What is Mechanism Design?

Explaining the research that won the 2007 Nobel Prize in Economics.

Alex Tabarrok | October 16, 2007

Two children are squabbling over how to divide a pie. We need a method to divide the pie fairly. Parents will already know one answer—one child cuts and the second child chooses. The second child will choose the larger half which gives the first child the incentive to cut as evenly as possible. The first-cut, second-choose solution is a simple example of an incentive-compatible mechanism. Leonid Hurwicz, Eric Maskin, and Roger Myerson received the Nobel prize in economics for their study of incentive-compatible mechanisms or, more informally, "mechanism design."

Mechanism design is a very general way of thinking about institutions. An institution or mechanism takes as input "messages" or "signals" from agents and it responds with an outcome. The idea of mechanism design is to create institutions that produce a desirable outcome while respecting the fact that agents have private information and are self-interested. It turns out that designing mechanisms that work well while respecting information and self-interest constraints is very difficult.

Ironically, the market, an undesigned mechanism, is the best example of a powerful incentive-compatible mechanism. Thus, in their explanation for the prize the Nobel committee wrote:

"These results support Friedrich Hayek's (1945) argument that markets efficiently aggregate relevant private information."

Mechanism design, however, is not simply a mathematical apparatus justifying the insights of Hayek. Leonid Hurwicz, the godfather of the field who is now in his nineties, was influenced by Hayek and by his opponent Oscar Lange. One can think of Hurwicz as trying to prove when the goals of Lange could work even taking into account the objections of Hayek.

It's long been known that markets are challenged by externalities, public goods, asymmetric information and so forth. Standard public finance theory says "thus government"—a clear example of the nirvana fallacy.

Mechanism design theorists at least take their challenge seriously, and thus try to design institutions that work under the same constraints as the market—i.e. institutions that respect information and self-interest constraints. The results have been mixed. Typically the mechanisms that work in theory are very complicated—far more complicated than the market or other mechanisms that we see used in practice. I see little hope that mechanism design will rescue the dreams of Lange, et al.

More realistically, I see mechanism design as a tool to make markets more powerful. In some situations, for example, mechanism design shows that public goods can be voluntarily provided. In other situations, mechanism design can make government more effective, but it will do so by making government more "market-like." Contracting-out of government services like garbage pickup, prisons, and roads, for example, can be carried out even farther if contracts are more carefully designed. The theory of mechanism design provides the template for thinking about the best possible types of contracts.

The most practical use of mechanism design to date illustrates my point. Mechanism design is the foundation for the sophisticated auctions that have been used to sell off broadcast spectrum. The moral here, however, is often misunderstood. The sophisticated auctions convinced governments that there was money in selling off spectrum, but the real gains came when spectrum, which was being wasted in government hands, was turned over to the private sector.

Overall, mechanism design increases our appreciation of markets, if only by showing how difficult it is to produce good outcomes while respecting the constraints that markets must satisfy. In a sense, mechanism design is to markets what genetic algorithms are to life. Theorists may one day design a better market mechanism or a better genetic code but for now the gains will come from using our deeper understanding to gently improve something that's already pretty marvelous.

Alex Tabarrok is an associate professor of economics at George Mason University, research director for the Independent Institute, and a research fellow at the Mercatus Center. He blogs at Marginal Revolution.

Comments

libertreee|10.16.07 @ 9:22AM|#

Among the first principles of economics are the notions that benefits and costs are subjective, and that human beings tend to place different subjective values on goods and services. The latter notion is why trade takes place. If you value your used car at $5000 and I (subjectively)value it at say, $6000, then gains from trade are possible. If we strike a bargain at $5500, it creates a mutually beneficial gain. This is what makes the economic world go around.
This year's Nobel Prize in economics was awared to three mathematicians (Leonid Hurwicz, Roger Myerson, and Eric Maskin)who clearly do not understand this freshman-level idea. They are "game theorists," and like many game theorists they seem mostly interested in mathematical games, not economic theory or economic reality.

The award was for "mechanism design theory," which purports to "correct" such alleged "failures" of markets as when "people hold back private information" or "when people refuse to divulge how much they are willing to pay for a good." As the Nobel committee said in its announcement, markets supposedly "fail" when "buyers and sellers privately know their evaluations" of goods and use this personal information to "gain personal advantage."

What a shocker. Everyone tries to get the best deal possible.

Naturally, this calls for more government regulation which, according to the Nobel committe's smarmy announcement, may not suffer from any information problems like the ones that supposedly plague private markets. Perfect politicians, imperfect markets.

What else should we expect from Swedish sociaists whose comrades believe Al Gore is the new Prince of Peace.

So sayeth Tom Di Lorenzo on LewRockwell.com. It is the Swedish Central bank that issues the Nobel Prize in Economics.

I have no dog in this hunt, but I thought since the Beltway Chicago School boys and the Austrians sometimes go at it I would submit this quote.

|10.16.07 @ 9:51AM|#

I was most curious about this part:

"An interesting special case of the above model occurs when the public good is excludable. If the public good is excludable then accept is a dominant strategy for each agent regardless of K. Consider a group of N agents and an excludable public good, like a bridge, which costs C to produce. Let the entrepreneur offer the agents an .F; S;K/ contract with the additional proviso
that only accepting agents can consume the public good if it is produced."

It seems like their model says, basically, if you can make an excludable public good, then risk of free-riders can be eliminated. But isn't that the definition of public good, i.e., something not excludable? If it were excludable, private markets are the most efficient valuator, and there would be no need for the public volunteerism.

Reminds me of Paul Riser's character in Aliens. "Come on, Ridley. If you do that, there's no exclusive rights for anyone. . ."

|10.16.07 @ 10:43AM|#

The recent article in Reason on October 5 by Ronald Bailey, "The Secrets of Intangible Wealth" highlights the first serious empirical study of the mechanism design issue by the World Bank in terms of institutions.

The study shows that intangible wealth is closely associated with institutions such as the rule of law, property rights and effective government.

It demonstrates, for example, that when a Mexican comes to the U.S., he or she has access to $418,000 of this type of wealth compared to $34,000 in Mexico.

This explains why, for example, attempts to provide aid or invest in parts of Africa are far more likely to fail than more developed countries with institutions necessary to support the same.

In this context, for all the flaws, institutions, including governments, enable (or at least are statistically associated with) much more trade and economic growth than less.

However, that trade and growth may be associated more with imposing stability and predictable expecations than it does to promote economic freedom and competition, particularly that competition that tends to "creatively destruct" existing monopoly strongholds.

Ed Crane of CATO recently stated at an Ayn Rand celebration speech that supply side economists had replaced freedom with growth in the sense of crowding out bad government with more private growth. If true, this could include institutions "mechanically designed" to promote economic growth which infringe on individual and economic freedom.

The studies of the Nobel Prize winners seem to generally support the World Band study, that institutions do support efficient outcomes, at least more than they suppress it, at least compared to other countries.

For a contemporary application of the complicated principles involved, consider the spector of intimidating Blackwater operatives roaming heavily armed around New Orleans post-Katrina on behalf of private property owners. Is that a private market solution to a public good failure? Are the incentives and the ability to avoid accountability any better or worse than the corrupt police department in the Ninth Ward?

David Rollins|10.16.07 @ 10:55AM|#

Also see Peter Boettke commentary in today's WSJ:
http://online.wsj.com/article/SB119249811353060179.html

|10.16.07 @ 11:03AM|#

It appears this award went to economists trying to revive the arguments for socialism that were refuted during this:

the Socialist Calculation Debate

That is, they're advancing theories that imply we can replace free markets with socialist government planning if we can just have sophisticated enough pricing systems, etc.

|10.16.07 @ 11:20AM|#

[snark]ah yes, nuanced view of markets = socialism, of course[/snark]

thoreau|10.16.07 @ 11:37AM|#

And the Libertarian Nobel Economics Prize goes to.....

(opens envelope)

Every blog commenter who ever shouted "DEMAND KURVE!" in the "Matt Damon!" voice from Team America.

We don't need no education!

Fluffy|10.16.07 @ 11:38AM|#

Libertree -

Yes, I have always had a huge problem with definitions of "market failure" that include things like unequal knowledge, location advantage, etc.

This is because economics as a science views the free market as an instrument for maximizing efficiency, and not as an arrangement necessary for justice.

For example: the fact that system participants have unequal amounts of information, and that the market rewards the participants with more information and penalizes those with less, is to me a positive feature of the market, because I think that it is just to reward preparation and punish mediocrity. But this messes up the equilibrium charts the economists love, and strikes them as therefore less than perfectly efficient. That makes them see a bug where I see a feature.

|10.16.07 @ 11:41AM|#

prolefeed,

In non-snark mode, it seems to a degree you are correct, the theory attempts to formalize the factors discussed in the Socialist Calculation Debate.

It, however, does not provide a clear advantage to either side of the debate, giving pro-market positions some empirical support, and providing strict demonstrations of how and why public-goods market failures can come about. (See the Nobel prize papers for an example- 16th century England beat France to some public goods innovations because France required consensual voluntary agreements prior to implementation).

The prize went to a tool, not a partisan political position.

|10.16.07 @ 11:43AM|#

[snark]if we define "success" as "that which the market produces" then of course there can be no market failures[/snark]

Timothy|10.16.07 @ 11:49AM|#

You pricks move, and I'll execute every mother fucking last one of you.

|10.16.07 @ 11:50AM|#

fluffy,

I think that it is just to reward preparation and punish mediocrity.

Market failure, if'n I understand this theory correctly, would be defined by a situation where your preparation is punished by my mediocrity. In the public good realm, consensual decisions by the community (that is, decisions without coercion for some members of the community), lead towards zero chance of being implemented as the number of agents increases, even in a situation where the aggregate economic health of each individual would benefit from implementation of the public good project.

Timothy|10.16.07 @ 11:53AM|#

Market failures exist where there are externalities due to lack of sustainable pricing mechanisms. If you think there aren't externalities, I'm moving in next door and playing nothing but Toybox at 5000db 24 hours a day.

COASE BARGAIN THIS MOTHER FUCKER!

|10.16.07 @ 11:59AM|#

[preemptive snark] The prize went to a tool...

NO, that was the PEACE prize[/preemptive snark]

thoreau|10.16.07 @ 1:09PM|#

The curriculum at a Libertarian Economics Department:

Econ 101: Micro. Make sure you get the Demand Kurve down. And Coase. That makes it easy to dismiss all sorts of pesky "facts."

Econ 102: Macro. All you really need to know is "Gold Standard!"

Stats 101: Correlation is not causation. There you go. Done!

For advanced students:

Econ 301: Atlas Shrugged. Yes, it's an economics textbook.

Econ 302: Von Mises, highly selective readings from Adam Smith, and bits of Friedman.

And that's it. That's all you need to know.

|10.16.07 @ 1:23PM|#

thoreau,

You are probably being a bit harsh on GMU, which is as close to a libertarian econ department as you are likely to find.

thoreau|10.16.07 @ 1:27PM|#

Jason-

My comment was directed at some of the comments in this thread, not at GMU. I admire GMU's econ department.

Theodore Dawes|10.16.07 @ 2:00PM|#

Typically the mechanisms that work in theory are very complicated

Once we finish completely computerizing our economic and legal processes, complexity will not be an issue. There might be some problems on the design side, but even that could be eventually automated. Run time of a complex economic mechanism would be almost instantaneous.

ed|10.16.07 @ 2:49PM|#

I'd like an award for pointing out that the Dismal Science leads to dismal threads.

Ed Clarke|10.29.07 @ 2:23PM|#

See letter to Nobel Prize Committee on
"Originality in Mechanism Design"