I don't mean to pre-empt Allin's reply to Alan's questions, but I'd like to
mention a couple of things in connection with this exchange. In response to
Allin's reply,
>The money wage divided by the CPI. Everyone knows there are
>serious index-number problems, which become more severe the
>more widely separated are the points in time between which
>comparison is made; the concept is not thereby rendered
>totally useless.
, Alan writes:
>OK, what are these problems, how severe are they
>and what use is it?
The central problem is that the underlying phenomenon which "real" measures
attempt to get at, i.e. the relative success of given decision makers in
attaining their objectives (e.g., "higher utility"), is invisible, since
people don't have their objective functions written on their foreheads.
Corollaries of this problem: consumer price indices are defined using
consumption "weights" derived from the market basket of a "typical
consumer." But these weights may change over time due to changes in
relative prices or (worse) changes in technology or underlying
preferences--thus Allin's comment that problems are likely to become
increasingly severe over time . Conversely, weights may stay the same and
yet quality improvements imply that consumers are better off even if the
measure of real wage doesn't change.
However, whatever the problems are, they are essentially *the same* whether
one attempts to measure empirically "the real wage" or "the value of labor
power". Both deal with aggregation problems (and thus the problem of
changing weights and product quality) and the attempt to measure something
not directly observable.
Indeed, they are worse, since unless one defines the value of labor power as
just equal to the real wage in value terms (in which problems in measuring
the former are at least as great as problems in measuring the latter), one
has the additional problem of distinguishing whether (for example) the value
of labor power has fallen, or the real wage has fallen below an invariant
value of labor power.
For what it's worth, in a world of heterogeneous capital goods, Marx's
notion of the "technical composition of capital" is incoherent in the
absence of an index compiled along the same principles as the price index,
and is thus also open to exactly the same measurement difficulties.
>In particular is it useful enough to
>be employed in the following type of
>phrase:
>'the rate of profit rises if with technological
>improvement if the real wage is constant?'
Sure it is, since it is clear *in theory* what is meant by constant real
wage, even if there are measurement problems in practice. Or, to parallel
the above comments, it is no more or less useful than the following type of
phrase: "the rate of profit rises with technological improvement if the
rate of surplus value is constant"