Equilibrium restored

一月 26, 2007

In his admirable Against the Flow: Reflections of an Individualist , Samuel Brittan observes that he will say nothing about Nash equilibrium - the idea that won John Nash the Nobel Prize for Economics.

Paul Ormerod's review of the book (January 19) makes this an excuse to air one of the numerous bees that buzz around in his bonnet. We are told that "extensive empirical research" shows that few people actually play according to Nash equilibrium, except where such play is "obvious".

Claims to this effect are often made by junior members of the behavioural school of economics founded by Daniel Kahneman and the late Amos Tversky, but they are false.

In the case of the Prisoners' Dilemma, there is extensive empirical research. More than half of all subjects do indeed fail to play the Nash equilibrium when experiencing the game for the first time. However, about 90 per cent do so after about ten games. This is typical of a large experimental literature that finds that experienced subjects play Nash equilibria in a wide variety of games with money payoffs.

The polemicists of the behavioural school contrive to overlook this elephant in the room by dishonestly quoting only results about inexperienced subjects or those in exceptional games such as the Ultimatum Game.

Ormerod's appeal to the authority of Vernon Smith - who shared a Nobel prize with Kahneman - is no more soundly based. Far from showing that economic theory is an empirical failure, Smith is famous for offering evidence to the contrary.

As for Brittan, he is sceptical about the wilder theoretical speculations of macroeconomists and financial gurus - and who can blame him? But he is far too sensible to decry the use that my colleagues and I made of the concept of a Nash equilibrium when we designed the telecom auction that made £23.4 billion for the British taxpayer.

Ken Binmore
University College London

请先注册再继续

为何要注册?

  • 注册是免费的,而且十分便捷
  • 注册成功后,您每月可免费阅读3篇文章
  • 订阅我们的邮件
注册
Please 登录 or 注册 to read this article.
ADVERTISEMENT