[ show plain text ]
In reply to Steve Keen's OPE-L 3886:
Steve: "I'm willing to retract the first part of what I said, since
Andrew says he concurs with my reading of Sraffa on this point."
Cool.
Me: "These problems [reswitching, capital reversing] arise from
simultaneous valuation. Get rid of the dogma that input prices equal
output prices, and you get rid of the problems. In light of the capital
controversy, the neoclassicists realized this, got rid of the dogma, and
nullified the reswitching critique."
Steve: "I find this difficult to interpret on a number of grounds. It
appears to argue that neoclassical economics is now a disequilibrium
theory. Unless you want to see game theory in this perspective, I find
that proposition difficult to swallow. Instead, the Arrow/Debreu GE
approach is completely devoid of time processes, working in the realm of
the formal existence of equilibrium, rather than any alleged time
processes which might get it there. But that is a side issue."
Me: Prior to the capital controversy, it was indeed "completely devoid
of time processes." That is no longer the case, strictly speaking --
though neoclassicists now smuggle in stuff like perfect foresight and
complete forward markets in order to make the present state of affairs
redundant to the determination of future states, in effect to abolish
time. Malinvaud's version of GE theory, for instance, explicitly
recognizes the difference between input and output prices, and computes
rates of profit (or interest) on the basis of input prices that differ
from output prices (i.e., rates of return are computed on the actual
capital advanced rather than the replacement cost of capital goods). A
very nice paper by Frank Hahn called "The Neo-Ricardians" (if I remember,
it was published in the CJE in the late 1970s, maybe 1977) likewise
overcomes the "Sraffian" critique of some neoclassical propositions by
allowing input prices to differ from output prices, thereby showing that
the "neo-Ricardian" static equilibrium results can be viewed as mere
special cases of a more general (and powerful) neoclassical theory.
Steve: "More importantly for our debate, this appears to accuse Sraffa of
what Andrew says he doesn't accuse him of: being "simultaneist". How else
can I read "These problems arise from simultaneous valuation. Get rid of
the dogma that input prices equal output prices, and you get rid of the
problems". Are you not accusing Sraffa of a "dogma"?"
Me: No, I'm not accusing him. Sraffa undertook (successfully) an
*internal* critique of neoclassical theory as it existed at the time. We
agree about this, no? The simultaneous pricing equations were those of
the neoclassicists. In recognition of the validity of Sraffa's (and
others') critiques, the neoclassicists revised their theory so that, as
it now exists (at least at the level of "high theory"), it is immune to
the criticisms.
Steve: "Why is this "long period" methodology "ridiculous" if you say
you read Sraffa as I do, to be considering "an economy which had been in
equilibrium for the indefinite past", and then showing that the
(neoclassical) presumed equilibrium of this dynamic system cannot, in
fact, be an equilibrium?
"If you read it as the charge that everything happens simultaneously, or
that dynamic processes can be ignored, then I agree it's ridiculous. But
firstly you say that you're not reading it that way (in the first half of
your reply), then you appear to do just that in your second half."
Me: I apologize. I should have been more careful. The "long period"
methodology can be employed in a "critical" manner or a "positive"
manner. I think its critical employment is perfectly legitimate. Sraffa
employed the long period method (more precisely: the static equilibrium
method) critically, in an internal critique of neoclassical theory that
had heretofore employed it in a positive manner. He showed that what the
neoclassicists of his time had taken to be the necessary results of their
theory were not in fact necessary results. Cool.
It is only the employment of static equilibrium or "long period"
methodology as an aspect of *positive* theory construction that I
consider ridiculous -- more precisely, logically invalid. It attempts to
deduce properties of the general case from the relations that happen to
hold (tautologically) in the special case of static equilibrium. It is
just not a logically valid move to pass from the special case to the
general case in that way.
Compare two static equilibria. First, any airplane A is in some airport
X, on the ground. Sometime later, it is in some airport Y, also on the
ground. "Ergo," airplanes cannot fly.
The "Marxist" variant of this is: airplanes need to reproduce
themselves. They do so by refueling. They refuel on the ground. Hence
airplanes that reproduce themselves do so only on the ground. But we
know that airplanes do reproduce themselves. "Ergo," airplanes cannot
fly.
Notice that the method is logically invalid even when the equilibria are
stable, as they are in the airplane case. Non-convergence to equilibria
only compounds the basic problem.
I have other complaints against static equilibrium methodology as well,
but its logical invalidity is the chief one.
Andrew Kliman
This archive was generated by hypermail 2b29 : Sat Sep 30 2000 - 00:00:05 EDT