21 Aug 2012

Putting the Cart Before the Horse

The use of Game Theory in economics have become quite fashionable. Game theory seems to offer an alternative to the Utilitarian ideas that came to nought with the discovery that aggregate demand doesn't follow the Law of Demand.

In particular the theories of John Nash became popular. Nash devised a series of game scenarios in which theoretically rational participants interacted in idea ways. It's thought that this branch of mathematics shows what ideal rational behaviour would look like and has thus been interesting to economists.

But it's all a farce. Human beings are not rational choice making machines. We make decisions using our emotional responses to situations and events. I've written about this elsewhere (Facts and Feelings) based on Antonio Demasio's book Descartes' Error. While no economist seems to be ready to take this on, the advertising profession has increasingly shown sensitivity to this fact in the design of advertising campaigns. Remember when cars were sold on the basis of engine size, fuel efficiency, or special features like limited slip differentials? None of that is relevant any more. Car ads speak directly to our emotions now, bypassing the rational completely: BMW sells "joy".

I wanted to learn a little more about Nash so I checked out a paper supposed to be a 'Classic' hosted on the website of the Proceedings of the National Academy of Sciences of the United States of America (PNAS).

An economist called Robert Weber is cited in this classic: "Nash's theory of noncooperative games should now be recognized as one of the outstanding intellectual advances of the twentieth century" but elsewhere we find it said:
The Nash equilibrium concept does not necessarily predict how people will behave in the real world. Rather, it provides a measure for how purely rational people might behave.

"The Nash equilibrium tells us what we might expect to see in a world where no one does anything wrong," Weber says. 
In fact the Nash equilibrium does not predict how people will behave in the real world except by accident. It creates a particular kind of fantasy about a "rational person". The definition of "rational" is left aside. We know the old Utilitarian definition of rationality as maximising one's own pleasure without regard for other people. This is seen as "rational" because of the conceits of a few Victorian English gentlemen. In fact if we met someone like this we would think they were crazy, a sociopath if not a psychopath.

In Nash's model humans were inherently suspicious of each other, selfish and constantly struggled with other people as competitors. And Nash did actually succumb to paranoid schizophrenia not long after he published these results. "Rational" in other words means inhuman or insane in these models!

But these people are so convinced that their worldview is Truth in an absolute sense that they put the cart before the horse and say things like:
"In some settings, people routinely deviate from the kind of behavior the Nash equilibrium predicts," Weber says. "These deviations have led to the discovery of pervasive psychological phenomena that pull people away from rational behavior." 
The assumption here is that the weird definition of reality included in Nash's models is normal, and that ordinary human beings deviate from that norm. Here I think we have to conclude that Weber is also a lunatic. Human beings have never conformed to the norms that Utilitarianism or Game Theory have sought to impose on them. To turn around and suggest that human beings are deviant for not following the models is madness. I can think of no other discipline which constructs arbitrary models in the abstract and then criticises phenomena which fail to conform.

Economists try to make people fit their theories, whereas scientists make their theories fit people. And this is fundamentally wrong. In grasping simple theories that give consistent answers, economists typically ignore the disconnect from the real world - the make a series of unwarranted assumptions that hide the implausibility of the theory.


And indeed Nash himself admits that his assumptions about people do not stand up to scrutiny in an interview with Adam Curtis for his film "The Trap".

The Economist anticipated some of my criticisms in 1998 - they say that a. people don't understand how economist use words like rational; and b. "rationality has proved a useful [simplification]." I think I do understand how they are defining rational and my point is that they define it in a way that is aberrant.

I'd have liked them to define "useful" since their models seem to be bloody useless at predicting the economy at present - we keep getting "shock" deviations from predictions these days. These "useful" models failed to predict one of the greatest perturbations of the economy since records began. The failure is on a scale that would render the theory irrelevant in any other discipline. It would be like suddenly observing anti-gravity, or neutrinos that really did travel faster than light. Such a result should signal the end of the old order, and an exciting race to provide the new more complete and accurate paradigm - in economics there are many contenders.



1 comment:

  1. When studying economics I was quite a fan of game theory, in particular how outcomes could arise that were counter-intuitive to ones economics would expect based on the usual assumptions. In particular Nash's insight in the prisoners' dilemma that agents acting under the economist's definition of 'rational' (ie selfish and utility maximising) could still bring about a result that made both agents worse off went completely against the economic principal that rationality increased welfare proposed by Adam Smith onwards. In my inexperience, this suggested a way that the economic theories I was learning could tie in to the real world that seemed to contradict them.

    Now I am beginning to learn that the reason some of these theories don't tie into the real world is because they are simply not correct, but I still have a begrudging respect for these economists who fight these incorrect foundations on the neoclassical's own terms, such as Stiglitz and Krugman. Whether Nash was doing this isn't clear, but in fictional interpretations of his life he seems to be proud of his proof that Adam Smith was wrong.

    ReplyDelete

Keep is seemly & on-topic. Thanks.