Les Schaffer wrote:
as a physicist, i have often been amazed at the diagrams in beginning
economics texts that show curves that intersect to define a price based
on supply and demand. what always strikes me is that the data points
("dots") lie right along the lines themselves, which i take to mean
that real data is never used to show a supply-demand curve. and i have
often wondered if such functional relations [S($), D($)] really exist.
the functional relationships are merely theoretical, but most
economists would say that the interaction between S and D leads to
changes in prices. Let me appeal to authority: Marx didn't reject
supply and demand. Rather, he saw S&D as presenting only a superficial
understanding of reality (cf. Roemer's failed effort to understand
Marx totally in terms of S&D).
there's what's called the "identification problem": If we just have
the quantity supplied qs = a + b*P (with a and b posited as constants
and P = price) and the quantity demanded qd = c + d*P (with c and d
posited as constants, d < 0), then with qd = qs, you can solve for
prices but not much more. You can't sketch out supply and demand
curves.
but if qs = a + b*P + c*X and qd = d + e*P + f*Y (e < 0) with X and Y
are variables lthat affect supply and demand (respectively) but are
unconnected in any way with each other, then statistical techniques
can be used. In simple terms, changing X shifts the supply curve,
tracing out the demand curve, while changing Y shifts the demand
curve, tracing out the supply curve.
on amazon i have been reading some reviews of books that michael p has
recommend to me. one review -- on the text "The Stock Market and Finance
From a Physicist's Viewpoint" -- says:
Lognormality is the last part of Osborne's book. The first chapters
are even more interesting. There, Osborne tears the `mathemology' of
Samuelson's Economics text to shreds by pointing out that the famous
supply-demand curve can't be constructed from any sort of data. The
main point is that price does not exist as a function of either
supply or demand. Example: suppose that 25 tomatoes are available
(supply). What's the price? Answer: anything or nothing
(nonuniqueness).
I don't get this. 25 tomatoes are likely available at some specific
price. It's the combination of prices and quantities that are the data
for the statisticians. It is more complicated, of course, if the
market doesn't clear (i.e., if the supply & demand aren't in
equilibrium), but the statisticians (a.k.a. econometricians) have ways
to make the data talk.
Even better, Osborne shows that one can obtain data
on both supply and demand as a function of price, so that discrete
(noncontinuous) supply and demand curves can be plotted for a given
commodity in a given market.
I don't get this.
What a pity that Osborne did not set
his mind to discussing `utility', because (as Mirowski points out in
"More heat than light) the differential form that defines utility is
generally nonintegrable, meaning that utility does not exist.
the concept has more problems than that. I see utility theory as
merely an unsuccessful effort to rationalize demand, i.e., why higher
prices of an item lead people to (almost always) cut back on purchases
of it (and vice-versa).
--
Jim Devine / "Segui il tuo corso, e lascia dir le genti." (Go your
own way and let people talk.) -- Karl, paraphrasing Dante.
