817. J. M. Keynes to Harrod , 23 August 1938 [a]

[Replies to 812 and 814 , answered by 819 and 828 ]

Tilton, Firle, Lewes. #

23 August 1938

My dear Roy,


I agree that you can ignore the distinction between over-production in consumption and over-production in capital goods, if you make it clear that you are assuming that what you call technological considerations involve an absolutely rigid relationship between capital requirements and income irrespective of the division of income between the production of consumption and investment goods respectively, and irrespective of the rate of consumption. But no-one would grasp that you were making so rigid an assumption from your present brief reference to technological considerations. In any case, of course, the assumption of a [b] rigid relationship is artificial. If it really was rigid, output in excess of the warranted output would be physically impossible. I believe that the best way of making your simplification would not be to introduce so artificial an idea, but to say that you are not here considering the consequences of the lack of balance between capital and consumption goods and are dealing with changes in the volume of output which preserves the balance. [1]

This does not deal, however, with my point that your argument requires the unstated assumption that t is much larger than s where t is the proportion of capital to income and s is the proportion of saving to income. [2] In your numerical example you are assuming that t is 50 times as large as s. [3] And, of course, in actual circumstances it is safe to assume, with a wide margin, that t is always much larger than s. On the other hand, this assumption is definitely required. If s was greater than t, then all your results would be precisely reversed. Moreover, what matters is, not the average relation of capital to income, but the marginal relation, and there is probably not nearly so much margin on this, at any rate in the short period. An increase of 10 per cent in income is not immediately devoted to an increase of 10 per cent in the consumption of house room and other capital-using consumption. I quite agree, of course, that all this is outside your simplified scheme. But the more remote your scheme is from reality, the more important it is to explain its peculiar assumptions. The numerical relationship assumed between t and s is not indeed peculiar, but extremely important.

I agree that I had not precisely grasped your definition of "warranted". My point of definition is met by explaining that by warranted you mean exactly [c] warranted. [4]

I believe you would make a stronger and more intelligible thing of the article if you were to re-write it de novo. I believe that you could make it shorter and much clearer by tackling it over again and not just tinkering at it. If you make it absolutely clear at the outset what simplifications you are making, and then proceed rigidly on those simplifications, it would be both shorter and clearer. It will also then be easier to indicate in brief outline the sort of way in which some at least of the simplifications could be removed.


Thank you very much for your notes on the discussion about my paper. [5] Some of the points made were sound, and there are, of course, all sorts of technical difficulties. But the advantage of the scheme is that it need not be tackled wholesale all at once. Experience can be gained and the problem settled gradually. The immediate step which I should advocate would be arrangements with the Canadian Government for the storage in this country of the large purchases which that government has committed itself to make in the course of the current season. However, I wish I could have been there to hear it all and to reply.


I have not escaped having to consider Tinbergen's stuff. The League of Nations people have sent me copies of the latest version of his book and asked for comments. [6] I enclose a copy of my reply. [7] So far as I can understand the matter, my present belief is that it is almost pure hocus. But it is all so obscure that I cannot confidently assert this. However, the enclosed is quite long enough to convey to you my first impressions.

Yours ever,

J M Keynes

R. F. Harrod Esq., Christ Church, Oxford.

  1. 1. Keynes's suggestion was accepted: see 1939:7 , pp. 18-19, and note 18 to the draft of the "Essay in Dynamic Theory" (essay 19 ).

    2. Letter 811 , [jump to page] .

    3. Letter 814 , [jump to page] .

    4. Harrod eventually accepted Keynes's suggestion: see 1939:7 , p. 16, and note 8 to essay 19 .

    5. Letter 812 . The paper referred to is Keynes, "The Policy of Government Storage of Foodstuffs and Raw Materials" (1938), read by Shove before the British Association.

    6. Tinbergen, "A Statistical Test of Business Cycles Theories" (1938) (later published as Statistical Testing of Business Cycles Theories, 1939).

    7. Keynes's letter to Tinbergen of 20 September 1938 is reproduced in Keynes's Collected Writings, vol. XIV, pp. 293-95.

    1. a. TLS with autograph correction, four pages on four leaves, in HP II-86. CcI in JMK EJ/1/5/309-12. Printed in Keynes, Collected Writings, vol. XIV, pp. 330-32. Reproduced by kind permission of the Provost and Scholars, King's College, Cambridge.

      b. Ts: «of rigid».

      c. Emphasis in original only.

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page