[OPE-L:3872] Re: Marxists AND Sraffians Misinterpreting Sraffa

From: Steve Keen (s.keen@uws.edu.au)
Date: Fri Sep 22 2000 - 16:47:23 EDT

[ show plain text ]

Dear Fred,

I can understand and sympathise with your reaction to Ajit's replies.
Unfortunately, I can also understand and sympathise with Ajit's reactions
to your arguments. I think a lot of the heat in this exchange emanates from
the manner in which you have approached Sraffa's contribution. Before you
take that as a criticism of yourself, this is precisely the same criticism
which I have levelled, in published papers, of "Sraffian economists"
themselves (I don't regard Ajit as a "Sraffian economist", by the way--as
I'll explain below).

The error in interpreting Sraffa which you share in common with the
"Sraffians" is to regard Sraffa's "Production of commodities by means of
commodities" (POCBMOC) as the foundation for a rival theory of economics
(to Marx in this instance; Sraffians regard it as a rival to all other
theories of economics).

That is not what Sraffa intended it to be. The Kaleckian economist Peter
Kriesler first emphasised to me the importance of the subtitle to POCBMOC:
"Prelude to a critique of economic theory". Sraffa, Peter emphasised, was
an incredibly pedantic individual who once argued for two weeks with
Maurice Dobb over the position of a comma in a sentence in the preface to
their collected works of David Ricardo. He had spent, effectively, a good
half of the years between 1926 and 1960 writing POCBMOC, and every word
counts--including those in the subtitle.

The import of that subtitle is that Sraffa never intended his magnum opus
to be the foundation for any theory--it was instead the basis for a
critique of economic theory. It is a "crucible", in which the mettle of
some (but *not* all) economic arguments can be tested.

This first theory in this crucible was neoclassical economics, and
specifically the neoclassical theory of income distribution. A key
component of that is the proposition that "factor returns" reflect marginal
productivity. To measure the "marginal productivity" of "capital", one has
to have an accurate measure of capital--yet the amount of capital is
calculated by adding up its money value, when according to neoclassical
economics, the money value itself should be based on the rate of profit.

Sraffa devised a means to calculate the amount of capital on the basis of
its embodied labour content--which apparently is independent of the rate of
profit. However, using the concept of the standard commodity, he then
showed that (a) there is a linear relationship between the wage rate and
the rate of profit when measured using the standard commodity; (b) the
valuation of the embodied labor content in a machine depends on the time at
which that labor was applied, and that factor must be multiplied by
(1+r)^n, where r is the uniform rate of profit and n the year in which the
labor was applied.

When you multiply these two factors together, you get a highly nonlinear
function which, nonetheless, tells you the "amount" of capital you have.
But the rub is that the "amount" of capital depends on the rate of profit,
in a highly nonlinear way. This shows that the neoclassical theory of
income distribution is untenable. If you would like a quick explanation of
that, you could check:


Steedman then put Marx into the crucible, and his conclusions have been the
source of some angst for Marxists since, including you. Now the problem
which Ajit is having with your arguments is that, in effect, you are
accepting the import of Steedman's critique by arguing that Marx's theory
determines surplus value only up to a "factor of proportionality", and yet
you then continue to attempt to build a theory on this foundation.

The reply that Sraffa's "theory" also has a factor of proportionality, and
that it can't explain the existence and role of money would be a cogent
rebuttal *if* Sraffa's work was in fact a "theory". But it isn't--it's a
testbed for theories, no more (Ajit is quite aware of this, which is why I
regard him as a scholar on Sraffa and Marx rather than a Sraffian or
Marxian). If you read Sraffa's 1926 piece, you can see that, if he had been
trying to build a theory of capitalism, he would have based it on a dynamic
analysis of production, not a static one. Consider the following quote:

"Business men, who regard themselves as being subject to competitive
conditions, would consider absurd the assertion that the limit to their
production is to be found in the internal conditions of production in their
firm, which do not permit of the production of a greater quantity without
an increase in cost. The chief obstacle against which they have to contend
when they want gradually to increase their production does not lie in the
cost of production–which, indeed, generally favours them in that
direction–but in the difficulty of selling the larger quantity of goods
without reducing the price, or without having to face increased marketing
expenses." (Sraffa 1926)

Sraffa talking about marketing expenses? That ain't POCBMOC!

The real question then is whether POCBMOC is a valid crucible in which to
place what various people allege is Marxian economics. Here, I expect,
we'll part company. I believe it is, and therefore I interpret quite
differently to you, Andrew et al who object to its conclusions.

One key difference there is how Andrew, and I think also you, characterise
Sraffa's analysis as "simultaneist", as though it alleges that everything
happens at once in the capitalist system when in fact, as we all know,
things take time.

Again, this is not exclusively a criticism of you or Andrew, because many
"Sraffian economists" make precisely the same error. They don't argue that
everything does happen at once, but they do argue that the long term
dynamic processes of capitalism will converge to Sraffian prices, etc. As
Steedman put it when he criticised Kaleckian pricing theory:

"The general point which is illustrated by the above examples is, of
course, that our previous 'static' analysis does not 'ignore' time. To the
contrary, that analysis allows enough time for changes in prime costs,
markups, etc., to have their full effects" (Steedman, 1992, p. 146)

This is, quite simply, wrong, as I argued in the paper I mentioned a few
days ago. It is only true if the dynamic processes of capitalism are
globally stable, in that any short term divergence from equilibrium will
dissipate over time. I showed in that paper that, when you work out the
dynamics properly, the resulting system is almost always "marginally
unstable" (it has at least one eigenvalue of 1), so that convergence to
equilibrium cannot take place.

I expect that you might take the argument above as support for the notions
you and others are developing, that Marx's system was dynamic, and it's
erroneous to apply static analysis (including Sraffa's crucible) to it.
Don't: there is a way of reading Sraffa's crucible which makes it a dynamic

This is that it is examining the validity of the alleged equilibrium of a
dynamic system. In his critique of neoclassical theory, Sraffa effectively
argued that he was examining the concept of the aggregate value of capital
in an economy which had been in equilibrium for the indefinite past. His
proof that the "amount" of capital depended on the rate of profit showed
that the alleged neoclassical equilbrium of a market economy was logically

The same perspective can be put on Steedman's critique of Marx. Even if
Marx's model is a dynamic one (which I believe it is by the way, though I
take a very different approach to it than you do), it will have an
equilibrium. Is this equilibrium logically tenable? The Sraffian crucible
says no.

Does TSS, or any other approach to Marx which preserves the proposition
that labor is the only source of value, escape Sraffa's crucible. I argue
no, but I'm not about to try to convince you or any other Marxists of that.

In conclusion, to reply to Ajit that:

>Readers, please note that Ajit has not answered my argument that Marx's
>theory is superior to Sraffa's theory in the sense that Marx's theory
>explains the necessity of money and Sraffa's theory does not.

is no reply to someone who knows his Sraffa well, as Ajit does--though it
might shake a less philosophically aware Sraffian. All it does is annoy
Ajit, and quite rightfully too, given that (I hope I'm not putting the
wrong words in your mouth Ajit!) he would never argue that POCBMOC was a
theory of capitalism.


At 02:20 PM 9/22/00 -0400, you wrote:
>This is a response to Ajit's (3833). I am sorry that I am falling further
>behind in responding to others. I hope to have more time next week.
>Ajit, to begin with, I really do not appreciate your condescending
>personal insults. I am trying to engage in serious debate and consider
>your points and criticisms seriously, and respond to them at some
>length. But you respond with insults. I have been trying to ignore the
>insults and carry on the debate, but unfortunately the ratio of insults to
>serious debate increased sharply in your last post. So, I ask you to
>please dispense with the insults and continue with the debate.
>On Mon, 18 Sep 2000, Ajit Sinha wrote:
>> _____________________Fred, basically your S is equal to (m.L - V), i.e. S =
>> (m.L - V), where according to you, you know your L and V but not m. Thus
>> S is neither known in absolute terms nor to any degree of
"proposnality". This
>> is so simple that i cannot believe I have to explain it to you so many
>Ajit, I am afraid that you haven't explained it even once yet. I have
>argued that Marx's theory concludes that the magnitude of surplus-value is
>proportional to surplus labor-time, with m as the factor of
>proportionality (i.e. S = m Ls). Why isn't this determination up to a
>factor of proportionality? Please be specific. What more is needed to
>make this equation determination up to a factor of proportionality? If m
>were determined, then the absolute magnitude of S would be
>determined. But if m is not determined, then the magnitude of S is
>determined up to a factor of proportionality. Why not?
>I would really appreciate some comments by other listmembers on this key
>specific point. How else are we going to resolve this dispute? Am I
>missing something or is Ajit? If Marx's theory concludes that S = m Ls,
>doesn't this determine S up to a factor of proportionality? If not, why
>> > Fred:
>> >
>> > Ajit did not respond to my argument in (3815) about COMPARED TO WHAT? So
>> > I repeat a part of that argument, and then continue.
>> _____________________
>> I'm not interested in comparing your theory to any theory. Even if only
>> theory existed in the world, my criticisms will stand. It basically states
>> that your theory is inconsistent and does not achieve what it claims to
>> achieve. It has nothing to do with the strength or weakness of any other
>> theory.
>> ___________________
>I have already answered your criticisms of "inconsistency" in previous
>posts. That is not what this post is about. This post is about Marx's
>failure to explain the determination of m. I acknowledge that the lack of
>determination of m is a weakness of Marx's theory. But all theories have
>weaknesses (elements of indetermination, etc.). So the question of
>weakness inevitably becomes a question of RELATIVE or COMPARATIVE
>weaknesses. If the weaknesses of other theories are even greater than
>Marx's lack of determination of m, then this lack is not a reason to
>reject Marx's theory.
>Readers, please note that Ajit has not answered my argument that Marx's
>theory is superior to Sraffa's theory in the sense that Marx's theory
>explains the necessity of money and Sraffa's theory does not.
>> > Fred:
>> >
>> > Ajit criticizes "my theory" because it does not explain the determination
>> > of m. However, if Marx's theory is weak, COMPARED TO WHAT? Sraffian
>> > theory cannot even explain money, period; i.e. it can not explain the
>> > necessity of money, why money must exist in a commodity-producing
>> > economy. At least Marx's theory can explain, as a logical deduction from
>> > the fundamental assumption of the theory (the "labor theory of value"),
>> > that money has to exist, in order to function as the special commodity in
>> > which all other commodities express their labor-value. This is not an
>> > ad-hoc explanation of the necessity of money, based on the "difficulties
>> > of barter", but a deduction from the fundamental assumption of the
>> > theory. In other words, the necessity of money is explained in an
>> > integrated way, along with the explanation of lots of other phenomena
>> > (e.g. conflict over the length of the working day), all derived from this
>> > fundamental assumption. I think this is a very significant theoretical
>> > accomplishment that no other economic theory has been able to achieve.
>> Your theory, and leave Marx out of it please, does not explain money
>> You just make definitional claims and say given given etc. In anycase, the
>> issue under debate has nothing to do with the explanation of money. You
>> that money is given and I have not raised any problem on that count anyway.
>> Why not keep the focus on the problem at hand?
>I am talking about Marx's derivation of the necessity of money in Section
>3 of Chapter 1. What is wrong with Marx's argument in this section? If
>there is nothing wrong, then this is a clear element of superiority of
>Marx's theory.
>> > Fred:
>> >
>> > Furthermore, it just occurred to me yesterday: Sraffa's theory does not
>> > explain absolute prices either, but only RELATIVE prices!
>> _________________
>> This occurred to you only yesterday! And you have been writing rhetoric
>> against Sraffa for how many years?
>I have read Sraffa many times (although I gladly acknowledge that I do not
>understand Sraffa's theory as well as Ajit and others), and I have known
>for a long time that Sraffa's system only determines relative numeraire
>prices. What occurred to me recently is the similarity between Sraffa's
>determination of relative prices and Marx's determination of surplus-value
>- in the sense that both are types of determination up to a factor of
>proportionality. In other words, the very weakness that you criticize
>Marx's theory for, Sraffa's theory also has. Therefore, this weakness is
>no reason to choose Sraffa's theory over Marx's theory.
>Instead of responding to this argument, Ajit insults my scholarly
>integrity by asserting that I have been writing about Sraffa without
>reading him.
>> You should know that price is a relative
>> concept. Marx's prices in terms of gold is also relative. It is a ratio of
>> exchange between two commodities.
>Marx's prices are indeed in terms of gold, but they are in terms of an
>absolute quantity of gold (or of money representing gold); e.g. 30
>shillings, 600 pounds, etc. in Marx's many examples). In this sense,
>Marx's money prices are absolute prices; they are not ratios of absolute
>Sraffa's relative prices, on the other hand, are ratios of absolute
>prices. The absolute prices are never determined, only the ratios between
>Ajit, what exactly are these absolute prices that appear in Sraffa's
>system of equations, prior to the selection of a numeraire and the
>determination of relative prices. How exactly are these absolute prices
>defined? In terms of gold or what? Could you please give me some
>references where Sraffa or others have discussed the definition of these
>absolute prices? Thanks.
>>> The Sraffian
>>> system of equations has an extra unknown. The system can be solved if
>>> the price of one of the commodities is set equal to 1 (i.e. is TAKEN AS
>>> GIVEN!). This one commodity, called the numeraire, is arbitrarily
>>> chosen and is not necessarily real money. (Indeed in a system of paper
>>> money, the numeraire commodity cannot be real money). Sraffa's
>>> innovation was to take as the numeraire the "standard commodity", which
>>> is a composite commodity with peculiar characteristics and which has no
>>> relation to real money at all. Sraffa's relative prices in terms of
>>> the ideal "standard
>>> commodity" have nothing to do with real world prices; they are only a
>>> solution to a logical problem with Sraffian theory ("the invariable
>>> measure problem" in order make prices invariant to changes in the
>>> distribution of income between wages and profits).
>> So now we know that you opened the PCMC for the first time only
yesterday. I
>> hope i can take credit for making you do so. The book is a hard one to
>> understand. It will take some time, but will put you on the right track.
>Readers, please note that my serious criticism of Sraffa's theory - that
>its numeraire prices, especially with the standard commodity as the
>numeraire, have nothing to do with real world prices - is met, not with a
>response to this criticism, but with another personal insult. The insult
>is not appreciated and the criticism remains unanswered.
>> > Fred:
>> > Therefore, we can see that Sraffa's theory is also determinant "only
up to
>> > a scalar multiple", just like Marx's theory. In this respect, so roundly
>> > condemned by Ajit, Marx's theory is no worse than Sraffa's theory. And
>> > the fact that Marx's theory is trying to explain real world prices and
>> > real world profit makes it preferable to Sraffa's theory, which is only
>> > trying to determine hypothetical numeraire-prices, which have no relation
>> > to real world prices and profit.
>> ____________________
>> Now you must know that for your theory "only up to a scalar multiple"
does not
>> hold. As far as Sraffa is concerned, it does not matter which commodity you
>> choose, take gold if you like, the price ratios will remain the same.
>This is similar to what I am saying about the role of m in Marx's
>theory. If m changes, all the absolute monetary magnitudes change
>proportionality, so the ratios among these monetary magnitudes remain the
>> > Fred:
>> >
>> > How can anyone who accepts such a highly unrealistic theory like Sraffa's
>> > criticize Marx's theory for failing to provide a complete explanation of
>> > the determination of the value of money in his theory of real world
>> > and profit? I think it is far better to have a partial explanation of
>> > reality than a partial explanation of hypothetical
>> > numeraire-prices. Especially when Marx's theory has such substantial
>> > explanatory power of important phenomena of real capitalist economies.
>> _______________________
>> Please leave Marx out of it, because, in my opinion, you don't understand
>> Marx. So leave the poor guy alone. Your above rhetoric only shows that you
>> have not understood Sraffa at all. But as i said, it will take time. Nobody
>> can understand Sraffa in one day. Cheers, ajit sinha
>Again, a serious criticism - that it is better to have a partial
>explanation of real world prices than a partial explanation of
>hypothetical numeraire prices - is met with an insult. Again, the
>criticism remains unanswered.
>Please see my next post for a response to Ajit's "please leave Marx out of
>I look forward to further discussion (without the insults, please).
Dr. Steve Keen
Senior Lecturer
Economics & Finance
University of Western Sydney Macarthur
Building 11 Room 30,
Goldsmith Avenue, Campbelltown
PO Box 555 Campbelltown NSW 2560
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 2b29 : Sat Sep 30 2000 - 00:00:05 EDT