Harrod, Hansen, and Samuelson on the Multiplier-Acceleration Model: A Further Note

This business of “influence” is a very elusive one.

—Dennis Robertson to Roy Harrod, 14 April 1950

In a recent issue of HOPE Arnold Heertje and Peter Heemeijer (2002) argued that Paul Samuelson's multiplier-acceleration synthesis (1939a, 1939b) was inspired not so much by Alvin Hansen, as Samuelson claimed and repeated two decades later (Samuelson 1959), but by Roy Harrod's pioneering book The Trade Cycle (1936), which Heertje and Heemeijer see as the genuine historical origin of the fluctuation mechanism based on the interaction of the multiplier and the acceleration principle. The authors argue, however, that in spite of his verbal declaration of intents Harrod failed to grasp the full implication of the mechanism he devised and had to rely on a number of exogenous variables in order to explain the turning points. The authors summarize the chapters of Hansen's Full Recovery or Stagnation? (1938) devoted to Harrod's book and to the interrelation between the multiplier and the acceleration principle, and affirm that Hansen trivialized this mechanism as an explanation of business cycles. Subsequently they examine Samuelson's reading of Harrod and Hansen, emphasizing Samuelson's criticism of the “unnecessary complication” Harrod introduced with his dynamic determinants (the exogenous factors referred to above) and his claim that Hansen reviewed Harrod's book favorably. While they agree on the critical line, Heertje and Heemeijer do not accept Samuelson's view that Hansen had “guarded admiration” for Harrod's book, and claim instead not only that Hansen's distaste for the book is evident throughout, but go so far as to affirm that Hansen's review “effectively disproves that he understood the fundamental importance such a combination might have had in business-cycle theory” (216). The article concludes by suggesting that Samuelson underestimated Harrod's role, that Harrod's model was based on too many exogeneities and that his name cannot therefore be attached to the model, and that Samuelson was not influenced by Hansen's writings but by the discussions in the Harvard workshops directed by Hansen, so that Hansen's name should not be attached to the model. The multiplier-acceleration model can therefore rightly claim to be Samuelson's own.

Samuelson replies by telling the story of how he formalized Hansen's imperfectly formulated 1938 thesis, commenting that adding Harrod's name to the list does not take any credit away from Hansen and that Harrod's merit lies in having closed the Keynesian model, thereby formulating a truly endogenous cycle model (Samuelson 2002).

The story is richer, however, than described here, and actually runs along rather different lines. I agree with Heertje and Heemeijer that Harrod supplied the first clear-cut and explicit theory of the cycle based on the interaction of the multiplier and the acceleration principle.¹ But this statement needs some qualifications. First, Harrod's original formulation was first published a few months before The Trade Cycle, in an obscure Danish journal called Anglodania (Harrod 1936a); the corresponding paper was read on 10 January 1936 before the British Union in Copenhagen. There Harrod boldly claimed that it illustrated “how economists explain the connection between growing wealth and growing fluctuation in trade.”

Second, Harrod's priority in discovery has long been recognized in the literature, first and foremost by Samuelson himself (1939b), and is recorded in monographs (for instance Shackle 1967, 264–71; Pugno 1992, 81; Besomi 1999b, 74), in the sectorial literature (Nicholas Kaldor [1940, 78], for instance, described his 1940 trade-cycle model as being essentially similar to the explanations based on the combined operation of the multiplier and an investment function, such as Harrod's and Michal Kalecki's), in various general articles on Harrod's contributions (for instance Weintraub 1978, 124; Brown 1980, 10; and Johnson 1979, 273), and dictionary entries (such as Kurz 1997; Beaud and Dostaler 1995; and Besomi 1997, 1999a). These lists...