Concerning depreciation:
While not claiming to be an expert on the subject, it seems to me that
Marx's views on depreciation are reasonably clear. Depreciation is that
part of the value of the fixed portion of constant
capital which is transferred "in aliquot parts, bit by bit" by living
labour to the new product. Because it is a transfer of the value of
constant capital inputs by living labour, it is "preserved value". So it is
not part of the output of "new value added" in the Marxian sense, therefore
also not part of surplus-value. While
depreciation write-offs enter into the value (and specifically, as Marx
says, the cost price) of the
new product, it is not part of "new value added" in the Marxian sense.
Marx further distinguishes between physical depreciation of fixed assets
and moral (economic) depreciation (technological obsolescence). That is to
say the value of fixed assets may alter over time not just because of
physical wear & tear, aging etc. but also because of competition, i.e.
their use-value relative to more advanced equipment produced by new
innovations.
The measurement difficulties would seem to lie in two areas: (1) the
valuation
of fixed assets (the fixed capital stock), (2) depreciation
schedules/methods.
Assets may actually be valued for depreciation purposes for instance at
historic
cost (acquisition cost) or at market value (what it would sell for in
today's market). Often it is difficult to know from official statistics
precisely what to make of the asset figures (book values) cited because a
mixture of different valuations are in reality applied by businesses.
Depreciation schedules (the amounts which may be written off annually for
tax purposes) are usually fixed by law, and there may be "tax incentives"
in the form of extra depreciation allowances. In the SNA national accounts
aggregate "consumption of fixed capital" some faux frais of production such
as insurance may also be included. The choices for the purpose of empirical
analysis seem to be either to stick as closely as possible to the actual
stock valuations and write-offs made, or to estimate fixed capital stocks
from benchmarks using gross fixed capital formation data, acording to some
variant of the perpetual inventory method. The difficulty then would be to
know which type of depreciation method to use, such as straight-line
depreciation or accelerated depreciation, and we may have to look at the
asset lives of stocks of fixed equipment and customary write-off
procedures.
Now why could the value of depreciation affect the estimated amount of
surplus-value, given that depreciation write-offs are conceptually not part
of surplus-value ? Presumably because if the write-offs are somehow
unrealistic or spurious in relation to the real value of fixed assets, then
depreciation can hide a fraction of total profits. The only way to resolve
this issue would seem to be to compare the available data on stocks,
investment, realised profits and consumption of fixed capital, and see what
difference it makes when different measures are used, and what measure
seems most realistic. For instance, what margin of error would be necessary
for the longer term trend in the rate of profit to be totally distorted ?
Presumably the object of the exercise is not to get a "perfect measure",
but something that allows us to identify the empirical trend, so that we
get an insight into the way capitalist development is actually moving over
time. When we try to obtain a measure, all we need to bear in mind is -
whatever the difficulties of measuring it - that fixed capital is not a
"theoretical entity" but a real value which we can estimate from the data
within certain limits of error. In my own (limited) work in this area I
often found that clever manipulations of asset and investment figures on
the basis of different theoretical assumptions do not really alter the
basic trend result very much.
Jurriaan.