Lots has been written already about this year’s Nobel prize in economics (here is a great set of links). But I can’t believe that while I was asleep no-one has drawn the obvious pithy parallel with this year’s Nobel peace prize. So let me do that first.

Hurwicz, Myerson and Maskin together founded the economic field of mechanism design. That field is premised on one, seemingly simple, but very important idea (the Revelation Principle): that if you want to get the truth out of somebody, if there is a way to do it, there is a direct way to do it; that is, if the truth can be outed, it can be outed conveniently. I presume I don’t have to beat you over the head anymore to get the Gore tie in; so my work is done.

Moving on, let’s take an example: while speculation abounded over the past week, a small set of people — the Nobel economics prize committee — knew the identity of this year’s prize winners. There were plenty of others who wanted to know and if betting markets were thicker could have made some money on that knowledge. If you had approached someone on the committee, they likely would have told you any old thing but the truth as the truth would be the only thing that could come back and be attributed to them and involve some sort of sanction (e.g., being fired from the committee).

So how might you have got that information out of the committee? One can imagine lots of elaborate schemes; not the least of which is allowing committee members to bet on prediction markets. Others could cause some sort of competition between them or perhaps resorted to torture. But what mechanism design says is: don’t bother. If you are going to get the truth out, just offer a committee member a payment or punishment for each thing they might have told you; probably, if their information turns out to be true then they get $x while if it is false they are penalised $y. Then you choose x and y so that it is worth their while to tell the truth rather than not (the incentive constraint) and that they will actually decide to accept the deal (the participation constraint). Anything more elaborate is a waste of time.

Now sometimes we have to be more elaborate for practical reasons. In auctions, we don’t ask someone their willingness to pay but ask them to bid instead — that is an indirect way of getting the truth. In regulation, we don’t ask a monopolist what their costs are but offer them a price such that they end up acting on the basis of their true costs. And, in pricing, we don’t ask consumers whether they really like a good or just sort of like it, but offer prices and products so that their purchases reveal the truth. The point is that mechanism design can cut through the crap to see if the truth can be revealed.

At MBS, my Advanced Game Theory course was premised on the idea that mechanism design was not abstract and the stuff of academic journals but actually useful for practitioners and managers. The insight was that you could use The Revelation Principle to sanity check any crazy market or pricing scheme you might be dreaming up. Basically, it was a whole course for MBAs on the work of Hurwicz, Maskin and Myerson. So don’t let anyone tell you this is all ivory tower stuff.

Indeed, it is even more practical than that. I use it in the house to reveal the truth from children. A typical claim made by them is: “I’m hungry. Can we have some ice cream?” To test this, I put forward the mechanism: “No, but if you are really hungry, I have some carrots.” That reveals the hunger truth every time.

Finally, I have met each of this year’s winners. Hurwicz gave lecturers at a summer school I attended in 1993; at that time, he received a boost in notoriety due to his work being front and centre in Milgrom and Roberts’ Economics, Organizations and Management; which remains the most accessible treatment today (especially Chapters 3 and 4).

Myerson was at an economics conference a few years later in Tokyo and struck me as enthusiastic, welcoming and generally a nice guy. His work and personal help allowed Catherine de Fontenay and I to write this paper; the paper I think is and will likely remain my best work. His obscure PhD work saved us years of time. We are perhaps the only people to cite it.

Maskin was lecturing at the same summer school as Hurwicz and also was Stephen King’s PhD advisor. (Stephen regularly practices mechanism design on business people at ACCC meetings). He has visited Australia several times. On one recent visit he told us about getting Albert Einstein’s house in Princeton. That has been remarked upon several times over the past day. But he also told us that he got rid of all the furniture. “It didn’t really go,” was the choice quote from that conversation. I had assumed to auction for what would have been an incredible amount of money. But no, he had just got rid of it! And this from a person who wins a Nobel prize for work on auctions. I was shocked then and remain so to this day. That aside, Eric Maskin shares with his co-winners a commitment to the profession and general all-round niceness. It is good to see.

2 Responses to A Convenient Truth

  1. [...] (16 October ): Of all the commentaries I’ve read over the last two days, only this one from Joshua Gans focuses on the revelation principle, the main result of modern mechanism design. Joshua also has a [...]

  2. Manas Gajare says:

    About x & y, how to choose them?
    There always is the risk of getting fired. So how to choose participatory constraint? Can you throw some light on that?

Leave a Reply

%d bloggers like this: