I really dont understand alll that is going on in this discussion since I am not an economist. But here are a couple of points:
1) Even if the assumption is true that there are unlimited human wants it does not seem to follow that one would have the slighest interest in satisfying the most that you could and the most efficiently. Plenty of wants we would not want to satisfy at all let alone efficiently For example some people have a want to copulate wiith corpses, to commit genocide etc. etc. I assume we are not interested in satisfying these at all or most of us enlightened ones!
2) In the area of welfare economics there is even passing notice of the absurdity of counting all preferences(wants?) in some welfare function. There is the problem of laundering preferences (Goodin, Elster, et al.) The problem is obiviously not solvable within economics except by the most ludicrous of
assumptions. For example, we willdiscount anti-social preferences and get on with it.. I guess like the preference in Saudi Arabia that some might have to allow women to drive or that some in the US might have to allow same sex marriage!
3) Some wants are related to addiction. The problem here is not primarily to satisfy them but to ensure that a person no longer has them at least in a manner that he or she cannot control.
Julio Huato <[EMAIL PROTECTED]> wrote:
3) Some wants are related to addiction. The problem here is not primarily to satisfy them but to ensure that a person no longer has them at least in a manner that he or she cannot control.
Julio Huato <[EMAIL PROTECTED]> wrote:
dd wrote:
> non-satiation is one of the von Neumann/ Morgenstern axioms.
> You need it in order to derive most of the fixed-point theorems
> that underpin utility theory, which matters more to people who
> think (unlike me but like a lot of college professors) that
> utility theory is really important to neoclassical economics.
It seems to me that you are referring to *local* nonsatiation, an
assumption that allows you to relax convexity in proving that a
general equilibrium is Pareto efficient. But the Sandwichman appears
to refer to *global* nonsatiation or even *global* monotonicity (more
of consumption of a good/bad leads always to more/less utility): a
preference set unbounded in the northeastern direction.
In the analysis, these stronger assumptions preclude "thick"
indifference curves. That's their role. So, the general equilibrium
theorems don't require that human wants be "unlimited," but only that
in the neighborhood of equilibrium, people can obtain more wellbeing
by consuming more of at least one good were they able to afford it.
But, and this is the troublesome thing here, one thing is the role of
an assumption intended to support a particular theoretical scaffolding
(general equilibrium welfare theories) and an entirely different thing
is to discuss whether and to what extent an idea corresponds to real
life. The latter is the kind of argument the Sandwhichman seems to be
engaged in with the libertarian fellow. (And this is also the topic
of Ted Winslow's posting, as I understand it.)
Apparently, "utility theory" means different things to different
people. Historically, the notion of utility had a psychological,
deux-ex-machina, individualistic, extra-social flavor. But that is
not inherent to the idea of utility as it is used in modern *theory*.
Very generally, utility describes a relation between actions (e.g.,
consumption, work) and human welfare. Obviously, in real life, human
welfare is to some extent a consequence of human actions or --
throughout history -- people have behaved as if it were. This
relation is what's involved in what we call "valuation." But, as
narrowminded as economists can be, I doubt there are many who think
the relation is straightforward in real life.
The question is, how do you get a hold of a mindbogglingly complex
relation in order to derive clear consequences from it? People can
only use the tools available to them. So theoretical economists make
simplifying assumptions: that the relation can be specified as a
*function* with convenient attributes. The usual objection here is
that the simplification *is* the problem: too much is given up in it.
This is a valid objection only if one has an alternative way to
improve upon the regular use of an utility function.
If we don't need a tractable way to map from actions to wellbeing,
then we don't need utility. But can we afford to give it up? No if
we want to *derive* the rules of aggregate economic behavior from more
fundamental "primitives" and not just *assume* such rules. I find it
curious that people who question the need of utility analysis are
frequently willing to accept without question much stronger
assumptions about aggregate behavior, e.g., a consumption or a saving
function. How can that be justified if simplification is the problem?
Julio
