Consider Schelling's work on residential segregation. He acheived a great
deal of insight about racial segregation by carefully studying the
implications of rule based behavior. The actors in his model are applying
rules to situations rather than searching for optimal, utility maximizing
behaviors. The same can be said about Axelrod's seminal work on repeated
prisoner's dilemma.

If we were to take Econ 101 seriously, Schelling's work is not real
economics because it doesn't mathematically deduce a behavioral outcome
from expected utilities. A mathematical proof showing that the Schelling
model leads to segregated areas of the checkerboard would probably be very
hard. Instead, Schelling did some experiments with an actual checkerboard,
which have been corroborated with extensive computer simulation. This is
unacceptable from the econ 101 perspective. 

My argument is *NOT* that we should dump the utility maximization approach
to economics, but that we should recognize that you can get good results
quickly by thinking about economic actors applying rules. An economic tool
box with two tools is probably better than a single tool, especially if
the new tool gives you purchase on interesting classes of problems.


On Sun, 29 Jun 2003, alypius skinner wrote:

> But would this really give us economic models more useful than the
> simplified ones currently used? Taking more factors into account may make
> the models so hard to maximize that there is no net gain in predictive
> accuracy.  Thoughts, anyone?
> ~Alypius
> > - Economists need to expand the repertoire of explanations. Economists
> > should learn how to model rule-based behaviors and interactions with the
> > same ease as they can calculate a Langrangian multiplier. Econ 101 should
> > start with a speech saying how people sometimes apply rules to economic
> > behavior and at other times they act like classical utility maximizers.
> > Students will then learn marginal analysis and models that embody rules
> > based behaviors.
> >
> > So there you have it. Some rambling thoughts...
> >
> > Fabio
> >
> >

Reply via email to