Re: Utility maximizing

["Alan G. Isaac" <aisaac@xxxxxxxxxxxx>]

Just a reminder that the role of the utility maximization hypothesis
in PK economics has been discussed before on this list.
There is a thread that started at the beginning of the summer,
which can be linked to at
   http://csf.colorado.edu/mail/pkt/may97/0132.html

I don't have time to hash through all this again, but in brief modern
economists have a derived (rather than primitive) notion of
utility maximization which essentially embodies:
i. stability of choices among sets of alternatives
ii. certain consistencies in these choices
The most obvious point of attack is (i),
since choice is often contextual.
But for this to be useful
(i.e., for us to end up having anything to say
about the economic behavior of individuals)
this becomes a call for a modification (through
attention to context) rather than a replacement
of the stability assumption.

At the psychological level,
I doubt you could find many economists
who have thought about the problem asserting
that there actually is something ("utility") that
is actively maximized.
However, there is a residual notion that people make
the (subjective) best of their perceived options.
This residual notion should also be retained by PK economists.

As soon as we say "best" we are within striking distance of
a maximizing formulation (ordering -> formal maximization).
The attacks on neoclassicism need to be substantive
rather than abstract: the role of uncertainty,
framing, preference shifts, preference endogeneity
to institutional environment, computational limitations,
etc. need substantive exploration.

At the bottom, the question is whether,
when PKs think about microeconomics, should they
treat individuals as selecting from the
(subjective) best of their perceived options.
The answer is yes.
If you look at the thread, I offer examples of
substantive areas where this matters.
I also offer a challenge to nay-sayers to offer
important examples of economic behavior where
it is more useful (for understanding or for prediction)
to deny this. So far this challenge remains met only
with some vague talk of satisficing, but the logical
difficulties in discriminating between satisficing and
maximization are well known.

--Alan G. Isaac