Andrew,
The problems in dealing with depreciation as we hold fast
to valuation using socially necessary labor time are not
easy. Perhaps, by listing them, we can at least begin
to find some solutions.
1. How do we figure the rate of profit on a particular capital
which uses fixed capital?
Here, as I have said, Marx gives us no clues. Perhaps we
need to start with the assumption that techniques are unchanging
and come to some sort of agreement. With unchanging techniques,
I agree that we need to assume a "natural" life span of fixed
capital.
If we use the idea of "present values" to compute the rate of
profit over the life of the fixed capital, then the amount of
profit or surplus value realized from any given period of
production will vary. That is, the simple separation of
depreciation and profit that occurs if we use straight line
depreciation loses its simplicity.
We could, of course, differentiate "our" view of the separation of
surplus value and depreciation from that of the owners of the
fixed capital. Given that we have assumed no change in techniques
as depreciation occurs, our sums of profit and our sums of depreciation
charges will not differ from theirs. However, in any given period
the depreciation charges as well as the profit observed will differ.
This is, perhaps, another kind of "transformation problem."
2. We then move to the difficulty of introducing technical change
as we consider the depreciation of fixed capital. For me and, I
think, for Marx, the "moral depreciation" of fixed capital that
occurs as technical change takes place shortens the life span
of fixed capital. Given capitalists allow for moral depreciation
as they decide whether or not to introduce a particular technique,
the life span used in calculating the anticipated rate of profit
would be its economic life span and not its "natural" one. To be
sure, neither the economic nor the natural life span can be known
*a priori*. Both are estimates. You use the latter and I, the
former.
I would argue that if machines generally prove to be profitable
for a period of 5 years even though they can used for 10 years,
both we and the capitalists that own the machines should use the
shorter time frame for rate of profit and depreciation calculations.
Given what we have done so far, it is clear to me that where we
start on this problem makes a great deal of difference. That is,
if I where to give you an example of a machine that is used for
5 periods with a given work force and tell you that the machine
has no value at the end of the 5th period, you could easily tell
me how much value was created and transferred in each period.
Further, given the initial value or price of the machine, you
could easily find the rate of return for the investment. I
could then "surprise" and say that the reason the machine
became unavailable for use in the 6th period was that while
the machine was technically useful, its use would have yielded
a lower rate of return than an investment in a new machine.
With this, I would have "tricked" you into including "moral
depreciation" charges as part of the overall depreciation
charge. Thus, given we have some agreement on the issues
in 1 above, perhaps "tricking" you is the solution. But
this would avoid what I see as the real difference in this
matter.
Your solution to the "moral depreciation" problem reminds me
of what we encounter and argue against when it comes to
technical change in circulating capital models. You recalculate
the value of the investment after it is made based upon the
changes in techniques that occur as the fixed capital is in use.
To be sure, you use the values of the previous period to make
this adjustment in a given period. Hence, I do not charge you
with using "simultaneous valuation." Yet, when you compute
the IRR or the rate of profit, you use the amount originally
invested without these adjustments. Why make the adjustments?
I assume that they are made in order to compute total value
of the output of each period.
Here, perhaps we run into a problem here in restricting our focus
to a particular capital rather than the economy as a whole.
Here let me attempt to put problem in your terms. Given that
techniques are changing as the machine is in use, the abstract
labor hours spent in the production process using the machine
will remain constant in each period for life of the machine
but the socially necessary labor hours will decrease as more
productive machines are introduced. To actually know how
much this decrease is you would want to know something of the
socially necessary hours for all processes used in a given
period. I do not intend to put words in your mouth but I
have simply surmised this as you have mentioned the
need to consider the entire economy in dealing with our
simple examples.
For now, I'm willing to assume a set prices for each period
and a constant MEL as we move from period to period. Maybe
we need a "simple" example of the entire economy so that
we can see what our differences are concerning "moral
depreciation."
John