111. D. H. Robertson to Harrod , 26 May  [a]
[Follows on from 110 , replies to 109 , answered by 112 ]
Trinity [College, Cambridge]
26 May 
Herewith some comments on your criticisms. They are rather scrappy, but I hope you may find them intelligible, even if not convincing. You seem to be the only person who has understood the book without being lectured to viva voce on it!  and I beg again to thank you for devoting so much trouble to it. I am v[ery]. interested, and pleased, to hear that in your lectures you've been thinking along the same lines.
Now for another matter. We want--or may want--to increase our staff here, and we have been conducting a flirtation with Robbins of the London School. Did you see anything of him during his temporary appointment at Oxford?  and if so, could you give us any confidential estimate of his utility as an economist, his power as a lecturer and teacher, and/or the general likelihood of his fitting in here? We haven't much to go on at present in the way of information,--naturally he hasn't been able to publish much so far. If he is a really good man we should like to secure him,--it would be good for us to have some fresh blood: but is he?
I hate being asked such questions, but if you could help us at all in forming a judgment, I should be very grateful.
1. Variations of output.
I think that perhaps you have a legitimate grievance against my terminology, but not against my argument!
Let us define as follows:--
"Capitalistic" variations in output are those which enlightened self-interest of capitalists tends to bring about. "Ideal" variations in output are those which enlightened self-interest of "group members" would tend to bring about. "Good" variations in output are those which in all the circumstances we ought to acquiesce in and promote. [b]
Then all I am saying is that, "ideal" variations being ruled out owing to the impossibility of here & now establishing a guild system, "capitalistic" variations must be regarded as "good". They are, I suggest probably "better" than those variations which the immediate self-interest of workmen, if they could dictate the scale of output, would bring about, because there is a danger that workmen would neglect their ultimate self-interest. In any case, I am assuming for purposes of this book that you have got to have "capitalism", & therefore a regulation of output in accordance with the imagined self-interest of employers: the best we can hope for is that their imagined self-interest shall be in accord with their real self-interest.
Of course it is open to you to say that there is no need to have "capitalism", and that some sort of socialism could and ought to be quickly established. But obviously that opens up an enormous field of controversy which I think I'm entitled, in a book on another subject, to decline to discuss.
But you may perhaps say that I oughtn't to use so strong a word as "justifiable" to describe a thing which is only "good" in this limited sense, and on the disputable assumption that a speedy and radical change in the constitution of society is impossible. I think, however, that I have guarded myself pretty carefully against a misunderstanding of the meaning of "justifiable" in the last sentence of p. 22.
2. Effects of various policies, pp. 25-26.
The cardinal point is that under no system can the new equilibrium rate of output be established unless either (i) more workmen are drawn into employment or (ii) existing workmen work longer hours, or both. (The employer cannot with his own hands make the 20% extra output). 1
Under policy (d), since there is no inducement to the workman in the form of a lowered cost of living, this result necessarily involves an increase in the total of money wage-bills. I.e. even if standard piece-rates, or standard time-rates for a normal day's work, remain unchanged, there will be more workmen getting these rates, and/or there will be overtime payments to be made.
But further, it appears to me that there will be a stronger tendency for money wage-rates to rise under policy (d) than there will be for them to fall under (a), (b) or (c). For under (d) the impulse towards a rise comes from the competition of employers bidding against one another for the services of workpeople, while under (a) (b) and (c) the impulse to a fall comes from a lowering on the part of the workman, owing to a fall in the cost of living, of the money supply-price of his labour,--a force, surely, more slow and doubtful in its operation.
Thus as regards wage-rates I don't admit symmetry between (a) (b) & (c) on the one hand, and (d) on the other: and if you look again at the second sentence on p. 27 you will see that I clearly affirm (whether rightly or wrongly) that under (d) there will be a rise in wage-rates.
But I admit that in strict accuracy I ought to have mentioned again (as I do on p. 21) the familiar point about the temporary transference from workmen to employers owing to a lag in the rise of money wage rates. What this really means is that [c] , if we stick to my definition of "appropriate" or "justifiable" increases of output as those which the enlightened self-interest of the capitalist will bring about, the "appropriate" increase under policy (d) will, so long as the lag persists, be greater than under the other policies, & we cannot properly use the same figure (in my example, 20%) to illustrate it. May I not be forgiven for not having re-introduced this further complication at this point?
If you will let me neglect it for the moment, let me give a numerical example of what I had in mind. Let us assume that the employer in equilibrium keeps of the proceeds, and hands over in wages. Let us assume further for simplicity that the whole of the new output is produced by the absorption of new workers, and none of it by overtime work on the part of old workers. Then we get:--
* In £
Haven't I given reason enough on p. 27 for supposing that the (d) equilibrium is attained with less friction than the (a) equilibrium?
3. Ch. V, App. I.
I accept your statement of the assumption on which my argument is based, though I don't at present follow your algebraic demonstration of what would happen on certain other assumptions. I will look at it again; but it may require a conversation some time to clear it up (I had to lecture for 2 hours to Pigou on my algebra before he understood it at all!) 
But I concede freely that my K - 1 hypothesis is purely arbitrary, and was adopted really because, for the first period of K days though not for subsequent ones, it gives results consonant with the Pigovian argument quoted in p. 61 note.  In his review in the forthcoming Journal, Pigou has proved to his satisfaction (though not entirely to mine!) that the total amount of Lacking done on my hypothesis is exactly midway between that which would be done taking the longest plausible hypothesis about the "period of restoration" and that which would be done taking the shortest hypothesis about the "period of restoration". 
This, if true, is pure luck for me!
2. L. C. Robbins was a temporary fellow at New College in Oxford in 1924-25.
3. A. C. Pigou, in his review article "A Contribution to the Theory of Credit", Economic Journal XXXVI, June 1926, pp. 215-27, acknowledged "personal help from Mr. Robertson", without which he doubted he could succeed in fully understanding the hard thinking compressed into the small space of Banking Policy and the Price Level (p. 215).
4. Reference is to Pigou, Economics of Welfare (1920), p. 672.
5. Pigou, "A Contribution to the Theory of Credit", 1926; the reaction speed in discussed on pp. 221-23. See also Pigou, Industrial Fluctuations (1927), pp. 144-45.
- a. ALS, one page, with three notes, each on one leaf, the first two of which written on both sides, in HP IV-990-1069d/5.
b. In the Ms, the repeated parts of these sentences are dittoed.
c. Ms: «means that».
top of page
Return to index of this section
Go to previous page
Go to next page