459. Harrod to J. M. Keynes , 1 August 1935 [a]

[Follows on from 458 , encloses 460 ]

as from Christ Church, Oxford

1 August 1935

Dear Maynard

I now come to a point which I think is of considerable importance and I should like you to give it your earnest attention. I have been reading Bks III and IV as far as ch 16, and I think your positive doctrine excellent and important and a great contribution, to which I can subscribe.

But in your critical part I think you have fallen into what I can only characterize as a confusion, and while there are only one or two passages which, if I am right, need to be corrected, I feel it has made you quite un-necessarily critical of Marshall and others. The error, as I conceive it, stands out most clearly in the last sentence of the 2nd para. on galley 64, viz. "; whilst the notion that the rate of interest ... whatever the rate of interest." [1] The view that I object to lies in the argument that because saving must always and necessarily equal net investment (which I accept) there is "no sense" in the view that interest is a price which equates the demand for saving in the shape of [b] investment to the supply which results from the community's propensity to save. Now I come on to your positive doctrine presently. What I am concerned to contend now is that the discovery that saving necessarily equals net investment has no relevance to the sensibleness or otherwise of the doctrine referred to above. (Of course in that doctrine the demand for saving must be conceived not as an absolute amount of investment but as a schedule by which the amount of investment is related to various rates of interest.) This doctrine makes perfectly good sense, but is open to the charge of being incorrect. I find no sense in saying that this doctrine makes no sense because in this case supply is always and necessarily equal to demand.

Take the case of supply & demand as applied to commodities. It is true that in certain cases there may be disequilibrium between supply and demand in the sense of production exceeding or falling short of consumption, the balance being absorbed in or released from stocks. Price movements may be thought of as bringing production into equality with consumption. But there are commodities in which production must always and necessarily be equal to consumption. And no one has thought of saying that the laws of supply and demand make no sense because of that. Take the number of private German lessons given in London . Here production is always and necessarily equal to consumption. That cannot be doubted. Yet there is also in this case no reason to doubt that the price of a German lesson is determined by the supply & demand functions in the ordinary way. 1 It seems to me that the discovery that saving = investment has deceived you into thinking you could derive from that an argument in favour of your general attack on the classical explanation of the rate of interest which is specious and indeed invalid.

In order to give you pause for thought, I should like to add that this was the most criticized part of your address in Oxford, [2] in which you brought out this argument. Frankly it convinced no one.

It seems to me you give the case away so far as that argument is concerned, when you say, as I think you do somewhere--I fear I havent noted the place, but perhaps I shall find it again-- [c] that this classical doctrine would be all right if there was some other mechanism always ensuring [d] a full employment position. Yet even in those circumstances it would still be true that saving necessarily equals investment. So it isnt that truth which invalidates the classical position. That seems to be right. If there was and it was known that there would continue to be full employment, the classical theory would come into its own again modified by something about liquidity preference.

By bringing in this as it seems to me quite invalid argument you will distract attention from the valid arguments you adduce for your position. I have referred to the passage in para 2 of ch. 14. The second para on galley 70 [3] wants re-modelling. And I dare say there are other passages: I will small-tooth-comb it for them once I know that you agree to the principle.

The notion that price is determined by supply and demand always rests on a cet. par. assumption: e.g. that there are no price changes of other things or that they are irrelevant. What you seem to me to have shown is that there are changes in other things which are so relevant and of such overpowering importance, that the old s[upply]. & d[emand]. analysis had better be put away. You have incidentally shown also that we know very little about the supply schedule of saving, i.e. amount of saving considered as a function of the rate of interest cet. par. which includes level of income being given. You have further shown that the level of income has an over-riding importance in determining the amount of saving (= amount of investment). You have further shown that the level of income is linked in a roundabout way (and \ the level of saving in a still more roundabout way) to the rate of interest. And this in such a way that the level of income and rate of interest are indeterminate unless you bring in another equation, which you do in fact, viz. the liquidity preference schedule. To say that in the special circumstances of this particular field of exchange, the ordinary supply & demand functions are no use, because the quantity is functionally linked to another variable outside the factors usually considered in supply & demand analysis (viz. total income) is very different from saying (and this is what I object to) that the supply & demand functions can in principle give no solution because in this case supply necessarily equals demand.

Note. What I mean by the ordinary supply and demand functions is this. Supply function relates amount people choose to save (= amount invested) to rate of interest. Demand function relates quantity of new investment undertaken (= amount saved) to rate of interest. In both cases cet. par. Ceteris includes level of [e] income. Now you are perfectly justified in saying that the amount of saving is so clearly related to level of income that to cover the level of income by the cet. par. clause is to refuse to examine the problem. But when you do examine this functional relation all sorts of funny things (including the very important one that variations in the propensity to save may be offset by variations in the level of income in such wise that they have no effect on the rate of interest) appear. But you are not justified in denying to classical theory a logical and water-tight view, albeit one which neglected the most important features in the situation. And in your best passages where you relate the classical doctrine to the economics of full employment you dont deny this.

I wont write more until I hear how you react to this. But I would like to write a good deal more if I hear that you dont react, because I feel that the matter is very important.



I transcribe some more notes on minor points.

Galley 43. Ch. 10. para 1. [4] Would you, please, insert (i) (ii) and (iii) after the 3 words "partly", as these 3 governing factors turn out to have an importance in the subsequent text which the reader doesnt realize they will when reading--and so he has to refer back and make it out for himself. 2

Galley 50. para 3. 2nd sentence. [5] "The greatest ... efficiency of capital." This is extraordinarily obscure where it stands and is, I suggest, un-necessary, as the point is covered lower down. 3

Galley 51. II. para 2. [6] I have noted:--"If the product is homogeneous I dont see where the difficulty lies." 4

Galley 52 top. [7] I have noted:--"But you yourself say that the rate of interest = current marginal efficiency of capital." 5

Galley 52. 2nd complete para. [8] Ref. to Marshall in first line. I have noted:--"In this traditional theory agrees. Where it differs from you is in finding the other determinant required in your (ii) and (iii) of ch. 10." 6

Galley 54. fragment of para. at top. [9] I have noted:--"But isnt this if foreseen reflected in the current rate of interest, and, if not foreseen, impotent?" 7

Galley 54. IV. 2nd para end. [10] I have noted:--"But there is a risk here even if a favourable change is equally likely." 8

Galley 61. [11] Does fortnightly settlement--but you know all about this!--really work this way? I had heard that the daily settlement of N.Y. had universalized & regularized trading on margins which our fortnightly settlement makes it possible for many speculators to avoid, so that our system really made for less speculation. 9

Galley 64. line 9 [12] something wrong. 10

Para. 2, early part, [13] I have noted:--"But of course modern classical theory combines these."

Later part. I have expanded my difference of opinion in enclosed letter. 11   [14]


As I am going away on Sunday, but destination is not absolutely certain, would you, if you write after Friday morning, post to Christ Church, which is most reliable forwarder. I will keep you informed of my movements.

  1. 1. Keynes, Collected Writings (1971-89), vol. XIV, pp. 470-71; General Theory (1936), p. 165.

    2. On 21 February 1935, Keynes spoke to the Oxford economists on "the marginal efficiency of capital and the rate of interest" (Keynes to Meade, 10 January 1935, in JMK PS/6/13; Keynes's engagement diary 1935, in JMK PP/41).

    3. Collected Writings, vol. XIV, p. 475; General Theory (1936), p. 177, second full paragraph.

    4. Collected Writings, vol. XIV, p. 444 (the paragraph referred to is actually the second); Harrod's suggestion was incorporated in the General Theory (1936), pp. 90-91.

    5. Collected Writings, vol. XIV, p. 459; General Theory (1936), pp. 135-36.

    6. General Theory, p. 138, first full paragraph; the passage remained unaltered.

    7. Collected Writings, vol. XIV, p. 461; General Theory (1936), top of p. 139.

    8. General Theory (1936), p. 140, first full paragraph.

    9. General Theory (1936), p. 143, last sentence of the full paragraph.

    10. Collected Writings, vol. XIV, p. 464 middle; Harrod's comment was accounted for by adding a sentence at the end of the first paragraph of p. 145 of the General Theory.

    11. General Theory, p. 160 first paragraph.

    12. Collected Writings, vol. XIV, p. 470, passage corresponding to General Theory (1936), p. 165 line 6 (the mistake was accidentally corrected in the variorum: the passage in the second proof actually concluded "... in such a way as to tend to the two towards equality."

    13. Collected Writings, vol. XIV, pp. 470-71. Harrod later resumed this criticism in letter 462 , [jump to page] .

    14. Here reproduced as letter 460 .

    1. a. ALS, one page, with ANI with pencilled annotations in Keynes's hand, five pages on five leaves numbered from the second, in JMK GTE/1/290-295. These documents are printed in Keynes, Collected Writings, vol. XIII, pp. 530-33.

      b. Ms: «shape investment».

      c. To facilitate legibility, dashes substitute here a pair of commas.

      d. Ms: «insuring».

      e. Ms: «of of».

1. "Price equals propensity to give German lessons with m[arginal]. e[fficiency]. of German Lessons.

But from propensity to save and schedule of m[arginal]. e[fficiency]. the rate of interest cannot be deduced." [Keynes's pencilled note in the margin and at the bottom of page].

"Suppose that whenever the price went down, the demand went down by the same amount as the supply" [Keynes's note at the top of the page].

2. Keynes ticked this sentence in the margin.

3. Keynes ticked this sentence in the margin.

4. "all products?" [Keynes's note in the margin].

5. "How relevant?" [Keynes's not in the margin].

6. "Yes" [Keynes's note in the margin].

7. "I don't think so" [Keynes's note in the margin].

8. Keynes ticked this sentence in the margin.

9. Keynes ticked this paragraph in the margin.

10. Keynes ticked this sentence in the margin.

11. Keynes ticked the last two paragraphs in the margin.

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page