Nobel Prize in Economics awarded to Roth and Shapley for matching algorithm work

http://www.nobelprize.org/nobel_prizes/economics/laureates/2012/popular-economicsciences2012.pdf

by traldan • 14 years ago

46 21 14 years ago

nirvana• 14 years ago

The so-called "nobel" prize in economics does not come from the estate of Alfred Nobel. It comes from the central bank. Thus it is not surprising to see it given to people that central banks would favor (often people whose theories try to justify inflationary policies.)

This one seems to be given for work that could be said to claim that central planning really can work after all.

equark• 14 years ago

They are creating decentralized mechanisms for free individuals to find individually and socially beneficial outcomes. The free market organized by the U.S. legal system is one such mechanism, but certainly not the only one.

zissou• 14 years ago

Both Shapley and Roth are microeconomists, so the fact that you are bringing up banking and inflation is a clear signal that you don't understand anything about economics. If you want to harp on Nobel Prize winning economists regarding their work on macroeconomic theory, you're 1 year too late.

Shapley and Roth deal with topics that fall, at the highest level, under the field of game theory, but more specifically market design. Market design relies heavily on mechanism design, which can also be called reverse game theory. In game theory, you're interested in determining the best strategy given a set of conditions (the players, their strategies, the timing of their moves, their payoffs and the information they have [or don't have]). On the other hand, mechanism design is concerned with what the conditions have to be in order to achieve a specific objective or outcome. Hence, mechanism design problems map well onto many problems of regulation.

For example, let's say there is a Senator, call her Mrs. X, who is interested in passing a bill that achieves outcome Y. The mechanism design solution is then the set of rules that lead to outcome Y occurring. Setting the right rules refers to Mrs. X drafting a bill that includes mechanisms such that every actor in the game HAS AN INCENTIVE to play the strategy/strategies that lead to outcome Y.

How does this fit in with market design? Well, besides generally using mechanism design to align the incentives of both the buyers and sellers, there is a deeper point to be made about market design here. All so often in economic theory it is assumed that a market already exists, and the analysis is then on understanding the inner functions of that market. However, what that type of analysis leaves out is how that market even came into existence in the first place. In what ways were incentives aligned such that both buyers and sellers found a match for one and other to begin with? In other words, why is this market here and how did it get here? These are the nature of the questions that market design tries to answer. By understanding how and why the actors in the market are doing what they are doing, you begin to understand how markets form. And if you understand how markets form, you can apply that know-how to the formation of new markets, or to improve old ones.

traldan2• 14 years ago

Great explanation, thanks for posting that!

001sky• 14 years ago

This is an interesting exposition, but (I would just add for generality) the nature of "Market" is quite undefined, ex ante.[1] There are only specific markets in specific places for specific things, with specific participants. The nature of a "free market," is also precisely that they are not "gamed" by external authority ex-ante.[2]

The study of "non-economic" (ie, meta) activity that has (determinant) economic impact is most properly referred to as Political Economy. Although many trained "economists" avoid this. The illusory notion that veracity lies in 'scientific' objectivity (and so with it, government funding), has something to do with it.

_______

[1] ie, It is only defined directly by identification of a contractual exchange between N parties (agents), or 'archeologcally' through recipt, contract, or other record. But such an exchange directed by fiat, for example, would not qualify.

[2] For the definitional seperation of Hierarchy from Market, see for example Coase or Williamson.

dreamdu5t• 14 years ago

Yup, and as such they should receive an award for "Game Theory" not economics.

john_horton2• 14 years ago

Their work has nothing to do with central planning in the sense of committee deciding on production quotas---it's actually about how to allocate indivisible resources (school positions, kidneys, residency slots etc.) in situations where using money isn't feasible and/or the market has broken down due to problems like congestion or lack of market thickness.

bo1024• 14 years ago

I think results like Shapley's and Roth's present an interesting commentary on the usual "philosophy of economics".

1. These results are all about finding good solutions to allocation problems without using money or in cases where money isn't or can't be applied.

2. These results do tend to show that a "centrally planned" solution can be better for everyone than a "free market" solution. But as you say, by central planning I really just mean that some entity takes in everyone's preferences and outputs a global allocation, with the nice property that everyone is happy with what they got and nobody can find a way to deviate from their assignment and be better off. One might argue that a "free market" would find the same solution eventually, but it might be slower.

These two results are sort of the opposite of the Invisible Hand stuff as well as the common assumption that, for instance, everyone values money the same (which underlies the justification for much of economics). So I think they're really cool in that respect.

john_horton2• 14 years ago

well put.

FYI - Here's a link to an interactive illustration of the deferred acceptance algorithm that I posted a few years back on HN: http://mathsite.math.berkeley.edu/smp/smp.html

cschmidt2• 14 years ago

Roth wrote quite a nice book about stable matchings:

http://www.amazon.com/Two-Sided-Matching-Game-Theoretic-Econ...

It is readable if you're interested in the problem. This book on the same subject is also good, although not by any Nobel prize winners:

http://www.amazon.com/Stable-Marriage-Problem-Algorithms-Fou...

0ren• 14 years ago

I find it interesting that they awarded the prize to Shapley after Gale's death[0]. I always thought that there was some rule that prevented them from awarding the Physics Nobel Prize to Aharonov after Bohm died[1].

[0] https://en.wikipedia.org/wiki/Stable_marriage_problem

[1] http://en.wikipedia.org/wiki/Aharonov–Bohm_effect

pranjalv123• 14 years ago

Several Nobels have been given out after a collaborator's death - probably the most famous being Watson and Crick's prize four years after Rosalind Frankin's death.

yairchu• 14 years ago

You are correct, Nobel prizes are not awarded to the dead, unless they die just after announcing the prize and before the ceremony (http://www.guardian.co.uk/science/2011/oct/03/nobel-prize-aw...). Gandhi was denied the prize for this reason.

In this case, it is simply the post title being misleading, as it's not the Nobel prize these folks won, but the Nobel memorial prize in economics, which is a relatively new prize which is not part of the 5 Nobel prizes which had been given by Nobel's estate for more than a century.

Edit: oh I see they didn't reward a dead in this case, misread your comment..

gmoore• 14 years ago

There is no Nobel Prize in Economics

http://www.alternet.org/economy/there-no-nobel-prize-economi...

guylhem• 14 years ago

Wasn't that due to Nobel girlfriend leaving him for an economist? Read something about that but can't find the surce at the moment (bad connectivity)

And I don't see that in the first 2 pages. Or was that about the math nobel prize ?

cschmidt2• 14 years ago

I believe that was the reason for the math nobel

traldan2• 14 years ago

Technically, yes, but HN has an 80 char limit in the title. ;)

thmcmahon• 14 years ago

Josh Gans provides a good summary on Roth and Shapley over at Core Economics http://economics.com.au/?p=9423

pm90• 14 years ago

I recognized their names from the first example in the classic book "Algorithm Design" by Kleinberg and Tardos. But there it was called Gale-Shapley Algorithm.

_delirium• 14 years ago

The algorithm is due originally to Gale and Shapley, in a 1962 publication in the American Mathematical Monthly, but Gale isn't eligible due to the rule against posthumous awards. Roth is included because he was the first to apply it practically in economic-matching settings, quite a few years later.

Nobel Prize in Economics awarded to Roth and Shapley for matching algorithm work

Related Stories