David R. Henderson  

The CEA's Mixed Thinking on Labor Market Monopsony, Part II

Is the data "rigged"?... AT&T, Time-Warner and the end ...

Yesterday, I posted on the basics of monopsony in the labor market, with a promise to critique the CEA report on monopsony, occasionally referring to Adam Ozimek's excellent, but shorter, critique.

In what follows, I go through the CEA report seriatim, with comments on highlights that I think are important.

Competing Less Aggressively

The CEA:

What the monopsonistic firm loses in reduced output and revenue, it more than makes up in reduced costs by paying lower wages. In other words, by recruiting less aggressively, paying less, and sacrificing some employment, employers with monopsony power can shift some of the benefits of production from wages to profits. (p. 2)

The CEA writers slipped in something important that does not follow from the monopsony model: "recruiting less aggressively." This is incorrect. In the standard perfectly competitive model that the CEA uses as a baseline to judge monopsony, employers don't compete aggressively at all. There's no need. They are price takers in the labor market and so they buy the amount of labor they want at the perfectly competitive wage. Because you can't compete less aggressively than zero, the monopsonistic employer cannot compete less aggressively than the employer in the perfectly competitive model.

Indeed, there's a strong case to be made that monopsonistic employers will recruit more aggressively. Look back at the difference between VMP and the wage, W (corresponding to Q*) that, the CEA and I agree, leads to underemployment. That gap is value not captured by employee or employer. So the employer has a strong incentive to compete aggressively for higher-supply-price workers whose supply price is still below VMP.

Incidentally, the CEA writers slip in the aggressive point again on page 3, but there, as on page 2, they don't justify their claim. As I noted above, one cannot justify their claim. You can't compete less aggressively than zero.

Monopsonists Will Discriminate on Wages

Unwittingly, the CEA lays out just how the monopsonistic employer will compete aggressively. The CEA writes:

Further, if employers with monopsony power are able to differentiate among workers' reservation wages, then they can also set wages that discriminate among their own employees. In the extreme case of "perfect" wage discrimination, firms can pay each worker the minimum he or she is willing to accept, regardless of the worker's skills or productivity. More generally, differing degrees of worker bargaining power across different groups of workers--for example by age, race or gender--may lead to varying degrees of wage depression, promoting within-firm wage inequality. For example, if women's job mobility is more constrained than men's by family responsibilities, then women will be more limited in their choice of employers and be more vulnerable to wage discrimination (Manning 2003, Ch. 7). (p. 3)

This is good reasoning. But note what it implies. It implies that the efficiency argument that the CEA made against monopsony--inefficiently low employment--is weaker than the CEA claimed at the start of the article. As I noted in yesterday's post, and as Adam Ozimek noted in his critique, to the extent that employers can pay people according to their supply prices, the gap between VMP and W lessens and the inefficiency lessens. Yet at no point in the 21-page report does the CEA recognize this tension.

Note something else: the CEA writers slip in the following: "For example, if women's job mobility is more constrained than men's by family responsibilities, then women will be more limited in their choice of employers and be more vulnerable to wage discrimination." That's possible. But notice what else "family responsibilities" imply: higher opportunity costs and, therefore, higher supply prices. So it's entirely possible that monopsonstic employers will compete for married women by paying more. That the CEA does not even mention this suggests that this report is a mix between an economic document and a political document.

The CEA plunges on:

To be sure, firms face a number of constraints in their ability to pay different wages to similarly qualified workers (or even to workers who perform different tasks), including legal constraints and concerns over internal equity or fairness.2 However, employers may be less constrained by equity concerns when workers lack good information about the wages of their coworkers (Card et al. 2012). Firms can also circumvent internal equity constraints or fairness norms by shedding activities to subordinate companies through subcontracting, third party management, and other organizational forms. Such "fissuring" of employment makes wage discrimination feasible by transforming wage setting within the walls of a business to a pricing problem among subordinate firms (Weil 2014). (p. 3)

Exactly. So yet again, the CEA has identified clever ways that profit-maximizing firms may use to reduce the inefficiency that arises from monopsony. But again, true to form, the CEA writers make no recognition, not even in a footnote, that this would reduce the inefficiency from monopsony.

Non-Compete Agreements

The CEA report has a number of statements about non-compete agreements. I won't quote them all, but here, I think, they score some points. Here's one quote:

Non-compete agreements are not always harmful to workers or to growth; by preventing workers with "trade secrets" from transferring technical and intellectual property of companies to rival firms, these agreements can be one means of facilitating innovation. However, employers also have other methods to protect their interests. And new evidence (discussed further below) suggests that the use of non-competes in the United States today extends well beyond cases where they are plausibly justified. In particular, the evidence shows that 30 million American workers are currently covered by non-compete agreements, and that these agreements are often imposed broadly on low-income workers or others with no access to trade secrets (U.S. Treasury 2015). In these cases, it is likely that the primary effect of these agreements is to impede worker mobility and limit wage competition. (p. 5)

Later the CEA reports:
Survey data suggests that in many cases, workers sign non-compete clauses without full information on what they are signing or how it will be enforced. A recent survey of electrical engineers finds that nearly 70 percent of respondents report that their employer presented them with a non-compete only after they had accepted the job offer, and nearly half of the time, the non-compete was presented to the employee on or after his or her first day of work (Marx and Fleming 2012). Further, Starr et al. (2016) find that these contracts are prevalent even in States where they are not enforced. Indeed, in California, which does not generally enforce non-compete agreements, 22 percent of workers report that they have signed one. The use of non-compete agreements where they are not enforced suggests workers are not well-informed, and raises the possibility of disparate impacts across workers with and without sophisticated understanding of the legal implications of these agreements. (p. 8)

This does seem, to, to use a technical term, suck. I do see a ray of hope here, though. To the extent it gets publicized, at least in my state of California, workers can sign these agreements knowing that they can't be enforced.

The Role of Employer-Provided Health Insurance

The CEA points out the existence of employer-provided health insurance can produce "job lock," making employees less mobile than otherwise. Then the CEA writes:

As discussed further below, the Affordable Care Act reduced job lock by providing workers with affordable non-employer sponsored health insurance options and banning private insurance policies from setting different coverage terms based on health status. The availability of non-employer sponsored health insurance may strengthen the bargaining positions of workers who do not leave their employer, since they can better leverage the option of leaving. (p. 7)

Irony, anyone? Obamacare has made such a hash of the individual insurance market that, if anything, it has probably increased job lock. Why leave a job you don't totally like if it has better employer-provided health insurance than what you can get under Obamacare?

The Role of Occupational Licensure

Here the CEA scores a home run. It writes:

One class of regulations that can present barriers to job mobility is occupational licensing laws (CEA, Department of Labor, and Department of the Treasury 2015). While licensing regulations can play an important role in protecting consumer health and safety, they also raise the cost of entering a licensed occupation. Today, roughly one in four U.S. workers requires a government license to do their job. For some of these jobs, the costs of obtaining a license can be significant while the health and safety benefits may be often minimal. In these cases, licensing can create unnecessary barriers to employment, restricting opportunities and depressing wages for those who are unable to obtain a license (CEA, Department of Labor, and Department of the Treasury 2015).

Because licensing restricts the supply of workers in a profession, licensed workers tend to earn higher wages at the expense of excluded workers. However, even workers who hold licenses can find their employment alternatives limited by existing licensing regulations, which often vary dramatically across States (Carpenter et al. 2012). In particular, the patchwork of State regulations and variability in State reciprocity make it harder for workers in licensed occupations to move across State lines (Kleiner 2015), and new data show that licensed workers are less likely than unlicensed workers to make such moves. (p. 7)

Keep especially in mind the statement above: "In these cases, licensing can create unnecessary barriers to employment, restricting opportunities and depressing wages for those who are unable to obtain a license." It will be especially relevant when we discuss the CEA's discussion of unions and the minimum wage.

Housing Restrictions

The CEA has been particularly good on this:

Other regulations--not necessarily in the labor market--can also present barriers to job mobility. For example, overly restrictive land-use regulations create costly barriers to housing development, limiting the availability of housing and increasing its cost (Furman 2015). In turn, higher costs of finding and purchasing or renting a new home can effectively narrow the labor market. (p. 7)

This post is getting overly long. Tomorrow I will finish with Part III.

Comments and Sharing

CATEGORIES: Labor Market , Regulation

COMMENTS (5 to date)
gmm writes:

I love this series so far, keep at it!
Isn't it wrong to say that all the firms are price takers and thus do not compete ? Isn't the price set through a continual process of competition between firms?

AGJ writes:


Besides a grammatical mistake that he made in that sentence, he's correct because he's describing the assumptions of perfect competition.

David R. Henderson writes:

AGJ answers correctly. And thanks. Part III will be today.
Thanks. Typo is fixed.

Aaron McNay writes:


With regards to non-compete agreements, I was a bit disappointed that the CES report, and your post, did not mention one possible justification that I can think of for their use: job training.

Both employers and employees would like to be able to train the employees if the cost of doing so is less than the gains in productivity. However, there is a potential collective action problem here. What happens if the employer provides the training, but the employee then moves onto another job? The employer bears the burden of the training costs, but does not receive any of the benefits. As a result, the employer does not provide the training, and a mutually beneficial trade is not made.

By preventing the employee from being able to move, a non-compete agreement eliminates the collective action problem. The employer provides the training and the employer and employee are made better off. There are more complications to this than presented here, of course. However, I still think it is an issue that is ignored too often when talking about these type of agreements.

This issue would also explain why there are non-compete agreements for low-skill jobs. In addition, there is at least some work that has been done on this issue and it finds some support of the idea that training increases in higher non-compete enforcing areas. The paper is here:
From the abstract:
"A change from non-enforcement to maximal enforceability increases firm-sponsored for training by at least 13.7% and reduces wages by 2.35% for occupations frequently found in noncompete litigation."

Like I said at the start, I am disappointed that the CES does not really even bring this issue up. They very briefly say: "This pattern of evidence casts doubt on the notion that non-compete agreements serve mainly to protect employers’ trade secrets and investments in employee training." However, they never talk about the training issue. At least, I could not find it.

David R. Henderson writes:

@Aaron McNay,
You’re absolutely right that I didn’t think this through. That’s an unforced error on my part. Thanks for pointing it out.

Comments for this entry have been closed
Return to top