Bob Hall (2010), a former colleague of mine at MIT, was one of the first of the handful of mainstream theorists who stepped up and examined the Great Recession using “a simple macro model that captures the most important features of modern models.” He emphasizes “realistic increases in financial frictions that… generate declines in real GDP and employment of the magnitude that occurred.” His analysis of real-shock costs provides a benchmark contrast to the GEM Project’s explanation of the huge 2008-09 welfare loss that focuses on the micro-coherent interaction of collapsing nominal demand and meaningful wage rigidity (MWR).
Financial frictions. From Hall: “The dominant view among macroeconomists today is that a financial crisis causes real economic activity to collapse by raising frictions.” His candidate friction for 2008-09 is measured by the jump in the spread between the Baa corporate bond and the 10-year Treasury note. The friction itself is rooted in well-known financial agency costs, featuring information asymmetries. After the Lehman bankruptcy, such costs helped push up the spread by more than four hundred basis points. Standard dynamic general equilibrium simulation tools were used to size macro implications of the higher cost of capital.
Model problems. Debilitating problems plague the capital-cost story. To begin, it is broadly understood that financial frictions, by themselves, cannot explain much cyclical behavior. From Hall: “…research generally shows that in standard neoclassical models, with normal preferences and technology and competitive markets, [financial-friction] shifts of realistic magnitude fail to deliver anything like the volatility seen in the U.S. economy.” Mainstream financial-friction models must find sufficient juice elsewhere. Hall provides the elsewhere by assuming two “… departures from the neoclassical benchmark”: (i) countercyclical pricing power and (ii) countercyclical movement of voluntary unemployment. His capacity to explain the Great Recession and its huge incidence of job separation is wholly dependent on those assumptions. He ignores that neither comes close to being consistent with rational behavior. He further ignores that each assumption is thoroughly rejected by all of the evidence. Both are easily recognized as no more than a Ptolemaic convenience.
From the perspective of stabilization authorities, the most objectionable aspect of Hall’s analysis is the broad range of critical facts that are shoved under the rug. Most notably, he provides little role for the contraction in nominal demand that occurred in the Great Recession. In generalized-exchange macroeconomics, almost all of the effects on real activity in the 2008-09 financial crisis result from the real shock’s propagation by weakening total spending interacting with MWR. Direct costs from increased financial frictions are easily seen to be relatively negligible.
In Hall’s model class, there is also no room for, and no mention of, the six million involuntary job losers, who accounted for most of the increased unemployment and that practitioners believe to have resulted from collapsing demand. (Surely paying attention to whomever orders employees to be laid-off does not inherently compromise model-builder principles.) Hall can make no useful distinction between nonstationary and stationary demand disturbances. He must pretend, counterfactually, that the former does not exist. The real-side credibility of stabilization authorities also has no place in his mainstream analysis. Even in the extreme 2008-09 instability, Hall must assume that interest rates exert the primary influence on investment outlays, despite investors making no secret about paying most attention to uncertain profit expectations and product demand. (Chapter 6)
Further indicative of a deeply flawed model, Hall achieves his realistic 16% and 23% reductions in real consumption and output by positing a nearly 40% cut in real wages as well as the destruction of half of the existing capital stock, wildly out-of-bounds assumptions that everybody knows did not come close to occurring. When did outrageous assumptions, conveniently tailored to yield specific results, that impose preposterous restrictions on other critical variables become blithely acceptable in stabilization model-building? Answer: When theorists and the editors who publish them became captive to the Ptolemaic impulse to defend at all costs their consensus macro model. Also telling is Hall’s concession that the theory “… cannot explain why GDP and employment failed to recover once the financial crisis subsided – the model implies a recovery as soon as financial frictions return to normal.”
The nature of the logical and factual shortcomings of Hall’s model, far from exhaustively covered in this post, is familiar from the decades-long effort to model coherent, continuous-equilibrium macro behavior while restricting exchange to the marketplace. A message microfounded in the GEM Project reflects what practitioners already know. The 2008-09 jump in the cost of capital did not induce the huge cutback in production and the millions of lost jobs. The sharp reduction in labor and capital utilization instead largely resulted from collapsing nominal demand interacting with the rational MWR. A more general message is that Hall’s effort was doomed from the start. Instability characteristic of the Great Recession cannot be adequately modeled within the coherent market-centric DSGE framework. What happened in 2008-09 was an episode of profound market failure, massively inconsistent with continuous general market equilibrium. The generalization of exchange permits that failure to be consistent with continuous general decision-rule equilibrium.
Reflecting mainstream willingness to ignore piles of contradictory evidence as well as debilitating analytical problems, Hall concludes that his model provides a promising, albeit “highly stylized”, explanation of the Great Recession. His analysis is made more Ptolemaic than promising by its inability to accommodate substantial market failure, rendering it powerless to come close to explaining what actually happened. Hall fully exploits today’s Ptolemaic Convention, identified in the GEM Project, that allows theorists to simply presume, without further justification, market clearing. By contrast, generalized-exchange modeling is not burdened with Hall’s logical and factual limitations while still maintaining its micro-coherence. The methodological choice matches Occam versus Ptolemy, evidence versus epicycles. When will mainstream theorists give up their increasingly debilitating Ptolemaic campaign to defend coherent market-centric DSGE modeling at the cost of explaining what actually happened in, and what to do about, the Great Recession?
Blog Type: Wonkish Chicago, Illinois