The GEM Project’s case for the criticality of a second (workplace) venue of price-mediated exchange boils down to a single question. Is rationality indispensable in useful macro model-building? Affirmative arguments cite clarity and persuasive power as well as the capacity to systematically interpret evidence and to distinguish among competing theories. Those virtues, however, have proved easy to ignore. A much more persuasive argument for the criticality of rational behavior is rooted in the guidance it provides in the construction of evidence-consistent macro models. Experience has shown that research roadmaps relying on irrationality are trouble waiting to happen. In this essay and the one that follows, the analytic necessity of rational behavior is illustrated by the checkered history of meaningful wage rigidity (MWR), as it morphed between rational and irrational foundations, in the development of stabilization-relevant macro theory.
During the early postwar decades, mainstream Early Keynesians (EK) suppressed wage recontracting, and derived involuntary unemployment, by imposing convenient assumptions that slow labor-price adjustment to labor-market disequilibrium. By their nature, the assumptions were irrational and at least somewhat misleading. That state of affairs greatly concerned EK theorists, causing them to label their modeling as provisional. More robust analysis must await the discovery of MWR foundations rooted in optimization and general equilibrium – two pillars of mainstream neoclassical economic theory.
In that context, Lucas’ watershed rational-expectations theory was a game-changer, rigorously rooting their New Neoclassical Synthesis research in rational behavior; harsh penalties were imposed for the continued practice of EK macroeconomics relying on irrational assumptions. New Keynesians had run out of patience with provisional irrationality and insisted that acceptable modeling be rooted in optimization and general market equilibrium even before the super market friction capable of rationally suppressing wage recontracting is identified.
It is not surprising, as it became increasingly clear that the hoped-for super market friction does not exist, many macroeconomists chafed at being confined to modeling that is badly inconsistent with the most important instability facts. (The huge effort assigned to the obviously silly attempt to make do with voluntary unemployment yielded little but embarrassment.) The desire to be policy-relevant led a large number of NK theorists to quietly reintroduce the EK class of irrational assumptions. For decades, the most popular MWR assumption has been Calvo’s (1983) rotating fixed fraction (ϋ) of wages that remain unchanged in each period. Other wages freely adjust to market conditions. The probability that a labor price in a given period is flexible is 1-ϋ. Calvo model-building guidance is unabashedly empty. Some unknown portion of the economy’s labor pricing is fixed for an unspecified time period. How firms are restricted is ignored. We have no idea why they differ from firms that are free to optimize. We have no idea about whose behavior – employers, employees, both – is truncated. How long periodic unemployment persists is overlooked. How is its incidence in the firm’s workforce determined? We know ϋ firms are paying wage rents during downturns but don’t know why profit-seeking does not mandate wage cuts. If employees effectively resist such cuts, we don’t know how.
Calvo produces involuntary job loss in the context of adverse disturbances in nominal demand. The price of that outcome is theorists being stuck with irrational, analytically empty labor pricing that provides no guidance to useful model-building. NK bad behavior here should not be brushed aside. Calvo’s mechanism does not functionally differ from the Early Keynesian free parameters that motivated Samuelson’s Neoclassical Synthesis. The only significant NK innovation is the harsh penalties imposed on early Keynesians’ for coddling irrationality, ill effects of which persevere. The hypocrisy of simply “forgetting” that the equivalent EK suspension of rationality led to that school’s extraordinary banishment from modern graduate-school curriculums and cutting-edge journals is breathtaking.
Aside from unacknowledged NK malice, the difference between the EK and NK regimes is patience. The former emphasized that their Neoclassical Synthesis was unfinished macro theory, putting microfounding MWR at the top of their research priorities.They believed that success on their top priority would provide rational guidance for the further construction of policy-relevant macroeconomics.
NK theorists, by contrast, increasingly argue that their New Neoclassical Synthesis and its attendant friction-augmented general-market-equilibrium (FGME) modeling is fundamentally finished theory. No core MWR research remains to be done; none would be useful. The presumption of settled theory has effectively erased wage-rigidity microfoundations from the macro academy’s research agenda. That erasure has been devastating to mainstream stabilization-policy relevance. The nonexistence of rational involuntary unemployment, which requires adequately microfounded MWR, apparently has not encouraged NK theorists engaged in empirical work to discontinue their use of misleading irrationality. A deeply troubling model-building question follows from that mess. What other powerful features of rational-behavior MWR are being suppressed by mainstream NK theorists’ satisfaction with their badly misleading labor-pricing assumption?
The second part of this post provides a practical demonstration of the criticality of rational behavior in guiding the construction of evidence-consistent macro models. It is an impressive exercise that effectively indicts modern NK reliance on irrational wage determination as a root cause of onerous model-building mistakes (most egregious, the continued truly unforgivable dominance of search/match market mechanisms in the explanation of mass unemployment) that have robbed consensus thinking of practical relevance.
Write a Comment