As promised last week, this post looks at some of the insightful, little known evidence that supports my claim that the GEM Project’s explanation of wage determination is how labor is actually priced in ubiquitous information-challenged workplaces. The Project’s narrative is more than just another theory. Simply put, it is the one and only real-world answer to the central macro quest to rationally motivate meaningful wage rigidity.
Many macroeconomists, working within the mainstream market-centric, general-equilibrium model class for most or all of their careers, have little grasp of the available labor-market evidence that is relevant to macroeconomics. They simply do not know enough labor economics, especially the careful modeling (rooted in the neoclassical tenets of optimization and equilibrium) of wage determination, work rules, and employee allocation inside large, highly specialized workplaces subject to costly, asymmetric information. They do not know that, in the U.S., original internal labor-market theorists (focusing on behavior inside complex establishments) dominated the labor-economics mainstream in the middle 20th century. Those pioneers uncovered fundamentally important evidence that has continued to be ignored, to their great detriment, by New Keynesian theorists. Two particularly interesting classes of data are illustrated here.
Employer testimony. Three examples will be cited. Most generally, employers are known to have long believed that dissatisfied workers, given the latitude, adversely alter their behavior on the job. An early Yankelovich poll of business leaders asked: Does job dissatisfaction lead to high turnover, tardiness, loafing on the job, poor workmanship, and indifference to customers and clients? Of the 563 respondents, 94 percent thought such an association does exist. (Katzell and Yankelovich (1975), p.114)
Getting closer to the critical economic question, Campbell and Kamlani (1997) surveyed 184 compensation executives from large firms, asking how much workplace productivity would decrease if wages were cut by 10 percent. The mean response was 20 percent. Nearly 7 out of 10 believed that the principal reason for the harmful effects was damaged worker loyalty.
In the third example, Bewley (1999a) conducted a careful survey in the U.S. during a period that includes the 1990-91 recession. He interviewed 104 business leaders, asking them why worker morale matters to them. The nature and incidence of their responses provide noteworthy support to generalized-exchange modeling:
Reason Percentage of Businesses Citing the Reason
Low worker productivity 89%
Poor customer service 14
Turnover 13
Recruiting 7
Absenteeism 4
Unionism 3
From Bewley (1999b, p.1): “Employers were reluctant to cut pay because they believed doing so would hurt employee morale, leading to lower productivity and current or future difficulties with hiring and retention. It was thought that these effects would in the end cost more than the savings from lower pay.”
Large-establishment employers tell a consistent story. They believe neither that worker OJB is wholly determined by technology nor that employees are inveterate shirkers. Over the past century, they have learned that employees strongly prefer fair treatment, that workers know how they are doing their jobs better than management, and that policies designed to influence labor willingness to adopt management goals significantly contribute to the pursuit of profit.
It is a huge mistake to ignore employer testimony. It is by far our greatest source of insight into how wage determination is done in modern highly specialized market economies. It is especially powerful in rejecting economist thinking that fails to capture critical influences that are not readily discernible from a strictly market-centric perspective. Another way of thinking about the same thing is to pay attention to what economists have long understood about large, highly specialized workplaces. In such circumstances, actionable knowledge about labor on-the-job behavior is inherently restricted by costly, asymmetric information. Employer testimony makes clear that efficient wage determination in such circumstances cannot occur wholly in the marketplace. If a theorist’s wage model does not include rational intrafirm employer-employee interaction, he or she is trying to fool someone.
Documenting wage rents. An extraordinarily consequential outcome of the GEM Project’s focus on rationally modeling information-challenged workplaces is the existence of labor pricing chronically in excess of market opportunity costs The best empirical work on such wage rents remains the 1989 Brookings paper, “Industry Rents: Evidence and Implications.” by Katz and Summers. (1989 was a long time ago. Labor rents, empirically or descriptively, have received very little attention during the mainstream run of New Keynesians, probably because their “settled” market-centric general-equilibrium macro theory cannot coherently accommodate them.) K&S used Current Population Surveys to demonstrate that many employees receive substantial wage rents for no other reason than they work in particular industries. K&S estimate the average wage rent among industries that pay premiums to be 28%; controlling for a broad range of worker characteristics (education, etc.) reduces their rent estimate to 15%. K&S also demonstrate that accounting for fringe-benefit or occupational differences tends to increase, not reduce, the wage differential.
Are the premiums caused by unionization? The evidence indicates similar-magnitude wage rents for nonunion workers. How about Adam Smith’s famous argument, i.e., relatively difficult working conditions being the cause of relatively high labor pricing? Available evidence indicates that industries paying labor rents tend to have better working conditions. K&S find a strong negative relation between an industry’s quit rate and its wage, which they show to be rooted in the labor-price premium rather than observed worker characteristics.
Here is the knockout punch. K&S use longitudinal evidence, focusing on worker movement from low-wage to high-wage industries, to identify a powerful characteristic of the U.S. labor market. Relocating employees immediately pocket 60 to 100% of the industry wage differential. New hires’ quick capture of the lion’s share of the existing premium destroys the mainstream argument that apparent labor rents reflect differences in unobserved human-capital. Job transfer itself cannot enhance workers’ intrinsic productivity. K&S finally investigate the nature of rent-paying industries. They are capital-intensive, experience relatively high rates of return, and invest more heavily in R&D. They are easily recognized as the large establishments, intrinsically restricted by costly asymmetric workplace information, that populate the rent-paying venue of the generalized-exchange model class.
The evidence on the existence, size, and location of wage rents has huge implications for how to adequately model highly-specialized economies. That NK modeling of labor pricing cannot accommodate such evidence helps explain why mainstream thinking is stabilization irrelevant.
Blog Type: New Keynesians Saint Joseph, Michigan
Write a Comment