The Bias Against Low-Wage Labor Markets in Merger Analysis

Why are labor harms so rarely pled in antitrust cases? One factor may be the bias against low-wage labor markets in merger analysis. This bias treats the exact same employee behavior as rational in high-wage markets and irrational in low-wage markets. This article concludes with recommendations on how to better account for labor effects more generally in antitrust.

Anant Raut¹

I. INTRODUCTION

Blistering critiques of the last five decades of antitrust enforcement have led us to a debate, several years in, over whether and how to improve the current approach. A voluble part of the ongoing discussion has been whether antitrust, content for decades to rely on price and output effects, should take into account more than just the end price for consumers, such as the impact that consolidation would have on employees of the merging parties. Scholars and critics from this neo-Brandeisian school of thought have called for antitrust to generally better account for these types of labor effects.

The landmark antitrust labor case would be a merger case where the competitive harm is to the employees of the post-merger entity. Neither enforcement agency has ever brought one, but there are signs that the FTC may be searching for just such a case to bring. But there is a bias built in to how traditional antitrust economics distinguishes the characteristics of “skilled labor” and “unskilled labor” markets. Left uncorrected, enforcers may focus too intently on mergers involving high-wage labor markets while ignoring the same competitive harm occurring in their low-wage counterparts.

By this author’s estimation, neither enforcement agency (the Department of Justice’s Antitrust Division and the Federal Trade Commission’s Bureau of Competition) has ever challenged a merger on the basis of post-merger employer monopsony power. Why haven’t labor harms been pled more? In part, because for many years labor cuts have been treated crudely, horribly, as a benefit of mergers. Terminating the livelihood of human beings is still referred to in the Orwellian argot of merger analysis as an “efficiency.”

True, the enforcement agencies have this decade brought nopoach cases, which involved agreements among several hugely successful companies to keep their best engineering or illustrating talent from being paid their fair market rates.² But these were straightforward Section 1³ price-fixing agreements where the prices were salaries rather than consumer goods. Similarly, I and others have argued that non-compete clauses where no trade secrets or intellectual property is involved should also be treated as a violation of the antitrust laws, but again, that falls under a traditional horizontal restraints theory.

The unicorn case, and certainly the most trailblazing, would be one where the enforcement agencies challenged a merger because after the merger the new company would be in a position to pay its employees less. It makes a sort of intuitive sense – it’s the thing employees fear second-most following a merger announcement, behind termination itself – yet it’s never been done. In traditional merger analysis, enforcement agencies are always focused on the competitive effects of the product of the merging companies – will the price of plane tickets go up once these two airlines merge? Will indie movies get shut out of distribution if this mainstream theater chain buys this independent movie chain? Will this global beer giant’s crappy midlevel beer get pricier once it buys out this craft brew that everyone likes? Never is the analysis on what happens to the employees of the companies themselves after they merge.

And here I’d like to preempt the terminology most often used and reject the categorization of employees as “skilled” and “unskilled” labor. Describing a market at the outset as “unskilled labor” begets the outcome you’re looking for, that any one of these workers is wholly replaceable by another one. But just because one set of jobs is more easily learned and has fewer credentialing requirements doesn’t mean that that worker lacks “skills”; that worker, in fact, may be acquiring skills that make that worker especially valuable to a competitor company in the same space, even the subject of a bidding war. A better, and certainly less offensive, pair of labels is high-wage and low-wage labor markets.

II. THERE HAS NOT YET BEEN AN EMPLOYER MONOPSONY MERGER CHALLENGE, BUT THERE LIKELY WILL BE

There’s nothing preventing the enforcement agencies from taking labor effects into account as part of the merger analysis. Without even changing the existing consumer welfare standard, where many of the fires of the current antitrust debate have raged, agencies absolutely can take labor effects into account, under plain vanilla monopsony power theory. A monopsony labor theory of harm can be summarized as follows: in the market for employee services, the post-merger entity will now have amassed sufficient market power to underpay employees relative to what they would be able to ask for in a more competitive market. Put another way, once Company A buys Company B, the combined company can pay its employees less (or reduce benefits or some other form of compensation) because it has the clout to do so.

But this author looked, and could not find a single instance in over two decades’ worth of merger challenges of either enforcement agency pleading a harm to the employees following the merger.⁴ Now the tide is turning, and I think that eventually we will. In October 2018 remarks to the American Bar Association, then-Director of the FTC’s Bureau of Competition Bruce Hoffman stated that the FTC was investigating monopsony, including labor monopsony issues, in three major merger investigations.⁵

Based upon traditional antitrust economic theory, I expect enforcers to look for a seminal case in a merger involving specialized, highwage labor like engineers or doctors, and assume that none exist in low-wage markets. The rationale under the traditional economic models used in antitrust is as follows: high-wage labor is more locked in and less likely to switch professions or geography than take a post-merger pay cut, whereas low-wage labor has a lot more low-wage labor options, so post-merger employers have no real leverage to cut their wages. This is the bias against low-wage labor markets in antitrust, and it is flawed because of real world work that demonstrates the opposite behavior in low-wage labor markets. Yet traditional antitrust treats the exact same choices made for the exact same reasons and treats the behavior in highwage markets as rational, and as anticompetitive effects, and the behavior in low-wage markets as irrational, essentially penalizing the worker for acting reasonably.

III.  ANTITRUST TREATS HIGH-WAGE LABOR MARKETS LIKE DUDLEY DURSLEY AND LOW-WAGE LABOR MARKETS LIKE HARRY POTTER

We can build what the hypothetical high-wage labor market employer monopsony case will look like. High-wage labor is a more attractive candidate for an enforcement agency because of the ways it can be more easily defined as a market. In this example, we’ll use doctors employed by a pair of merging hospitals or health systems. In defining high-wage labor as a market, the enforcers will plead the following assumptions: doctors have invested so much time in their credentialing that they are more likely to accept a small pay cut than switch professions. Enforcers will also plead that it is unreasonable to expect these doctors to move to another location to practice.

We can anticipate these arguments from how the enforcement agencies have pled the next closest thing, harms to physician reimbursement rates from potential health plan consolidation.⁶ Look at the assumptions built into the calculation of post-merger market power in these cases.

See, e.g. U.S. v. Aetna and the Prudential Insurance Company, Complainant 27: “A small but significant decrease in the prices paid to physicians by these buyers would not cause physicians to seek other purchasers of their services or to otherwise change their activities (away from providing physician services towards other uses or leisure) in numbers sufficient to make such a price reduction unprofitable.” Further, “[a] small but significant decrease in the prices paid to physicians would not cause physicians to relocate their practices outside of the Dallas and Houston markets in numbers sufficient to make a price reduction unprofitable.”⁷

The argument was repeated nearly verbatim six years later in U.S. v. UnitedHealth Group, Inc. and PacifiCare Health Systems, Inc., Complaint at 33: “A small but significant decrease in the price paid to physicians would not cause physicians to seek other purchasers of their services or to otherwise change their activities (away from providing physician services) in numbers sufficient to make such price reduction unprofitable.”⁸

In contrast, traditional antitrust is unlikely to even see employer monopsony harm in low-wage labor markets because of its unhesitating embrace of the classical economic argument that low-wage labor will always get reallocated more efficiently into the vast, undifferentiated pool of general low-wage jobs. There is no harm because there is no market power, and there is no market power because the low-wage workers can swim back into the sea of low-wage labor and swim back out into completely different low-wage positions. As a result, employers daren’t wage monopsony power.

As a result of these assumptions, when enforcers see high-wage labor accepting a pay cut rather than switch jobs after a merger, the high-wage workers are victims, the pay cuts are unavoidable, and the result is an antitrust harm. By contrast, when low-wage workers do the same, they are people with better options making decisions not in their self-interest, the pay cuts are avoidable, and the result is not an antitrust harm.

IV. WHY THIS APPROACH IS FLAWED

These disparate outcomes stem from flawed assumptions about both labor markets.

First, on the high-wage labor side, the assumptions that lead to a conclusion of post-merger employers having wage power are overly broad and not borne out in real world data. Take the physician employee example above. Surveys of medical school graduates on their 10-year reunions have found that fewer than 100 percent of them were still doctors. A casual perusal of some medical school alumni magazines demonstrates a variety of seemingly fulfilling non-medicine career choices: some choose to go into hospital administration, some leave practice for full-time academia, and some switch professions entirely, joining even more lucrative fields like hedge funds or the pharmaceutical industry, or obtain yet another degree and defend their colleagues against medical malpractice cases. In addition, while complicated, it is far from pioneering for a nurse or doctor to move to a new area to practice.

But putting aside how much conventional antitrust analysis may overstate how much market power a post-merger employer may have in certain high-wage markets, it definitely understates how much may occur in low-wage markets. This is a reality that classical economics misses, but behavioral economics does not.

Low-wage labor markets aren’t the amorphous miasma of interchangeable jobs that classical economics would have you believe. On the contrary, there are a number of reasons why low-wage labor, facing a pay cut, may stay in the same job rather than take a better-paying low-wage job elsewhere. Some of these factors, as detailed in an October 2016 report by the White House Council of Economic Advisors, include information asymmetries and search frictions.⁹ Job seekers may not know about all relevant job openings at a given time. And while a considerable industry including LinkedIn and recruiters and headhunters has built up around matching qualified high-wage candidates for high-wage openings, there has been less profit and motivation to do so on the low end, where we see less efficient methods of communications such as job fairs and word of mouth.

Nor may people in low-wage jobs seek out higher-paying opportunities where you expect them to. There is a “stickiness” to jobs that contradicts assumptions about labor movement. So-called job lock can result from any number of factors. Prior to passage of the Affordable Care Act, workers with pre-existing conditions were afraid to switch jobs and either lose their insurance or have their premiums go up considerably. Others can only take certain jobs that work with their childcare arrangements. One study¹⁰ has found that U.S. job seekers are 35 percent less likely to apply to a job 10 miles away than one within their own zip code.¹¹ As economist David Autor sums it, “It turns out the real world is not that close to that best-case scenario. People don’t move really readily; they have skills that are specific to their industry, they have attachments to their jobs, it’s wrapped up in their identity. And then the shocks, because they’re so geographically concentrated, they’re highly, highly disruptive.”¹²

A management-level professional at a factory may under classical economic theory be considered “unskilled,” and therefore willing to switch jobs. But that job may be “sticky” because that senior-level employees has now earned the shifts that s/he wants, the vacation time that s/he wants. So it becomes much easier to understand why that professional, facing a 5 percent pay cut (the test in antitrust), would be willing to accept that rather than start over with no seniority at an entirely new company.

But under traditional antitrust economic theory, the same choices made for the same reasons under the same circumstances are treated as rational behavior in high-wage labor markets, and irrational behavior in low-wage labor markets. Classical antitrust theory penalizes the low-wage laborer for making the exact same jobs choices as the high-wage laborer, resulting in a blind spot of antitrust analysis, whereas a behavioral approach that takes into account real world behavior should find the same competitive harm.

V.  WHERE THIS ARGUMENT FITS WITH THE OTHER WORK BEING DONE ON ANTITRUST AND LABOR

Some of the other work on labor antitrust has focused on hidden concentration at the employer level, resulting in coordinated effects and wage depression. This work is important in uncovering additional reasons that low-wage labor markets are not acting as predicted. However, the critique I am leveling is at a flaw within application of the consumer welfare analysis itself, and the way it protects and penalizes the exact same behavior in two types of labor markets.

Other arguments have focused on whether antitrust needs to adopt a public interest standard in order to account for labor effects. South African law, for example, requires its competition authorities to consider the effects of a merger on employment. Competition authorities divide the work force into three main categories – unskilled, semiskilled, and skilled – and appear to be focused more on retraining the employees who are going to lose their jobs and making sure they are not forgotten and left to their own devices as pawns of greater corporate machinations. While I have no objection to taking labor effects into account that way, my argument is more that of “treat them the same” under existing law not

“find a way to make sure they get taken care of.”

VI.  SOME WAYS TO MAKE A BAD SITUATION BETTER

What can enforcers do to better take into account labor effects? I offer five policy prescriptions that range from taking better account of labor effects in merger analysis to better protecting low-wage labor markets generally.

• Establish an interagency labor task force to examine these issues. The White House or Congress should mandate the creation of an interagency labor task force. The task force would include the DOJ Antitrust Division and the FTC’s Bureau of Competition and could include, among others, principals from the Department of Agriculture, Department of Commerce, and the National Economic Council. An interagency group would force the two enforcement agencies to harmonize enforcement approaches on this issue, and obtain input from peer agencies with greater experience in labor or the effects of consolidation on low-wage labor markets. Precedent for such interagency working groups already exists – the Committee on Foreign Investment in the United States (CFIUS) is an interagency working group that is convened as needed to evaluate and approve certain foreign investments in the United States.
• Bring in labor specialists. The enforcement agencies should permanently hire labor specialists, to help with all of their investigations, but in particular to help correct any bias against low-wage labor in merger analysis. This means hiring labor lawyers, labor economists, and behavioral economists.
• Consider labor effects in every merger case. While there is no “checklist” for merger evaluation at the enforcement agencies, and theories are considered anew even in mergers in the same industry or parties from a prior attempt taking another run at the altar, there can be a tendency to dismiss certain types of harm out of habit, and stick to the product market effects. The remedy could be as simple as an internal directive from each agency requiring case teams to include a paragraph in their recommendation memos affirming that they considered labor effects and how this supports challenging or not challenging the merger.
• Bring a labor effects case in a merger with no downstream product market effects. There will be a temptation to bring an employer monopsony case in a merger with downstream product effects, so that the effect on high-wage labor can be added as a secondary count to a more traditional core complaint – that the merger will result in price or output effects on the product market. For example, the primary harm alleged in the two DOJ complaints cited above was that the merger would raise the cost of health insurance; the labor harms were secondary. The problem with doing it this way is that judges may rule on the mainstream argument without addressing the labor harm. To truly advance this issue, the agencies have to move forward without a safety net and bring a labor harm-only case. The paradigmatic case would be one where, for example, two coal mines in a two-coal mine town merge – presumably not enough market power to affect the market for coal, but enough to affect the local labor market for coal miners.

Legislatively limit Amex to prevent abuses of the gig economy. At this point, everyone is still kind of feeling out exactly what the decision in Ohio v. Amex¹³ means for their corner of antitrust law. However, it could theoretically worsen conditions for low-wage labor by enabling, for example, a ride-sharing platform like Uber that brings together riders on one side of the platform with gig drivers on the other side of the platform to justify harming the driver side of the platform because it would in some way enhance competition on the rider side of the platform. The solution would require defining the boundaries of Amex to prevent this kind of abuse, either by passing legislation or bringing subsequent enforcement actions to limit the reach of the holding.

1 Anant Raut is a former Counsel to two Assistant Attorneys General of the Department of Justice’s Antitrust Division. He also served as a special advisor to the Office of Vice President Joseph Biden and President Obama’s National Economic Council on stimulating market competition and real wage growth for low- and middle-income workers. He was previously a staff attorney in the Federal Trade Commission’s Bureau of Competition. Mr. Raut is a graduate of Yale University and Harvard Law School. These comments
reflect his personal views.

2 See Press Release, “Therapist Staffing Company and Two Owners Settle Charges that They Colluded on Rates Paid to Physical Therapists in Dallas/Fort Worth Area” (July 31, 2018), available at https://www.ftc.gov/news-events/press-releases/2018/07/therapist-staffing-company-two-owners-settle-charges-they; United States v. Knorr-Bremse AG and Westinghouse Air Brake Technologies Corporation, No. 1:18-cv-00747 (2018); United States v. Lucasfilm Ltd., No. 1:10-cv-02220 (2010); and United States v. Adobe Systems, Inc., et al., No. 1:10-cv-01629 (2010).

3 15 U.S.C. 1.

4 The search was somewhat cursory, due to the author’s lack of a grant or junior associate to whom to farm the work out to.

5 Feinstein, D. & A. Teng, “Buyer Power: Is Monopsony the New Monopoly?,” Antitrust, Vol. 33, No. 2 (Spring 2019). Unrelated, but no less interesting, is the fact that Bruce Hoffman once poked my infant son in both eyes while misperforming a magic trick.

6 Because the impacted physicians are not employees of the merging entities but third parties selling them a service, these cases are classic monopsony buying power cases, not employer monopsony cases.

7 Complaint at 29, U.S. v. Aetna, Inc. and the Prudential Insurance Company of America, No. 3-99 CV 1398-H (N.D. Tex. 1999), available at https://www.justice.gov/atr/case-document/file/483516/download.

8 Complaint at 33, U.S. v. UnitedHealth Group, Inc. and Pacificare Health Systems, Inc., 1:05 CV 02436 (D.D.C. 2005), available at https://www.justice.gov/atr/case-document/file/514011/download.

9 “Labor Market Monopsony: Trends, Consequences, and Policy Responses,” Council of Economic Advisers Issue Brief (October 2016), available at https://obamawhitehouse.archives.gov/sites/default/files/page/files/20161025_monopsony_labor_mrkt_cea.pdf.

10 Marinescu, Ioana, & Roland Rathelot. 2018. “Mismatch Unemployment and the Geography of Job Search,” American Economic Journal: Macroeconomics, 10 (3): 42-70.

11 This author has passed on applying for jobs that would require switching onto one additional Metro line.

12 Aleem, Z., “‘Another kick in the teeth’: a top economist on how trade with China helped elect Trump,” Vox (March 29, 2017), available at https://www.vox.com/new-money/2017/3/29/15035498/autor-trump-china-trade-election.

13 Ohio v. American Express Co., 138 S. Ct. 2274 (2018).