Monday, June 23, 2008

The Game Theory Boom

I'm linking to a nice paper [Nicola Giocoli: Three Alternative (?) Stories on the Late 20th-Century Rise of Game Theory] asking about the reasons for the sudden boom in the use of Game Theory in economics in the eighties. His bottom line is Harsanyi's development of tools for thinking about games of incomplete information, so-called Bayesian games. But he also gives some importance to the following circumstance: The Chicago School completely dominated Antitrust policy in the eighties. I remember in Grad school in the late 70's the suffocating growth of Panglossian Industrial Organization thinking. This was the stuff that would look at any apparent example of restraint of trade - resale price maintenance, eg - and find that it was rational and efficient. Predatory pricing was irrational and could never happen; the only real entry barriers were put in place by the dead hand of the state. Laissez faire; laissez allez, in short. Economists dissatisfied with this picture - in particular, thinkers associated with Stanford's Graduate Schoool of Business such as Paul Milgrom and John Roberts - used Harsanyi's Bayesian techniques to provide rigorous counter-examples to the Chicago conventional wisdom: to show how predation might make sense, how non-state entry barriers - such as building excess capacity, might well succeed, and so on.


Myrtle Blackwood said...

"The principle argument was that in order to sustain a fallacious economic system, whose foundation is but shifting sands whipped by the ebb & flow of monetary inflation, that system needs an utterly absurd sequence of myths to be widely accepted as ideology, promoted by a trusted harlot. The result is like a crowd of mindless zombies uttering mantras like people devoid of brains, but whose bodies move enough to cast their next order to purchase stocks or bonds. FOREX traders do not qualify as zombies, and therein lies a problem...

Death of Bretton Woods II by Jim Willie CB, editor of the "Hat Trick Letter". Apr 27, 2007

I keep going back to this article. China had given notice to 'the world' in early 2007, Jim Willie said, that it was going to drop the purchase of US Treasuries and start purchasing oil and other commodities and the companies that mine them. said...


While IO has been the most important area of economics that game theory has been used in, I think this piece overstates the importance of this particular set of arguments. I would suggest that it was the spread of the use of non-cooperative game theory into many other areas of economics, which really took off in the 1980s, that made it into the standard piece of micro it is now, with most grad programs in econ spending the second half of the year in micro, if not more, on it.

BTW, it is not just the Harsanyi part, but also the Selten part, the infinite repeated games part. So, it is the Bayesian (Harsanyi) version of the sub-game perfect equilibrium (Selten extension of one shot deterministic Nash equilibrium) that is the standard workhorse for this theoretical approach.

BTW, it was Selten himself who understood that real agents do not backwardly induct as required by subgame perfect equilibrium and suggested to Werner Guth that he should test it experimentally. Guth came up with the famous ultimatum game experiment to do so.