to criticize assumptions. If you can get at least as strong a result as this, with less assumptions, then by all means, go ahead. Otherwise, as long as theory is insightful and contributes productively to the discussion, there's no reasons not to bring it up.

Lastly, information asymmetries are not transaction costs, in any meaningful sense. Michael G. makes the point:

http://yetanothersheep.blogspot.com/2008/04/what-are-transaction-costs.html

[Shane Taylor]

Why assumptions are fair game:

Alan Musgrave (1981). "'Unreal Assumptions' in Economic Theory: The F-Twist Untwisted", Kyklos 34 (3) , 377–387

Some assumptions are domain restrictions, and they must be assessed to determine where a theory does and does not apply.

[Gabriel]

Shane,

We can disagree about method, but as a practical matter of everyday science and moving things forward, methodological navel-gazing is very destructive and counterproductive. There's a reason for why there's no progress in philosophy.

It seems to me, again, as a matter of practicality, to interpret Megan McA.'s comments not as some claim to literal truth, but rather as a potentially useful perspective. -- That a perspective is limited and limiting is obvious. Picking at assumptions offers nothing extra.

[Karmakin]

Reason:The real problem is that we have a term for that:

Blackmail.

Quite frankly, I think a society which relied on blackmail would break down fairly quickly. Now, personally, if you're talking about anti-noise bylaws, I think they're misdirected. Music in a bar after 10 is bad, but music coming out of a church (which is MUCH louder) at 9 in the morning is fine? Erk.

Sorry. I'm just a bit bitter. (Used to work the backshift, lived next door to a church)

I don't really see either how information asymmetry is actually a transaction cost. Some you can, but I don't see those as IAs in the first place...like doing research online to find reliability ratings or whatever.

However, say if you're buying an iPod. You don't have the information on if it was already bought and taken back, if it was dropped in transit, or whatever.

The IA simply can't be overcome. Thus, consumer protections will always be needed. We can argue about the degree on the merits, but already, the ideological libertarian standpoint goes down in flames.

[Karl Steel]

On the contrary, it can be "good to think with" (to borrow the phrase of a theorist I can't recall at the moment).

Levi-Strauss on food.

[Bloix]

Coase says whatever the legal rule may be, the parties can bargain to an efficient result. i.e., if the legal rule is that the neighbor has a right to quiet, then the music-player can pay the neighbor to put up with the music. If the utility of music to the music-player is higher than the utility of quiet to the neighbor, the music-player will be willing to pay the neighbor's price. If not, then not.

On the other hand, if the legal rule is that the music-player has a right to play music, then the neighbor can pay the music player to turn it down. If the utility of quiet to the neighbor is higher, etc.

Of course, in the real world, relationships between neighbors don't work this way. It would be highly offensive to make an offer of payment. Libertarians don't understand that, or disapprove of it - any relationship that's not a cash relationship is irrational.

But let's take a business context where the Coase rule makes sense, so we can see how McArdle misunderstands it. Suppose I am remodeling my clothing store. The workers' power tools are making a lot of noise, and it's affecting business at your coffee shop next door. If you have a legal right to quiet, you can threaten to shut me down. I can offer to pay you to compensate for lost business if you will put up with my noise for a week. If you demand more than I am willing to pay, I will have to do the work at off-hours or have the workers use hand tools or put up noise-barriers - all at a dollar cost to me. On the other hand, if you have no legal right to quiet, you can offer to pay me to have the noisiest work done at off-hours, etc. Either way, we will reach a result that's "efficient" in the sense that by bargaining we determine whether the dollar cost to me of not making noise is lower or higher than the dollar cost to you of the harm to your business.

But Coase does not say that the legal rule "doesn't really matter." To the contrary, the legal rule determines who has to pay. Coase says that whatever the legal rule is, bargaining can yield an efficient result. He never says that the rule "doesn't matter."

[HCG]

Gabriel said: ""Preference maximization" is "utility maximization"

Sorry, but in consumer choice theory preferences are taken as fixed, and exogenous. You cannot maximize something that is fixed. Utility is what varies as a function of the level of consumption, preferences fixed, and therefore can be optimized. The confusion comes from the fact that most people use the word preference as a synonym for choice or welfare.

[donna]

Was Megan annoyed she didn't get invited to the party?

[Megan McArdle]

It doesn't matter in the very narrow context of the argument, which involved a counterparty who was arguing that there was no such thing as a normative model of distributive justice. Obviously, distributional questions matter--which is one of the reasons I am against vesting the right to enjoy low density housing on people who are already lucky enough to own homes.

[DrDick]

*sigh* Let me say this once again. There are no "free markets" in the legal sector. The only truly "free" markets are black markets. All other markets exist within strict regulatory environments including property and contract law, tort law, legally defined weights and measures, and all the rest. The Coase theorem assumes such a regulatory environment in the form of "well-defined property rights."

[Mark Shapiro]

I remember being struck by the beauty of the Coase theorem. Only two men in the world, one harms the other, and yet they can resolve the matter without violence! This bucolic scene is even better than the bible: Genesis without the bloodshed. Better yet, even with only two people, they enjoy price discovery.

So sad that the case global warming doesn't satisfy all the Coase assumptions.

[Shane Taylor]

"We can disagree about method, but as a practical matter of everyday science and moving things forward, methodological navel-gazing is very destructive and counterproductive. There's a reason for why there's no progress in philosophy."

Are the national accounts destructive and counterproductive to economics? There, assumptions and methodology are constantly debated. The current work (though delayed due to budget constraints) on R&D and health satellite accounts are all about reassessing assumptions and methodology.

[Susan of Texas]

Ms. McArdle can't admit she is wrong, since her sinecure at The Atlantic depends on her air of infallibility. When someone corrects her, as people frequently do, she simply says they don't understand.