## 516. Harrod to J. E. Meade

, 19 January 1936[a]

[Replies to 515 , answered by 517 ]

With regard to your last note:--

(i) I agree that in recovery there is a possible behaviour of the interest rate which would make this all right and I must put that in. [1] Paradoxically the interest rate is required to fall as the boom proceeds.

(ii) I have no doubt you are right about the relation of to the elasticity of AC and I thank you for your careful demonstration. [2] But the trouble is that I dont know anything about the elasticity of AC a priori, I have a strong feeling that my view is likely to be right for any probable behaviour of MC and I dont want to make my argument depend on a problematic property of AC which I cant make appear probable. I should prefer your proposition about the elasticity of AC to come as a corollary rather than as the base of my reasoning. I fear I cant see light at present. What do you think of the following?

Suppose that MC is rising at and after the point of intersection in constant geometrical ratio, p. Let output at the point of intersection be x and x+n at a subsequent point. Let AC = y at the point of intersection

since y = z (where z = MC), and successive marginal costs are , etc.

If AC was to increase in the same proportion as MC between n and n+1

but this is greater than the actual value of since 1 and .

Of course the whole thing depends on a geometrical ratio. I am worried about the whole thing. But I am in hopes of thinking through to something!

You appreciate that I am not challenging your solution: I myself feel that it does not give me the approach I want because the elasticity of AC is a mysterious quantity of which I feel I know nothing. Do you?

- 1. In the
final version of The Trade Cycle, the changes in the rate of
interest required to maintain a steady advance are extensively
discussed (Harrod
1936:8 , pp. 112-19). The original typescript not having been
found, it is however impossible to ascertain the extent of the
changes triggered by Meade's comment (letter
515 , [jump to
page] ).
2. Letter 515 , [jump to page] .