----------------- HES POSTING ----------------- 
 
[Summary, by BD: this discussion taking place on the PKT list originated from Steve Keen
posting there my "advert" of the Harrod Home Page which was posted here a few days ago.
Barkley Rosser commented that my claim on the cover page that Harrod "contributed
fundamental concepts to ... economic dynamics (the multiplier accelerator theory of the
trade cycle)" was too bold, as Kalecki came first. Participants in the discussion evoked
other "contestants" to priority, including Giblin 1930 -I pointed out that, on PKT list,
that although Giblin has some place in the history of the multiplier I haven't found a
trace of the multiplier-accelerator model in his piece. Below you'll find my recent answer
to Barkley Rosser's last message (which was posted here as well). One of my messages,
however, did not come here, and on an aspect (whether or not Kalecki in 1935 had a
multiplier) I would like to hear the opinion of HES-ers.]
 
I argued as follows: 
 
***** 
As the remark that Kalecki pre-dated Harrod: firstly, Kalecki's paper, although published
in 1935 (Econometrica, and a less mathematical version in French in the Revue D'Economie
Politique) was read in 1933 at the Econometrician's meeting, and was previously published
in Polish. I would, however, hesitate in calling it a multiplier-acceleration model. The
acceleration principle is surely there (Kalecki indeed got it from Aftalion), but I am not
so sure about the multiplier. A fundamental ingredient of the multiplier is definitely
present, namely, the inversion of the causal nexus between saving and investment (or, in
Kalecki's version, profit and expenditure). There is also a formula involving the
equivalent of 1/s. But this is obtained in quite a different way from the multiplier:
Kalecki says that profit depends on the capitalists' expenditure, and expenditure in part
depends on profits (there is a constant part + a variable part), and if the relationship
is linear we have a coefficient the denominator of which amounts to the saved part of
variable profit. Although the quantification of the effect of increased expenditure on
consumption looks very much like a multiplier, it only spreads increases in consumption,
not total expenditure, while for JMK the effect is global, as it includes (and indeed is
centred on) investment.]
 
***** 
 
Here's my reply to Barkley Rosser: 
 
>     2)  Although one reads frequently that R.F. Kahn 
>"discovered" or "invented" the multiplier (according 
>to Keynes, anyway), I do not know where he published 
>that result, if he even did. 
 
 "The Relation of Home Investment to Unemployment", Economic Journal XLI, June 1931, pp.
173-198.
 
On the Multiplier literature see for instance H. Hegeland, The Multiplier Theory, Lund
1954 (which ignores Giblin and Hawtrey). There's a chapter in Shackle's Years of High
Theory (and before that his article in EJ 1951); on Hawtrey see R. Dimand, Hawtrey & the
Multiplier, Hist. Pol. Economy, 1997.
 
>     3)  Although he did not work out a mathematical model, 
>J.M. Clark not only had the idea of the accelerator, but 
>understood that it was associated with a multiplier effect 
>as well prior to both Kalecki and Harrod.  The source 
>is J.M. Clark, "Capital Production and Consumer-taking: 
>A Further Word," Journal of Political Economy, Oct. 1932, 
>pp. 692-93.  In particular, although he sees an initiation 
>of a change in investment from a decrease in consumer 
>demand, he then notes that the decrease in investment, 
>"in turn reduces purchasing power, unless offset by opposite 
>movements elsewhere, and results in a positive decrease 
>in consumers' demand." 
>     If that is not the multiplier effect interacting with the 
>accelerator effect, I do not know what is. 
 
One should be cautious with such statements. Recognizing that a decrease in investment
affects purchasing power is NOT a theory of the multiplier: Kahn recollected (if I
remember rightly, in his Mattioli lecture -which I don't have here for reference) that
that was a well-known relationship (see for instance the discussions surrounding the
Liberal Yellow Book: everybody recognized that public expenditure gives rise to SOME
secondary employment, but people kept discussing on HOW MUCH of such employment was
generated); the multiplier theory consisted in QUANTIFYING such a relationship (and the
leakages associated with it: Mr. Meade's relation). If that's not a multiplier-accelerator
relationship (and even less a multiplier-accelerator theory of the cycle), it's a
(Wicksellian?) cumulative process -which almost everybody in the 1920s put at the basis of
their trade cycle theories (and which is, by the way, how Haberler considered Harrod's
theory -in a letter to him in 1937).
 
This comment of mine, however, is solely based on the above citation. If one looks at the
section on Clark in Hegeland's book cited above, one sees that in other of his writings
Clark clearly dealt with the multiplier in connection to the trade cycle: Hegeland refers
to Clark's Economics of Planning Public Works (Public Works administration, Washington
1935) and "Cumulative effects of changes in aggregate spending as illustrated in public
works" (Am. Ec. Rev,. 1935), none of which I have read.  So that case may still be open.
 
>     4)  The Kalecki model has certain mathematical problems. 
>It is a sort of multiplier-accelerator model, but not a full blown 
>one. 
 
It would seem (also in conjunction with 5) below) that you consider a "model" only if
formulated without mathematical problems. I would consider instead a multiplier-
accelerator trade cycle model a model (in the general sense of an intellectual
construction for explaining some phenomenon, not necessarily of a mathematical kind
-that's a restrictive view of what a model is: it isn't, for instance, what Keynes was
thinking of in his famous discussion with Harrod in July 1938) which associates distinct
and recognized "multipier" and "acceleration" effects. I doubt (and not for mathematical
reasons) that Kalecki had a fully developed multiplier (although there were some
ingredients of it in his 1935 Econometrica paper). (and I can't remember having read of
mathematical problems in Kalecki's model, in spite of having studied it fairly thoroughly
some time ago: it was subject of the limitations of linear models, and people like Frish
pointed out some implications of that immediately after publication, but that was common
of all mathematical models at the time)
 
>     5)  In his original paper in REStat in 1939, Samuelson 
>only mentions Alvin Hansen (who wrote in 1938) as an 
>influence on him.  He reiterates this in a piece in 1959 
>in REStat.  However, he discusses Harrod's role in his 
>less well known paper from later in that year, "A Synthesis 
>of the Principle of Acceleration and the Multiplier," Journal 
>of Political Economy, 1939, Dec. 1939, pp. 786-797.  He 
>provides the quote from Clark in there that I just noted.  He 
>gives a lot of credit to Harrod (and does not mention Kalecki), 
>but also declares that "Upon rigorous analysis his exposition 
>will be found to abound with non sequiturs and 
>oversimplifications. On the whole Harrod's intuition surpasses 
>his reasoned conclusions." 
>     In short, Harrod clearly played a very important role.  But,  
>he would appear to be neither the inventor of the idea of 
>combining a multiplier with an accelerator, credit for which 
>should probably go to J.M. Clark, nor of producing a fully 
>consistent and rigorous model that does so. 
 
I indeed referred to the JPE article. I am fully aware of the limitations of Harrod's book
(which, nonetheless, I consider a great book for what he attempted: far superior to all
the more famous books and articles Harrod wrote later); but nonetheless the multiplier-
accelerator model is there: the two principles are clearly expounded, individually first
and in their interrelation later. Harrod was fully aware of what he was doing: he
explicitly stated it in the 1936 book and in correspondence.  Which I still think we can't
say of the other "claimants" to the title of priority.
 
The point is: looking for forerunners is always a dangerous business, because one is
tempted to look for vague similitudes (which one can always find, somewhere) and attribute
them some properties which were not originally there but which we, with hindsight, can now
recognise as containing some of the ingredients we are looking for. Such similitudes
should be seen as a problem, not as a discovery: why did such author almost get there, and
could not do the final step? What presuppositions, historical conditions, theoretical
background etc. shaped the author's thought to bring him/her to that point and not beyond
it?
 
Daniele Besomi 
 
------------ FOOTER TO HES POSTING ------------ 
For information, send the message "info HES" to [log in to unmask]