A reply to Bruce's ope-l 3163.
First, I want to acknowledge that (a) I agree with the manner in which Bruce
applies Marx's category of "release of capital" to my figures; (b) I am happy
with Bruce's category of profit + release of capital, which he terms
"deployable surplus revenue"; (c) I think the term "deployable surplus
revenue" is very apt (though not "deployable surplus," because I think it is
important to distinguish between surplus-value, surplus-product,
surplus-labor, and surplus-revenue); (d) I accept that the ratio of deployable
surplus revenue to the replacement cost of capital is a constant 25 0n my
example; and (e) I think this ratio is a meaningful one in some contexts.
The rest of this post will attempt to clarify the TSS calculations. The posts
by Duncan, Fred, and now Bruce are leading me to believe that we haven't
explained the calculations carefully enough, and that this may be causing some
confusion. I hope this process of clarification will show why I cannot accept
Bruce's numbers, and therefore cannot accept the conclusions he derives from
them. I would like to ask Bruce to read my comments and then indicate how, if
at all, he would modify his numbers and/or conclusions in light of them.
Before I respond in any deeper sense, I think it is best that we make sure
there is no misunderstanding.
Bruce writes: "What I don't accept, though, is that the dollar magnitudes
here directly represent equal magnitudes of labor-time. Andrew believes they
do, and that's the way he accounts in labor-time terms." I am guessing from
this, and the rest of Bruce's post, that he thinks I presented certain money
value figures (along with amounts of corn and living labor) as data (givens),
assumed that $1 = 1 labor-hour throughout all time, and then derived certain
labor-time value figures from the data and the assumption. I think Bruce
considers this an illegitimate procedure, because the relation between money
and labor-time is endogenous and therefore cannot be arbitrarily assumed in
this context.
I agree. But this was not my procedure. I determined labor-time values
independently of both money prices and the labor-time money relation. This is
necessarily the case.
Let me explain in more detail. I'm going to stick to the single-sector,
circulating capital case for simplicity. There are two ways (at least) of
doing the TSS value calculations. BOTH ways of doing the calculations have
this in common:
(a) begin with a given sum of constant capital measured in *labor-time* (400
labor-hours),
(b) assume, or determine by means of actual data, the labor-time extracted
during the period from workers in capitalist production (100 labor-hours),
(c) stipulate that the value of constant capital is (exactly) preserved in
production and transferred to the value of output,
(d) stipulate that the labor-time extracted from workers in capitalist
production equals the labor-time value added,
(e) therefore add (a) and (b) to obtain the labor-time value of output of the
period (500 labor-hours),
(f) assume, or determine by means of actual data, the share of the output of
the period ploughed back into means of production next period (100%)
(g) stipulate that no value is created or destroyed in exchange
(h) therefore equate the share of the *value* of output ploughed back into
constant capital-*value* next period to (f) (100%)
(i) multiply (e) by (h) (500 labor-hours times 100%) to determine the sum of
constant capital of the next period (500 labor-hours).
(j) set (a) equal to (i). Continue through the list to compute the results
of the next period of production.
This is a recursive algorithm. Note, however, that in each period (b) and (f)
can change from period to period (in my example, they did not) and therefore
so can everything else.
The stipulations (c), (d), and (g) are theoretical propositions that I
interpret Marx as having held. From these propositions, together with an
initial condition (the original (a)), the other data ((b) and (f)), and the
purely mathematical operations ((e), (h), (i), and (j)), the results follow as
night follows day. Thus, everything depends on whether one accepts (c), (d),
and (g) (as an interpretation). Unless one interprets (c) as stipulating the
revaluation of inventories --- and it is clear to me it stipulates exactly the
opposite --- there is simply no revaluation to be found in this procedure.
When one works with labor-time, and with aggregates, and not money figures and
per-unit magnitudes, one sees this immediately.
Now note that the above procedure begins with a quantum of labor-time,
transfers quanta of labor-time, adds quanta of labor-time, and subtracts
labor-time not ploughed back. It is *all* done purely in labor-time, and the
amounts are determined from data and theoretical propositions. They are NOT
deduced from any money magnitudes and they are NOT deduced from any postulated
relationship between money and labor-time sums.
Note also that NONE of the above permits the determination of any money sums,
or of the money/labor-time relation.
OK. As I mentioned before, there are two ways of doing the TSS calculations.
The first is to *assume* a (constant or variable) money/labor-time relation,
and use it together with the *already-determined* labor-time magnitudes to
obtain money sums. Thus, the order is:
(1) take theoretical propositions and data
(2) derive labor-time value magnitudes from them
(3) *assume* a money/labor-time relation
(4) derive money magnitudes from (2) and (3).
This was the procedure I used in my example.
In the other procedure one takes money sums as data and uses them together
with the *already-determined* labor-time magnitudes to obtain the
money/labor-time relation. In this case, the order is:
(1) take theoretical propositions and data
(2) derive labor-time value magnitudes from them
(3') take money magnitudes
(4') derive the money/labor-time relation from (2) and (3').
As an illustration of the latter, I'll rework my example, this time NOT
assuming $1 = 1 labor-hour, but assuming I know the money output prices (and
therefore the money-values of the constant capital of next period). I'm going
to assume that the money-value of corn is a constant $100 per bu.
The top rows are the first-calculated labor-time values. The middle rows are
the given money values of output and constant capital, and the derived
money-values added (subtract money-value of constant capital from money-value
of output to get money value-added). The bottom rows are the derived
money/labor-time relations (monetary expressions of value).
year C V+S C+V+S
0 400 hrs. 100 hrs. 500 hrs.
$400 $100 $500
$1/hr. $1/hr.
1 500 hrs. 100 hrs. 600 hrs.
$500 $125 $625
$1/hr. $1.0417/hr.
2 600 hrs. 100 hrs. 700 hrs.
$625 $156.25 $781.25
$1.0417/hr. $1.1161/hr.
3 700 hrs. 100 hrs. 800 hrs.
$781.25 $195.31 $976.56
$1.1161/hr. $1.2207
The nominal profit rate would be a constant 25%, given v = 0. The labor-time
profit rates are as I reported before. If one uses the monetary expressions
of value to deflate the money-value figures, of course, one arrives back at
the labor-time profit rates.
In sum, I can't accept Bruce's numbers, or his conclusions, because they are
premised on the notion that my original example assumed a set of money sums
and a constant monetary expression of value in order then to arrive at
labor-time value sums. Actually, I first determined the labor-time value sums
and assumed a constant MEV in order then to arrive at money sums. It is
impossible, therefore, for a constant money sum in my example to represent a
variable amount of labor-time.
As I've just shown, I *can* work the example by assuming a set of money sums.
But I still can't, and didn't, use them to determine labor-time sums. Again,
I agree with what Bruce seems to say, that it is improper to assume money sums
and an MEV to get labor-time values.
It is crucial to note that, in the reworked example with a variable MEV and a
constant money-value per unit of corn, the labor-time values remained just as
they were before, when the MEV was constant and the money-value per unit of
corn varied. A labor-hour is a labor-hour is a labor-hour, and in Marx's
theory, and hour of labor always adds the same value (in labor-time terms).
Variations in money values and/or the MEV can't change that.
Andrew Kliman