Lucas’ expectational equilibrium, price rigidity, and descriptive realism

ABSTRACT

Robert Lucas' ([1972b] 1981a) article on the neutrality of money represented the first effective challenge to Samuelson’s neoclassical synthesis methodological separation between static microeconomic optimisation and macroeconomic dynamics. Lucas rejected disequilibrium price dynamics, as expressed by the Walrasian tâtonnement and auctioneer mechanisms. Lucas’ new treatment of equilibrium as an expectational concept, determined by the rational behaviour of information processing agents, was not restricted to market clearing competitive economies. Lucas’ effort to compare alternative rational expectations models of price stickiness (including his 1972 original formulation) led him to stress the notion of descriptive realism of the models’ main assumptions, which played an important role in his original discussion of model robustness.

KEYWORDS:

• Robert Lucas
• economic dynamics
• expectational equilibrium
• price rigidity
• descriptive realism

JEL codes:

• B22
• B41
• E32

Acknowledgements

I have benefitted from helpful comments by two anonymous referees, Peter Galbács, David Laidler, Beatrice Cherrier, Uskali Mäki, Roger Backhouse, Edward Nelson, Francesco Sergi, Lars Jonung, Joaquim Andrade, Charles Goodhart and Muriel Dal Pont Legrand. I would like to thank Guido Erreygers for bibliographical support. Funding from CNPq (Brazilian Research Council) is gratefully acknowledged.

Disclosure statement

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

Notes

1 Lucas (1983) published a correction to his 1972b article, acknowledging that, on the basis of correspondence with Jean-Michel Grandmont, the uniqueness of rational equilibrium was not established in that overlapping generations model, unless very strong assumptions were made. Multiple equilibria in rational expectations OLG models would become a main feature of sunspots business cycles formulated by Grandmont (1985) and others (see Cherrier & Saîdi, 2018).

2 Lucas’ and others’ usage of the term ‘neoclassical synthesis’ did not square exactly with Samuelson’s original meaning, which focused on the adoption of Keynesian full-employment policy and the ensuing validity of the orthodox theory of resource allocation.

3 Walrasian tâtonnement was also rejected at the time by a set of Keynesian authors (A. Leijonhufvud, R. Clower; R. Barro and H. Grossman, E. Malinvaud, J.-P. Benassy, among others) who established, together with D. Patinkin, ‘disequilibrium macroeconomics’ as an alternative research strategy. That was quite distinct from Lucas’ equilibrium macroeconomics (see Backhouse & Boianovsky, 2013). Lucas (1987:, pp. 51–52) referred critically (if imprecisely) to that literature as ‘fix-price’ models obtained by ‘dropping the assumption of market clearing while retaining all other aspects of the Walras auction scenario.’

4 As pointed out by Wade Hands (2012:, pp. 110–111), Samuelson’s neoclassical synthesis distinguished between the ‘behavior’ of the Walrasian agent and the ‘behavior’ of the Walrasian auctioneer, with ensuing disequilibrium dynamics. Lucas, on the other hand, modeled the equilibrium of the Walrasian market in the same way as the equilibrium of the Walrasian consumer, with permanent equilibrium of the price system and no tâtonnement adjustment. Hence, equilibrium in agents and markets meant essentially the same thing in Lucas’ economics.

5 This is somewhat reminiscent of the role of stability analysis in Samuelson’s ‘correspondence principle’.

6 Despite the textual evidence concerning Lucas’ misgivings about the Walrasian tâtonnement and auctioneer ‘automatic’ or ‘mechanical’ devices (see also the opening quotation at the outset of this paper), some commentators have argued that those concepts are central to Lucas’ framework (see e.g. Van Zijp, 1994, p. 68; De Vroey, 2016, pp. 157–58; 184; 352).

7 Lucas’ reliance on mathematical modeling did not prevent him from expressing admiration for aspects of other economists’ works which did not deploy that approach – e.g. Hayek’s business cycle analysis, Friedman and Schwartz’s historical narrative of the Great Depression, and particularly Friedman’s formulation of the natural rate of unemployment hypothesis.

8 Samuelson had been since the late 1960s impressed by Peter Medawar’s (1969) definition of science as the ‘art of the soluble’, which influenced his treatment of tractability issues in economics (see Boianovsky, 2022, section 2).

9 The Brazilian economist Mario H. Simonsen anticipated some main features of Clower’s cash-in-advance model in the early 1960s. See Boianovsky, 2002 and Walsh (2010:, p. 100).

10 As Lucas recollected, ‘the writing [of the 1978 paper with Sargent] was mine’ (Hoover & Young, 2011, p. 28).

11 As argued by Mäki and Marchionni (2012:, p. 195), theoretical progress in economics comes from the interplay between beliefs about matters of causal fact and the ‘available research technology.’ Hence, it is driven and constrained by both ‘ontological convictions and conventions about tractability.’

12 In a letter of March 20 1990 to Lars Svensson (held in the Lucas Papers, Duke University; quoted from Duarte, 2012, p. 226, n. 39), Lucas recalled how, by the early 1980s, he was still very ‘hostile to the idea of pre-set prices.’

13 De Vroey’s (2016:, p. 262) assertion that Lucas ‘came to endorse Kydland and Prescott’s’ real business cycle model is not accurate. In a letter of October 22 1990 to Prescott, Lucas stated: ‘I don’t agree with your remark that ‘persistence’ is a difficulty with monetary-shocks business cycle theories. Monetary shocks must work because people react to them as if they were taste/technology shocks, because they can’t tell the difference or because they are locked in to certain decisions. If so, then any theory of the persistence of the consequences of actual tastes/technology shocks (like yours and Finn’s) should be adaptable without change to monetary shocks’ (quoted from Duarte, 2012, p. 201; original in the Lucas Papers, Box 26, David M. Rubenstein Rare Book and Manuscripts Library, Duke University).

14 As suggested by Hoover (1988:, p. 227), Lucas’ endorsement of Knight’s distinction led to the view that, if the model was fully specified, uncertainty reflected a ‘residual’ which cannot be economically analyzed. However, such a residual may be economically relevant, as discussed below. Keynes too, of course, distinguished between uncertainty and risk; see Gerrard, 1994 on Lucas and Keynes on uncertainty. Interestingly enough, unlike Lucas, Friedman ([1962] 1976: 282) rejected what he perceived as Knight’s distinction and reaffirmed instead J. Savage’s notion of ‘personal probability.’ See also LeRoy and Singell’s (1987) criticism of the general interpretation, followed by Lucas, that Knight’s distinction was based on whether or not the situation is such that agents are able to form subjective probabilities. According to LeRoy and Singell, Lucas was closer to Keynes’ than to Knight’s distinction.

15 See also Laidler (2010). Likewise, the 2008 financial crisis was outside the scope of Lucasian macroeconomics, as he acknowledged while mentioning how those events ‘swept away’ some of the premises of his approach to business cycles (Lucas, 2013a, 2013b: p. xvii). Lucas eventually found Cass and Shell’s (1983) ‘sunspots’ approach (self-fulfilling prophecies and multiple equilibrium) helpful in explaining bank panics, after showing initial skepticism about that literature in the 1980s and 1990s (Lucas & Stokey, 2011; see Cherrier & Saîdi, 2018).

16 See also Lucas and Nicolini’s (2015, p. 49) claim about the ‘realism’ of their model of money demand and liquidity, in which currency and different deposit types are treated as alternative means of payments.

17 See Chakravartty (2007, section 7.2, especially pp. 187-92) on the consistency between realistic descriptive assumptions and the unrealistic models in which they are embedded. See also Mäki (2009, section 4) on how the mechanisms in operation in imaginary or unrealistic models may be similar to those operating in real situations. From that perspective, an unrealistic model captures ‘significant truth if it contains a mechanism that is also operative in real systems.’

18 See also Galbács’ (2020, chap. 4) argument, that Lucas’ point about the costs of gathering information indicated the ‘realism’ of his monetary misperception hypothesis and illustrated how Lucas’ abstract monetary models attempted to render the causal, ‘real’ behaviour of agents. Blanchard and Fischer (1989, p. 367) and other New Keynesian authors found relevant affinities between Lucas’ notion of small information costs and the menu costs approach to price setting.

19 Interestingly enough, Fischer (1987:, p. 648) claimed that the main implication of Lucas’ ([1972b] 1981a) supply function – that only unanticipated changes in the money stock had real effects – ‘is shared by sticky wage theories … and turns out not to distinguish’ Lucas’ macroeconomics ‘from other approaches.’ But that was Lucas’ point in his 1989 paper. Indeed, the aggregate supply function derived from contracts models is the same as Lucas’ supply function, even though the mechanisms are different (see Blanchard & Fischer, 1989, p. 519; Mankiw, 2010, p. 386).

20 Lucas (1996) must have been one of the very few Nobel lectures featuring passages of self-criticism. For instance, Lucas (1996:, p. 676; first italics in the original) asked: ‘Why is it that people cannot obtain that last bit of information that would enable them to diagnose price movements accurately? In reality, up-to-date information on the money supply does not seem at all that hard to come by.’ Moreover, though the evidence indicated that monetary surprises had real effects, ‘they do not seem to be transmitted through prices surprises’ as in Lucas ([1972b] 1981a) (Lucas, 1996, p. 679). Lucas’ frustration with the substantive results – as opposed to the methodological contributions – of his misperceptions model increased over the years. By 2011 he commented on how his 1972 article did not succeed in providing a positive theory about the observations called Phillips curve. ‘So we had what we call price stickiness, which seems to be central to the way the system works. I thought my [1972] model was going to explain price stickiness and it didn’t. So we’re still working on it; somebody’s working on it’ (Hoover & Young, 2011, p. 34).

Additional information

Notes on contributors

Mauro Boianovsky

Mauro Boianovsky is a historian of macroeconomics and its methodology. He has published a number of articles, chapters and books internationally. He holds a PhD in economics from Cambridge University, and served as president of the History of Economics Society.