Tyler Cowen explores economic ideas that should be popularized. Let me take this opportunity to help popularize what I think is one of the pillars of economic theory and the fruit of the information economics/game theory era.
When we notice that markets or other institutions are inefficient, we need to ask compared to what? What is the best we could possibly hope for even if we could design markets from scratch? Myerson and Satterthwaite give the definitive answer: even the best of all possible market designs must be inefficient: it must leave some potential gains from trade unrealized.
If markets were perfectly efficient, whenever individual A values a good more than individual B it should be sold from B to A at a price that they find mutually agreeable. There are many possible prices, but how do they decide on one? The Myerson-Satterthwaite theorem says that, no matter how clever you are in designing the rules of negotiation, inevitably it will sometimes fail to converge on such a price.
The problem is one of information. If B is going to be induced to sell to A, the price must be high enough to make B willing to part with the good. And the more B values the good, the higher the price it must be. That principle, which is required for market efficiency, creates an incentive problem which makes efficiency impossible. Because now B has an incentive to hold out for a higher price by acting as if he is unwilling to part with the good. And sometimes that price is more than A is willing to pay.
From Myerson-Satterthwaite we know what the right benchmark is for markets: we should expect no more from them than what is consistent with these informational constraints. It is a fundamental change in the way we think about markets and it is now part of the basic language of economics. Indeed, in my undergraduate intermediate microeconomics course I give simple proof of a dominant-strategy version of Myerson-Satterthwaite, you can find it here.
(Myerson won the Nobel prize jointly with Maskin and Hurwicz. There should be a second Nobel for Myerson and Satterthwaite.)
9 comments
Comments feed for this article
October 26, 2010 at 1:20 pm
How do Students Game the Course Allocation System at HBS? « Cheap Talk
[…] constraints. This may not lead to ex post efficient allocations (see Jeff’s timely post about Myerson-Satterthwaite!) but it can improve ex ante welfare. There is a large literature on […]
October 26, 2010 at 2:36 pm
The difficulty of optimal trading explained simply | Economics and Mechanisms
[…] economists this topic is known as the Myerson-Satterthwaite Theorem. Over at Cheap Talk, Jeff Ely has a nice explanation in plain language. This entry was posted in Mechanism Design, microeconomics. Bookmark the permalink. ← […]
October 26, 2010 at 9:11 pm
Les liens du matin (69) « Rationalité Limitée
[…] « Popularize: The Myerson – Sattherthwaite Theorem » – Cheap […]
November 2, 2010 at 9:43 am
vic
I think this is a difficult one to popularize as it arises from mechanism-design rather than from empirical stuff like voting and preference revelation. Your manner of putting it- ‘If markets were perfectly efficient, whenever individual A values a good more than individual B it should be sold from B to A at a price that they find mutually agreeable. There are many possible prices, but how do they decide on one? The Myerson-Satterthwaite theorem says that, no matter how clever you are in designing the rules of negotiation, inevitably it will sometimes fail to converge on such a price.’- begs the obvious question as to why the two agents shouldn’t get into the mechanism design themselves- i.e. playing a mixed stratgy game to determine who gets to extract the potential surplus subject to sealed bids re. its ultimate carve up.
I don’t see how anything at all could be said about this in advance- though I’d be grateful for a signpost to anything in the literature on the revelation principle to indicate otherwise.
November 2, 2010 at 10:06 am
vic
Sorry, I should have clarified- if mechanism design is internalized then there is an exponential complexity problem which I think defeats impossibility theories in advance.
Again, grateful for any signposts on current thinking on this.
November 2, 2010 at 1:27 pm
vic
On reflection, I can’t see this is an important result that could or should be disseminated in a popular form. It encapulates no folk wisdom or tried and tested emirical hueristic. Mechanism design aint natural in the way that Game theory is natural.
I am reminded of Morishima’s attempt to reconcile Socialism and Capitalism by means of ‘rational income distribution’ founded upon shadow prices- i.e. solving the dual.
I notice that Dutch notions like Brouwer’s- or Van den Berg’s phenomenological metabletics- really start to look sexy- if we Anglo-Austrio-Americans geld ourselves in this fashion.
Ultimately rationality cant be both Epimethean and Promethean- there is nothing to be made popular here because there is no notion not already contained in things like the second-best theorem and the meliorative effects of the law of large numbers.
Or am I missing something?
November 2, 2010 at 4:47 pm
jeff
Dear Vic: I have the distinct impression that you are not missing a thing. 🙂
November 7, 2010 at 4:25 pm
Introducing Roger Myerson « Cheap Talk
[…] Recently, Jeff mentioned that Roger and Mark Satterthwaite should get a second Nobel for the Myerson-Satterthwaite Theorem which identifies environments where it is impossible to achieve efficient allocations […]
May 16, 2011 at 1:03 am
Efficient Market Mechanisms « Cheap Talk
[…] some background, start with the bilateral trade problem of Myerson-Satterthwaite. We know that among all DSIC, budget-balanced mechanisms the most efficient is a fixed-price […]