790. Harrod to J. M. Keyne s , 6 July 1938 [a]

[Replies to 787 , answered by 791 ]

Christ Church, Oxford #

6 July 1938

Dear Maynard

I am very glad you found the address interesting. [1] With regard to the length, I had in mind that the audience will have printed copies in their hands. Consequently one can and indeed ought to read it at a much quicker pace.

I am not sure that I agree altogether with your hostility to the idea of economics as a natural science; tho' no doubt it has its own special modes of procedure.

With regard to Schultz, I intended to indicate that his particular work is probably more or less waste of time. [2] But this is only to the extent that one admits that the theory of value as set out by Marshall, Walras, Cassel etc. is played out. This still occupies the central position in any textbook [b] of economics. If on the other hand these equations of static theory are of importance, surely they should be given provisional quantitative significance, even if this had to be re-done from time to time. Just as, surely, one wants to know the value of the Multiplier, even if one also knows that it is subject to both cyclical and secular change. At present I should prefer to endow someone [c] to research into the value of the multiplier, than into demand schedules of commodities, because it seems to me that the former type of theory will take us further. Your letter seems to suggest that we should be content with the concept of the Multiplier and not bother about the quantitative filling out.

While, therefore, I agree more or less about Schultz, I feel that Tinbergen may be doing very valuable work, in trying to reduce this part of theory to quantitative terms. e.g. in the model the rate of interest and the marginal efficiency of capital are joint determinants of investment. But if he can show over a wide range of countries and time that the influence of changes in the rate of interest is very small compared with the influence of changes in the m[arginal]. e[fficiency]. of capital--which is precisely what he does think he has shown [3] --this, without impugning the validity of the model, may affect our judgment with regard to various matters. Of course the most important function remains that of getting a suitable model.

The matter has a personal interest as well as a purely academic one for me. I have now got my "dynamic" theory, I think, into a much better form than I had it in my book. [4] I hope to send you a short article quite soon--it is what I mentioned in a former letter. [5] But I begin to feel that the time has come when I ought to soil my fingers by doing some of this sort of statistical work myself or supervise others in the doing of it. [6] I make a statement about the "acceleration principle". Then Tinbergen comes and says that the facts do not suggest that it had the influence I ascribe to it. [7] Surely one ought not to leave the matter there. To Tinbergen the statistics merely suggest a negative result; to me, if I applied his technique, they might suggest a refinement of my concepts or a re-stressing of the importance of one at the expense [d] of the other.

I confess all this frightens me a little. I imagine the possibility of a vast amount of time-wasting. But I have a feeling that one ought to take the risk. The trouble with the statisticians I feel is that they are too divorced from theory to derive the full benefit of their own experiments with figures.

We have a sort of minor Tinbergen here in the form of Marschak. It is awfully funny to watch Hubert Henderson's reactions to his statistical findings. He thinks the whole thing obvious nonsense. Me on the other hand he tells that my theoretical work is entirely divorced from the facts. [8] He himself is content with what you once called wise-cracks, very good ones I think. But he happens to be a shrewd person. If there is to be a developing subject with a lot of workers, competent but not outstandingly inspired, who want to find systematic work to do, more or less prescribed by the state of the subject--as in other sciences--I should have thought a mixture of Tinbergen and pure theory was the right answer. Otherwise the ordinary competent researcher finds nothing to do but to write a history of the milk marketing board, or to indulge in the mathematical but rather fruitless refinements of the green publication of the L.S.E. [9] How do you react to all this?



I wonder what the chances are of your appearing in person in Cambridge? Good, I hope!

  1. 1. Harrod, "Scope and Method of Economics" ( 1938:15 ), presidential address to the British Association for the Advancement of Science, Section F.

    2. "Scope and Method of Economics" (Harrod 1938:15 ), p. 401. A reference to Schultz in this connection can also be found in "Mr. Keynes and Traditional Theory" (Harrod 1937:4 ), p. 74.

    3. See for instance J. Tinbergen, An Econometric Approach to Business Cycle Problems, Paris: Hermann, 1937, p. 25.

    4. Harrod, The Trade Cycle ( 1936:8 ). The discovery of the new formulation, to become the "Essay in Dynamic Theory" (Harrod 1939:7 ) was also announced to Marschak and to Robertson: see note 3 to letter 788 .

    5. Letter 778 , [jump to page] .

    6. See letter 788 , [jump to page] .

    7. J. Tinbergen, "Statistical Evidence on the Acceleration Principle", Economica V, May 1938, pp. 164-76.

    8. See e.g. letters 524 , [jump to page] , and 526 , [jump to page] ; for a similar complaint see letter 563 , [jump to page] .

    9. The Review of Economic Studies.

    1. a. ALS, eight pages on four leaves, in JMK CO/11/269-76; photocopy in HP II-200. The letter is printed in Keynes, Collected Writings, vol. XIV, pp. 297-99.

      b. Ms: «text book»

      c. Ms: «some one».

      d. Ms: «expence».

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page