Over the past four weeks, the GEM Blog has pounded mainstream NK theorists. They deserve it, in no small part because their market-centric dynamic stochastic general-equilibrium model class was useless in explaining the 2007-09 Great Recession. Even worse, there appears to be little consensus resolve to reconstruct NK theory to correct its flaws and become stabilization-relevant.
The failed market-centric DSGE model was not always the go-to NK theory, suggesting that a brief look at the evolution of NK thinking may be interesting. One of the first attempts to provide perspective on NK theory as a school of thought is Robert Gordon’s “What Is New-Keynesian Economics?” in the Journal of Economic Literature (1990). He understood that New Keynesianism was a generational shift away from then-textbook Early Keynesian theory. Stopgap or not, a new generation came to believe that it should not be acceptable to model macro market-disequilibrium by assuming meaningful wage rigidity (defined by its capacity to rationally suppress wage recontracting). MWR had long been the lynchpin of EK stabilization-relevant analysis, reversing neoclassical real-to-nominal causation and combining with adverse demand shocks to generate involuntary job loss and evidence-consistent contractions in employment, output, and income.
Original New Keynesianism
In the first phase of their campaign to supplant EK thinking, New Keynesians did not compromise their predecessors’ goal to explain the periodic absence of market clearing. Their key to EK replacement was instead to convince the academy that fitting critical evidence is less important than macro-micro coherence. Foundations in rational behavior organized by market-centric general equilibrium became the primary test for macro model adequacy. From Gordon: “A central theme of both new-classical and new-Keynesian macroeconomics is that accurate empirical predictions are necessary but not sufficient conditions of an acceptable theory. In addition, a theory must have microeconomic foundations in the behavior of utility-maximizing and profit-maximizing individual agents.”
Original NK theorists notably continued to use the EK business-cycle framework that features interacting demand shocks and nominal price rigidities. They were optimistic that they could microfound sufficiently powerful nominal product-price rigidities in rational market frictions. From the beginning, they were less confident about the existence of a super friction that could rationally suppress nominal wage recontracting.
The new generation had learned the lesson of their EK mentors as well as their own experience; microfounding MWR is really hard. Not wanting to bet the ranch on succeeding where the giants of the profession had failed, original NK theorists quickly floated the idea that the EK emphasis on labor pricing was a model-building error. Again from Gordon (p.1124): “… only price stickiness, not wage stickiness, is a necessary condition for business cycles in real output, given the particular path of nominal aggregate demand. There are no arithmetically necessary implications of nominal wage rigidity for the cyclical behavior of output or employment.” That convenient argument is wrong. Effective rebuttal begins with the basic fact of economic theory that, absent MWR, cyclical models are unable to produce involuntary job loss and, as a result, are restricted to producing recessions so mild as to be close to trivial.
The uncomplex theory here is surely known to trained economists. Rational employees must respond to wage reductions from their market opportunity costs by quitting, voluntarily moving to the alternative, now better-paying positions. Involuntary job loss plays no role. Moreover, if workers are somehow receiving wage rents (for example, as a result of an adverse shift in aggregate nominal demand), they must accept any pay cut, in lieu of job loss, that does not violate their opportunity costs. Forced job separation continues to play no role. The introduction of involuntary job loss into micro-coherent macro modeling requires the textbook labor-pricing story to be altered in two fundamental ways. First, for whatever reason, a significant share of employees rationally receive wage rents. Second, if macroeconomics is to be stabilization-relevant, firms’ capacity to offer labor-price reductions that reduce or eliminate those rents in lieu of job loss must be rationally suppressed, implying circumstances in which excess labor supply cannot induce wage cuts. The two-part wage rigidity is named “meaningful”. Forced job separation requires the existence of meaningful wage rigidity.
The early literature, much of it centered around Barro, makes clear that many New Keynesians understood that no rational market friction would be up to the task of suppressing wage recontracting. Fundamental problems with the super-friction’s existence include that recontracting occurs outside of the marketplace and produces immediate, large consequences which workers cannot rationally delay dealing with.
Modern New Keynesianism
The modern phase was ushered in with the famous New Neoclassical Synthesis that celebrated the NK adoption of RBC general-market-equilibrium mechanics and the NK acceptance of a diminished role of nominal demand. The market-centric DSGE model class becoming mainstream signaled the academy’s jump from emphasizing the useful explanation of business cycles to focusing on defending micro-coherent general market equilibrium as settled macroeconomics.
Making the difficult case for market-centric DSGE as settled theory is critically supported by the use of mainstream gatekeeping to stifle research and debate on microfounding MWR. That NK theorists don’t know how to rationally suppress labor-price recontracting provides the greatest challenge to the modern assertion of settled theory, motivating the apparent mainstream desire to push MWR research and debate off center stage. That push has had two important outcomes. First, it has been surprisingly effective. Illustrative is the 2016 Handbook of Macroeconomics having only two mentions of involuntary job loss, the rational existence of which requires MWR. (Neither mention attempts to integrate forced layoffs into NK modeling.) Searching for wage recontracting comes up empty, as do efficiency wages and wage rent.
The second outcome is rooted in troublesome facts that kept introducing doubt into the wisdom of banishing MWR. Embarrassing facts about the nature of unemployment eventually forced NK theorists to reinstate EK reliance on assuming effective downward labor-price inflexibility, especially in empirical work. The NNS became the mainstream roadmap for macro model-building at about the same time it was becoming clear that MWR, not menu or marginal-cost product pricing, is a necessary condition for stabilization-relevant macro theory. But NK theorists, after insisting on rigorous adherence to rational general-market-equilibrium, had no better success microfounding MWR than their EK predecessors. Many eventually accepted that there is no market super friction capable of rationally suppressing wage recontracting and that MWR must instead be assumed.
It is important that the reinstatement, and its fundamental violation of NNS rules of the game, has occurred without apology or notice. In particular, there has been no restoration of research on MWR microfoundations to its EK top-priority status. As a result, the defining difference between the original and modern schools is that today’s New Keynesians have chosen to divert attention away from that failure rather than emphasize correction of it. New Keynesians have chosen to accept the MWR assumption but reject the need to understand the MWR concept. That twisted state of affairs makes it much easier for the macro academy to insist that mainstream market-centric general-equilibrium is settled theory.
The Big Problem
Here’s the rub. The absence of microfounded MWR has deprived mainstream theorists of crucial model-building guidance in their attempts to construct stabilization-relevant macroeconomics. Such analysis became much more complex after the global industrial revolutions, tempting market-centric theorists into many wrong turns. An accurate roadmap is a must. In an example that particularly bugs me, the better map would have steered modern macroeconomists away from the hugely wasteful wrong turn that resulted in the widespread use of labor search/match theory to “explain” involuntary job loss. I marvel that so many NK theorists tolerated that embarrassment. In not reverting back to the honesty of their EK forebears, NK theorists have for some time now been severely damaged the prospects for stabilization-relevant macroeconomics.
Blog Type: New Keynesians Saint Joseph, Michigan