I don't want to pass over the new debate that has started on
IVA and the like without clearly welcoming Duncan's [3073]
which, I think, opened up a new chapter of the discussion
as is clear from the subsequent flurry of exchanges.
I do agree wholeheartedly with the following point and I
wanted to state this agreement regardless of the subsequent
evolution of the debate:
"We now have the advantage of being able to focus on the
real point of dispute, which is the definition of value
added. This issue arises only when prices are changing,
which clarifies the importance of non-stationarity to the
TSS position."
I think this hits the nail right on the head. My objection
to simultaneous valuation has always included the fact that
value added emerges as a result instead of a starting point.
Whatever the surrounding metaphysical argument, in the actual
mathematics of the thing, prices and profits are not determined
by hours worked but by the maximum eigenvalue of a fixed-point
equation. Value added is then an 'interpretation' of this and
has no independent standing. As a consequence, if we alter the
magnitude of value added by, for example, changing the number
of hours worked without changing the real wage received, the
profit rate is as far as I can see in most presentations
unaffected. For Marx it has always seemed to me rather crucial
that if people work harder, the capitalists make more profits.
Otherwise what is absolute surplus value about?
In a dynamic framework there is no such predetermination of
anything. The only thing which is determinate is the past,
and the reason for this is that it *is* the past, and
therefore cannot be altered by the subjective desires of the
consumers of the product.
For a dynamic formulation it appears at first sight that
nothing determines the price of outputs or the magnitude of
profits. These are completely arbitrary, which has led David
Laibman to exclaim that without simultaneous determination,
there is no theory of anything. I can understand his
exasperation but in my view the equations are indeterminate
because life is indeterminate, and only on the basis of such
indeterminacy can live reality be captured.
Nevertheless, a system with very many degrees of freedom is
not a system with no laws at all, or there would be no laws
of physics. In my view *value* definitely is determinate,
and it is given at any moment by the fact that a a
determinate amount of dead labour given by the price of
inputs and the existing monetary expression of value, is
combined with a determinate amount of living labour given by
the time taken to work up this dead labour into a new
product.
The question for me is this: in order to progress from
stocks and prices at a given point in time, to stocks and
prices at the next point in time, we have to identify some
fundamental and persistent property of the system in terms
of which this progression can be described. We require,
moreover, a definition of this fundamental property, value,
which ensures that the value emerging from production is
independent of the price for which the produced good is
subsequently sold.
Without such a quantity, we cannot for example make any
statements about surplus, or distribution, hence
exploitation or intercapitalist competition, since we cannot
say *what* is being distributed as a result of subsequent
changes in the price of the product and hence of the profits
of the various sectors.
In the simultaneous framework, value is a structural
property of the entire system independent of its history.
The ratio, therefore, between the value added in monetary
terms and the value added in terms of labour hours, is
determined by the technology of the system and the
assumption of an equal profit rate.
In a dynamic framework, on the contrary, there is no such
determination. We cannot assume an equal profit rate and we
cannot make any assumption at all about future prices. All
we know is the given value of the money that was paid for
the inputs to production including the wage.
As I see it, therefore, there is no way to make values
determinate - no way to arrive at the value of the product -
except by adding to this pregiven (because past) value, some
quantity either equal to or derived from the actual hours
worked to produce the product.
We cannot calculate value added on the basis of the sale
price of the product because we do not know what this sale
price is, at the moment when the product has only been
produced but not sold. We are at the C' stage of the circuit
of reproduction, not the M' stage.
If we attempt to define value added in purely monetary
terms, then it seems to me we cannot adequately distinguish
between purely nominal changes in the price level, and
genuine technical change. Indeed in a certain sense isn't
this the whole point of having a theory of value?
My own inclination is to say that the monetary expression of
this value added is simply obtained by multiplying the hours
worked by the monetary expression of value that held prior
to production. So, if at time t the value of money (=mev)
was $1 per hour and workers work for 10 hours, then the
monetary expression of value added to the product at time
t+1 is $10.
Differences may arise from the following. It may be that the
actual sale price of the product is more than $10 in excess
of costs. In this case, I would say that the monetary
expression of value has changed, but the issue that now
needs to be discussed is, by how much has it changed?
My position has always been that the new monetary expression
of value can only be calculated by dividing the *total*
price of all stocks of all commodities by the *total* value
of the same collection of commodities.
The problem is the following. Suppose, for example, that
prior to production there were in existence stocks whose
value was $1000 in value. We know that this represents 1000
hours because the mev was $1 per hour.
Now suppose that some of this stock is worked up into a new
product with the 10 hours of labour, so that the new value
of stock including sales stock, measured in hours, is 1010
hours. Suppose that all prices now double, due to a sudden
inflation brought about, say, by an external devaluation
relative to some other currency (this is a future world
where either US finance capital can no longer determine
exchange rates or US technology lets me send E-Mails without
using the $ sign for every currency in the world) In that
case, the new price of this stock of goods will be $2020.
I think it is very hard, and leads to many problems, to say
that the labourers have added $1020 to the product and that
the value of money is therefore $102 per hour. I would
prefer to reason as follows:
A)measured in hours, the workers added 10 hours to create a
new stock of all goods whose total value is 1010, as
previously stated.
B)measured in pre-inflation money, these workers added $10
to create a stock whose value is $1010
C)a new monetary expression of value comes into existence
at the point that the market establishes a new price for
the commodities of which this stock consists (the M'
phase of the circuit) . This is given by the ratio
between the dollar price of all stocks now in existence,
and their value in hours, that is, $2020/1010 or $2 per
hour
D)measured in post-inflation money, the pre-production
stocks were worth $2000, since the exchange rate of old
for new dollars is 2/1. In this same measure, the workers
have added $20.
I think this leads to a fully determinate calculation of the
value of money that holds for any arbitrary sales prices,
that provides for a clear distinction between value and
price, and which makes no assumptions concerning
stationarity at all. I know, however, that this result
contradicts both the New Solution definition of the value of
money and Alejandro's. I wrote a short piece on this for the
1995 EEA but it seems to have got buried.
My feeling is that it would be very fruitful to discuss this
further because I think it may turn out to be an issue which
can be resolved by a more careful definition of terms. I
think the distance between us in terms of the general way we
conceive of the value of money is quite small, and I think
as I have said before that the concept of the value of money
introduced by the New Solution is absolutely seminal.
Now, as regards IVA I strongly agree that this is a very
important concept but I would urge caution to distinguish
monetary effects from genuine technical change. Indeed, I
think the great merit of a value-based approach
I find nothing to disagree with in the following point of
Duncan's:
"But it seems to me that the nub of the issue is, in
accounting terms, whether or not one includes the IVA in
value added, or in terms of the labor theory of value,
whether or not one attributes the change in the money value
of inventories over the production period to the
expenditure of living labor.
"I hope that Alan will accept my claim that excluding the
IVA from value added is not the same as assuming input
prices are equal to output prices, nor does it violate
Alan's principle that the money paid for the inputs ought
to equal the money received for the inputs. In the
equations we clearly distinguish p(t) from p(t-1), so input
prices are not being assumed to be equal to output prices.
The equations clearly reflect the fact that the money paid
by the capitalist for the inputs is p(t-1)a, the same as
the money received by the producers. The issue is whether
or not in applying the principle that it is living labor
that adds value to the product, we should count the IVA as
part of the value added or not."
I haven't followed through all the implications of this
assertion in the subsequent debate but as it stands I
completely agree. Living labour is unaltered by changes in
asset values. Moreover we need to account separately for
changes in asset values precisely in order to measure the
distributional impact of these changes. And I don't think
that accounting for stock revaluation runs counter to the
dynamic principle though we do need to distinguish pure
price effects, brought about by changes in the value of
money (mev) from real value effects brought about by
technical change.
Indeed the concept of asset revaluation is central to my own
understanding of the course of the business cycle, of the
causes of moral depreciation, and of the reasons for uneven
development.
In the last chapter of our book, I derived differential
equations that can be written as follows:
vX + p'<K> = pA + L
pX+ p'<K> = pA + L + E
where X is a matrix giving the flow rate of outputs, A a
matrix giving the flow rate of inputs, L the flow vector of
hours worked and E the flow vector of differences between
the realisation price of goods and the value embodied in
them. E summarises all the behavioural particularities of
the system. It sums to zero. K is the general matrix of
capital stocks of all kinds including production goods,
sales stocks, monetary stocks and stocks of secondhand
consumer durables. <K> is the diagonal matrix obtained by
summing K across commodities. P' here means the time
derivative of price.
The profit equations are as a result modified as follows:
S = L - V + p'(K - <K>)
Profit = L - V + p'(K-<K>) + E
where S is the vector of flow rates of surplus value
generation and 'Profit' is the vector of flow rates of
profit generation. The sum of S is the same as the sum of
profits, and the rate of profit is therefore independent of
E, which is why I claim that the profit rate does not depend
on any particular behavioural assumption.
The difference between these equations and the standard
value equations, apart from the lack of an equal profit rate
assumption, is the term p'<K> which I term the 'stock
revaluation' term. The rate at which value is added to any
product is diminished by the rate at which value is being
transferred to or from the total stock of that good, due to
changes in the price of the capital stocks brought about by
technical change or relative prices. The rate at which any
capital makes a profit is then modified by a redistribution
term p'(K-<K>) which transfers value from all those owning
stocks that are rapidly declining in value, to those ownning
stocks that are slowly declining or rising in value.
In the case of general technical advance, all elements of p'
are negative and gross value produced in each sector is
correspondingly smaller. However, value added is not
affected. So I agree with Duncan.
Moreover the loss of value is greater for those capitalists
who possess stocks of rapidly-depreciating goods, for
example those who buy hi-tech goods from the North. I should
maybe have called this extra term the 'moral depreciation'
term since, as far as I can see, it quantifies this concept
exactly. My point in the moral depreciation debate is that
moral depreciation must be completely independent of current
value added, or we end up with the idea that value can arise
out of nowhere. So this is close to what you say.
If the analysis is pursued it is found that the equation
K'=I
in value terms, where I is that portion of the surplus S
that is invested, is exactly true. Hence the capital stock
will rise as long as the capitalists invest any part of
their surplus. The equation r = S/K is also found to
represent exactly the price rate of profit if the monetary
expression of value remains constant.
[Incidentally if we add some accelerator condition on
investment, eg that I' is a proportion alpha*S of surplus
given by the difference between r and some target profit
rate, then we get a stable business cycles in which the only
variable is the profit rate]
Now, if the monetary expression of value in my sense
changes, I found the following seems to hold: if the
monetary expression of value is m then the equations have to
be modified to read
R = r + m'/m
Here R is the money rate of profit and r is the 'real' rate.
That is, the money rate of profit is increased by an
inflation term equal to the proportionate rate of increase
in the monetary expression of value.
The monetary term corresponds to the fact that I can make a
profit in money terms merely by holding onto goods whose
price is rising.
We can thus clearly distinguish by this means that part of
the profit rate due to 'real' factors from the purely
monetary effects of inflation
This also, I think, helps explain what happens in the
business cycle. As Marx notes, during booms there is a
generally rising rate of all prices (m' positive) caused by
excess demand. This raises the observed, monetary profit
rate above its real rate, which is in any case high because
the preceding slump has wiped out capital values.
At a certain point, the rate of profit begins to fall
because of the rise in capital stock. But as long as
monetary inflation persists, that is until the effect of
declining profit rates is manifested in a slackening of
investment, this is not immediately perceived by the
capitalists, who therefore overinvest.
Now, at some point a general fall-off in demand begins.
However, there is now negative feedback. For, once a decline
in prices sets in, the sign of the m'/m term reverses and
there is a sudden sharp decline in realised money profits.
This, I think, might help account for the catastrophic
nature of the onset of the slump phase.
The underlying real profit rate is in any case already low,
so that profits fall below a level at which fresh investment
on a large scale can continue.
The consequential fall in the prices of investment goods
wipes out both fictitious capital values in stock markets
and excessively valued productive assets, so that as Marx
puts it, the slump performs the function of readjusting
prices down to real values. Because of the dynamic term in
the price and value equations there is a knock-on effect on
profits as stock values collapse. At this point there is
genuinely negative investment first because even though the
capitalists do repurchase inputs they do so on a reduced
scale, and second because this renewal is offset by
declining asset values.
Finally capital prices have fallen to the point where the
real profit rate begins to recover, and the cycle restarts.
So I think that the equations are quite useful. There are no
particular behavioural assumptions in them as they hold for
any arbitrary E. My personal view is that this is rather
important since, if they depended on any particular
behavioural assumptions, they would be less generally
applicable.
However, one word of caution is required: the above is my
own view and not a general TSS view. In order to derive the
above equations, I argue that in the process by which social
or market values are formed from individual values, capital
stocks enter the averaging process along with current
production. This rather controversial view is not generall
shared. Nevertheless, differences on this question do not
affect the elements of qualitative agreement on the rate of
profit, because our difference with Okishio is a consequence
of dynamic valuation and arises no matter what method of
dynamic valuation is employed. There is in fact a spectrum
of such valuations which are possible and I think you will
see just about every possible variant from different
'TSSers'.
Our original purpose in calling our EEA conference was to
discuss out these questions among ourselves, but we were
joined by a lot of other folk and it seemed just as
important to discuss with them, too. I'm glad we did.
There really is no TSS orthodoxy!
Alan