With this post GEM Blog the nears the end, at least for now, of its two-month rant on the centrality of Ptolemaic modeling in modern macro research. To reiterate, Ptolemaic research is defined by its primary objective being the defense of the mainstream model class. The analogy used to help understand the modern debasement of stabilization-relevant macro theory is the geocentric theory of the solar system. In the second century A.D., Claudius Ptolemy wrote about the motion of the planets, putting the earth at the center of an astrophysical system that dominated astronomy for centuries. Being fundamentally wrong caused the original Ptolemaic system to take on its now characteristic Rube-Goldberg (defined as “deviously complex and impractical”) look, featuring tiny circles in which planets would dance as they transversed their larger orbits around the earth. The goofy epicycles were critical to the defense of the consensus model, providing free parameters that allowed the geocentric model to be recalibrated to fit incoming evidence.
What follows illustrates Ptolemaic woes in the macro academy with three American Journal of Economics: Macroeconomic articles that attempt to explain the crucial issue of cyclical joblessness in the context of mainstream market-centric general-equilibrium theory. As usual, names are named. In chronological order, they are Blanchard and Galí, “Labor Markets and Monetary Policy: A New Keynesian Model with Unemployment,” AEJ:M (2010); Bils, Chang, and Kim, “Worker Heterogeneity and Endogenous Separations in a Matching Model of Unemployment Fluctuations,” AEJ:M (2011); and Elsby and Michaels, “Marginal Jobs, Heterogeneous Firms, and Unemployment Flows,” AEJ:M (2013).
The capacity of this post to critique three longish analyses is rooted in their mutual reliance on search-match-bargain (S/M/B) modeling, which has become the canonical framework used to introduce equilibrium unemployment into macro modeling and was the generally critiqued in last week’s blog. Each of the three articles is conveniently subject to the same set of debilitating problems that badly damage its capacity to explain macro instability. Indeed, accepting the search framework as the go-to model of cyclical labor behavior is one of two fundamental mistakes that typically condemns mainstream macro theory to stabilization-irrelevance.
The central defect in S/M/B modeling is its inherent market centricity. It turns out that nonintuitively restricting optimizing price-mediated exchange to the marketplace prevents the rational existence of meaningful wage rigidity (MWR), defined by its capacity to suppress wage recontracting. Understanding the role of MWR, Lucas and Barro famously acknowledged that involuntary job loss (IJL) cannot exist in the consensus macro model class. In a classic Ptolemaic response, they advised macroeconomists to get along without it. Modern search theory does just that. In it, job separation is wholly comprised of voluntary quits; unemployment wholly results from voluntary job search and matching.
Each AEJ:M paper incorporates an innovation into familiar S/M/B mechanics. The enhancements, unfortunately, recall the Rube-Goldberg epicycles used in the original Ptolemaic scholars’ defense of indefensible mainstream thinking. B&G (2010) introduce a bit of indeterminant labor pricing by assuming idiosyncratic needs by both firms and jobseekers, complicating the matching function. However, for routinized jobs, characterized by high ratios of specialized to general human capital and for which probationary periods provide superior information to applicant interviews on new-hire productivity, B&G wage indeterminacy cannot much matter. Moreover, given that their individually bargained labor price does not suppress recontracting, the market-indeterminant wage is downward flexible, implying another channel producing the nonexistence of involuntary job loss (IJL) – a heavy blow to B&G stabilization relevance from which their model cannot recover. Finally, if large, specialized firms carefully construct internal wage structures in order to manage unit labor costs, and they do, entry-level wages must be part of an existing structure. That helps explain the actual circumstances in which entry-wages are posted, not negotiated. B&G simply ignore that devastating fact, soldiering on with their Ptolemaic construction a Rube-Goldberg mechanism that defends the mainstream model class. They ignore that practitioners and government-collected economic evidence both reject their model.
The two other papers tell similar Ptolemaic stories. BCK (2011) promise interesting endogenous job separations but only deliver job quits that are stirred up by arbitrary heterogeneous wage shocks complicated by wealth effects. Endogenous job loss wholly comprised of voluntary quits cannot much interest stabilization authorities. E&M (2013) promise stabilization relevancy rooted in “very intuitive” labor pricing. Instead, perhaps indicating some difficulty with the meaning of “intuitive”, they assert a process, surely ruinous if some hapless firm actually tried it, “in which a firm negotiates with each of its workers in turn, and where the breakdown of a negotiation with any individual worker leads to the renegotiation of wages with all other workers”. Endless costly negotiation cannot inform rational wage-setting arrangements, helping to explain why the E&M “intuition” does not pass practitioners’ laugh test. Let’s be clear. I am not objecting to constructing Rube-Goldberg models; theorists must have latitude, especially on long-standing, difficult problems. But I do object to passing such models off as an intuitive, useful descriptions of actual labor pricing. They don’t come close.
Taken together, the AEJ:M innovations pretty much trash what is broadly known about hiring, firing, and compensating routinized employees in highly specialized firms. They are little more than modern versions of Rube-Goldberg epicycles that divert attention from the damage to business-cycle modeling caused by the absence of involuntary job loss in mainstream market-centric thinking. We know that more than three-quarters of the increase in unemployment during the Great Recession resulted from forced layoffs. We know that much of the huge welfare loss of the Great Recession resulted from that job loss. Yet we blithely persist in modeling the Great Recession absent involuntary job loss.
Since the early Keynesians were banished from macro-research debate and dissemination decades ago, the major lesson learned is that consensus micro-coherent, market-centric general-equilibrium theory cannot accommodate nominal wage rigidity capable of rationally suppressing wage recontracting. Preservation of the mainstream model class requires getting along without MWR and IJL. In their place, we have modern epicycles associated with the rise to keystone status of the S/M/B model class and the huge theorist investment in getting voluntary job quits to dance, even if ever so slightly and unconvincingly, to a counter-cyclical tune.
I admit that dancing on that particular pin head does not much interest me. But shouldn’t everybody object to passing off research that cannot accommodate the most important recession evidence as policy-relevant. How does what is happening today, with mainstream theorists pushing stabilization policy-makers to accept their S/M/B modeling as useful descriptions of actual macro instability, differ from high-stakes fraud?
Blog Type: New Keynesians Saint Joseph, Michigan