368. Harrod to G. Haberler , 8 September 1934 [a]

[Replies to 367 ; the exchange continues at 378 ]

Christ Church, Oxford [b]

8 September 1934

Dear Haberler

I am sorry that I have given you all this trouble. I honestly thought that what I wrote [1] was simple; if it does not seem so to you, it is probably because we are approaching the problem from different points of view.

What I think Hayek does is this. He implicitly assumes that we have a stationary society and considers the effect of injecting new credit after a period of stationariness [c] . Whereas I am considering a regularly advancing society. It is the difference between the dynamics of starting a train that has been stationary and the dynamics of the train when in motion at uniform velocity. The two problems are really quite different. I am concerned with the second. [2] It is possible that Hayek's contention about the shift in the ratio of the demand for producers' and consumers' goods [3] is true when you are regarding the transition from a stationary to an advancing condition--I dont pronounce on that as I have not considered it fully enough--but I am clear that it is not true in the case of the regularly advancing society.

It is because, I think, you are considering the transition from stationariness to progress that you say that "the unit period means obviously the period before money has been injected, but after (or in) which the volume of production has risen." [4] This cannot be so. As I am not considering the transition from stationariness to advance but a cross section in a period of advance, the injection of new money and the increase in the volume of production take place in each and every unit period. The two processes are simultaneous.

You ask why I assume that the consumers wish to "hoard" money in the unit period. [5] (I should not use the word hoard of course). They wish to add to their stock of money. This is implied by the assumption of stable velocity. 1 This is the simplest assumption. 2 Of course it is not necessary. And I myself point out that if velocity rises sufficiently, increased credits are not necessary for the maintenance of stable prices.

You suggest that I assume my conclusion in my premise, when you ask whether I mean that consumers spend (1-q)x less than they ought (why this "ought"?) in order to keep prices stable. This is not the right way to put it. The banks create new money if velocity is being kept constant. (If V is rising, the banks create less or do not create or even withdraw). They create sufficient new money to keep prices stable (I assume). Now suppose they assume that V is constant and suppose this assumption to be correct, it follows that consumers in the unit period desire to add (1-q)x to their monetary holding. 3 If they do not so desire, then V is not stable and the banks must inject less than x new money. We must make our assumptions consistent. One assumption is that the banks add such amount of money as taking V to be what it is keeps prices stable. Next assumption is that the bank adds x units. 4 This implies stable V. If we now assume that V is not stable we contradict ourselves. If V is stable consumers will wish to add to their monetary holdings in the way I describe. If they do not wish to do this prices will tend upwards, which since money and goods have increased in proportion implies that V is rising.

Mark you, I do not say absolutely that anyone wishes to add to his monetary holdings; the banks are assumed to adapt their policy to what people do wish.

I am going on talking a little more because I am terribly afraid I havent made myself clear.

You might object that there is some ambiguity in this idea of wishing to hold more money. If the banks create new credits, one lot of people or other must hold more money de facto, since there is more money going about. What is the special virtue in my assumption of wishing to hold more money?

If x new units are introduced, x new units must be held. When I say that people wish to hold them, I merely mean they behave in such a way as not to raise V. If they do not wish to hold them, they must all the same, but in resisting, so to speak, the holding of new money, they raise V.

Now, you may say, is not this exactly what they will do? But if they do this, i.e. not wish to hold the new money, i.e. behave in such a way as to raise V, the banks will not inject so much new money. They will only inject so much as to keep prices stable.

Oh but, you may say, if the banks do not inject any new money, there will be no new money for people to resist holding, and V will not raise. Quite so. Then the banks will inject some new money, but only that amount which, having regard to the tendency of people to resist holding it and the consequent raise in V, keep prices stable.

Actually in the cross section of a regularly advancing society, V will probably not be rising at all--unless for technical reasons connected with banking--i.e. people will wish to add to their monetary holdings in my sense.

Some other question you raise:--You write "`income receivers disburse in the purchase of securities s - (1 - q)x in the unit period.' Why do you assume that if they hoard they spend less on securities but as much as before on consumers' goods?" Again your language is couched in terms of the transition from a stationary to an advancing state. s is defined as the difference between income and expenditure in each unit period. It follows that s - (1 - q)x is the amount spent in the purchase of securities.

You refer to the case "when q is zero, i.e. all members of the business community are overdrawn." There is no specific point about this. It is possible that the business community holds no money. In that case the whole of x goes to the consumers.

All this argument of Pt II is very negative. I only want to urge that stable prices and increasing income are quite consistent with the natural rate being kept equal with the market rate. Moreover I believe that in his contra case made out in the Monetary Theory, Hayek has forgotten one book-keeping entry. [6] He just did not pause to consider that a certain quantity of saving is held off the capital market by reason of its being absorbed in the holding of new money.

I read Prices & Production as soon as it came out before reading the other book and I at once made a marginal note that this fallacy seemed implicit in it. [7]

Of course I may be wrong! But I am not yet convinced that I am. I apologize for being so long.

Yours very sincerely

R. F. Harrod

  1. 1. Harrod, "The Expansion of Credit in an Advancing Community" ( 1934:8 ).

    2. Harrod later used the same analogy in his rejoinder to Robertson's "Mr. Harrod and the Expansion of Credit" (1934): Harrod ( 1934:11 ), p. 478.

    3. Hayek, Prices and Production (1931), in particular pp. 51-55.

    4. Misquote from letter 367 , [jump to page] .

    5. Letter 367 , [jump to page] and note 4 .

    6. Hayek, Monetary Theory and the Trade Cycle (1933), in particular pp. 216-26.

    7. Next to the first paragraph on p. 24 of his copy of Prices and Production, Harrod noted: "From this foolish fallacy most of the subsequent errors seem to follow". This is the only annotation in the whole volume (in HCN Roybooks).

    1. a. ALS, six pages on three leaves, in GH Box 66.

      b. On paper headed Birr Castle, King's County (crossed out by Harrod).

      c. Ms: «stationary-ness» (also in the following occurrences).


1. Taken with the banking policy of stable prices [Harrod's note at the foot of page].

2. Haberler marked the last three sentences with two vertical lines in the margin.

3. Haberler marked Harrod's conclusion with a vertical line and a question mark in the margin.

4. Where x/X = rate of increase of income [Harrod's note at the foot of page].


Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page