========================================================================
A response to Ajit's ope-l 4639.
Ajit: "Okay, I'm going to try it again."
With, I note, a whole new set of questions. I answered his
next-to-last set of questions in my ope-l 4635 (posted by Jerry),
at the end of which I wrote: "Can you identify any internal
inconsistency in this reasoning, Ajit? If so, please tell us what
it is. If not, please state that you cannot."
His ope-l 4639 does neither. I reiterate my requests.
For the record, "this reasoning" refers to the following, which I
wrote in the same post:
"if *all* the determinants of the output prices of 1996 and 1997
are the same, then the output prices of 1996 and 1997 must be
the same.
"But the input prices of 1996 are NECESSARILY given, already
existing, at the start of 1996, and not determined during 1996.
That is because the output prices of 1995 were determined during
1995, and, since the output prices of 1995 ARE the input prices of
1996, the input prices of 1996 were determined in 1995.
"This is not a complete theory of price, but it is a necessary
element of any theory of price that (a) does not permit future
prices to determine past prices and (b) does not permit the same
commodity to have two different prices at the same place and time.
"Notice that it does not imply that prices are arbitrary. The output
prices of any year are fully determined by a set of events occuring
before and during that year, including those events which occurred
before that year and which determined the input prices of that year."
Now, my reasoning implies that if output prices depend in part on
input prices, then, even when all the non-price determinants of output
prices of two different periods are the same, the input and output
prices of each of these two periods can differ. Let X indicate a
vector of non-price determinants, and P a vector of prices. Then we
can write
P[1996] = g(X, P[1995])
P[1997] = h(X, P[1996])
Quite obviously, g and h are different functions.
Now, let us imagine that 1995 prices were determined as follows:
P[1995] = f(X', P[1994]) = P[1994],
where X' is a vector of non-price determinants that differs from X.
Note that, by assumption, we begin with stationary prices --
P[1995] = P[1994]. I think Ajit would agree that if X and X' differ,
P[1995] will not equal P[1996] in general. Good. Then P[1996]
differs from P[1995] because both vectors in f differ from both
vectors in g. And then P[1997] differs from P[1996] because
one vector, the vector of input prices, differs in g and h.
This was all explained by Alan Freeman 13 years ago in his
chapter in _Marx, Ricardo, Sraffa_.
I consider this a definitive rebuttal to the key point Ajit has
been making all along, which he stated in ope-l 4419 as follows:
"By the way, you didn't get the point. The point was that if all other
things, leaving prices out, remain constant from period zero
[i.e., 1996 -- AJK] to one [i.e., 1997 -- AJK], then any reasonable
person would expect prices to remain constant as well. Why?
because there is no reason for prices to change. Why? Because a
reasonable person would think that prices are DETERMINED by
various other factors in the world, and if they are all constant then
there is no reason for prices to change. You do not have to ASSUME
that prices remain constant too. It IMPLIES that prices would also
remain constant.
On the basis of two ideas which I consider, if not reasonable, then at
least true -- (a) the output prices of one period are the input prices
of the next; (b) the output prices of any period depend partly on the
input prices of that period -- I have just demonstrated that what Ajit
has been claiming is incorrect. To maintain the reasonableness of his
position, he must reject (a), (b), or both, and defend the notion that
his rejection is reasonable.
Indeed the notion of "leaving [input] prices out" of the determination
of output prices seems to imply rejection of (b). I think it would
make for a very interesting empirical study if a decent-sized sample of
reasonable persons were asked whether they think that changes in
gasoline prices are unaffected by changes in oil prices. Another
interesting empirical study would be to determine whether this is the
case or not.
It would also be interesting to determine whether reasonable persons
think that (a) is true or false. I think that Gerard Dumenil and
Dominique Levy are reasonable persons. In their recent paper, "The
Conservation of Value: A Rejoinder to Alan Freeman" (March 27, 1997),
they acknowledge, on p. 15, that "A commodity cannot have a price as
the output of one production period, and another as the input of the
next period, since there is only one transaction."
On to some of the new questions that Ajit raises in ope-l 4639.
Ajit: "Wouldn't you agree that in a capitalist economy we find that
prices of most of the commodities remain more or less stable for
reasonable amount of time?"
No. Between 1967 and 1993 in the U.S., there was only one year in
which the CPI-U rose less than 3%. The prices changes, are, moreover,
extremely broad-based. Nor are relative prices stable. Let's examine
a very-low inflation year, 1993, during which the CPI-U rose 3.0%.
In that year, the disaggregated percentage changes in prices, for
major commodity groups, were
energy 1.2%
food 2.2
shelter 3.0
apparel & upkeep 1.4
transportation 3.1
medical care 5.9
fuel oil - 0.9
electricity 2.0
utility (piped) gas 6.2
telephone services 0.7
The range of percentage changes is 7.1 percentage points. For these
unweighted figures, the mean is 2.48; std. deviation, 2.21; and
coefficient of variation, 89%.
Ajit: "When prices begin to fluctuate drastically, it amounts to
some kind of crisis. In normal situation people, i.e. producers and
consumers, capitalists and workers, go about their daily business
with a sense of stability in the prices. This particular empirical
fact ..."
What empirical evidence do you have that this is the "sense"
people have? The people I know think prices always change, so
they hold off buying things in expectation of discounts. In the
U.S., such behavior is more and more common, since the practice
of varying the prices has become more and more common. The
airlines are continually playing with airfares, and clothing
and department stores are continually varying prices.
Ajit: "This particular empirical fact has led most of the economic
theorists to ask the question: what is it that determins these prices,
because they don't seem to be arbitrary."
Which prices are "these prices"? Actual prices, which are not
stationary? The prices that people "sense," which you haven't
shown are stationary either? Or imaginary prices corresponding
to an imaginary stationary state?
The answer to what determines the latter is extremely simple:
the imaginary stationary prices do not change unless technology,
real wages, or relative profit rates change because they are
*postulated* not to change; these imaginary prices are
"determined" solely by technology, real wages, and relative
profit rates because that has been *postulated* to be the case.
For instance, Sraffa fails to prove that, in an economy with a
surplus, there is only one set of exchange ratios that permits
simple reproduction to take place together with equalized rates
of return on capital advanced. Nor does he prove that there is
a unique uniform profit rate in such a case. I demonstrated this
in my paper "The Okishio Theorem: An Obituary," which I know
that Ajit knows, because he was a presenter on the same panel at
the ASSA in January at which I presented it.
Ajit: "Now, in your example, do the initial prices of all the
previous year you are quoting remain the same or differ from
each other? In either case, you need to answer *why* they either
remain the same or differ."
I have done so, in the most general terms possible, above.
Andrew Kliman (AX)