The American Economic Journal: Macroeconomics was launched in 2009 during the most perilous U.S. macroeconomic crisis since the 1930s depression. The timing emphasizes that AEJ:M editors’ competence in helping guide the development of coherent, useful macro theory matters. Unfortunately, at least so far, they have not been up to the task. None of the AEJ:M articles published since its inception has significantly advanced the stabilization-relevance of macro theory. Most do not try.
The journal’s 2014 best-paper winner, “Marginal Jobs, Heterogeneous Firms, and Unemployment Flows” (Winter, 2013) is illustrative. Its authors, Michael Elsby and Ryan Michaels (E-M), promise a coherent, stabilization-relevant model constructed on firm-size heterogeneities that features “very intuitive” labor pricing. Such claims stir hope that, upon reading the article, is dashed. E-M turns out to be just another AEJ:M effort that suffers badly from market-centricity, producing results that are Ptolemaic rather than useful. The most critical evidence is characteristically, and conveniently, ignored. The “intuitive” description of labor pricing does not pass practitioners’ laugh test. The authors (p.8) 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”. I am not objecting to constructing that model. I object to passing it off as an intuitive description of labor pricing. It isn’t close.
The theory. E-M seek to enrich the search/match/bargain (S/M/B) model class. They introduce “a notion of firm size” by endowing their analysis with kinked employment-adjustment costs that they associate with firm-size heterogeneity. Such costs reflect “… the (partial) irreversibility of separation decisions…. While firms can shed workers costlessly, it is costly to reverse such a decision because hiring (posting vacancies) is costly.” (p.7) E-M, working within the coherent market-centric DSGE framework, run afoul of a core mainstream problem. Their take on the S/M/B model is inherently incapable of accommodating essential characteristics of macro instability. As a result, it generates business-cycle properties that never occur in highly specialized economies. Most damaging, their counter-cyclical movements in job separation and unemployment must be wholly voluntary. Nominal demand disturbances induce trivial employment/output effects.
Moving beyond their model’s unhappy business-cycle properties, E-M claim significant correspondence with a number of facts from cross-section analysis. First, the share of smaller (fewer than 20 employees) firms is countercyclical. Second, “for each establishment size class considered, [the model] broadly matches the co-movement with unemployment over the business cycle observed in US data.” Third, “in our final quantitative application, we evaluate the model’s ability to account for the observation that workers employed in larger firms are often paid higher wages… The magnitude of the size-wage effect implied by the model is mediated by two competing forces, as noted by Bertola and Garibaldi (2001). On the one hand, the existence of diminishing returns in production might lead one to anticipate a negative relation between employer size and wages. On the other, larger firms also tend to be more productive. Quantitatively, the latter dominates, generating one quarter of the empirical size-wage effect.”
Hiring-cost red herring. Modern S/M/B theorists assign critical significance to hiring costs, denoted here by C. The basic idea is that such cost makes employers vulnerable to hold-up after an employee obtains a job. If workers quit sooner than expected, the firm incurs a charge. It is rational to pay labor a rent equal to bC, such that 0<b<1. In E-M, b is idiosyncratically bargained by the firm and each employee, introducing a wedge between the rational wage paid and labor’s opportunity costs. Note that wage recontracting, a pillar of coherent market-centric DSGE modeling that prevents Barro’s nightmare of dollars lying around on sidewalks, is not suppressed. E-M labor pricing is sufficiently downward flexible to prevent involuntary job loss.
In the generalized-exchange model class, the shaky S/M/B story falls apart. Worker hiring costs for routinized, nonsupervisory jobs are properly identified to be inherently small. The amortized surplus to be divvied up is too tiny to drive any important macro process. (Chapter 4) Practitioners know this; macro theorists should learn it. More important, large firms already pay endogenous wage rents as a result of their rational efforts to manage labor input in circumstances of costly, asymmetric workplace information. (Chapters 2, 3) Those rents effectively discourage voluntary quits, preventing hiring-cost hold-up. In the large-establishment venue, significant posting costs are little more than a Ptolemaic instrument, used to support mainstream market-centric theory in the face of a pile of contrary evidence.
Fundamental problem. The GEM Project microfounds the exceptional power of the old-fashioned model class endowed with meaningful wage rigidity. It effortlessly, coherently, and intuitively explains all the cyclical and cross-sectional evidence cited (and mangled) by E-M, as well as the even more crucial data (e.g., 6 million involuntarily lost jobs in 2008-09) that they conveniently ignore. The Ptolemaic hoops the authors must jump through explain the absence Occam’s razor in consensus macroeconomics. Most fundamentally, the market-centric DSGE model class that occupies the macro mainstream and therefore the content of AEJ:M cannot coherently suppress wage recontracting and therefore has no access to labor pricing that actually occurs in highly specialized economies. The problem is fatal to New Keynesian and AEJ:M aspirations for stabilization-relevance. Given the rules of engagement mandated in the consensus New Neoclassical Synthesis, mainstream NK theorists seeking stabilization relevancy will always fail.
Here is an instructive challenge to the editors and readers of the AEJ:M. Identify the most stabilization-relevant article ever published in that journal. An article is “stabilization-relevant” if it would have been helpful in guiding Bernanke’s thinking as he designed and executed the Fed’s successful efforts to halt and reverse the acute 2008-09 instability. The GEM Project Blog will be happy to publish nominations and supporting arguments. My bet is there are no serious candidates for stabilization relevancy. If I am right, that’s a big problem.
Blog Type: Wonkish Saint Joseph, Michigan