837. Harrod to J. M. Keynes , 18 September 1938 [a]

[Replies to 833 , answered by 839 ]

Christ Church, Oxford #

18 September 1938

Dear Maynard

"The very reasonable assumption that consequences succeed their causes in point of time does not make a theory into a time-lag theory." [1]

Not necessarily! But it may. The introduction of a lag into an otherwise smoothly working system may set up an oscillation. Tinbergen reviews a number of theories of this sort in Econometrica 1935. Kalecki, Lundberg & others have been working on them. [2] I think it is really only doing systematically and with the help of a sine curve what Dennis does laboriously with his day by day analysis. [3] I take it that the more mathematical part enables one to detect the quantitative implications of one's own theory more easily and with an exactitude that it would be superhumanly laborious to get by the sort of methods Dennis uses.

The key to Tinbergen's book is I think to be found in the part where he has eliminated all his variables except profit and profit lagged. [4] I havent got his book with me but I remember that it comes towards the end.

I dont want it to be given an imprimatur as a final verdict--which he himself would not wish--but only as a study on valuable lines.

You deny that his book is in effect a study of the oscillatory consequences of assuming a lag. I think it is and T[inbergen] agreed when I put it to him.

You get an oscillation by assuming a certain dependence of investment on profit lagged--to reduce the matter to its simplest terms. Now investment does not depend entirely on this, but on other things also. Problem: to discover the comparative degree of its dependence on profit lagged. The object of all the multiple correlation is to get some sort of approximation to this degree of dependence. Now it is arguable and you may argue that he would have got just as good a result by inventing a figure out of his head. And I do not deny that it may be so. All the same it is up to the investigator to try and put a better figure and he would claim that he had made the attempt in the most thorough way he could with the data at his disposal.



  1. 1. Letter 833 , [jump to page] .

    2. J. Tinbergen, "Annual Survey: Suggestions on Quantitative Business Cycle Theory", Econometrica 3:3, July 1935, pp. 241-308; M. Kalecki, "A Macrodynamic Theory of Business Cycles", Econometrica 3:3, July 1935, pp. 327-44; Kalecki, "Essai d'une Théorie du Mouvement Cyclique des Affaires", Revue d'Economie Politique XLIX:2 (1935), pp. 285-305 (see letter 484 ); E. Lundberg, Studies in the Theory of Economic Expansion, London: King, 1937 (Harrod had reviewed Lundberg's book for the Zeitschrift für Nationalökonomie: 1937:12 ).

    3. See for instance Robertson, Banking Policy and the Price Level (1926), and "Saving and Hoarding" (1933).

    4. Refers to a preliminary version of Tinbergen, Statistical Testing of Business Cycle Theories, volume II (1939); in the final version, the passage is to be found in section 6.3, in particular pp. 140-47, where Tinbergen discusses the significance of difference equations for trade cycle theory.

    1. a. ALS, four pages on two leaves, in JMK CO/11/286-89. Photocopy in HP II-200. The letter is printed in Keynes, Collected Writings, volume XIV, pp. 304-05.

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page