New Keynesian theory, the go-to model in modern macroeconomics, is a quagmire of seemingly insoluble puzzles. Readers of this Blog know that much of the difficulty is rooted in the New Neoclassical Synthesis, the academy’s governing directive for acceptable research. In particular, the NNS mandates use of friction-augmented general market equilibrium, which produces much more mild recessions than consistent with the evidence. The best-known NK shortcoming is the employment-volatility puzzle: FGME modeling generates much smaller increases in joblessness in cyclical downturns than does the real world. Additional business-cycle characteristics that escape the analytic grasp of NNS theorists include meaningful wage rigidity, involuntary job and income loss, chronic wage rents that motivate unemployment persistence, chronic market inefficiency, money non-neutrality sufficient to justify discretionary management of total nominal spending, an evidence-consistent trade-off between unemployment and inflation, extreme instability and depression rooted in rational behavior, rational-behavior stagflation, and more. Simply put, the mainstream NNS framework does an embarrassingly shoddy job of explaining policy-relevant features of business cycles.
This post’s selective look at unresolved mainstream puzzles considers the adverse employment and production fluctuations generated in the FGME model class, concluding it can never produce policy-relevant instability. At the heart of that problem are inherent limitations of NK search/match modeling that is almost always used in FGME analysis to provide behavioral content to labor-market behavior.
Mainstream job-loss modeling. NK research on downward nominal wage rigidity, a necessary condition of involuntary job loss, is fatally focused on identifying one or more market frictions that rationally suppress labor-price recontracting. Decades of failure indicate that the hoped-for super friction does not exist. Illustrative of that continuum of disappointment is the relatively recent attempt by Blanchard and Gali (2010 )to come up with a plausible super friction.
Blanchard-Gali’s fate is sealed with their assignment of a central role to the familiar labor wedge, asserting that market frictions induce a wage band within which any real labor price is consistent with private efficiency: MRS≤W≤MPL. MRS denotes the marginal value of worker time; W is the wage paid; MPL is labor’s marginal product. Blanchard-Gali manufacture a bit of wage indeterminacy by assuming away the bargaining power that informs determinate labor-wedge pricing.
Instead, their starring role is given to hiring costs that are increasing in labor-market tightness, which introduces a search/matching type of unemployment into the continuous-equilibrium model. That is their market friction of choice. This version of the labor wedge, however, captures little of what is widely known about labor pricing after the Second Industrial Revolution; and the Blanchard-Gali analysis suffers from inattention to that available knowledge. Most important, they simply ignore that rationally suppressing wage recontracting is a necessary condition of the rational existence involuntarily lost jobs, which we know accounts for most of the increase in unemployment in recession.
Constructing a mechanism that suppresses labor-price recontracting while preserving the optimizing, continuous-equilibrium economic method has proven to be the most durable problem in macroeconomics. Its resistance to solution eventually led to the expulsion of Early Keynesian thinking (Modigliani, Hicks, Samuelson, Okun, Solow, and many other really good economists) from graduate-school curriculums and cutting-edge journals. In setting aside recontracting problem, Blanchard and Gali follow modern NNS model-building practice. The question then becomes why are New Keynesians who assert policy-relevancy allowed to get away with chronic malpractice. Setting aside recontracting does not make its central role in actual behavior disappear. The Blanchard-Gali model cannot accommodate involuntary layoffs and is, as a result, fundamentally misaligned with the true nature of cyclical downturns.
Moreover, Blanchard-Gali err in following the standard practice of anchoring their analysis in the Mortensen-Pissarides search/matching framework, which posits an exogenous constant job-separation rate and consequently makes their theory even more inappropriate for the coherent analysis of job loss. (Despite those fundamental limitations, they claim labor-market realism as an important virtue of their model.) Because of their exclusive focus on marketplace decision rules, search/match theorists never come close to providing useful explanations of forced job loss and, more generally, the cyclical behavior of production, employment, and unemployment.
Market-centric production fluctuations must be sufficiently small to ensure no associated forced job loss. Given such tiny output contractions, it is difficult to imagine how discretionary aggregate demand management could improve on market solutions, which is the RBC argument against activist monetary policy. It is interesting that the Blanchard-Gali simulations produce “large and inefficient movements in unemployment” resulting from monetary policy designed to stabilize inflation. They don’t appear to understand that their assumption-based empirical conclusion cannot coherently follow from their NNS-class model.
Really bad news. The job-loss problem for mainstream macro thinking is more general than challenging the usefulness of the Blanchard-Gali paper. It demonstrates the entire NNS-model class can never be stabilization relevant. That conclusion follows from the inconsistency between (i) wage recontracting in lieu of being laid-off required by general market equilibrium and (ii) involuntary job- and income-loss produced by real-world instability. The successful New-Classical insurgency taught us that rational-behavior macro research was badly served by Early-Keynesian satisfaction, to the degree that it existed, with the use of free parameters to generate meaningful (i.e., capable of supporting forced job loss) wage rigidities. But New Keynesians do greater damage by refusing to acknowledge that recontracting forces meaningful labor-price rigidities in market-centric models to require free parameters.
For a host of reasons – analytic convenience, reluctance to depreciate NNS-specific human capital, difficult-problem avoidance, devotion to incrementalism, etc. – leading macro theorists continue to assert the dominance of NNS modeling that, absent free parameters, neither usefully supports stabilization policymaking nor recognizes the constructive management of aggregate nominal spending. Would any macroeconomist, reasonably informed and in good conscience, claim that the Blanchard-Gali model would have improved the Fed’s understanding and stabilization-management of the severe 2007-09 recession that generated nearly six-million job losers?
Blog Type: New Keynesians Chicago, Illinois