668. Joan Robinson to Harrod , 13 May [1937] [a]

[Replies to a letter not found; follows on from 666 , continues at 671 ]

3, Trumpington Street, Cambridge

13 May [1937]

My dear Roy

I took a glance at your review [1] & was quite overwhelmed by your flattering remarks. If you take out the critical bit I shall really be afraid of the evil eye. But I think you are right to take it out, for we ought to get it cleared up ourselves before bothering the public.

(1) When I say amount of capital, I mean the total amount--not per head or per unit of output. The measurement involves of course an Index [number] problem, but it seems to me that that is inevitable when changes of technique are involved. The best measure I think is in terms of cost units at a given date i.e. two stocks of capital are equal if they cost the same to produce at the same date.

It is because you deal in capital per head (period of production) instead of stock of capital that we have got at cross purposes. I think my method more flexible & generally useful. e.g. how would you deal with the Classical case i.e. employment fixed and the rate of interest variable? (In my case rate of interest is fixed & employment variable, but my method deals with Classical case just as easily)

(2) You define a neutral invention as one which would, with given rate of interest, leave capital per head unchanged. i.e. it is an invention which leaves m.p. of capital the same, & raises m.p. of labour only. This is the border line case between Pigou neutral & capital saving regions--it [b] lies within Hicks capital saving region. [2] It requires no change in capital per head to give equilibrium with a constant rate of interest & it is for this reason that you do not have to make use of elasticity of substitution in your argument. But surely you are wrong in saying it leaves relative share of capital unchanged? It leaves absolute share unchanged. Capital per head is same, r[ate] of i[nterest] is same \ income going to capital per unit of labour is the same. But product per unit of labour is raised (because technique has improved) \ real wages are raised & relative share of capital is reduced. Assuming all wages consumed (saving done by capitalists only) we get equilibrium with

1) same employment

2) same stock of capital

3) larger total income

This is an important case for mapping out effects of inventions, but I think it a pity if you insist on calling it "neutral" as Hicks has got it first (not to mention Pigou) & tagged the term to something else.

(3) I think you are right in your review about my wage-subsidy case [3] --but to go to the bottom of it would take me very deep, & as it is a bit of a rag anyway I thought I'd leave it like that.



  1. 1. R. F. Harrod, "Essays in the Theory of Employment" ( 1937:9 ; see note 1 to letter 666 for context). Robinson's notion of neutrality is discussed on pp. 328-29. Reproduced by kind permission of the Provost and Scholars, King's College, Cambridge.

    2. References are to A. C. Pigou, Economics of Welfare (1924), p. 632 (part IV, chap. IV, § 4), and J. R. Hicks, The Theory of Wages (1932), pp. 121-27. Robinson later illustrated this statement by means of a diagram: letter 671 , [jump to page] .

    3. Harrod, "Essays in the Theory of Employment" ( 1937:9 ), p. 327.

    1. a. ALS, four pages (the last two numbered) on three leaves (the first written on both sides), in HP IV-1089-1107. The front page is marked "?1937" on top.

      b. Ms: «--lies».

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page