Summer 1927, D3/12/3: 49–51, A4.17.i–iii, reference to Pigou, A. (1927) The laws of diminishing and increasing costs. Econ. J., June, 188–197
Pigou ‒ Laws of incr. + dim. costs, E. J. 1927, § 3‒5
It is not at all clear where prof. Pigou finds a difficulty with di incr. costs industries, in which, which [with] change of output, the value of factors of production is changed. His argument may be put thus: suppose demand for silver increases, inferior mines are worked, royalties increase and wages remain unchanged ‒ how shall we know if cost is increased or decreased? if we measure them in terms of mines, they may be diminished, if in terms of labour they may be increased. “There would be no real significance in the result ...; for its nature would depend, not upon the facts, but upon an arbitrary choice of technique. There is no way, whether on lines familiar to students of index numbers or otherwise, by which the difficulty can be successfully overcome” (191‒2).
P. is talking of real costs (191): if by this he means “total disutilities” one does not see why in the present case the measurement should be more difficult than in any other case, since they have to be measured in “utils” or some unit of that sort: of course, [it] this is impossible in all cases to measure a total of various sorts of disutilities of different individuals, but this case makes no difference. If the reply is that in the present case there must be an increase, not only in real costs, but also in rents, and the two increases are likely not to be proportional, there are two answers: 1) that (values) of factors are changed, but in any case in which, through a var change in the amount [of commodity] produced, a change in the proportions in which different factors are employed in that particular production takes place (although their rewards per unit be unchanged) = so that and as in the reward of each factor there may be a different proportion that is rent and not cost, the total real costs may be diminished [increased] when the total value (ast total factors used per unit of product is diminished (this total being measured in terms of any one factor, according to P.'s method). 2) that if this is the difficulty, it is irrelevant whether the whole argument on impossibility of finding a non-arbitrary unit of measure is irrelevant, because whether there is or not such a unit the difficulty remains equal.
If by real costs P. means expenses of production (which, from the context seems more probable) one does not see why they should be measured in terms of one factor ‒ in terms of silver mines or of hours of labour. Since the eq reduction to unity is made through their exchange value (and not according to any other, psychich or physical, standard) of the obvious unit of measure is money (i.e. [purch. pow. ofver] commodities in general). One does not see the use, at this time of the day, to raise the old question: if 1a was worth 1b, and is now worth 2b, has a risen or b fallen in value? rather, the question raised in the article is worst and seems to be: is a+b worth more or less than before? Of course a definite reply can only be found by reference to [all] other commodities, i.e. money.
[The reference to “index-numbers” suggests that the objection to the use of money may be that, being changed the value of silver mines, its purch. power is changed: but of course this would belong to our old friend “the second order of magnitudes”.]