The vicissitudes in the interpretation of Ricardo s
The vicissitudes in the interpretation of Ricardo\'s new machinery chapter are well illustrated by the evolution of Paul Samuelson\'s assessment of Wicksell\'s account. Samuelson (1947) was one of the first to mention Wicksell\'s demonstration of the production optimum conditions, made as part of Wicksell\'s criticism of Ricardo, as discussed above. In his well-known article about induced technical progress and economic growth, Samuelson (1965, pp. 353–354) stated that “Wicksell gives the first modern discussion of technical change and distribution”. In particular, “Wicksell corrected what appears to be one of Ricardo\'s rare outright errors – namely the view that an invention might harm all of a competitive society.” Moreover, “in resurrecting Ricardo\'s discussion, Wicksell called attention to the passages in which Ricardo is all but recognizing” that labour\'s share is affected by a shift in the marginal productivity of labour. In two articles written in the 1980s, however, Samuelson (1988, 1989) changed his mind about Ricardo\'s machinery question. He argued that, although Wicksell was right in insisting that perfect kv1.3 inhibitor is Pareto optimal, he was mistaken to hold that against Ricardo. In Samuelson\'s new reading, the introduction of machinery leads to a reduction of output because of Ricardo\'s assumption that population adjusts to long-run wages determined at subsistence level. Labour-saving innovation can lower the market-clearing real wage at the old labour/land densities, which causes population and the labour force to drop. The long-run outcome is a reduction in output, as claimed by Ricardo, which is, however, consistent with Pareto optimality since labour supply adjusts. Samuelson (1989, p. 47) acknowledged that “strictly speaking, we cannot find in Ricardo\'s words what would pass today for an entirely satisfactory proof of his contention”, but claimed nevertheless that Ricardo\'s “basic intuition is on the mark” (there is no mention of population change or subsistence wage in Ricardo\'s chapter 31). Samuelson (p. 52) concluded that “Ricardo is right. Wicksell (and Kaldor and…) are wrong”. In 2003, while working on the paper that would be published as Boianovsky and Hagemann (2005), I corresponded with Samuelson about his criticism of Wicksell\'s criticism of Ricardo (the letters are in my possession; copies are also held in the Paul A. Samuelson Papers, David M. Rubenstein Rare Books & Manuscript Library, Duke University). In correspondence of 22 January 2003, Samuelson wrote that “Ricardo is right to say ‘an invention can lower Q’. And Knut would be wrong to doubt this and do so on the grounds that ‘Pareto-optimality à la perfect competition precludes lower Q from the same old land and the now-lower quantity of labour input.’ Anyone can make a mistake, but why did Wicksell make the same mistake from 1890 to 1924?” In our correspondence, I did not ask him about his change of opinion between mid-1960s and late 1980s, but observed that Wicksell did not overlook the role of population change in his criticism of Ricardo, since he took into account the possibility of wages falling down or even below subsistence level. In a letter dated 30 April 2003, Samuelson retorted: “I never contended that. Gratuitously, Wicksell interpreted Ricardo as believing that a viable invention could somehow lower what today we call ‘total factor productivity’. So too with Kaldor, Schumpeter, Stigler, et al. In my readings Ricardo\'s imperfect exposition did not assert that.” Hence, after his 1965 endorsement of Wicksell\'s criticism, Samuelson (1988, 1989) sought elements of internal consistency in Ricardo\'s new chapter 31. As it is well known, that chapter expressed Ricardo\'s recantation of his previous opinion on the machinery question. Just like Samuelson, Wicksell too admired Ricardo\'s analytical powers, but he was less concerned in finding internal coherence in Ricardo\'s approach to machinery. Having at first interpreted Ricardo as a supporter of the so-called “theory of compensation”, Wicksell eventually realized that the central message of the controversial chapter 31 was the negative impact of machinery on output and employment. Instead of trying to fit that chapter in Ricardo\'s long-run theory of distribution – as Samuelson would do later – Wicksell criticized it from the perspective of neoclassical marginal productivity theory with given factor endowments. It was in that context that Wicksell, under the impact of Pareto\'s not fully developed claims about optimality of free competition, reformulated the machinery question to bring out the distinction between maximization of aggregate output and of total utility in welfare economics.