814. Harrod to J. M. Keynes , 21 August 1938 [a]

[Replies to 811 , answered by 817 ]

Christ Church, Oxford [b]

21 August 1938

Dear Maynard

Many thanks indeed for your comments. They are most useful and I shall go through my text in the light of them with a view to getting greater clarity.

With regard to your main criticism [1] which is based on the distinction between capital goods and consumption goods what I feel is this. The distinction is clearly of great importance for the theory; but I have not made it. I feel that if I did make it I should have to embark on a rather elaborate development. You will probably agree that my article is already quite long enough for the Journal! There are other things I have neglected, e.g. the knotty question of gross investment vs. [c] net investment, i.e. replacements vs. depreciation allowances, and the influence of interest (though here I agree that a saving clause must be inserted, to explain that that subject is held in reserve [2] ). If I am to cover all this ground, I must run to a book. But I should very much like this preliminary article to appear. It might stimulate thought and be helpful to some people who are working out theories of lags but have no notion how to deal with the trend. And I would justify my claim, which I have made in a number of places, including my Presidential address, that one ought to be able to formulate a simple fundamental equation in which growth figures as an unknown variable. [3] Hence my desire for publication.

But of course I have to show that this temporary neglect of an important distinction does not involve me in a fallacy. And here I think your criticism is not altogether sound. Your statement that over-production of consumption goods does not make for expansion [4] is plausible, but not I think right in relation to the definitions and argument of the text. C, you understand, is for me simply production minus consumption. All output is lumped together and no distinction is made between fixed and liquid capital. Now to see whether my view is right let us suppose that there is no fixed capital. All production is production of consumable goods but some of these have to be held in warehouses, shops etc. as a reserve.

Now let us suppose the normal warranted position to be this:--2% growth, 4% of income saved which goes solely to increase liquid stocks by 2% (stock on this assumption = annual consumption ¥ 2). Now suppose that in period 2 (base period 1 = 100) 103 units of output are produced instead of the normal warranted 102. If all had gone on the warranted line saving would have been 102/25 = 4.08. Actually saving will be 103/25 = 4.12. But the saving required to increase liquid stocks in proportion to consumption is 6. \ liquid stocks will not be increased sufficiently to balance the increased turnover. (They will be increased by 4.12 instead of by 6.) \ there will be a force making for expansion. Thus though the overproduction of goods (overproduction in the sense of production in excess of the unique warranted level) consists entirely of production of consumption goods, none the less we have a situation making for expansion.

Now though I fully grant your point that if there is a lack of balance between capital and consumption goods, it may make all the difference which way this lack of balance goes, I hold that I am justified in omitting the treatment of this topic in the preliminary framework of thought here set forth. My argument is simplified to the highest possible degree. You may say that the whole thing will be altered when we consider a possible lack of balance and others may say the same with regard to time lags. I should still claim that I have given the most general possible framework within which these complications may be more easily considered. I also claim that the "acceleration principle" works in precisely the same way with regard to each category or capital, fixed and liquid, considered separately; \ we may lump them together always remembering that a lack of balance between them may set up forces not yet taken into account.

What I will do is to state much more explicitly that I am neglecting the capital vs. consumption goods issue. [5]

Some of your later remarks suggest doubt as to your understanding and my lucidity. [6] Of course if actual > warranted, entrepreneurs will be satisfied, they will be more than satisfied. If actual > warranted you do not get overproduction in the ordinary sense. That happens when warranted > actual. That is the fundamental paradox of the dynamic equilibrium.

When actual departs from warranted, the stimulated change of conduct does not bring about adjustment. On the contrary it leads to greater mal-adjustment. The adjustment can only come when the warranted rate is distorted to a new level (save for the one case when the actual rate hits the ceiling of full employment). The distortion of the warranted rate is related to the phenomena you have yourself discussed, profit inflation with its abnormally high saving and the squeezing out of saving in the slump etc. [7]

I feel that in this article I ought to aim at the minimum of theoretical development in the interest of the maximum of clarity.

Your paper [8] was a huge success. I will write shortly more about that. [9]



  1. 1. Letter 811 , [jump to page] .

    2. Letter 811 , [jump to page] . See notes 13 , 17 , and 38 to the draft of the "Essay in Dynamic Theory" (here reproduced as essay 19 ) for a description of Harrod's alterations.

    3. Harrod, "Scope and Method of Economics" ( 1938:15 ), pp. 403-4. Harrod had already advanced this claim in a letter to Keynes: see letter 799 , and in particular note 5 for further references.

    4. Letter 811 , [jump to page] .

    5. Having received further suggestions by Keynes (Letter 817 , [jump to page] ), Harrod inserted an additional paragraph at the end of section 5 of the "Essay in Dynamic Theory" ( 1939:7 , pp. 18-19): see note 18 to essay 19 .

    6. Letter 811 , [jump to page] : at first, Keynes interpreted the notion of "warranted" as analogous to the "justified" rate of advance in The Trade Cycle (Harrod 1936:8 ). He eventually understood Harrod's point: letter 817 , [jump to page] .

    7. See Keynes's, Treatise on Money for the notion of profit inflation in connection with the credit cycle (in Collected Writings, vol. V, pp. 249 and 264-65; vol. VI, pp. 144-45), and the General Theory (1936) for the proposition that the propensity to consume decreases as people become more affluent (in Collected Writings, vol. VII, pp. 96-98). Harrod had already referred to this "fundamental psychological law" in the course of the correspondence on The Trade Cycle (letter 652 , [jump to page] ). It should be noted that in the draft of the "Essay in Dynamic Theory" (but not in the final version) Harrod explicitly referred to the shift to (or away from) profits (essay 19 , [jump to page] and [jump to page] ) in this connection.

    8. J. M. Keynes, "The Policy of Government Storage of Foodstuffs and Raw Materials" (1938). The paper was read by Shove before Section F of the British Association.

    9. Harrod reported the comments on Keynes's paper in letter 812 which is, however, dated 20 August, that is, the day before the present letter. No other letters by Harrod regarding Keynes's paper have been found.

    1. a. ALS, four pages on two leaves, in JMK EJ/1/5/305-8. Printed in Keynes, Collected Writings, vol. XIV, pp. 328-29.

      b. Written on King's College letterhead.

      c. Ms (here and in the following occurrences): «v».

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page