An interim message of the on-going series of posts on Ptolemaic woes is bad news. Mainstream macro modeling is trapped in practical irrelevance by the profession’s gatekeeper acceptance of the presumption of market clearing. The troubled nature of consensus market-centric general-equilibrium modeling is never more on display than during efforts to explain the 1930s depression. But the subject is irresistible; the effort continues to be made. Ben Bernanke (1995, p.1) got it right: “To understand the Great Depression is the Holy Grail of macroeconomics.”
Market-centric general-equilibrium narrative. Edward Prescott, the bold, take-no-prisoners leader of RBC theorists, has interestingly separated the U.S. experience in the 1930s into two phases: the extreme instability that marked the onset of the Great Depression (1929-1933) and the subsequent weak recovery throughout which total output remained below its trend (1933-39). The Nobel-Prize winner then sets aside modeling the rapid contraction of the extreme-instability phase and focuses instead on the ensuing stagnation. From Prescott (p.27): “[During 1933-39], there was an important change in the rules of the economic game. This change lowered the steady-state market hours. The Keynesians had it all wrong. In the Great Depression, employment was not low because investment was low. Employment and investment were low because labor market institutions and industrial policies changed in a way that lowered normal employment.” He is paying particular attention to original work by Cole and Ohanian, using the continuous-market-equilibrium neoclassical growth model to analyze 1933-39 macrodynamics.
The critical evidence in the Cole-Ohanian analysis is that total factor productivity (TFP) contracted 18% during 1929-1933 and returned in 1936 to its trend level. Positing technological regress measured by the TFP disturbance to be the impulse mechanism, they simulated real output during the Great Depression. They conclude that TFP “explains” 40% of the de-trended drop in real GDP in 1929-33 but is reduced to an insignificant role in the subsequent 1933-39 stagnation. The most critical fact is the weak recovery in employment, the de-trended level for which remained below its 1929 level throughout 1933-39, indicating the existence of a powerful impediment to the downward adjustment of wages. That unsurprising conclusion is shared by Bernanke’s depression model and the GEM Project.
From Cole-Ohanian (2002, p.30): “The marginal rate of substitution between consumption and leisure is 41 percent below the wage rate, and factor prices differ considerably from their implied marginal products. The wage rate substantially exceeds the marginal product of labor, and the return to capital is below the marginal product of capital. Between them, these data suggest that some factor raised the wage above its market-clearing level, and that this high wage prevented households from satisfying their marginal rate of substitution condition.” From their market-centric general-equilibrium perspective, the two theorists finger New Deal legislation, especially the National Industrial Recovery Act, as the culprit. Manufacturing wages, reflecting large establishments that were subject to the NIRA, experienced a 7% jump in 1934 (the year in which the NIRA became effective), remaining above their trend line for a decade. Meanwhile, real wages in non-manufacturing sectors, which were mostly transacted in small firms that were not covered by the NIRA, remained below their trend line throughout the decade.
Cole and Ohanian then cobble together a multi-sectoral, insider-outsider model with imperfect competition and wage bargaining. They posit that insiders set the wage, restricted to exceed the market-clearing rate and subject to the firm’s reservation profits. In the NIRA, firms that codetermine wages and working conditions with their workers’ union were permitted to collude on production. Their implicit bargaining model, rooted in insider market frictions, provides a weak, misleading explanation of wage codetermination between unions and management. Perhaps not surprisingly, the generalized-exchange model class featured in the GEM Project provides a much more powerful, policy-relevant, and intuitive bargaining theory. (Chapter 7)
The GEM alternative. Prescott’s separation of the extraordinary 1929-39 market failure into two distinct phases is unnecessary and deceptive. Generalized-exchange modeling easily integrates the Great Depression’s onset contraction and subsequent weak recovery into a coherent continuous-equilibrium macrodynamic process. In the GEM narrative, the extreme-instability output loss of 1929-33, centrally associated with a nonstationary demand disturbance (Chapter 6), rationally morphs into the extended period of stagnation, characterized by GDP growth that is chronically inadequate to reduce high unemployment. (Chapter 5)
The Project identifies conditions necessary for chronic stagnation, drawing from its coherent two-venue labor-pricing model that is much more powerful than Cole and Ohanian’s cobbled-together bargaining theory. Rational exchange-generalization cleans up problems that are a constant embarrassment to RBC analysts. Two-venue modeling incorporates the extreme-instability phase (1929-33, occurring prior to the “important change in the rules” motivated by the passage of the NIRA) into its overall analysis of the Great Depression by microfounding meaningful wage rigidity and its channel through which the outsized nominal demand disturbance induced same-direction outsized changes in employment and output. The generalized-exchange theory critically accommodates the huge jump in involuntary job loss produced in the deep contraction. (Chapter 2) The GEM framework additionally models the inherently slow adjustment of large-establishment employee reference standards and labor-pricing to the nonstationary reduction of nominal profits and the increased incidence of permanent job downsizing. (Chapter 3) The more powerful two-venue approach microfounds the critical role of nominal demand interventions, the basis for policy that effectively addressing the costly stagnation, and explains how the sharp jump in early-1940s war-related spending finally ended the Great Depression. (Chapter 5) An important message of the Project is that Prescott’s assessment is wrong: Keynesians got it right. In both phases of the Great Depression, employment was low largely because nominal spending, particularly investment, was weak.
In closing, perhaps we should cut the RBC theorists some slack. Like almost everybody else, they work within the market-centric general-equilibrium framework. But, unlike early Keynesian model-builders, they refuse to violate the micro-coherence of their mainstream model class. The RBC School consequently occupies a kind of high-ground that is more upfront than market-centric stabilization-relevant analysis that pretends to be micro-coherent. Unfortunately, the RBC trademark coherent market-exchange centricity can never align with the actual stability of highly specialized economies. As a result, it cannot be used to construct stabilization-relevant models of costly macro disruptions. Leading mainstream theorists are not dumb; they must have figured out that their market-centric general-equilibrium model will never explain periodic extreme instability. It would help in the development of more useful macro theory if they would stop claiming otherwise.
Blog Type: Wonkish Saint Joseph, Michigan