On Keynesian Economics and the Economics of Keynes after fifty years

Abstract

Axel Leijonhufvud’s On Keynesian Economics and the Economics of Keynes (1968) definitely belongs to the category of “classic” books. Its message—decentralised economies are prone to large increases in unemployment since communication failures prevent the optimal coordination of private decisions—is by now well understood. In this paper, we argue that even though most commentators correctly identify this message, they overlook two crucial aspects of Leijonhufvud’s demonstration. These aspects relate to what the author calls the “aggregative structure” and the “transaction structure” of macro-models. We show that the former type of structure plays a central role in the emergence of unemployment in the “Economics of Keynes”, while the latter type of structure explains why unemployment is “involuntary”.

Keywords:

• History of macroeconomics
• Leijonhufvud
• Keynesian economics
• coordination failures

JEL CLASSIFICATION:

• B22
• E12

Acknowledgements

We would like to thank Michel De Vroey, Hans-Michael Trautwein and two anonymous referees for valuable comments and suggestions. The views expressed in this article are ours and do not necessarily reflect those of our respective institutions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 References to Leijonhufvud (1968) will be restricted in the following to page numbers, indicated after the quotations.

2 Neither the aggregative structure nor the transaction structure is considered in De Vroey (2016). Backhouse and Boianovsky (2013) make no reference to the transaction structure, and consider the aggregative structure only briefly in a single paragraph (57).

3 Leijonhufvud himself has recognised this weakness: “there is a lot in my 1968 book on aggregation which readers, I suppose, often fail to see the purpose of” (Leijonhufvud in Snowdon 2004, 123).

4 For instance, at the beginning of Chapter III, Leijonhufvud stresses that “one of the most important issues of modern monetary theory revolves around the ‘aggregation over transactors’ problem. A seldom examined, though highly questionable, convention of contemporary monetary theory is the assumption that ‘real’ aggregate demand is independent of the ‘real’ value of intra private-sector financial contracts. Redistribution of real net worth between creditors and debtors, due to price-level movements, is assumed to have no qualitatively predictable effects on aggregate demand, output, and employment. These variables are assumed to depend, apart from current income, only on the variables contained in the private sector’s consolidated balance sheet” (Leijonhufvud 1968, 116, italics in the original).

5 In Chapter III, Leijonhufvud also discusses the aggregative structure put forward by James Tobin in the model of his 1955 paper. There, Tobin proposed to lump together consumption goods and capital goods in a single commodity, and money and bonds in a single financial asset. Leijonhufvud makes clear that Tobin’s distinction between “physical” and “financial” assets is the relevant one to deal with the “risks of price level changes” (148), which are “the most significant in the context of his work.” On the other hand, he argues that the distinction between “short” and “1ong” assets involved in Keynes’ model is the relevant one to deal with a “change in ‘the’ rate of interest” (ibid, italics in the original), which is “the characteristic event which does loom large in his thinking about the capital accounts” (ibid.).

6 Leijonhufvud explicitly refers to Lawrence Klein and his Keynesian Revolution book.

7 So much so that “Keynes saw a need to argue that the interest-elasticity of investment will not be infinite” (167, italics in original).

8 Indeed, the analysis of the interest-elasticity of investment has led to the conclusion that “a decline in the interest rate, in Keynes’ model, does imply a very considerable increase in the demand price for capital goods” (166). Furthermore, since “adjustments in ‘securities’ markets are assumed to proceed faster than in other markets” (329), a fall in the interest rate generates an increase in the relative price of capital and consumption goods.

9 The system that Keynes assumed “is a system wherein the social function of production is eternal and the individual households, in comparison, ephemeral. ‘In the Long Run we are all dead,’ but production goes on and the capital stock is maintained and handed down from generation to generation.” (258)

10 “The present value of net worth increases in greater proportion than the present cost of the old consumption plan, and the consumption plan can thus be raised throughout.” (45)

11 Interest rates have an impact on consumption and saving only through the wealth effect because “Keynes regards the substitution effect of interest changes as ‘open to a good deal of doubt,’ as ‘secondary and relatively unimportant’—a phrase which in his works means, in effect, that the relation is ruled out of consideration.” (196)

12 Again, under the assumption (made by most “Keynesians”) that the substitution effect is “secondary and relatively unimportant.”

13 Leijonhufvud stresses that Keynes, in several passages of the General Theory, seemed to assume a purely “inside” money system. This ambiguity notably led Johnson (1965) to claim that Keynes’ theory “is guiltless of the charges brought against it by Pigou and elaborated by Patinkin and others if interpreted as applying to an inside-money world” (8). Nevertheless, Leijonhufvud concurs with Patinkin (1965) on the fact that “Keynes’ system does contain some ‘outside’ money” (1968, 325). He therefore recognises that Keynes’ theory could be prone to the criticism raised by Patinkin. The distinction between inside and outside money, due to the seminal contribution of Gurley and Shaw (1960), is very important in Leijonhufvud’s eyes. This relates to his strong interest (that we have pointed out previously) for the debt-deflation mechanism. In Chapter II, dealing with the “dynamic structure” of macro models, he especially notes that a “purely inside money system might ‘implode’ if the initial shock is heavy enough to set off a chain of defaults” (86). He then cites Irving Fisher’s 1933 paper and argues that “One of the most disturbing features of the consolidated balance-sheet models, which dominate contemporary monetary theory, is that they are incapable of generating the debt-deflation process sketched by Fisher” (ibid, n.10). However, the ‘no distribution effects’ assumption made for the whole book prevented him from pursuing the analysis further.

14 In the literature, the Keynes-effect usually refers to the increase in investment—and only in investment—stemming from a general fall in prices. However, Leijonhufvud defines this effect as “an effect on aggregate demand (i.e., on both consumption and investment) due […] to all-around deflation” (325, italics in the original).

15 Both models, however, involve inside and outside money.

16 Especially by Tobin (1958).

17 The concept of ‘elasticity of expectations’ was coined by Hicks (1939).

18 “The non-money store of value is augmentable. The higher its current price, the higher the flow-rate of net output.” (373).

19 “The faster the rate growth of the capital stock, therefore, the faster the rate at which the ‘real price’ per unit will decline.” (373–374).

20 More generally and as suggested by Hicks in Value and Capital—certainly one of the major sources of inspiration of Leijonhufvud—“it is uncertainty of the future, and the desire to keep one’s hands free to meet that uncertainty, which limit the extent of forward trading under capitalism” (Hicks 1939, 139). “Owing to the limitations of forward trading, [the model of a ‘Spot Economy’] is not really a very drastic simplification of reality,” whereas the model of a ‘Futures Economy’ “can have no claims to be a good approximation to reality, for it would only be in a world where uncertainty was absent and all expectations definite, that everything could be fixed up in advance” (Hicks 1939, 140). Accordingly, Hicks concentrates his analysis on Temporary Equilibrium (which does not extend to the coordination of intertemporal decisions) rather than on Equilibrium over Time, requiring that “the change in prices which occurs is that which was expected” (Hicks 1939, 132) and hence assuming perfect intertemporal coordination, through futures markets or else through perfect foresight.

21 One of us has suggested a formalisation of the economics of Keynes, at least partly along Leijonhufvud’s lines (Dos Santos Ferreira 2014).

22 Even some prominent names belonging to Keynes’ “Circus,” such as Joan Robinson (1937) or Richard Kahn (1976), have interpreted the concept of “involuntary unemployment” in this downgraded sense.

23 Barro and Grossman (1976) pointed out the influence of Leijonhufvud’s book on this approach: “earlier efforts by other authors, especially Patinkin (1965) and Clower (1965), along lines which seemed promising had apparently made little impact on the profession. However, the publication of the argumentative book by Leijonhufvud (1968) did succeed in stimulating interest in these issues.” (xi)

24 “The way in which I (and Robert Clower) couched the ‘microfoundations of macro’ problem was in some degree responsible for the attempts by Barro and Grossman, Benassy, Malinvaud, Frank Fisher and others to construct ‘Keynesian’ models on neo-walrasian optimizing foundations. This did not seem a promising way to go and I took no part in this development” (Leijonhufvud 1998, 175).

25 “Note that there are two types of theoretical structures that will not allow you to analyze the EDF (Effective Demand Failures) question: (1) The fix-price GE (General Equilibrium) models, commonly called ‘Keynesian disequilibrium’ models. (2) The New Classical models. The question ‘what ED’s (Effective Demand) govern price adjustments’ does not make sense either in models where prices do not adjust or in models where there are no meaningful excess demands. So the EDF question is an example of a question, central to Keynesian theory, that is excluded from analysis by current methodologies” (Leijonhufvud 1989, 18, parentheses added).

26 See Trautwein (2020) for a review of Leijonhufvud’s criticism towards New Keynesian Economics.

27 The various papers of Stephen Morris and Hyun Song Shin, starting in the mid-1990s (see e.g., Morris and Shin 1995, 1997), as well as Woodford (2002) and Sims (2003), initiated these developments. Angeletos and Lian (2016), in a chapter (entitled “Incomplete Information in Macroeconomics: Accommodating Frictions in Coordination”) of the second edition of the Handbook of Macroeconomics, provide a survey.

28 Note that references to Leijonhufvud were already scarce in the ‘coordination failures’ literature of the 1980s (exemplified by Diamond 1982 or Cooper and John 1988).

29 See, in particular, Leijonhufvud (1993, 1995, 2006). One of us has provided a more detailed account of Leijonhufvud’s evolutionary perspective (Clerc 2019).

30 See the conclusion of Trautwein (2020) for references.