Andrew's OPE:1301 of 04/03 very conveniently lets me explain
my position, to which he refers. I won't take up time and space
justifying it but it might be useful for folks to know what
it actually is.
Andrew writes
=============
"In the case of constant capital that lasts more than one
period, I interpret Marx (e.g., Vol. I, p. 318 Vintage) as
saying that a fractional share of the pre-production
reproduction cost of the capital is transferred to the value
of the product. For example, imagine a machine that lasts 4
periods and which has the following values: Vo = 40, V1 =
28, V2 = 16, V3 = 12. Then the value transferred to the
products that emerge at times 1, 2, 3, and 4 is VT1 = 40/4 =
10, VT2 = 28/4 = 7, VT3 = 16/4 = 4, and VT4 = 12/4 = 3. In
sum, a value of 24 has been transferred through "regular"
depreciation. The remaining value of 40 - 24 = 16 is moral
depreciation which, as I interpret Marx, the firm does not
recoup in sales (cet. par.), i.e., a capital loss."
==============
I only need talk about the period 0 to period 1 transition
to explain my view.
Agreed, I think, is the following:
=================================
At the beginning of period 0 the capital stock of the
capitalist, who owns four machines each worth $40, is $160.
At the beginning of period 1 the capital stock of the
capitalist, who owns four machines now each worth $28, is
$112.
Hence the decline in the value of the machines is $48.
The bit that isn't agreed
==========================
Of the $48 decline in capital stock, one part is due to
physical wear and tear and one part is due to moral
depreciation, that is, the cheapening of machines of this
type due to technical progress externally to this process of
production.
I will take the two extreme cases first in order to
illustrate the problem. Then I will deal with the
intermediate and fully general case.
Extreme case 1
==============
The whole of the decline in value is due to
material wear and tear. That is, $48/$160 = 300f the
material body of the machine has been used up or worn away.
If the life of a machine is three and one-third periods of
production, for example, exactly 300f it is used up in
each period.
In this case, it seems to me that there should be no dispute
since moral depreciation does not arise. Assuming the
wearing-away of the the machine is distributed evenly over
time (and no reason has been given for any other assumption)
it must contribute an equal amount of its value to the
product in an equal time.
Therefore, if 300f the machine has been used up, then 30%
of its value should pass into the product (whatever is being
made with the machine). Therefore, what it passes to the
product is equal to $48, the whole of the loss in value. On
the basis of Andrew's calculation above, the value passed
into the product is *less* than that accounted for by the
wearing away of the machine, which I find a very odd
conclusion.
Andrew appears to argue (from the numbers) that the machine
wears out in an uneven manner, as testified by the uneven
sequence of falls in value. But at the beginning of period 1
we do not *know* what is going to happen in the next 3
periods, a dispute we have had occasion to take up with the
simultaneists in another context. We only know that they
have lost so much through wear and tear *so far* that they
are now worth only $112, and that this wear and tear has
accounted for all of the decline in value. Therefore, all of
this decline in value has passed into the product.
Consequently, in the next period $12*4 = $48 passes into the
product, in the next period $4*4 = $16, and if the machine
passes away completely at the end of the next period, the
whole remaining value of $4*12=$48 would pass into the
product in this last period. Total $48+$48+$16+$48 = $160;
under my assumptions there is *no* moral depreciation so I
would not accept that the moral depreciation is $16. Gil is
free to argue that my answer follows from my definition.
Indeed it does, but we must surely be able to deal with this
limiting case where there is no moral depreciation.
Far be it from me to say my calculation is closer to Marx's,
but in this case, where there is no moral depreciation,
there ought to be very little room for dispute. It seems to
me that Marx is categorical and indeed almost pedantic about
the workings of wear and tear and its relation to turnover
time. I am quite surprised, in a sense, to find that in this
case we come up with a different value transfer.
Much more controversial, and this is what I have until now
always assumed to be the substance of the discussion we need
to have, is the case where the decline in value is not
wholly due to wear and tear but some part of it is due to a
general cheapening of machines of this type.
Hence
Extreme case 2
==============
None of the decline in value is due to material wear and tear.
The machines are in the same mint and perfect condition as the
day they were bought. [This is not so far-fetched: I think it
occurs in the case of computer programmes, which never wear
out. The whole of the decline in the value of software,
considered as an independent use-value, is moral depreciation].
In this case, since none of the 'machine' has worn out, no
value is passed into the product.
However, the capitalists *experience* the decline in value
as *if* the machine had physically worn out. They therefore
wrongly record it in their accounts as a cost, exactly as if
it had worn out. The arguments in the accountancy profession
about whether to record software as a current or a capital
item are almost comical.
This is thus a false, fetishised representation of what has
really happened. Something which is not in fact a cost, has
been recorded as if it were a cost. We have to penetrate
behind this fetishised form and determine what this falsely-
represented sum, really is.
Really, I would argue, it is a value transfer. We can see
this by comparing the cost of a *new* machine of the same
type, at the beginning of period 1, with the machines in the
hands of our capitalist. The capitalist could take each
machine, sell it as new (which in fact it is, since it is
unblemished) and get $28 for it. That is, with no material
wear and tear, a new machine costs in fact $28.
This means that the market, or social value, of one machine
is $28. To simplify matters I ignore price-value deviations
(pace Gil) and assume goods sell at their (social) values.
The machine, when purchased, cost the capitalist $160. That
was its value then. Assuming no change in the monetary
expression of labour, the capitalist has lost $48 due to the
effects of circulation alone. This $48 has vanished because,
and only because, the market forms a uniform price and hence
a uniform value for machines of this type.
This $48 has been transferred elsewhere. Who has received
it? The producers of new machines of the same type, who have
driven the average or social value down to $28. They in turn
will actually be producing these machines for even *less*
than $28. Suppose for example eight of these new machines
have been produced.
In society there remain not eight, but twelve machines in
existence. $28 being the social value of one such machine
(and they are all indistinguishable from each other), the
total social value of these twelve machines is $28*12=$336.
This represents the total dead and living labour which has
gone into the production of the total stock of 12 good-as-
new machines. $160 went into this stock last year and $336-
$160=$176 this year. But when the labour was actually
expended is irrelevant. The final result is that there are
12 machines containing labour equivalent to $336.
$160 of this is the dead and living labour which went into
the production of the first four machines. Therefore the
labour which went into the production of the eight new
machines must be $336-$160=$176.
There being eight such machines, the dead and living labour
in each such machine must be $176/8 = $22.
This is the *individual* value of new machines. The new
producers, once the weight of the existing stock of machines
pressing on the market is swamped by the weight of new
production, will eventually (assuming no further change)
have to sell them for $22 each, just as the price of new
computers rapidly sinks after they are introduced at an
initially high price, once they become the norm instead of
an innovation.
For now, however, they can make a machine comprising $22 in
dead and living labour, and sell it for $28, that is, they
make a superprofit of $6 per machine.
The *depreciation* of $48 in the existing machines thus has
an exact counterpart in the $48 superprofit earned by the
new manufacturers. This has the same status as rent,
interest, commercial profit or any other 'levy' on the
general pool of profit - it is a transfer from one
capitalist to another.
The original capitalist *records* and *perceives* this as a
cost of $48, that is, as part of the constant capital
involved in making the product. But really, it is a
*deduction from profits*, a 'technical rent' which allows
the new producers to make a higher-than-average profit.
Again, in this extreme case I also differ from Andrew, since
he has it that the ('old') machine transfers $10 to the
product even though it has not suffered any material
alteration at all, that is, even though its turnover time in
use-value terms is effectively infinite.
How do we distinguish moral from material depreciation?
=======================================================
This already illustrates a very important point: Marx,
wherever he speaks of moral depreciation or the cheapening
of constant capital, always says that the change in value
originates 'outside the process itself', that is, outside of
the process that uses the machines.
But most of the illustrations that have been employed in the
discussion on depreciation, do not attempt to quantify which
portion of the decline in value originates outside, and
which inside, the process of production which uses the
machine. Thus Andrew (like, I think, many others in the
discussion) gets the *same* value transfer to the product,
regardless of whether the decline in value originates
inside, or outside, the process using the machine. I think
this cannot be right. Whatever moral depreciation is, it
must be possible to distinguish it both quantitatively and
theoretically from material wear and tear, and the value
transferred to the product therefore *cannot* be determined
without knowing what proportion of the decline in value, is
due to what cause. I cannot accept that the *same* answer is
valid in all cases. Otherwise, in what wise is moral
depreciation different from any other depreciation?
I have always founds this strange since, when I study
Andrew's (much-neglected) work on the Okishio theorem, the
equations I find there seem to me to make this necessary
distinction, which is why (in my opinion) he gets the same
curve and the same figures for the rate of profit as I do.
Case 3: a mixture of moral and material depreciation
====================================================
Now pass to the intermediate, and general case. This is
where a *part* of the depreciation is material, and a
*part* of the depreciation is moral.
In this case we must know the relative proportions by some
other means, or we cannot properly estimate the value
transferred to the product. But I think this is quite easy
to ascertain. What we should do is compare the 'resale' or
scrap value of the machine in the capitalists' hands, with
that of a new machine of the same type. If, for example, a
new machine could be found on the street for $32, then I
would say that of the total decline in value of $12 (from
$40 to $28) in each machine, $8 is the effect of moral
depreciation, to be calculated as in the second part of this
post, and the remaining $4 is the effect of material wear
and tear, to be calculated as in the first part of this
post.
An alternative which Andrew might find more acceptable,
since I think he prefers to think in terms of the natural
physical life of the machine, is the following: if we know
what this natural physical life is (on reasonable
assumptions about intensity of use) then we can say that in
each year, the value passing into the product through
material wear and tear is $160 divided by this life (that
is, as Marx says, the total purchase price divided by the
turnover time) and the rest is moral depreciation.
Alan