[OPE-L:4989] Re: Comments on 3 recent debates

From: Steve Keen (s.keen@uws.edu.au)
Date: Tue Feb 20 2001 - 07:38:04 EST

As well as endorsing the thrust of Allin's superbly argued post, I'll
embellish two points he made in his first 2 paragraphs (well, that's how
this post started; I ended up getting carried away--blame a good wine and a
good dinner).

These are the propositions that "it's a secondary matter whether the shift
to the new equilibrium GDP occurs in one period (instantaneous multiplier)
or takes several periods", and "Drewk's changing prices of production
converge (ceteris paribus) to the ones found via the method of simultaneous

It is also a secondary issue whether convergence actually occurs. The
method of simultaneous equations locates the equilibrium of (strictly
speaking a linear representation of) a dynamic system; that equilibrium
itself can be either stable or unstable.

This clarification is worth making, because some may be inclined to reject
Allin's analysis on the basis that prices don't converge to the equilibrium
in the long run: in the long run, everything changes.

That, strictly speaking, is irrelevant to the question of whether the set
of simultaneous equations themselves are properly specified. Whether
defined in terms of Price=k or dP/dt=k, if they express the equilibrium of
the related dynamic system, then they will be internally consistent--they
will have a unique solution (whether or not that solution is stable; those
who don't understand this qualification should either consult a basic text
on Ordinary Differential Equations, or read my ROPE 1998 paper).

If on the other hand they are internally inconsistent, then equivalently
the dynamic system they describe will be ill-formed: it will either lack
determinate variables, or it will have too many equations in too few
unknowns (be overdetermined).

Coming from a dynamic modelling perspective, this is why I can't buy the
entire TSS program: if Marx's transformation of values into prices doesn't
work in a system of simultaneous equations, then it won't work in a dynamic
system as well (this isn't to say that the static answer to a question is
always identical to a dynamic answer--again something you should consult
ROPE 1998 if you don't understand).

Any dynamic model which purports to show consistency--for instance one
which argues that Marx really meant to include the *prices* of the means of
production in his calculations of values (or record the inputs in price and
the outputs in value, if that's TSS's methodology), even though he
occasionally said that the *value* of the means of production is paid for
and transferred to the output--will in reality be either overdetermined, or
have too many degrees of freedom.

On the Copernican tag, whether applied facetiously or otherwise, I am more
inclined to seeing TSS as Ptolemy's Last Gasp.

While I probably should apologise for the arrogance of this post, I also
can't resist a "party political" plug. If not even dynamics or TSS can save
Marx from inconsistency, then we have to accept there *is* an
inconsistency, and properly locate its source.

Sraffians have erroneously claimed that the source of simply a lack of
understanding of linear algebra, and with that have thrown 95% of Marx, his
philosophy and sense of dynamics, out the window. They have thrown the baby
out with the bathwater.

On the other hand, the TSS attempt to "save" Marx does little more than
preserve him as a Wax Museum relic. While I can easily accept that TSS's
(and Fred's, and Rakesh's, and whoever else has recently claimed to have
solved or proved the irrelevance of the transformation problem) methodology
generates a set of numbers which let us work out prices with multiple
sectors with differing capital to labour ratios and differing rates of
technical change and depreciation and whatever else we want to allow change
each period... and which show a tendency for the rate of profit to fall...
I cannot believe that their methodology in any way lets you do anything
else. There are many other aspects of capitalism we should be able to
analyse, but in the end, all we can do is "solve" or "prove irrelevant" the
transformation problem. TSS et alia have thrown the baby out, but kept the

There is a "third way" (hence the party political bit above): Marx's
underlying philosophy contradicts the labour theory of value. Marx's
Copernican Revolution should have been to free revolutionary thinkers from
the shackles of a simple but metaphysical and technically impossible
presumption that all value can ultimately be resolved into labour alone
(see Bose's *Marx on Exploitation And Inequality*, Oxford University Press,
Delhi, 1980). It was his Ptolemain failure to be unable to reconcile
himself to having found a way to disprove a proposition he had always
passionately believed). Please why can't Marxists take the time to work out
which is baby, and which is bathwater.

I'm quite willing to elaborate on what I believe the baby is, but I have
absolutely no expectation that anyone else on this list will be interested
in discussing what that may be.

At 11:39 PM 2/19/01 -050s0, you wrote:
>A few comments on matters recently debated.
>1. Price of production and technological change.
>I'm with Fred in that I don't see simultaneous equations as evil: I
>think they can provide some useful information, even if in fact
>technical change is going on more or less continuously.  But I find
>Fred's specific argument against Drewk a bit scholastic.  The "smoking
>gun" (OK, not Fred's phrase) is a set of quotations showing that Marx
>held that prices of production change _only if_ there's a change in
>labour productivity somewhere in the system, or if the wage changes.
>But Drewk's tables show prices of production changing period by period
>even when there's no change in technology or the wage, therefore Drewk
>is inconsistent with Marx.  Well, suppose we change the example.
>Keynes held (this is a simplification, of course) that aggregate
>output changes _only if_ there's a change in autonomous expenditure.
>We bump up investment (I) at some time t and let the multiplier go.
>On some interpretations of the multiplier we'll find aggregate output
>changing for many periods after the change in I, as GDP converges
>towards the new equilibrium.  Can we immediately deduce that this is
>inconsistent with Keynes's concept?  Surely not: it's a secondary
>matter whether the shift to the new equilibrium GDP occurs in one
>period (instantaneous multiplier) or takes several periods.
>Similarly, Drewk's changing prices of production converge (ceteris
>paribus) to the ones found via the method of simultaneous equations.
>Drewk's prices of production are analogous to the multi-period
>multiplier.  Change in these prices is _triggered_ by a change in
>technology or the wage (and I think _only_ by such things, though I
>certainly wouldn't claim to speak for Drewk), but the change is spun
>out over time.  So I don't see much force in the smoking gun.
>2. Two sources of discrepancy between values and prices of production.
>Rakesh and Fred were working at this one not so long ago.  There's a
>real problem of inconsistency in Marx's statements here.  We have
>statements, highlighted by Alejandro and Fred, where Marx says (I
>paraphrase): "value = cost-price plus surplus value; price of
>production = cost-price plus profit". The Marx manuscript brought
>forward by Alejandro puts this in algebraic notation, giving it added
>definiteness.  But we also have statements where Marx says (again, I
>paraphrase; the quotations are well known): "There are two reasons for
>divergence between values and prices of production: (1) the price of
>the means of production employed in the production of the given
>product differs from the value of those means of production, and (2)
>the profit realized in the sale of the product differs from the
>surplus value embodied in that product."
>I'll refer to the first sort of quotations as the "cost-price plus"
>statements, and to the second sort of quotations as the "double
>divergence" statements.  These two sorts of statements sometimes occur
>in close proximity in the texts of Marx that we have available to us.
>The obvious problem is that if "cost-price plus" is a proper statement
>of Marx's view then "double divergence" is nonsense: there is one and
>only one source of divergence between price of production and value,
>namely the discrepancy between surplus value and (equalized) profit.
>It's certainly tempting to suppose that one or other of the sets of
>statements in Marx must be "not really what he meant to say"
>(otherwise his views were flat-out incoherent).  My preferred
>interpretation is that "cost-price plus" was not really what he meant
>-- or at least, that these statements only hold good on the assumption
>that the means of production (inputs) were purchased at prices equal
>to their values.  I do _not_ claim that it's clear from context that
>whenever Marx issued a "cost-price plus" statement he was in fact
>assuming that input prices were equal to input values (rather, I think
>there is a degree of inconsistency in the text -- although of course
>it's a text that Marx never prepared for publication).
>I prefer this interpretation, despite its problems, because the
>alternative is worse, namely that Marx was just gibbering when he made
>his "double divergence" statements.  I haven't seen a rationalization
>of these statements, from the proponents of "cost-price plus", that
>makes any sense to me.
>3. Copernican revolutions
>It seems to me egregious hubris for the champions of any new theory
>(or new interpretation of an old theory) to describe their own work as
>a "Copernican Revolution".  Let future intellectual historians judge
>whether such phrases are applicable.  As a practical matter, nobody is
>able to hasten the day when his own theory is so acclaimed by others,
>by so advertising it himself.
>Allin Cottrell.
Dr. Steve Keen
Senior Lecturer
Economics & Finance
Campbelltown, Building 11 Room 30,
School of Economics and Finance
s.keen@uws.edu.au 61 2 4620-3016 Fax 61 2 4626-6683
Home 02 9558-8018 Mobile 0409 716 088
Home Page: http://bus.macarthur.uws.edu.au/steve-keen/

This archive was generated by hypermail 2b30 : Thu Mar 01 2001 - 14:01:39 EST