702. Harrod to D. H. Robertson , 8 October 1937 [a]

[Replies to a letter not found; follows on from 701 , answered by 717 ]


Sorry, this transcription cannot be displayed as the copyright holder did not grant permission to proceed with the electronic edition.

This document, however, can be read in The Collected Interwar Papers and Correspondence of Roy Harrod, edited by D. Besomi, Cheltenham: Elgar, 2003.


Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page


Ch[rist]. Ch[urch, Oxford].

8 October 1937

My dear Dennis

I left a postcard for you on a book case when I left London, and I have a feeling it wont get posted. What I said was this:--it is no doubt easy to talk to the young about time preference and roundaboutness; that is the danger and temptation. My point is that it is wrong. It is like the old statement of Ricardo's that rent does not enter into the cost of production. [1]

Of course the young men wont spot that it is wrong. Why, we ourselves are only just spotting it now.

I think I demonstrated in my last that according to the orthodox view, the rate of interest must always be zero. [2] You did not reply to this.

The fallacy can be approached in a number of ways. I try this one. The orthodox (I dont know whether this is the best word to describe it) theory is wrong because it is one dimension out. When we consider the supply of other factors, we say what price is necessary to elicit a flow of so and so much per unit of time. £y will elicit x man hours of unskilled labour, £z will elicit w acres of land etc. [b] The amount elicited is not a lump sum but a flow which is conceived to go on so long as the conditions are unchanged. To meet this we have a demand schedule, depending on tastes and technique of production. So and so much will be demanded at various prices. Again the amount demanded is a flow. x man hours per week will continue to be employed at this price, week after week, so long as conditions remain the same.

Now turn to saving. I will not bother for the moment with Maynard's difficulty about supply. The crucial trouble is on the side of demand. With tastes and technique the same there is no demand in equilibrium. There may be of course some delayed projects still to be done; but this is a once over demand, which will not go on week after week; when it is satisfied, there is nothing further.

You may now be tempted to say--well, then the rate of interest will fall to stimulate demand sufficiently to absorb the supply. Quite so; but again this stimulus will only--tastes and technique being the same--provoke a once over demand. When that is exhausted the rate will have to fall again. In fact we come to this. Given tastes and technique and the supply schedule, there is no sense in the question what rate of interest will allow demand to absorb the supply. We have to ask what rate of fall in the rate of interest will allow demand to absorb the supply. But this is not at all on all fours with our treatment of the other factors. We do not say what rise or fall in wages or rents is required to get labour and land used (assuming constant supply schedule) but what level of wages or rents. This is what I mean by saying that the orthodox theory of interest is one dimension out.

Put it the other way round. We may ask at what rate must factors other than saving increase or alternatively at what rate must capital using inventions proceed in order to absorb the current offer of savings at a given rate of interest. But we do not ask at what rate must the factors other than labour increase or alternatively what labour using inventions must occur in order to continue to absorb a given quantity of labour at a given price.

Put it another way. You may perceive that to bring interest theory into line with the others, one must consider the whole quantum of capital including all past savings, however this may be measured. You have your demand for the use of so and so much capital as a whole per week which is on all fours with your demand for so much labour per week. There will be a demand schedule corresponding to various rates of interest. This gives you a particular rate related to each quantum of capital. It does not give you a particular rate related to a given amount of addition to the capital. Which comes back to the same point. Given your tastes, technique, time-preference and all the rest of it, what comes out of the hat isnt a rate of interest at all, but a rate of fall in the rate of interest.

Why not, you may say? But I say that you cannot admit the concept of a rate of fall in the rate of interest in your Pt I and not talk about any other rates of increase and decrease. You have now to consider the economy as a whole, with the tendency to increase or decrease in all its parts. In fact you are in the dynamic system.

Have I at all succeeded in making plain why I conceive the orthodox theory to be definitely fallacious? It seems to me to contain one piece of definitely wrong thinking, an intellectual mistake, which now that it is exposed must be admitted. And if you ask how it is that so many minds have been guilty of it, I reply precisely because each mind received it with the authority of many minds and did not pause to examine it.

Do you think I ought to write a short note for E.J. on this point? But perhaps you still believe that there is nothing in it!



P.S. The orthodox theory determines the rate of interest and the volume of saving at one blast (x and y on intersection of supply & demand curves). In my book there were things called dynamic determinants which were supposed to determine not indeed the rate of interest but the volume of saving. [3] What about these things? They were such things as rate at which inventions make production more roundabout, proportion of income saved etc. (N.B. prop[ortion]. of income saved is not to be confused with amount of saving in ordinary analysis: my thing has different dimensions: constant prop[ortion] growing amounts) [c] . Now this analysis by dynamic determinants is not something which can be added onto the Pt I theory; it washes it out entirely and replaces it. Do you in fact think that my dynamic determinants are all nonsense? Otherwise you have to sacrifice the orthodox account. I may say that in spite [d] of a number of fingerposts in my book pointing at these dynamic determinants, no review that I have seen has paid any attention to them. [4] I have no doubt it is my fault for being too cryptic.

1. D. Ricardo, On the Principles of Political Economy and Taxation, in The Works and Correspondence of David Ricardo, Cambridge: Cambridge University Press, 1951, vol. I, chapter 3. Harrod had pointed out that this statement of Ricardo was fallacious in International Economics ( 1933:10 ), p. 27n, and already criticized it in his lectures on Ricardo (in HP V-103, section on "Theory of Value") with reference to Marshall's letter to Edgeworth in Pigou, Memorials of Alfred Marshall, pp. 436 and 68-69.

2. Letter 701 , [jump to page] .

3. Harrod, The Trade Cycle ( 1936:8 ), pp. 88-101.

4. In fairness, Harrod should have acknowledged Hansen's accurate description of the dynamic determinants, also in their relation to the static determinants: see A. H. Hansen, "Harrod on the Trade Cycle", Quarterly Journal of Economics LI, May 1937, in particular pp. 517-21. Hansen's review was brought to Harrod's attention by Keynes in June 1937: see letter 674 .

Hugh Gaitskell also accurately stressed the role of the dynamic determinants in Harrod's theory; his review, however, was published after this letter was written ("The Trade Cycle. By R. F. Harrod", Economica NS IV:16, November 1937, pp. 472-76; see in particular p. 475).

a. ALS, four pages on four leaves, numbered from the second, in HP IV-990-1069d/53.

b. Ms: punctuation missing.

c. Ms: there is no closing bracket.

d. Ms: «inspite».