Antitrust Division Deputy Assistant Attorney General Michael Kades Gives Remarks at Annual CRA Brussels Conference 2023

Remarks as Prepared for Delivery

Thank you, Fiona, for the kind introduction. It is a pleasure to be here.

Given who is speaking and the expertise in the crowd, I fear the next 20 minutes will be obvious and basic (and devoid of humor).

But, that is the point. As an antitrust enforcer for most of my career, I am frustrated. Too often, the government has to litigate cases that should be obvious. The government constantly has to prove that water makes things wet.

There are many explanations: outcomes in litigation have a high variance, the law is not strong enough, the courts are too conservative, defense bar attorneys are better in the court room, the government is too theoretical, the government is too simplistic, we need less economics in the courtroom, we need more economics in the courtroom.

Yes. That was your joke, so if you did not laugh for the rest of speech, don’t blame me.

There is no single solution, but I want to focus on one specific area. If antitrust fights gave scars, I would have a deep one from the fight over merger enforcement in the hospital industry. Beginning in the 1980s, the government lost nine challenges in a role. Yes, it was corrected. But the empirical literature on the harms from horizontal consolidation in hospital markets is overwhelming.

A second scar would be from the fight over pay-for-delay patent settlements. After a ten-year battle with a serious of launches, the Federal Trade Commission convinced the Supreme Court (by a single vote) that a company with substantial market power paying a potential competitor not to enter the market can be anticompetitive.

In both those examples, the economics was clear. Economists found the results puzzling and were often surprised that the courts were relying on principles that were no longer valid, if they ever were.

So, one challenge is that antitrust law, how it is practiced, often clings to principles long after they have lost currency in the relevant field.

If we analogize courts to a policy making arm, the problem is not unique. In many areas, principles accepted by policy makers as indisputable turn out to have been rejected.  

Economic propositions that were once taken as an article faith just a few decades ago now resemble discarded theories refuted by newer research: the minimum wage reduces employment or addressing inequality stifles growth.

That reevaluation is occurring in antirust policy as well driven, but not exclusively, by academic research. That research is driving changes in enforcement Good – and that is the critical word – economic analysis is and will continue to be critical for effective antitrust reform.

It should be obvious that the good research works it way into policy, but it turns out to be uncommon. As basic as it is, it is worth reviewing where and how economic research is affecting antitrust enforcement.

The 1990s and early 2000s were a heady time. Democracy had triumphed over Communism. The tech boom seemed to prove yet again the vibrance of western-style capitalism.

On economic policy, there was broad agreement. Clintonomics seemed to incorporate conservative critiques of big government while finding a role to address social and economic prices.

Antitrust law could be seen in the same light. In law schools, treatises, and in the bar, many celebrated a purported consensus one that mixed Chicago School and Post Chicago School ideas. There were debates still to be had, but they were on the margins and technical. In the late 1990s. I would have agreed. Spoiler alert – I was wrong.

Whatever consensus had existed was shattered by 2015. A broad public debate developed. No longer was antitrust enforcement a backwater involving a small group of experts.

Members of Congress began giving major speeches (or even wrote books) on the dangers of increasing market power and lax antitrust enforcement. The Economist and Time Magazine published stories on the growing monopoly threat. Congress investigated large Tech companies. Wildly divergent public figures from John Oliver to Tucker Carlson raised concerns about the lack of competition in U.S. economy.

The president of the United States created a competition council and discussed competition in his State of the Union.

I want to put the U.S. antitrust debate into a larger context. What is happening in antitrust discussions is part of a broader reevaluation. Take two principles – the minimum wage and the tension between growth and equality.

In 1946, George Stigler asserted confidently that the potential for the minimum wage to harm workers was “unusually definite for questions of economic policy.” Endorsed by many economists, that conclusion became a mainstay of microeconomics for decades.

 This assumption remained largely unchallenged until studies repeatedly failed to validate its findings. In 1994, David Card and Alan Krueger published an article assessing the impact of New Jersey’s 80-cent increase in the minimum wage. They compared employment outcomes for fast food workers in New Jersey with Pennsylvania. The authors found no evidence that New Jersey’s increase in the minimum wage dampened employment in the fast-food industry; in fact, relative employment in New Jersey actually rose.

Thirty years later, few economists would contend unequivocally that raising the minimum wage is always bad policy. To the contrary, a review of the economic literature concluded that the effect of the minimum wage was so “muted” “that the minimum wage raises wages much more than it has any effect on jobs.”

With respect to economic growth, Professor Arthur Okun described redistribution programs as a “leaky bucket,” asking the question how much loss of water – of economic productivity – society was willing to tolerate to reduce inequality and poverty. That question haunted social welfare programs for decades. Yet new evidence suggest that the tradeoff is more complicated. As a 2014 OECD report concluded, “income inequality has a negative and statistically significant impact on medium-term growth.” 

It would be shocking if those ideas – so central to policymaking from the 1980s to the mid-2010s – were questioned, limited, and even discarded, but the antitrust principles of the same provenance remained unscathed.

Indeed, many principles that have carried and continue to carry weight with many who practice antitrust law or judges who decide cases no longer should be accepted. Today, I will look at three of these principles that should no longer have currency in the law, identify the research, and how enforcement is changing.

The leader of the Chicago School, Robert Bork, contended that “changing market conditions and the temptation to ‘cheat’ frequently result in outbreaks of price competition that either destroy the cartel or must be repaired by further meetings and agreements.” This fragility, Bork posited, made cartels “easy to detect” and tacit agreements unlikely. In his view, “[i]f uncertainties and temptations break down even detailed and specific agreements on prices, what chance has an unspoken, detailed mutual restraint of surviving market pressures? The best guess seems to be none.” This admitted “guess” grew into a virtually indisputable presumption.

Even during this precept’s heyday, some scholars questioned whether cartels were really that unstable and competition that certain. And subsequent scholarship by Margaret C. Levenstein & Valerie Y. Suslow, among others, found that cartels have considerable longevity. Some cartels lasted decades. Often, only government enforcement terminated the cartel, not cheating or entry. 

Two recent cases emphasize the ability of even a large number of competitors to coordinate over a long period of time. For example, the Antitrust Division has settled to date with four poultry processors over allegations that they collaborated, including during regular in-person resort trips, with their competitors in making decisions about worker compensation. The conduct persisted for at least 20 years without detection.

The Antitrust Division has recently brought an information-exchange case, alleging that the defendant has orchestrated for many years information exchanges among all the major chicken, pork, and turkey processors to exchange highly granular information related to price, output, cost (including labor), and other sensitive aspects of competition. The Division alleges that conduct began decades ago. That case is now ongoing.

As with traditional cartels, Bork asserted confidently “that we are entitled to be highly skeptical of the entire theory oligopolistic interdependence” and that he was “not persuaded that such behavior occurs outside of economics textbooks.” This critique, coupled with a broader attack on structural evidence in merger enforcement, drove the focus on unilateral effects.

The unilateral effects section of the Guidelines expanded from a brief paragraph in the 1982 Guidelines to “the theory . . . most often pursued by the Agencies.”  Unilateral effects theories increased from “less than twenty percent of merger investigations at the start of the 1990s . . . to well over seventy-five percent of merger investigations by 2010.”

Not to be misunderstood, I am not a criticism of unilateral effects. The theory and its adoption by courts is an important victory for antitrust enforcement. It just should not be an either/or situation.

Recent empirical work has confirmed increasing market concentration can result in oligopoly price leadership. For example, one study found that, after the Miller-Coors merger in 2008, the two major American beer giants (ABI and MillerCoors) started pursuing oligopolistic pricing where, every year, one would publicize prices to take effect in a couple months that the other would follow. Ultimately, the study concluded that the price leader’s prices after the merger were about 9%. Similarly, a literature review of mergers in oligopolistic markets—which mostly examine airlines, banking, hospitals, and petroleum concluded that “[t]he empirical evidence that mergers can cause economically significant increases in price is overwhelming.” “Of the 49 studies surveyed,” which of course consisted mostly of mergers not challenged by the enforcement agencies, the authors found evidence of merger-induced price increases in 36.

We are beginning to see the impact of this work. Judge Pan, in United States v. Bertlesman, found the merger would raise the likelihood of tacit collusion relying, among other things, on the history of collusion in the industry. The Proposed Merger Guideline incorporate recent learning and research on coordinated effects.

A third area where economic realities have overtaken antitrust dogma is labor. Antitrust discounted the likelihood that workers or suppliers were victims of anticompetitive conduct or mergers. One argument is that labor markets are highly competitive, and some argued that monopsony power lowers cost to consumers. 

Relatedly, In the past 40 years, courts in the United States are increasingly willing to enforce non-competes, whatever their breadth or questionable justification. Emblematic of this view, one prominent judge in the United States advocated for enforcement of such covenants on the ground that the employee is “an adult of sound mind who made an unequivocal promise, for which he was doubtless adequately compensated,” and that “[i]f covenants not to compete are forbidden, the employer will pay a lower wage, in effect charging the employee for the training.”

Those assumptions, highly competitive markets, benefits to consumers, noncompetes are a fair bargain and employees are adequately compensation all suggest that labor market are relatively free of anticompetitive conduct.

It turns out those assumptions were wrong. A recent literature review by Professor Alan Manning found market power among employers to be common: “All of these studies, including those with high quality research designs, show that firms have considerable monopsony power, even in online markets that, a priori, one might have thought would be very competitive.” A study by Professors Elena Prager and Matt Schmitt concluded that mergers dampened wages for specialized healthcare workers. 

Nor has the theoretical support for monopsony power held up. Even if we set aside the legal rule that courts may not weigh benefits in one market against harms in another, monopsony power is unlikely at best to benefit the buyers downstream from the monopsonist. As Professors Scott Hemphill and Nancy Rose explain “when a merger suppresses competition between rivals for a seller’s business, ordinarily we may expect, in addition to the harm to workers or other sellers, inefficiency and consumer harm.” Increased classical monopsony power will harm have no effect on downstream purchasers if that market is competitive or exacerbate harms if the company has market power in the downstream market.

The research on noncompetes is more substantial. In the United States, noncompete agreements have become common even in low-wage fields, such as fast-food workers, which lack the traditional justifications for noncompetes. Several studies have shown that increased enforceability of noncompetes depresses – not increases – wages for employees.

Consistent with the literature, the Division has been active in addressing anticompetitive conduct affecting workers and suppliers.

• Last year, the Antitrust Division prevailed in a labor-based merger challenge – blocking Penguin Random House acquisition of Simon & Schuster.
• As I discussed above, we also brought a case alleging a 20-year agreement among chicken processors to collaborate on wage and compensation decisions. Those have led to consent decrees.
• Last month, the Antitrust Division brought its first enforcement action challenging a de facto noncompete. This case reflected the many ways in which real-world experience departed from certain assumptions around noncompetes. The workers subject to the noncompete were contract farmers, not highly skilled workers or salespeople. The farmers received nothing for agreeing to the clause. And, as we alleged, the noncompete suppressed contract farmers’ wages by stifling competition among processors for their services.

If, as you listened to the last third of this speech, you thought something along lines of “Well, that point is obvious.” Exactly that is the point. Effective antitrust enforcement requires expunging ideas that, if they were ever valid, no longer are. Otherwise, we can’t hope to tackle the more complicated problems. I will now return you to regularly scheduled programing on cutting edge antitrust issues.