More on New Keynesian Macro Silliness

Print/Save PDF

An on-going task of the GEM Blog is to monitor the American Economic Journal: Macroeconomics (AEJ:M), flagging the more egregious use of neoclassical general-market-equilibrium modeling to explain actual behavior in modern, highly specialized economies. The exercise is always discouraging, but things are not as bad as they could be. In my experience, few policymakers, and none who are good at their jobs, take seriously academic analysis that does not come close to aligning with important evidence.

The exercise, however, does suffer from repetition. It ends up asking the same two questions over and over again. Why do authors,  ostensibly concerned with significant issues, keep coming up with useless stuff? And why do respected editors of this reputable journal keep publishing that useless stuff? I reject the answer that comes most quickly to mind. The editors, almost all of them, are not dumb. They know deep down that their mainstream market-centric general-equilibrium theory, enriched with rational market frictions, is inherently incapable of usefully explaining the nature of business cycles, notably including the millions of forced layoffs that always occur in recessions. As already indicated, most editors silently hope that stabilization authorities, in the serious business of mitigating the welfare costs of actual instability, do not take what they publish seriously.

Weak Job Recovery from the Great Recession

Earlier this year, Sylvain Leduc and Zheng Liu (hereafter L&L) published “The Weak Job Recovery in a Macro Model of Search and Recruiting Intensity” in the AEJ:M (2020, 12(1)). From the L&L summary: “We show that cyclical fluctuations in search and recruiting intensity are quantitatively important for explaining the weak job recovery from the Great Recession. We demonstrate this result using an estimated labor search model that features endogenous search and recruiting intensity. Since the textbook model with free entry implies constant recruiting intensity, we introduce a cost of vacancy creation, so that firms respond to aggregate shocks by adjusting both vacancies and recruiting intensity. Fluctuations in search and recruiting intensity driven by shocks to productivity and the discount factor help bridge the gap between the actual and model-predicted job-filling rate.”

If you are looking for a laugh in your own perusal of the unfolding macro literature, stay alert to the use of the neoclassical job search/match model class to explain what happens in recessions. It is without fail silly, and the examples of authors willing to engage in this particular, egregious debasement are plentiful. Once you read “we introduce a cost of vacancy creation, so that firms respond to aggregate shocks by adjusting both vacancies and recruiting intensity”, you know that a train wreck is coming.

Doesn’t everybody understand that profit-seeking managements that have laid-off a significant share of their workforce are looking in recovery to recall those employees, not recruit new applicants to those jobs. Specific human capital matters. L&L give adjusting vacancies and recruiting intensity the starring roles in part because their market-centric modeling cannot accommodate chronic wage rent and, consequently, therefore the prevalence of recalls – not new hires – in cyclical recovery. The standard practice of NK theorists in the frequent occurrence of being unable to accommodate important phenomena is to, without apparent shame, ignore them. Involuntary job loss in recession and recalls in early recovery do not exist in the L&L world. Labor action is confined to the marketplace and must, consequently, center on worker job search and firm recruitment. In highly specialized economies, that analysis never produces useful description of macro instability.

Generalized-exchange modeling easily remedies Leduc-Liu’s inability to explain the aftermath of the Great Recession. Its microfounded chronic wage rent, which it associates with firms most likely to experience forced layoffs and is enabled by the Project’s microfounding of meaningful wage rigidity (MWR), is a necessary condition for the existence of the factual, broad recall of most laid-off employees. Moreover, rational-behavior GEM modeling identifies inadequate demand as the principal cause of the weak recovery from the Great Recession, with the overhang of debt/bankruptcy as an important contributor to inadequate total spending. The GEM theory provides useful guidance for the management of the recovery from the pandemic shutdown. The L&L theory provides harmful policymaker guidance.

The Leduc-Liu reliance on mainstream NK market-centric macro modeling prevents their analysis from having policymaking relevance. Their loyalty to the evidence-innocent theory, however, does answer the second question posed above. Embattled editors continue to publish silliness to protect their own human capital. Misleading stabilization policymakers is apparently acceptable collateral damage.

Sticky Wages and the Willingness to Work

Also this year, Zhen Huo and José-Víctor Ríos-Rull (hereafter HRR) published “Sticky Wage Models and Labor Supply Constraints” in the American Economic Journal: Macroeconomics (2020, 12(3)). From the HRR summary: “In sticky wages models (either à la Calvo or à la Rotemberg), labor is solely determined by the demand side. However, a change of circumstances may make labor demand higher than agents’ willingness to work. We find that workers are required to work against their will between 15 percent and 30 percent of the time (with 5 percent wage markup, less with higher markups and in Rotemberg models). Estimating models with the minimum of the demand and supply of labor instead of the demand-determined quantity yields different and unappealing properties. Hence, special attention should be paid to possible violations of the labor supply constraint.”

The “willingness-to-work” problem tackled in the paper is neither adequately explained nor very important, but the HRR analysis does illustrate how mainstream sticky-wage modeling badly distorts whatever it touches. A brief description of the go-to Calvo assumption will suffice here. In a 1983 paper, he posits that a fixed fraction (ϋ) of wages remains unchanged each period, permitting the remaining labor pricing to be recalibrated. The probability that any given wage will be adjusted in any given period is (1- ϋ), implying strong restrictions on rational behavior. At any given time, a significant share of all firms must ignore the time lapse since wages were last changed, the size and nature of changes in market conditions, and movement in the price level. Calvo’s staggered wage stickiness enables tractable aggregation and is, despite its dependence on arbitrary, thoroughly impractical free parameters, the market-friction mechanism of choice in New Keynesian models.

Calvo’s free parameter prevents wage adjustments in circumstances of mutual benefit and can never be consistent with optimizing behavior. His broadly used mechanism must be, by its nature, very short-term and particularly runs afoul of model coherency requirements in its crude suppression of wage recontracting. In the HRR article, the convenient, badly inaccurate assumption about downward wage rigidity prevents the authors from having anything useful to say about real-world problem of workers being required to work longer workweeks than they prefer. By contrast, the highly specialized firms whose information-challenged workplaces are rigorously modeled in the GEM Project turn out to be most vulnerable to the HRR willingness-to-work problem. Generalized-exchange modeling provides an insightful description of that predicament that is consistent with the available evidence. The HRR problem, of course, remains relatively unimportant, especially relative to the huge costs of instability that usually occupy this Blog. Outside of providing an application of the failed NK market-centric general-equilibrium theory, there is little reason for this paper to be published.

Blog Type: New Keynesians Saint Joseph, Michigan

Write a Comment

Your email address will not be published.