Wednesday, August 18, 2004
Irrational number-crunching
This is further to Jaideep’s recent post about the role of emotion in economics. The fact that people are irrational is not new, even for economists: Adam Smith wrote about it as long ago as the 18th century. The problem is one of technology: it’s just extremely difficult to model a marketplace in which the actors are heterogeneous and irrational [1]. To make the analysis more tractable, classical economic theory assumes that people are uniform and rational – more like Mr. Spock than Capt. Kirk.
This is not an entirely arbitrary assumption. We can invoke the ‘efficient markets hypothesis’ to argue that arbitrageurs will always iron out any local inefficiencies (‘irrationalities’) so that the market as a whole behaves rationally, and is thus amenable to classical analysis. Unfortunately, the efficient markets hypothesis fails to hold in many real-life scenarios: for instance, when irrational traders outnumber rational ones by a significant margin. As Keynes famously put it, “the market can remain irrational for longer than you can remain solvent”.
But I digress. Coming back to the topic of behavioural economics, the great breakthrough achieved by Kahneman, Tversky and others was to realize that although people are irrational, they are often irrational in consistent ways. And these persistent biases can be isolated empirically, using ingenious experiments that measure how people react to risk and reward. The results of these experiments (‘asymmetric utility functions’, for example) can then be plugged into analytical models of irrational behaviour. Voila!
For further reading, the Economist has a comprehensive report on the history and ideas current in this field. (The report was published in late 1999, at a time when theories of irrational behaviour were in vogue – I wonder why). Also of interest is the introductory chapter of Camerer, Loewenstein & Rabin’s review of the field, ‘Advances in Behavioral Economics’, which can be found here.
[1] As an aside, the advent of cheap computation has spawned a new field called ‘intelligent agent theory’, in which individual (possibly heterogeneous and irrational) actors interact in simulated marketplaces. Call it ‘experimental economics'.
This is not an entirely arbitrary assumption. We can invoke the ‘efficient markets hypothesis’ to argue that arbitrageurs will always iron out any local inefficiencies (‘irrationalities’) so that the market as a whole behaves rationally, and is thus amenable to classical analysis. Unfortunately, the efficient markets hypothesis fails to hold in many real-life scenarios: for instance, when irrational traders outnumber rational ones by a significant margin. As Keynes famously put it, “the market can remain irrational for longer than you can remain solvent”.
But I digress. Coming back to the topic of behavioural economics, the great breakthrough achieved by Kahneman, Tversky and others was to realize that although people are irrational, they are often irrational in consistent ways. And these persistent biases can be isolated empirically, using ingenious experiments that measure how people react to risk and reward. The results of these experiments (‘asymmetric utility functions’, for example) can then be plugged into analytical models of irrational behaviour. Voila!
For further reading, the Economist has a comprehensive report on the history and ideas current in this field. (The report was published in late 1999, at a time when theories of irrational behaviour were in vogue – I wonder why). Also of interest is the introductory chapter of Camerer, Loewenstein & Rabin’s review of the field, ‘Advances in Behavioral Economics’, which can be found here.
[1] As an aside, the advent of cheap computation has spawned a new field called ‘intelligent agent theory’, in which individual (possibly heterogeneous and irrational) actors interact in simulated marketplaces. Call it ‘experimental economics'.
