Second contribution to the ENE critique: Reply to Davidson, part 1

ABSTRACT

On first encounter, the ergodic/nonergodic (ENE) approach has apparent plausibility. Although concerned by some of its problems for many years, it was only after more concentrated reflection on both its parts and their combinations that I became aware of its manifold deficiencies, some of which I outlined in my previous critique (O’Donnell, 2014). In this paper, facilitated by Davidson’s (2015b) rejoinder, these criticisms are deepened, broadened, and strengthened. Because the debate deals with fundamental matters in several disciplines, a considerable amount of investigation, unpacking, and logical dissection is required to clarify the argumentation beneath the compressed and seemingly smooth surface of the ENE position. For this reason, my reply is divided into two parts. This contribution primarily examines the central role of framing in ENE arguments, and clarifies the various misunderstandings and misrepresentations to which it leads. The subsequent contribution provides more detailed discussion of mathematical, stochastic, and methodological issues.

KEYWORDS:

• Ergodic/nonergodic distinction
• post Keynesian economics
• uncertainty

JEL CLASSIFICATIONS:

• B31
• B41
• D80
• E12

Notes

¹While this is the first opportunity PD has had to respond to my critique, it is far from his first opportunity to respond to my arguments about Keynes and uncertainty. See Davidson (1991, 1996a, and 2003–4), for example, as well as the JPKE mini-symposium suggestion.

²Although one or two statements in my initial critique took some ENE assertions on trust, further investigation reveals that revision is necessary.

³All unreferenced page numbers in this paper refer to PD’s rejoinder (Davidson, 2015b).

⁴The breadth of my proposition was also noted in my critique (O’Donnell, 2014, p. 203), although less directly.

⁵Every serious student of Keynes is aware of his 1937 QJE paper, and it is simply wrong to suggest ignorance on my part; see, for example, O’Donnell (1982, pp. 206–209; 1989, pp. 259–262; 1991, pp. 28–33; 2013, pp. 128–129).

⁶See also Davidson (1991, p. 70): “I have tried to focus and enlarge Keynes’s view of uncertainty in terms of more recent developments of the relevant probability theory popular with my fellow economists.”

⁷Obviously this process can continue forever, Keynes being successively forced into the frameworks of ever more “modern” theories.

⁸The substitution originates in the misquotation from Keynes’s QJE paper (p. 5). The misquotation is not an isolated instance, however, having previously occurred in Davidson (2011, p. 38).

⁹Keynes’s discussion of insurance in chapter 3 of the TP is actually much more complicated than the mere (Knightian) application of (numerical) probabilities.

¹⁰Similar necessities are imposed on other writers (pp. 6–7, 9).

¹¹Reframing exercises are exactly what orthodoxy has performed, from Hicks onward, in trying to capture the GT, the result being a series of caricatures. ENE, although highly sympathetic to Keynes, finishes up doing the same in a different guise.

¹²See also Davidson (2002, p. 41): “An axiom is defined as ‘a statement universally accepted as true … a statement that needs no proof because its truth is obvious’”; Davidson (2009, p. 6): “An axiom is defined as a universal truth that need not be proved”; and Davidson (1984, p. 561). To my knowledge, sources are never given for these definitions.

¹³Keynes’s explicit statements (in either answer) directly contradict PD’s claim that in 1936 “Keynes … could not precisely label the classical theory equivalents to the ‘axiom of parallels’ that had to be thrown over” (Davidson, 2000, p.16n10). This claim derives, not from the GT, but from the reconstructed ENE axiomatization.

¹⁴See chapters 3 and 18 of the GT.

¹⁵In monetary production economies, the labor market alone does not determine the real wage.

¹⁶Recall that Keynes was a classical economist for many years.

¹⁷Relaxing assumptions was a common practice of Keynes, one example being his treatment of money wages: initially assumed constant for expository purposes, then made flexible in pursuit of generality, the outcome being a theory embracing all possibilities without becoming entangled in ENE conceptual minefields.

¹⁸This is very clear in the discussion surrounding the page 4 quotation above. In the apt phrase used by Uffink (2006, pp. 63–64), this is “victory by assumption.”

¹⁹See also Part 2 of this reply.

²⁰See Davidson (2015a) for his most recent commentary, which contends that Samuelson’s “arrogance” in believing he had discovered the foundations of economic analysis was the primary factor in marginalizing post Keynesian theory. This is far too simplistic; many more powerful factors are involved, including possibly the intellectual strategies of some post Keynesians.

²¹See Samuelson (1968, pp. 1, 4–5). To my knowledge, Ehnts and Alvarez (2014) are the first to make this key point.

²²See Samuelson (1968, pp. 5, 12). Lucas (1981, p. 271) likewise accepts that New Classical analysis is “patently unreal,” again divorcing the analysis from universal truth. Despite regularly citing both these papers, PD nevertheless persists with the claim that “most mainstream economists today, following Samuelson and Lucas, accept as a universal truth the existence of a predetermined [ergodic] reality” (Davidson, 1996b, p. 479). See also Davidson’s (2000, p. 16n10) citation of Blanchard’s remark that long-run money neutrality is based more on faith than empirical science. If one is to use open acknowledgments of untruth by the orthodox in the critique of orthodoxy, one cannot simultaneously say that the orthodox base their theories on universal truths.

²³Rather surprisingly and confusingly, Davidson (1980b, pp. 154–155) classifies neoclassical synthesis Keynesians (including Samuelson) as borderline or partial post Keynesians.

²⁴Schefold (1980) clarifies why they are understandably missing from the 1973 edition, but regrettably missing from the 2007 and 2013 editions.

²⁵For PD’s request, see http://web.archive.org/web/20031018220417/http:// archives.econ.utah.edu/archives/pkt/1995m02/msg00177.htm.

²⁶This was later recast as the claim that Keynes’s general theory was built on the “smallest axiomatic foundation that could be applied to the real world” (Davidson, 2006, p.193n1).

²⁷See also Keynes’s two main departures from orthodoxy underlined in the QJE paper—uncertainty and its ramifications, and a theory of output as a whole.

²⁸The issue of restrictive assumptions returns briefly in Part 2 of my reply. Further difficulties in the ENE treatment of general theories are explored in another paper.

²⁹See also O’Donnell (1989, pp.175–176 and n32), including Keynes’s 1939 letter to Gardiner Means, which allows further assumptions to be added to general theories to develop applications to particular situations.

³⁰See also Part 2 of my reply. Anthropogenic climate change is another example. In science and elsewhere, Keynes’s TP provides a good guide in these matters.