Tuesday, October 16, 2007

Mechanism Design: 2007 Nobel Prize

My campaign to reduce use of the mechanism metaphor in Economics has taken a hit with the awarding of the Nobel Prize to "mechanism design." The scientific background paper provided by the Swedish Academy drew a lot of "bunk" comments in the margin of my copy. On p. 1 it says, "Some markets are free of government intervention ...." Government and all markets form a nexus. The reference to Hayek and the aggregation of all relevant information is still questionable. p. 2, "The theory thus helps to justify financing of public goods through taxation." I thought the case was already made by Pigou. The assertion ignores the interests of what I call unwilling riders. p. 5, "private information precludes full efficiency." No Kidding!
p. 7 re: Groves and Clarke condition that "there are no income effects" makes the theory trivial. p. 8, the condition that agents be "expected utility maximizers is unreal. Economics must be the only field where prizes are given to people with conflicting views since Kahneman received the prize for showing expected utility to be false. The further conditions of "the set of possible allocations was unidimensional and the agents had quasi-linear references" is also unreal.
p. 9, "the probability of funding a public-goods project tends to zero as the number of agents increases." Conclusion: "classical Pareto-efficiency is incompatible with voluntary participation." Amazing what rigor produces! Every school child knows that the decision rule for most direct popular votes on public production/taxes are passed with majority votes. Some think this acceptable and others support politicians who promise to lower all taxes. The two sides are unlikely to alter their positions as a result of this rigorous analysis. If these results only "provide a rigorous foundation for Samuelson's (1954) negative conjecture about public goods" who needs it?
p. 19 with reference to social choice rules. Noting that indeterminate multiple equilibriums are to be expected, "The final outcome can then depend on negotiations and bargaining among the voters." Reference is then made to Schelling who observes that outcomes depend on "social and psychological factors." Heaven forbid that the world is explained by psychology and not mathematical deduction!
I can only conclude that the prize was given for rigor for rigor's sake and not for anything useful.

2 comments:

Ed Clarke said...

Your reservations on Clarke-Groves mechanisms notwithstanding, you may be interested in a letter to the Nobel Committee on:
"Originality in Mechanism Design"

Ed Clarke said...

The comment appears on Ed Clarke's weblog:
http://ed-clarke.blogspot.com